1 commit 16860d448f11ba405050bd1ff02dc63a4b26280c
2 Author: Chris Wilson <chris@chris-wilson.co.uk>
3 Date: Sat Sep 29 10:40:51 2012 +0100
5 uxa: Only destroy the bufmgr if it exists
7 Be wary in case the Screen is destroy before we initialise bufmgr.
9 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11 commit 13f7a9e267fbefd3326569e89b840bef31d77996
12 Author: Chris Wilson <chris@chris-wilson.co.uk>
13 Date: Fri Sep 28 09:25:12 2012 +0100
15 sna: Restore LLC cache level after using a buffer for scanout (uncached)
17 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19 commit 8bfd31e9bb13bcb7f12e4147bec0da87b8e87dde
20 Author: Chris Wilson <chris@chris-wilson.co.uk>
21 Date: Thu Sep 27 18:12:11 2012 +0100
23 sna/dri: Improve handling of flipping with no outputs
25 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27 commit dddb6e4e63d5cc8b2a1f2ae3ff287922d30460b9
28 Author: Chris Wilson <chris@chris-wilson.co.uk>
29 Date: Thu Sep 27 17:17:30 2012 +0100
31 sna: Attempt to restore the current mode if pipe is disabled on DPMS on
33 If we attempt to power up the pipe through a DPMS request, but it was
34 previously disabled due to an error, first try re-enabling.
36 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
38 commit 88cfd23379950d1fe4e682519c48ef6e3091c2f3
39 Author: Chris Wilson <chris@chris-wilson.co.uk>
40 Date: Thu Sep 27 16:16:39 2012 +0100
42 uxa: Update mode->fb_id after completing pageflips
44 As intel_crtc_on() depends upon the current value for determining if the
45 pipe is active, we can only change the value afterwards.
47 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
49 commit 83990c649cbd3cb563d08f04e1c02232c44997aa
50 Author: Chris Wilson <chris@chris-wilson.co.uk>
51 Date: Thu Sep 27 14:58:26 2012 +0100
53 Remove the last couple of references to mibstore
55 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
57 commit cb0b72c4fbb7b1bc941522d8679fef856f2f8936
58 Author: Chris Wilson <chris@chris-wilson.co.uk>
59 Date: Thu Sep 27 14:53:46 2012 +0100
61 uxa: Fixup the last couple of commits after realising I didn't have uxa enabled
63 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
65 commit feef53005b57d6b4489a15f8d0db8245dcb0cfaa
66 Author: Chris Wilson <chris@chris-wilson.co.uk>
67 Date: Thu Sep 27 13:36:55 2012 +0100
69 uxa: Do not attempt to pageflip whilst not master
71 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
73 commit 5a45cbacb777e478d8fbda9223b0fb5c705d7249
74 Author: Chris Wilson <chris@chris-wilson.co.uk>
75 Date: Thu Sep 27 13:31:27 2012 +0100
77 uxa: Confirm the pipe is alive before flipping
79 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
81 commit edc5b7f741a4bb8e9a58b5eb2201fa9304f94ef7
82 Author: Chris Wilson <chris@chris-wilson.co.uk>
83 Date: Thu Sep 27 13:27:20 2012 +0100
85 uxa/dri: Confirm Drawable is still attached to an active pipe before flipping
87 If we schedule a flip to happen on a far distant vblank, it is possible
88 for us to disable the pipes with a modeset, whilst maintaining the fb
89 size, or with a DPMS command. If the pipe is disabled and we still try
90 to flip, then we trigger an EINVAL error from the kernel.
92 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
94 commit eb667378543f97d7c7e7767daddbd9b39e07c1f8
95 Author: Chris Wilson <chris@chris-wilson.co.uk>
96 Date: Wed Sep 26 15:12:57 2012 +0100
100 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
102 commit 557ad78d02dfff3db9e7fc21f05230a8fe482f7c
103 Author: Chris Wilson <chris@chris-wilson.co.uk>
104 Date: Wed Sep 26 15:02:17 2012 +0100
106 Kill mibstore pretend backwards compatibility
108 Adam Jackson was kind enough to point out that mibstore had been a no-op
109 since xorg-1.4 and the earliest we claim to support is xorg-1.6.
111 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
113 commit 367f57faf2c422da026a25c099426dc750cf3759
114 Author: Chris Wilson <chris@chris-wilson.co.uk>
115 Date: Wed Sep 26 14:06:18 2012 +0100
117 sna/video: Force integer promotion to avoid u16 overflow for frame size
119 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
121 commit 34772cb0e724a880a8c6db936c0bb5dd129398f2
122 Author: Chris Wilson <chris@chris-wilson.co.uk>
123 Date: Wed Sep 26 12:09:39 2012 +0100
125 sna/dri: Handle pending flips when drawable disappears.
127 The async case needs special handling as the current front buffer may be
128 pending a flip onto the scanout which needs to be completed even if the
129 Drawable is destroyed.
131 Fixes regression from commit ad877abdc70b842afd202b2f97892d97a0b8d151
132 Author: Chris Wilson <chris@chris-wilson.co.uk>
133 Date: Sat Jul 7 17:48:52 2012 +0100
135 sna/dri: Attach the chain of frame events to the Window, not pixmap
137 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
139 commit addfa653cf51fa45c7eae49be1fd722aaadf97d6
140 Author: Chris Wilson <chris@chris-wilson.co.uk>
141 Date: Wed Sep 26 08:59:18 2012 +0100
143 uxa: Prevent attempting to call intel_mode_fini() before PreInit runs
145 In the Ubuntu xserver it is apparently possible for the Screen to be
146 deleted prior to being fully initialised. Make sure we don't crash in
149 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55346
150 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
152 commit b880d8e0a9a32acd50333f72d695268c4819c29e
153 Author: Chris Wilson <chris@chris-wilson.co.uk>
154 Date: Tue Sep 25 16:59:20 2012 +0100
156 sna/gen5: Add WA reference
158 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
160 commit 74f930fd80c3f97a1b6213e9e79e02f8f51c64b9
161 Author: Chris Wilson <chris@chris-wilson.co.uk>
162 Date: Tue Sep 25 15:41:51 2012 +0100
164 sna: Make sure the batch is flushed before setting DPMS
166 Just in case on the off-chance that the kernel modifies the pipe and we
167 had a pipe-dependent operation in the batch.
169 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
171 commit 0a00c3765c3ba464bb4e9cde12e964c979421ed6
172 Author: Chris Wilson <chris@chris-wilson.co.uk>
173 Date: Tue Sep 25 13:50:33 2012 +0100
175 sna: Accommodate changes to the Damage extension not unwrapping the GC
177 commit 4dc2a76740d921c824a4d8193f39dd373475f02a
178 Author: Adam Jackson <ajax@redhat.com>
179 Date: Thu Sep 20 17:57:00 2012 -0400
181 miext/damage: Only wrap into the GC ops chain if there's a listener (v3)
183 leaves the GC->funcs still pointing to the damage layer as it calls into
184 the driver functions, so we need to be careful to restore the correct
185 function table after we manipulate the GC for fallbacks.
187 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
189 commit 8f5a6d107bb30fa218ff3c1d07532ca834b82edc
190 Author: Chris Wilson <chris@chris-wilson.co.uk>
191 Date: Tue Sep 25 13:36:12 2012 +0100
193 Add miInitializeBackingStore() to compat-api.h
195 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
197 commit dee133a722fb137767a78b244a312620f52a51b1
198 Author: Chris Wilson <chris@chris-wilson.co.uk>
199 Date: Tue Sep 25 13:22:40 2012 +0100
201 sna: Remove unused reference to mibstore.h, removed in future Xorg
203 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
205 commit f0dbfc20aaa9b47329438ce42150770d4d7934c3
206 Author: Chris Wilson <chris@chris-wilson.co.uk>
207 Date: Tue Sep 25 13:22:21 2012 +0100
209 sna/dri: Fixup a DBG statement not to print gibberish
211 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
213 commit c6008068372709c73034163eddc902b47bf87d24
214 Author: Chris Wilson <chris@chris-wilson.co.uk>
215 Date: Sun Sep 23 21:42:31 2012 +0100
217 sna: Check against op->dst.bo rather than priv->cpu_bo for composite upload
219 In this case, we may also be handling an unattached pixmap, so avoid the
220 deferences of the sna_pixmap unless we are sure it will exist.
222 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
224 commit 5ed840881c26e90eb8e00521b6d77b0ea514de5e
225 Author: Chris Wilson <chris@chris-wilson.co.uk>
226 Date: Sun Sep 23 20:29:21 2012 +0100
228 sna: Check that the CPU bo exists before declaring it is busy along upload
230 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55251
231 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
233 commit a858afc66c1fa2eec65a7041e991f2266f82deca
234 Author: Chris Wilson <chris@chris-wilson.co.uk>
235 Date: Sun Sep 23 13:49:40 2012 +0100
237 Silence a couple of potential compiler warnings
239 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
241 commit 9326acc2917109f06dda809107c8fa5a2273c3d2
242 Author: Chris Wilson <chris@chris-wilson.co.uk>
243 Date: Sun Sep 23 13:40:04 2012 +0100
245 Allow compilation of a separate i810 driver
247 Allow --enable-ums-only as a counter-option to --enable-kms-only in case
248 the distribution wishes to enable a non-root KMS driver but also offer
249 a separate UMS driver for i81x.
251 On the second pass, use "--enable-ums-only --disable-uxa --disable-sna"
252 to get the trimmed down unaccelerated i810 support.
254 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
256 commit 0afb7efe8c48b5fc839e2137e870bea0f5fb3c9c
257 Author: Chris Wilson <chris@chris-wilson.co.uk>
258 Date: Sun Sep 23 09:41:59 2012 +0100
260 sna: Avoid overflows when translating the box16 extents during a copy
262 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
264 commit 0fc6e5820e4543d52bcf8d0285ca6f69b5213831
265 Author: Chris Wilson <chris@chris-wilson.co.uk>
266 Date: Sun Sep 23 09:32:49 2012 +0100
268 sna: Add missing protection against int16 overflow when copying
270 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
272 commit c2ca1c5abca687adb08d6e137bc3b70bdecb083d
273 Author: Chris Wilson <chris@chris-wilson.co.uk>
274 Date: Sun Sep 23 08:56:50 2012 +0100
276 Reorder INTEL_XVMC define so that we pick up Option "XvMC"
278 Otherwise it will remain disabled.
280 Reported-by: Giacomo Comes <comes@naic.edu>
281 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
283 commit 92dbedc6138b923aa473935013ecb0346280c4d3
284 Author: Chris Wilson <chris@chris-wilson.co.uk>
285 Date: Sat Sep 22 08:56:49 2012 +0100
287 sna: Force the stall before trying to upload into a busy CPU bo
289 Under the circumstances where we can not instead stream the write into
290 the GPU bo, we need to use the busy CPU bo.
292 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54978
293 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
295 commit 3e1be265cf950976b5929b14a9dad0664deaa2c1
296 Author: Chris Wilson <chris@chris-wilson.co.uk>
297 Date: Fri Sep 21 17:37:58 2012 +0100
299 sna: Force an inplace upload if already wedged
301 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
303 commit b8967aff382c1b6bef2335dea51c979a3f0800c7
304 Author: Chris Wilson <chris@chris-wilson.co.uk>
305 Date: Fri Sep 21 12:16:27 2012 +0100
307 sna: Do not query for the NULL edid property
309 If the EDID blob id is set to 0, that means that it does not exist and
310 so we can safely skip it.
312 References: https://bugs.freedesktop.org/show_bug.cgi?id=55193
313 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
315 commit 0be1d964713ca407f029278a8256d02d925dc9da
316 Author: Chris Wilson <chris@chris-wilson.co.uk>
317 Date: Tue Sep 11 21:48:24 2012 +0100
319 sna: Use inplace X tiling for LLC uploads
321 Based on a suggestion by Chad Versace (taken from a patch for mesa).
323 This allows for a faster upload of pixel data through a ShmImage, or for
324 complete replacement of a GPU bo.
326 Using a modified version of x11perf to upload to a pixmap rather than
327 scanout on an IVB i7-3720qm:
330 40000000 trep @ 0.0007 msec (1410000.0/sec): ShmPutImage 10x10 square
331 4000000 trep @ 0.0110 msec ( 90700.0/sec): ShmPutImage 100x100 square
332 160000 trep @ 0.1689 msec ( 5920.0/sec): ShmPutImage 500x500 square
335 40000000 trep @ 0.0007 msec (1450000.0/sec): ShmPutImage 10x10 square
336 6000000 trep @ 0.0061 msec ( 164000.0/sec): ShmPutImage 100x100 square
337 400000 trep @ 0.1126 msec ( 8880.0/sec): ShmPutImage 500x500 square
339 However, the real takeaway from this is that the overheads for
340 ShmPutImage are substantial, only hitting around 70% expected efficiency,
341 and overshadowed by PutImage, which for reference is
343 60000000 trep @ 0.0006 msec (1800000.0/sec): PutImage 10x10 square
345 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
347 commit d853064e7eebc5719645c12605782f995131a6fe
348 Author: Chris Wilson <chris@chris-wilson.co.uk>
349 Date: Thu Sep 20 22:43:26 2012 +0100
351 sna/gen3+: Trim the target extents to the CompositeClip
353 When computing the active region with of a composite operation with
354 unknown extents we try to simply use the whole Drawable. However, this
355 needs to be clipped otherwise it may trigger assertion failure with an
358 References: https://bugs.freedesktop.org/show_bug.cgi?id=55164
359 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
361 commit 853beff4306d5a168e310af82d0ddf9db319fdce
362 Author: Chris Wilson <chris@chris-wilson.co.uk>
363 Date: Wed Sep 19 08:30:59 2012 +0100
365 sna/dri:Add some DBG and assertions to validate names passed back to clients
367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
369 commit 8c1e60a15431b193b57d94a27b8e8de7c5a7e683
370 Author: Chris Wilson <chris@chris-wilson.co.uk>
371 Date: Tue Sep 18 15:01:52 2012 +0100
373 sna: Add some asserts useful for tracking an xserver drawing bug
375 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
377 commit cd23ac69ffb670468b2993242ce0d596081bbca4
378 Author: Chris Wilson <chris@chris-wilson.co.uk>
379 Date: Tue Sep 18 13:39:33 2012 +0100
381 sna/gen7: Add some ring switching sanity checks
383 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
385 commit 22c84d02ade83c39940d76fc616a9122e535606a
386 Author: Chris Wilson <chris@chris-wilson.co.uk>
387 Date: Tue Sep 18 12:01:49 2012 +0100
389 sna: Remove special casing of drawing depth=1 glyphs
391 It turns out that they were not so special after all, and the "fast
392 path" was missing the important handling such as validating the GC. As a
393 result we ended up trying to render the glyphs through a planemask of 0
394 and similar garbage resulting in corruption. An example given by Kaus
395 Dittrich was the use of rotated glyphs in gnuplot.
397 Reported-by: Klaus Dittrich
398 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
400 commit 71fa350325bb06a6141f89ef14431d96f0c4956a
401 Author: Chris Wilson <chris@chris-wilson.co.uk>
402 Date: Mon Sep 17 13:40:24 2012 +0100
404 sna: Check that we do not keep a GPU bo around after flushing a ShmPixmap
406 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
408 commit 88502b32d5d2b826fc08344392c98300341410e1
409 Author: Chris Wilson <chris@chris-wilson.co.uk>
410 Date: Sun Sep 16 09:02:46 2012 +0100
414 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
416 commit 395d738b5086f24659b56b4b3536bffaebad8a3e
417 Author: Chris Wilson <chris@chris-wilson.co.uk>
418 Date: Sat Sep 15 21:18:35 2012 +0100
420 sna: Improve handling of probe failure during initialisation
422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
424 commit fe78904a45979d47bb092ce37d321854e2418344
425 Author: Chris Wilson <chris@chris-wilson.co.uk>
426 Date: Thu Sep 13 22:15:14 2012 +0100
428 Add basic support for ValleyView
430 Bind to the ValleyView SDV for verifying the render routines.
432 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
434 commit 3b9259f7cedc179617bc24b3912d2d8d75c5a824
435 Author: Ian Stakenvicius <axs@gentoo.org>
436 Date: Fri Sep 14 21:04:42 2012 +0100
438 configure: Make udev dependency optional
440 In order to support buildbots where the udev headers may exist on the
441 build system but not the target, we need explicit control over optional
444 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54942
446 commit d995705fb01842652a79076cbecee4392f653bfe
447 Author: Chris Wilson <chris@chris-wilson.co.uk>
448 Date: Fri Sep 14 19:32:06 2012 +0100
450 sna: Only discard CPU damage if we completely overwrite its extents
452 If we are performing a clipped copy, then we must be careful not to
453 completely discard the CPU damage as it may not be entirely replaced.
455 Reported-by: Roman Jarosz <kedgedev@gmail.com>
456 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54937
457 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
459 commit deacab87e4ba7fe09f0f16568dc2848949f2707a
460 Author: Chris Wilson <chris@chris-wilson.co.uk>
461 Date: Fri Sep 14 17:01:13 2012 +0100
463 sna/dri: flip->next_front.bo is owned by the DRI drawable not by the flip
465 Be careful not to delete the reference we presume we hold as it is
466 borrowed from the DRI drawable.
468 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
470 commit 5369408d596a0611d4f76333dac2e1c93e63e8b6
471 Author: Chris Wilson <chris@chris-wilson.co.uk>
472 Date: Fri Sep 14 16:12:00 2012 +0100
474 sna/dri: And the hunt for the use-after-free continues...
476 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
478 commit 8e9d64203428b344433ac8f86dc1f06d2c32529a
479 Author: Chris Wilson <chris@chris-wilson.co.uk>
480 Date: Fri Sep 14 14:06:38 2012 +0100
482 sna: Fix typo for sna_render_picture_flatten()
484 Pass it the correct value for the pixmap height, and not its 'y'
487 Reported-by: Eugene Rosenzweig <ugn@outlook.com>
488 Bugzilla: https://bugs.gentoo.org/show_bug.cgi?id=434860
489 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
491 commit 6ee9969f0e99975a24e113af9591909d9023d8af
492 Author: Chris Wilson <chris@chris-wilson.co.uk>
493 Date: Fri Sep 14 13:04:50 2012 +0100
495 sna/dri: Update AsyncSwap for recent changes in tracking DRI drawables
497 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
499 commit d2dbb991384f2b820a714eba262691ed97a6a22a
500 Author: Chris Wilson <chris@chris-wilson.co.uk>
501 Date: Fri Sep 14 13:04:14 2012 +0100
503 sna/dri: Add a bunch of assertions for hunting a use-after-free
505 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
507 commit d8756091474aeb44e5d8eed6b20b65c23b6574e7
508 Author: Chris Wilson <chris@chris-wilson.co.uk>
509 Date: Fri Sep 14 12:17:28 2012 +0100
511 sna/dri: Version bump for prototypical AsyncSwap
513 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
515 commit 0fc3d020f3086b5c6ecef1fed1aaedef7aa52b01
516 Author: Chris Wilson <chris@chris-wilson.co.uk>
517 Date: Fri Sep 14 09:49:00 2012 +0100
519 Skip adding unknown special modes
521 In light of the discussion on how to add stereo modes it was revealed
522 that we would happily add modes with unknown and unhandled strange
523 flags. Adam Jackson mentioned that he has plans to fix that with some
524 upcoming work, but as a first step we can simply eradicate them whilst
525 pondering how to support the stereoscopic vision of the future.
527 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
529 commit 35d81d97b38d05f132a4823e57fff4012e966504
530 Author: Chris Wilson <chris@chris-wilson.co.uk>
531 Date: Thu Sep 13 22:32:47 2012 +0100
533 sna/gen7: Defensively program default GT values
535 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
537 commit 3fbdedbf9d509c5ca58ae0a01fe8e54dcc990cf8
538 Author: Chris Wilson <chris@chris-wilson.co.uk>
539 Date: Thu Sep 13 22:10:08 2012 +0100
541 sna: Fix analysis of source extents for BLT composite
543 After we have computed the source offset vector for the transformed
544 source bo, we need to use that with respect to the destination rectangle
545 to verify that the source sample is wholly within bounds.
547 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
549 commit 6c7f998f7bda4f4f1286e079412d5d3a5b2eb073
550 Author: Chris Wilson <chris@chris-wilson.co.uk>
551 Date: Thu Sep 13 20:41:23 2012 +0100
553 sna: Fix BLT composite offset
555 Reported-by: Jiri Slaby <jirislaby@gmail.com>
556 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54868
557 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
559 commit 2575cd0d236b4e1694e3185a487ebfd8bfe6499f
560 Author: Chris Wilson <chris@chris-wilson.co.uk>
561 Date: Thu Sep 13 20:08:37 2012 +0100
563 sna: Avoid corrupting the CPU bo when compositing fallback data through BLT
565 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
567 commit 5781de2e5ab0d7e8e62965bda714789879555ed4
568 Author: Chris Wilson <chris@chris-wilson.co.uk>
569 Date: Thu Sep 13 19:53:29 2012 +0100
571 sna/gen3: Tidy NO_COMPOSITE debugging
573 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
575 commit dd1432b2c00c94ac75af4740b66b5cd1a573e261
576 Author: Chris Wilson <chris@chris-wilson.co.uk>
577 Date: Thu Sep 13 19:47:29 2012 +0100
579 sna: Do not attempt to change tiling if wedged
581 The caller will just have to live with the current tiling, which should
582 be fine in most cases, in other cases the gpu is wedged...
584 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
586 commit 8c7dd2219fa777bf6354c4e0ef38a2f09fe09675
587 Author: Chris Wilson <chris@chris-wilson.co.uk>
588 Date: Thu Sep 13 17:14:23 2012 +0100
590 sna: Prefer to use indirect uploads for very small updates
592 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
594 commit d87a56ed1789e0c5058b302bb930d7e952ff3e5e
595 Author: Chris Wilson <chris@chris-wilson.co.uk>
596 Date: Thu Sep 13 09:57:30 2012 +0100
598 sna: Teach sna_replace__xor() how to tile large uploads
600 This path is hit using eog+cairo-1.10 and a large image, e.g.
601 http://marsrovers.jpl.nasa.gov/gallery/press/opportunity/20120705a/PIA15689_Greeley_Pan_wDeck_L257F.jpg
603 Reported-by: Michael Laß <bevan@bi-co.net>
604 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54808
605 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
607 commit 58a96f0f684fe0d7d1a7890c630539ef8b065d1e
608 Author: Chris Wilson <chris@chris-wilson.co.uk>
609 Date: Wed Sep 12 23:53:44 2012 +0100
611 sna: Also remove bogus assertion from sna_replace__xor()
613 Like the assertion in sna_replace() this is now incorrect due to the
614 automatic flagging of large bo with flush.
616 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
618 commit 6069e78238ca57e3433d528680fe04766a1e5307
619 Author: Chris Wilson <chris@chris-wilson.co.uk>
620 Date: Wed Sep 12 16:22:49 2012 +0100
622 sna: Keep a very small, short-lived cache of large buffers
624 As we now regularly retire and so discard the temporary large buffers,
625 we find them in short supply and ourselves wasting lots of time creating
626 and destroying the transient buffers.
628 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
630 commit 1e2a03ad19664e846ece4110e53d26babb5a2159
631 Author: Chris Wilson <chris@chris-wilson.co.uk>
632 Date: Wed Sep 12 16:15:22 2012 +0100
634 sna: Propagate busyness when creating the proxy
636 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
638 commit e28f5a2537821dc170a7ef78bf33ecbd3c032da0
639 Author: Chris Wilson <chris@chris-wilson.co.uk>
640 Date: Wed Sep 12 15:21:27 2012 +0100
642 sna: Add some DBG as to why kgem_check_bo() flushes
644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
646 commit 1af1dae1feefd6567017170bab241570fb528e5c
647 Author: Chris Wilson <chris@chris-wilson.co.uk>
648 Date: Wed Sep 12 15:20:49 2012 +0100
650 sna: Ensure tiling upload buffers are trimmed to fit in the cache
652 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
654 commit e5f137807c318588f546960668345eef34159e26
655 Author: Chris Wilson <chris@chris-wilson.co.uk>
656 Date: Wed Sep 12 14:11:43 2012 +0100
658 sna: Avoid fallbacks to shadow pixels if FORCE_GPU is in effect
660 References: https://bugs.freedesktop.org/show_bug.cgi?id=54808
661 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
663 commit 0d17208a66a7e54d4106f8a4034d3a928e28bb62
664 Author: Chris Wilson <chris@chris-wilson.co.uk>
665 Date: Wed Sep 12 14:04:50 2012 +0100
667 sna: Avoid readback inplace if the target is unmappable
669 We have to use the tiling indirect path, or else fail.
671 References: https://bugs.freedesktop.org/show_bug.cgi?id=54808
672 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
674 commit 4b4abdaae94d164d5d0b2755907e76b9cbe0c988
675 Author: Chris Wilson <chris@chris-wilson.co.uk>
676 Date: Wed Sep 12 13:47:26 2012 +0100
678 sna: Flush after operating on large buffers
680 As we know that such operations are likely to be slow and consume
681 precious GTT space, mark them as candidates for flushing.
683 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
685 commit 15911f533d6a7ef40e42ba5921fac7c62b290f8b
686 Author: Chris Wilson <chris@chris-wilson.co.uk>
687 Date: Tue Sep 11 14:59:53 2012 +0100
689 prime: Reorder probe messages so that the driver prefix is correct
691 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
693 commit bc73acbd4da53bc50752c0413adcd0ce876e0a03
694 Author: Chris Wilson <chris@chris-wilson.co.uk>
695 Date: Mon Sep 10 18:55:04 2012 +0100
697 uxa: Error out after failing to become DRM master during PreInit
699 As we set the card fd to -1 upon failure, the error here is
700 non-recoverable, so make sure it gets propagated and we abort loading
703 Bugzilla: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/962892
704 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
706 commit b0d14071f7b60729c223af925935227393fbd3ee
707 Author: Chris Wilson <chris@chris-wilson.co.uk>
708 Date: Mon Sep 10 13:53:45 2012 +0100
710 sna: Workaround issue with global glyph privates and shared ZaphodHeads
712 Under ZaphodHeads we end up with multple screens accessing the common
713 sna_glyph_key and so cause conflicting updates and erroneous references
714 into the screen-local texture atlases.
716 Two approaches can be tried here. Transition to a screen-specific
717 private key introduced with xorg-1.13, or to move the glyph cache (and
718 the rest of the gpu state tracker) down into the device private rather
719 than screen private. This is neither of those, but a workaround to avoid
720 reusing the incorrect entries from shared screens.
722 Reported-by: Stephen Liang <inteldriver@angrywalls.com>
723 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54707
724 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
726 commit cfa0c6162bbd2f6edab1301b6aed886af7e8ab38
727 Author: Chris Wilson <chris@chris-wilson.co.uk>
728 Date: Mon Sep 10 10:50:03 2012 +0100
730 sna: Fixup possible_crtcs for ZaphodHeads
732 As the possible_crtcs is a bitmask of the available crtcs exposed to the
733 Xserver, we need to adjust it for the limited view given by Zaphod mode.
735 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
737 commit 95c71d1626dfbd97733e7e7e03fdde3af833446e
738 Author: Chris Wilson <chris@chris-wilson.co.uk>
739 Date: Sat Sep 8 20:12:32 2012 +0100
741 NEWS: Fix version reference 'bugs since 2.20.6'!
743 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
745 commit 0659711dfa9c9fc52adeaba766049887d5b6f040
746 Author: Chris Wilson <chris@chris-wilson.co.uk>
747 Date: Sat Sep 8 20:01:32 2012 +0100
751 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
753 commit 2c8e48f8149499040342836491333402fd05b762
754 Author: Chris Wilson <chris@chris-wilson.co.uk>
755 Date: Fri Sep 7 23:00:47 2012 +0100
757 sna: Remember that LineDoubleDash involves updating the GC between segments
759 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
761 commit ed7b8db24921dc0bb6ea59dacf35ea41a61d59bf
762 Author: Chris Wilson <chris@chris-wilson.co.uk>
763 Date: Fri Sep 7 22:12:17 2012 +0100
765 sna: Pass the hints down to the PolyPoint fallback
767 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
769 commit 6c9df48fa8f1ac93103ce771ae28f47faa321619
770 Author: Chris Wilson <chris@chris-wilson.co.uk>
771 Date: Fri Sep 7 22:05:19 2012 +0100
773 sna: Fix PolyPoint fallback through a complex clip
775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
777 commit 7d76eb155ed78343abfa15b7c8af87fc3d4211e4
778 Author: Chris Wilson <chris@chris-wilson.co.uk>
779 Date: Fri Sep 7 18:06:25 2012 +0100
781 sna: First scan for a backlight associated with the device
783 The goal is to find the right interface in a multi-GPU system.
785 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
787 commit 2de7d2d8150d648815feb74a3ae2239b908b971e
788 Author: Chris Wilson <chris@chris-wilson.co.uk>
789 Date: Fri Sep 7 16:56:33 2012 +0100
791 sna/prime: Fixup attaching the randr14 scanout pixmap
793 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
795 commit 448808d8a4e78afa7fe90e318d6d7afdbff66190
796 Author: Chris Wilson <chris@chris-wilson.co.uk>
797 Date: Fri Sep 7 14:31:52 2012 +0100
799 prime: Allocate a GPU_SCREEN when not primary to fulfil the output sink role
801 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
803 commit 3dce661ef3df5558fe9c9d85324ab3395a37b7ea
804 Author: Chris Wilson <chris@chris-wilson.co.uk>
805 Date: Fri Sep 7 13:48:12 2012 +0100
807 sna: And remember to update the ScreenPixmap after resize
809 Fixes the regression from the previous commit
811 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
813 commit 0c39c363d7235d03f27337c731c0ffe3369bf713
814 Author: Chris Wilson <chris@chris-wilson.co.uk>
815 Date: Fri Sep 7 13:07:52 2012 +0100
817 sna: Protect sna_crtc_resize() against early invocation with no RootWindow
819 If the outputs are resized very early on, in CreateScreenResources, then
820 we will not yet have created a RootWindow and so trying to change its
821 pixmap is a futile effort.
823 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
825 commit d7879a5939cabcd8b804e19fc422d2022ab7e3a4
826 Author: Chris Wilson <chris@chris-wilson.co.uk>
827 Date: Fri Sep 7 12:18:08 2012 +0100
829 sna/dri: Fix the double-buffer pageflipping path
831 Notably, we need to remember to exchange the front/back buffers after
834 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
836 commit 2630c81937115602faa352fca369d89fbb926b33
837 Author: Chris Wilson <chris@chris-wilson.co.uk>
838 Date: Fri Sep 7 11:23:02 2012 +0100
840 sna/dri: Make sure we set the SCANOUT flag when creating bo for pageflips
842 In case we should be creating an untiled framebuffer, we need to make
843 sure we honour the scanout alignment.
845 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
847 commit aeef03520f1b2a4b8638ff1ed4fbf19c9b35e7c8
848 Author: Chris Wilson <chris@chris-wilson.co.uk>
849 Date: Thu Sep 6 11:46:53 2012 +0100
851 sna/prime: Correct the pinning flag when importing the prime bo
853 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
855 commit 9bdf46d6c1a29b7b9e3d35e39a2e789334c0e1be
856 Author: Chris Wilson <chris@chris-wilson.co.uk>
857 Date: Thu Sep 6 11:38:08 2012 +0100
859 sna/prime: Distinguish forms of pinned pixmap
861 So that we can allow prime to replace the backing bo of currently
862 exported pixmaps through DRI2.
864 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
866 commit 39c4be2b0be66ed83bf059d3007f4143325e340d
867 Author: Chris Wilson <chris@chris-wilson.co.uk>
868 Date: Thu Sep 6 11:28:49 2012 +0100
870 uxa: Distinguish forms of pinning
872 This allows us to replace the prime bo to meet sharing requirements, but
873 still maintain the integrity with other users.
875 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
877 commit 031fa1c1c1129e486a02ffb5b3b9071f03b60048
878 Author: Chris Wilson <chris@chris-wilson.co.uk>
879 Date: Thu Sep 6 11:08:23 2012 +0100
881 uxa/prime: Mark the pixmap as pinned once exported (similar to flink)
883 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
885 commit e5d3f3818660cb875ce2ef4721c51bc95317b78d
886 Author: Chris Wilson <chris@chris-wilson.co.uk>
887 Date: Thu Sep 6 10:54:57 2012 +0100
889 sna: Add is-pinned checks to changing the pitch on an existing bo
891 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
893 commit f2affe403baea78b9c94e3d726d1b9d8a0004f35
894 Author: Chris Wilson <chris@chris-wilson.co.uk>
895 Date: Thu Sep 6 10:24:04 2012 +0100
897 sna: Apply the minimum 256 pitch to CREATE_USAGE_SHARED pixmaps as well
899 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
901 commit fbe4080816191ee97347128444f56bcc34858481
902 Author: Chris Wilson <chris@chris-wilson.co.uk>
903 Date: Thu Sep 6 08:57:34 2012 +0100
905 prime: Align pitch of shared untiled buffers to 256 bytes
907 In order for nvidia to handle the buffers we are currently generating,
908 they need to have a pitch alignment of 256 bytes. Make it so.
910 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
912 commit c5fb025267bf1019907dfcc60f0efef2691be436
913 Author: Chris Wilson <chris@chris-wilson.co.uk>
914 Date: Thu Sep 6 08:55:46 2012 +0100
916 sna: Correct assertions for adding damage that bypasses the region
918 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
920 commit 1f5d5a37e57e63fa6e5b336a4847ce4422b89044
921 Author: Chris Wilson <chris@chris-wilson.co.uk>
922 Date: Wed Sep 5 17:17:58 2012 +0100
924 Add PlatformProbe to handle sharing of device entities
926 Reported-by: Nick Bowler <nbowler@draconx.ca>
927 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54561
928 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
930 commit ddde40afc081f47a3b6b694aeb21682c240c9562
931 Author: Chris Wilson <chris@chris-wilson.co.uk>
932 Date: Wed Sep 5 16:48:01 2012 +0100
934 sna: Harden against initialisation failures
936 References: https://bugs.freedesktop.org/show_bug.cgi?id=54561
937 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
939 commit 4627dc36fa56ab971cd90088bd29a597dd12dc02
940 Author: Chris Wilson <chris@chris-wilson.co.uk>
941 Date: Wed Sep 5 16:17:17 2012 +0100
943 Revert "sna: Flush the batch before a render operation if the GPU is idle"
945 This reverts commit ad57ac07a273bf376b74884de47d8ee1e7129fb8.
947 These checks end up being too frequent and not allowing us to batch
948 sufficient commands to offset the overhead of batch submission. Hmm.
950 commit 886f3cef4101a674bb62656cf7f3046643cae4b1
951 Author: Chris Wilson <chris@chris-wilson.co.uk>
952 Date: Wed Sep 5 15:27:21 2012 +0100
954 uxa: Fix cut'n'paste error in Option "Backlight"
956 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
958 commit 6211d0914e2ed1cdd4cc2e170ee16129fc9f0d73
959 Author: Chris Wilson <chris@chris-wilson.co.uk>
960 Date: Wed Sep 5 12:33:08 2012 +0100
962 uxa: Flush batch immediately after queuing a WAIT_FOR_EVENT
964 We still encounter hangs with kernel-3.5 with the culprit being a wait
965 on a disabled pipe. As we thoroughly check before that the pipe is still
966 disabled and flush before a modeset, the only possibility that remains
967 is that DPMS is disabling the pipe before we submit. Close that race by
968 always submitting the batch immediately after a WAIT_FOR_EVENT.
970 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
972 commit 6a1423547769c24d15f6475628981d14a76a4bb2
973 Author: Chris Wilson <chris@chris-wilson.co.uk>
974 Date: Wed Sep 5 12:15:52 2012 +0100
976 sna: Initialise the empty boxes field of the region before subtracting it
978 Fixes regression from commit 38fb77af757318e5fb6f605b37306ce4585b11a5
979 Author: Chris Wilson <chris@chris-wilson.co.uk>
980 Date: Wed Sep 5 08:23:34 2012 +0100
982 sna: Don't upload ignored cpu damage
984 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
986 commit dff25e5ec4071a0404f82760e8deec3f99f4a0a9
987 Author: Chris Wilson <chris@chris-wilson.co.uk>
988 Date: Wed Sep 5 11:05:28 2012 +0100
990 sna: Drop master after discarding framebuffers
992 As Imre Deak pointed out in the previous patch, drmModeRmFB only works
993 when we hold the DRM master, therefore to prevent a leak of the
994 framebuffer across server reset we need to defer dropping master until
995 after we release our scanouts and modes.
997 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
999 commit 46981d01700c1159bfb6bc0aebc938ff1d447a0f
1000 Author: Imre Deak <imre.deak@intel.com>
1001 Date: Fri Aug 31 14:31:51 2012 +0300
1003 uxa: fix leakage of FB when closing the screen
1005 Calling drmModeRmFB is only allowed in DRM master mode. Since leaving
1006 the VT also drops master mode we need to remove the FB before calling
1009 This is only a real leak in case of a server reset, otherwise the server
1010 process will exit anyway and the kernel will clean up the FB.
1012 Signed-off-by: Imre Deak <imre.deak@intel.com>
1013 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1015 commit 55cef974a5dad3fda1922648fa27bcf5bb32ea03
1016 Author: Chris Wilson <chris@chris-wilson.co.uk>
1017 Date: Wed Sep 5 09:38:47 2012 +0100
1019 sna: Review validity of damage when discarding CPU bo
1021 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1023 commit 587499bbf55b7eb0e1848822a792d535a8a3db1b
1024 Author: Chris Wilson <chris@chris-wilson.co.uk>
1025 Date: Wed Sep 5 10:56:18 2012 +0100
1027 sna/video: Use the scanout flag and FB id for sprite framebuffers
1029 So that we can use the same teardown path as normal scanouts.
1031 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1033 commit 913adacc5400f94bc754f22375447e6f90ccd510
1034 Author: Chris Wilson <chris@chris-wilson.co.uk>
1035 Date: Wed Sep 5 09:20:55 2012 +0100
1037 sna/gen7: Always emit a stall when flushing the texture cache
1039 Reported-by: Reinhard Karcher <reinhard.karcher@gmx.net>
1040 References: https://bugs.freedesktop.org/show_bug.cgi?id=54488
1041 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1043 commit bdfedb46bd5bb3b96543a3ff553abd8ed0774fad
1044 Author: Chris Wilson <chris@chris-wilson.co.uk>
1045 Date: Wed Sep 5 09:15:12 2012 +0100
1047 sna: Use async upload only if the last render was not CPU
1049 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1051 commit 67b75ae6f66e2d8e97b0aaf70d66b6de30f9e67e
1052 Author: Chris Wilson <chris@chris-wilson.co.uk>
1053 Date: Wed Sep 5 08:46:39 2012 +0100
1055 sna: Tweak upload path to remove redundant code
1057 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1059 commit 38fb77af757318e5fb6f605b37306ce4585b11a5
1060 Author: Chris Wilson <chris@chris-wilson.co.uk>
1061 Date: Wed Sep 5 08:23:34 2012 +0100
1063 sna: Don't upload ignored cpu damage
1065 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1067 commit 2eb32f711ddd65a55c3c21f813b41aca49635a22
1068 Author: Chris Wilson <chris@chris-wilson.co.uk>
1069 Date: Wed Sep 5 00:22:04 2012 +0100
1071 sna: Add DBG control to disable CPU bo
1073 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1075 commit 19e170aa6fd3652cc2e983a291f82350dca27e52
1076 Author: Chris Wilson <chris@chris-wilson.co.uk>
1077 Date: Tue Sep 4 23:52:42 2012 +0100
1079 sna: Fix comparison of memcpy overlap to include x-offsets
1081 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1083 commit 7895416eabfc5bd361d0d0ff92c570e2f6af47e0
1084 Author: Chris Wilson <chris@chris-wilson.co.uk>
1085 Date: Tue Sep 4 20:30:53 2012 +0100
1087 sna: Rearrange use_cpu() tests for composite ops to avoid syncing CPU bo
1089 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1091 commit ad57ac07a273bf376b74884de47d8ee1e7129fb8
1092 Author: Chris Wilson <chris@chris-wilson.co.uk>
1093 Date: Tue Sep 4 20:06:49 2012 +0100
1095 sna: Flush the batch before a render operation if the GPU is idle
1097 Experiment with pushing those first commands earlier.
1099 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1101 commit 6325d7e29a8bacede321362c20bf3acd33c48651
1102 Author: Chris Wilson <chris@chris-wilson.co.uk>
1103 Date: Tue Sep 4 14:51:54 2012 +0100
1105 NEWS: fixup missing links to bugs referenced in 2.20.6 release notes
1107 Reported-by: Adam Jackson <ajax@redhat.com>
1108 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1110 commit bc309bad39035649ee377c1d5af9099491a8ea43
1111 Author: Chris Wilson <chris@chris-wilson.co.uk>
1112 Date: Tue Sep 4 14:22:15 2012 +0100
1114 sna: And remember to flush the damage to the output slaves...
1116 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1118 commit 68d207588a177afa4e999260bfddb4d6dba1029b
1119 Author: Chris Wilson <chris@chris-wilson.co.uk>
1120 Date: Tue Sep 4 13:20:21 2012 +0100
1122 sna: Improve handling of output offloading
1124 In particular, don't forget to flush when we only have offload slaves
1125 and no native pixmaps.
1127 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1129 commit 8cf7ac776b9b47dabd5ab141e5a5385c44d3f309
1130 Author: Chris Wilson <chris@chris-wilson.co.uk>
1131 Date: Tue Sep 4 13:55:34 2012 +0100
1133 sna: Fix a typo in an error message
1135 s/achieve/retrieve/ otherwise it is nonsense.
1137 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1139 commit be83b98d299c74101357e7bf170b1451aec2e1f2
1140 Author: Chris Wilson <chris@chris-wilson.co.uk>
1141 Date: Tue Sep 4 11:23:58 2012 +0100
1143 sna: Add some DBG to the prime paths
1145 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1147 commit d0684d55e3a62c231c92a68403cf100c9e18351b
1148 Author: Chris Wilson <chris@chris-wilson.co.uk>
1149 Date: Tue Sep 4 11:05:43 2012 +0100
1151 sna: Fixup CREATE_USAGE_SHARED
1153 The DRI2 code tries to create pixmaps with non-zero width/height,
1156 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1158 commit b6ced7a2267f9429ad38949dda39c843e2119e99
1159 Author: Chris Wilson <chris@chris-wilson.co.uk>
1160 Date: Tue Sep 4 10:05:46 2012 +0100
1162 sna: Fix typo for detecting prime ioctls
1164 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1166 commit df68723baae71498de95924c72d4f23fa7fc7fdf
1167 Author: Chris Wilson <chris@chris-wilson.co.uk>
1168 Date: Mon Sep 3 20:43:48 2012 +0100
1170 sna: Port prime interfacing
1172 Preliminary prime support.
1174 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1176 commit e3ad18036b4dca29744ecfa4694006f01dd3fd18
1177 Author: Chris Wilson <chris@chris-wilson.co.uk>
1178 Date: Mon Sep 3 20:53:30 2012 +0100
1180 sna: Enable platform probing
1182 Completing commit 0768ac4d195214825137152893deb74a87fcd11e
1183 Author: Dave Airlie <airlied@redhat.com>
1184 Date: Wed Jul 25 16:11:23 2012 +1000
1186 intel: add platform probing support.
1188 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1190 commit 3dc644b2a959fc559e1138b332ed42d7235de42f
1191 Author: Сковорода Никита Андреевич <chalkerx@gmail.com>
1192 Date: Mon Sep 3 13:44:21 2012 +0400
1194 uxa: Fix Backlight option support.
1196 Signed-off-by: Сковорода Никита Андреевич <chalkerx@gmail.com>
1198 commit a972affe0c78b23a119d5dd14bb4446c89122af1
1199 Author: Chris Wilson <chris@chris-wilson.co.uk>
1200 Date: Sun Sep 2 17:43:33 2012 +0100
1202 sna/gen6+: Redirect fills if the destination is too large for 3D
1204 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
1205 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54134
1206 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1208 commit 18d26076c778c20eb589b638fc47fa847793f149
1209 Author: Dave Airlie <airlied@gmail.com>
1210 Date: Tue Sep 4 07:19:12 2012 +1000
1212 intel: fix return value for no pEnt
1214 Probably never gets hit but shuold return FALSE,
1216 pointed out on irc by Lekensteyn
1218 Signed-off-by: Dave Airlie <airlied@redhat.com>
1220 commit d14ff42f2a205542df2ef723c6151d18db2bea8b
1221 Author: Dave Airlie <airlied@redhat.com>
1222 Date: Thu Jul 26 10:43:29 2012 +1000
1224 intel: query kernel for caps to setup scrn capabilities.
1226 This queries the kernel for prime support before advertising
1229 Signed-off-by: Dave Airlie <airlied@redhat.com>
1231 commit 6705d8237aca90964449e4dbee97b4f62b87c28b
1232 Author: Dave Airlie <airlied@redhat.com>
1233 Date: Thu Jul 26 10:37:04 2012 +1000
1235 intel: add pixmap tracking and scanout support. (v2)
1237 This adds support for pixmap tracking and scanout of
1240 v2: do dirty updates after uxa block handler, check if kernel
1241 can flush vmap for us so we don't have to.
1243 Signed-off-by: Dave Airlie <airlied@redhat.com>
1245 commit 69827126abdfa289417b55fe7db8ae0535037185
1246 Author: Dave Airlie <airlied@redhat.com>
1247 Date: Wed Jul 25 16:22:57 2012 +1000
1249 intel/uxa: add pixmap sharing support.
1251 Signed-off-by: Dave Airlie <airlied@redhat.com>
1253 commit 0768ac4d195214825137152893deb74a87fcd11e
1254 Author: Dave Airlie <airlied@redhat.com>
1255 Date: Wed Jul 25 16:11:23 2012 +1000
1257 intel: add platform probing support.
1259 This allows the driver to be loaded by the platform loading code.
1261 Signed-off-by: Dave Airlie <airlied@redhat.com>
1263 commit 164ae7a4f9fd3e538f1b506031d297088fc0c659
1264 Author: Chris Wilson <chris@chris-wilson.co.uk>
1265 Date: Sun Sep 2 17:19:12 2012 +0100
1267 man: Malformed "Backlight" section
1269 Reported-by: Matthew Monaco <dgbaley27@0x01b.net>
1270 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54397
1271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1273 commit 67b44104168650473a24e69056707e0b35f1713e
1274 Author: Chris Wilson <chris@chris-wilson.co.uk>
1275 Date: Sun Sep 2 12:47:36 2012 +0100
1277 sna: Cache the temporary upload buffer when copying boxes
1279 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1281 commit 4b558281e04e4d4febfc361632a90f8a45080c49
1282 Author: Chris Wilson <chris@chris-wilson.co.uk>
1283 Date: Sun Sep 2 12:23:59 2012 +0100
1285 sna: Don't retire when searching for a snoopable buffer without hw support
1287 If the hw/kernel doesn't support snoopable buffers, then it makes little
1288 sense to search for one, and force a retire in the certainty of not
1291 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1293 commit d933f3a7965c9aa70f70477be3bfe94d5ded948b
1294 Author: Chris Wilson <chris@chris-wilson.co.uk>
1295 Date: Sun Sep 2 10:29:49 2012 +0100
1299 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1301 commit 62e7e69d88f7e5c8b0f0931fe887ee0c3bddb43f
1302 Author: Chris Wilson <chris@chris-wilson.co.uk>
1303 Date: Sun Sep 2 09:55:07 2012 +0100
1305 sna: Tweak placement rules if we already have a CPU bo
1307 If we have a CPU bo, consider if it may be quicker to render to it then
1310 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1312 commit f837807cc257fbedd35fa6101dc0b4a57e5ad78d
1313 Author: Chris Wilson <chris@chris-wilson.co.uk>
1314 Date: Sat Sep 1 10:36:09 2012 +0100
1316 sna/dri: Remove busy-wait spin from vblank chaining of swapbuffers
1318 The issue being that, due to the delay, the chained swap would miss its
1319 intended vblank and so cause an unwanted reduction in frame throughput
1320 and increase output latency even further. Since both client and server
1321 have other rate-limiting processes in place, we can forgo the stall here
1322 and still keep the clients in check.
1324 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54274
1325 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1327 commit b56e8c5105c858452ca4eabf15b298fc06dfd3c8
1328 Author: Chris Wilson <chris@chris-wilson.co.uk>
1329 Date: Sat Sep 1 12:15:47 2012 +0100
1331 sna: Nullify gpu_bo after free in case final release is deferred in destroy
1333 As we may defer the actual release of the pixmap until after completion
1334 of the last shm operation, we need to make sure in that case we mark the
1335 GPU bo as released to prevent a use-after-free.
1337 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1339 commit c4d994014160d8c946af731196a908991c77d9f9
1340 Author: Chris Wilson <chris@chris-wilson.co.uk>
1341 Date: Sat Sep 1 10:44:24 2012 +0100
1343 sna/gen2+: Add the missing assertions in case the drawrect is invalid
1345 Only the later gen had these useful assertions, add them to the rest
1348 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1350 commit 25c1b478a957ceb0474e7121840b7eed1ba2e140
1351 Author: Chris Wilson <chris@chris-wilson.co.uk>
1352 Date: Sat Sep 1 09:49:34 2012 +0100
1354 sna: Discard cached CPU upload early when clearing with a solid fill
1356 Otherwise we end up considering the GPU bo as a real target, causing
1357 confusion and failed asserts.
1359 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1361 commit 530b1d1516595cf14c5112d8833b870cd50eca46
1362 Author: Chris Wilson <chris@chris-wilson.co.uk>
1363 Date: Thu Aug 30 22:21:57 2012 +0100
1365 sna/dri: Use common routine for testing busyness after flush
1367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1369 commit 00d8c776b3607dbdab32c1126f91a7a38b8065f6
1370 Author: Chris Wilson <chris@chris-wilson.co.uk>
1371 Date: Thu Aug 30 21:47:17 2012 +0100
1373 sna/dri: Hold a reference to the vsync blit
1375 Fixes regression from
1377 commit 96a921487ef00db03a12bec7b0821410d6b74c31
1378 Author: Chris Wilson <chris@chris-wilson.co.uk>
1379 Date: Mon Aug 27 21:50:32 2012 +0100
1381 sna: Track outstanding requests per-ring
1383 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1385 commit b2a6c74c2159c9968c19400d61a11f4773724b4a
1386 Author: Chris Wilson <chris@chris-wilson.co.uk>
1387 Date: Wed Aug 29 14:51:40 2012 +0100
1389 sna: Remove inconsistent assertion
1391 As we now may not prefer to use the GPU even if all-damaged and clear,
1392 asserting that if we choose to use the CPU if clear is now bogus.
1394 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1396 commit deaa1cac269be03f4ec44092f70349ff466d59de
1397 Author: Chris Wilson <chris@chris-wilson.co.uk>
1398 Date: Tue Aug 28 22:23:22 2012 +0100
1400 sna: Align active upload buffers to the next page for reuse
1402 If we write to the same page as it already active on the GPU then
1403 despite the invalidation performed at the beginning of each batch, we do
1404 not seem to correctly sample the new data.
1406 References: https://bugs.freedesktop.org/show_bug.cgi?id=51422
1407 References: https://bugs.freedesktop.org/show_bug.cgi?id=52299
1408 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1410 commit 9e0305a3928f253ded6c8e141a4dd42be3952eb9
1411 Author: Chris Wilson <chris@chris-wilson.co.uk>
1412 Date: Tue Aug 28 20:49:20 2012 +0100
1414 sna: Discard inplace flag if we create a CPU map for the upload buffer
1416 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1418 commit f3e122554e88da0351bfb9a7a722f0715553689b
1419 Author: Chris Wilson <chris@chris-wilson.co.uk>
1420 Date: Tue Aug 28 21:27:04 2012 +0100
1422 sna: Propagate the request to flush rather than directly submit the batch
1424 The subtly is that we need to reset the mode correctly after
1425 submitting the batch which was not handled by kgem_flush(). If we fail
1426 to set the appropriate mode then the next operation will be on a random
1427 ring, which can prove fatal with SandyBridge+.
1429 Reported-by: Reinis Danne <reinis.danne@gmail.com>
1430 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1432 commit b5c8efe4309248e62d94d80b37a70775284ae985
1433 Author: Chris Wilson <chris@chris-wilson.co.uk>
1434 Date: Tue Aug 28 08:33:00 2012 +0100
1436 sna: Make sure we reset the domain tracking when exporting DRI bo
1438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1440 commit 96a921487ef00db03a12bec7b0821410d6b74c31
1441 Author: Chris Wilson <chris@chris-wilson.co.uk>
1442 Date: Mon Aug 27 21:50:32 2012 +0100
1444 sna: Track outstanding requests per-ring
1446 In order to properly track when the GPU is idle, we need to account for
1447 the completion order that may differ on architectures like SandyBridge
1448 with multiple mostly independent rings.
1450 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
1451 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54127
1452 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1454 commit 26c731efc2048663b6a19a7ed7db0e94243ab30f
1455 Author: Chris Wilson <chris@chris-wilson.co.uk>
1456 Date: Mon Aug 27 20:50:08 2012 +0100
1458 sna: Ensure that we create a GTT mapping for the inplace upload buffer
1460 As the code will optimistically convert a request for a GTT mapping into
1461 a CPU mapping if the object is still in the CPU domain, we need to
1462 overrule that in this case where we explicitly want to write directly
1463 into the GTT and furthermore keep the buffer around in an upload cache.
1465 References: https://bugs.freedesktop.org/show_bug.cgi?id=51422
1466 References: https://bugs.freedesktop.org/show_bug.cgi?id=52299
1467 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1469 commit 2cbf88980ede50370b97f32e565dea33db16ac44
1470 Author: Chris Wilson <chris@chris-wilson.co.uk>
1471 Date: Mon Aug 27 20:07:17 2012 +0100
1473 sna: Force the stall if using a busy ShmPixmap bo with PutImage
1475 As we will stall in the near future to serialise access with the
1476 ShmPixmap, we may as well stall first and do a simple copy using the
1477 CPU in this highly unlikely scenario.
1479 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1481 commit 705103d77e6b80d796a4535cade96cb6e9ebece3
1482 Author: Chris Wilson <chris@chris-wilson.co.uk>
1483 Date: Mon Aug 27 20:07:19 2012 +0100
1485 sna: Add a modicum of DBG to kgem_is_idle()
1487 Print out the handle of the bo we just checked.
1489 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1491 commit 8218e5da2b177ca9cd0e2b1e7dbe114e5ef2ebf0
1492 Author: Chris Wilson <chris@chris-wilson.co.uk>
1493 Date: Mon Aug 27 19:36:03 2012 +0100
1495 sna: Fix crash with broken DBG missing one of its arguments
1497 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
1498 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54127
1499 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1501 commit 593f549b45fbb1528472feed51207d84901d142e
1502 Author: Chris Wilson <chris@chris-wilson.co.uk>
1503 Date: Mon Aug 27 16:43:19 2012 +0100
1505 sna: A little more DBG to try and identify ratelimiting ops
1507 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1509 commit 71ac12e9b6ed00c28993637aafd5186a2ba26256
1510 Author: Chris Wilson <chris@chris-wilson.co.uk>
1511 Date: Mon Aug 27 15:49:59 2012 +0100
1513 sna: Assert that the bo is marked as 'flush' when exported to DRI clients
1515 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1517 commit cf64c8ce758cfa5d3bcd1b7626ff94cce7a84636
1518 Author: Chris Wilson <chris@chris-wilson.co.uk>
1519 Date: Mon Aug 27 13:59:56 2012 +0100
1521 sna: Upload PutImage inplace to a fresh GPU bo if the device doesn't snoop
1523 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1525 commit 3c6758fc4a50ecfce9ed317fec669cc48addedcf
1526 Author: Chris Wilson <chris@chris-wilson.co.uk>
1527 Date: Mon Aug 27 11:58:05 2012 +0100
1529 sna: Flush the batch if it references a ShmPixmap and the GPU is idle
1531 This helps minimise the stall when syncing with the GPU before sending
1532 the next reply to the Client.
1534 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1536 commit f21079bad6e8316baf5d0295d6e7a809041bce06
1537 Author: Chris Wilson <chris@chris-wilson.co.uk>
1538 Date: Mon Aug 27 11:25:28 2012 +0100
1540 Revert "sna: Cleanup composite redirection after substituting the BLT"
1542 This reverts commit 5a5212117e7a73ce3fffb87c60a505a849e38c36.
1544 The clean up is in effect too early, as this is during preparation and
1545 the actual work is already being correctly done at the end.
1547 commit 414e87255cdee6eb556703ddefd194af71b985ed
1548 Author: Chris Wilson <chris@chris-wilson.co.uk>
1549 Date: Sun Aug 26 19:47:31 2012 +0100
1551 Add Option "Backlight" to override the probed backlight control interface
1553 The automatic selection may not correspond with the correct backlight
1554 (such as in a multi-gpu, multi-panel device) or the user may simply
1555 prefer another control interface. This allows them to override the
1556 chosen interface using
1558 Option "Backlight" "my-backlight"
1560 to specify '/sys/class/backlight/my-backlight' as the interface to use
1563 Suggested-by: Alon Levy <alevy@redhat.com>
1564 References: https://bugs.freedesktop.org/show_bug.cgi?id=29273
1565 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1567 commit d4f7c58186849374cd929e20fa49ea2e93939a69
1568 Author: Chris Wilson <chris@chris-wilson.co.uk>
1569 Date: Sun Aug 26 17:01:55 2012 +0100
1571 sna/gen5+: Use the common methods for choosing the render targets
1573 This should afford us much more flexibility in where we render.
1575 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1577 commit 3c88b5f693c29b990d69f96508b121ce97a7209e
1578 Author: Chris Wilson <chris@chris-wilson.co.uk>
1579 Date: Sat Aug 25 11:49:30 2012 +0100
1581 sna: Use a fast span emitter for mono trapezoids without damage or clipping
1583 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1585 commit 8e10a5b348a37feadcf935ec7694e46cc0802bdf
1586 Author: Chris Wilson <chris@chris-wilson.co.uk>
1587 Date: Sun Aug 26 14:53:12 2012 +0100
1589 sna/gen6+: Do not call sna_blt_composite() after prepping the composite op
1591 As sna_blt_composite() will overwrite parts of the composite op as it
1592 checks whether or not it can execute that operation, it will lead to a
1593 crash as the normal render path finds the op corrupt. (The BLT
1594 conversion functions cater for the cases where we may wish to switch
1595 pipelines after choosing src/dst bo.)
1597 Reported-by: rei4dan@gmail.com
1598 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1600 commit cbbe7727e766a5ee8767673feb6c8cdec38a7051
1601 Author: Chris Wilson <chris@chris-wilson.co.uk>
1602 Date: Sun Aug 26 14:43:47 2012 +0100
1604 sna/gen6+: Simplify prefer_blt_bo
1606 As we already check the tiling state, so all we need to then check is
1607 that the pitch is within the BLT constraint.
1609 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1611 commit 0c15824a8143a288716d2eacf03252cc54eb9466
1612 Author: Chris Wilson <chris@chris-wilson.co.uk>
1613 Date: Sun Aug 26 14:29:41 2012 +0100
1615 sna: Add some DBG to kgem_is_idle()
1617 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1619 commit d432983421286d343f7c487c12c7244b711f5a66
1620 Author: Chris Wilson <chris@chris-wilson.co.uk>
1621 Date: Sun Aug 26 13:43:35 2012 +0100
1623 sna: Add some DBG to BLT composite substitute to show if redirection is used
1625 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1627 commit 5a5212117e7a73ce3fffb87c60a505a849e38c36
1628 Author: Chris Wilson <chris@chris-wilson.co.uk>
1629 Date: Sun Aug 26 13:38:23 2012 +0100
1631 sna: Cleanup composite redirection after substituting the BLT
1633 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1635 commit 335821d588460c253b2ba2c8616a7c46e5ad0150
1636 Author: Chris Wilson <chris@chris-wilson.co.uk>
1637 Date: Sun Aug 26 13:29:04 2012 +0100
1639 sna/gen4+: Check for allocation failure for the clear solid bo
1641 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1643 commit b5c77a6aaf520e331e82409b2592911cb1bb3100
1644 Author: Chris Wilson <chris@chris-wilson.co.uk>
1645 Date: Sun Aug 26 12:48:18 2012 +0100
1649 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1651 commit 454cc8453af1852758c3396dbe303c13c5c1be27
1652 Author: Chris Wilson <chris@chris-wilson.co.uk>
1653 Date: Fri Aug 24 08:48:12 2012 +0100
1655 sna: Submit the partial batch before throttling
1657 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1659 commit 0e1e83ed4952f620e9422e58f955a5aea406e300
1660 Author: Chris Wilson <chris@chris-wilson.co.uk>
1661 Date: Fri Aug 24 00:59:31 2012 +0100
1663 sna: Allow the batch to be flushed if the GPU is idle upon a context switch
1665 Submit early, submit often in order to keep the GPU busy. As always we
1666 trade off CPU overhead versus concurrency.
1668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1670 commit 5059db0697c5516f1538f7062937664baf7b1c2e
1671 Author: Chris Wilson <chris@chris-wilson.co.uk>
1672 Date: Fri Aug 24 00:21:07 2012 +0100
1674 sna: Correct a pair of DBG messages
1676 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1678 commit c5b46e411a2c738c5ae55bffb9b3d460249f5c24
1679 Author: Chris Wilson <chris@chris-wilson.co.uk>
1680 Date: Thu Aug 23 17:36:10 2012 +0100
1682 sna: Tidy up users of __kgem_bo_is_busy()
1684 A lot of callsites had not been converted to the common function.
1686 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1688 commit 82e91327d57e03d2117638165f298a50b946fcaa
1689 Author: Chris Wilson <chris@chris-wilson.co.uk>
1690 Date: Thu Aug 23 15:59:00 2012 +0100
1692 sna: Use a temporary userptr mapping for a large upload into a busy target
1694 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1696 commit b286ffa6beccb8fe341c464a4fb9f2af98541263
1697 Author: Eric S. Raymond <esr@thyrsus.com>
1698 Date: Thu Aug 23 12:42:18 2012 -0400
1700 Fix seriously malformed list syntax on intel(4).
1702 Signed-off-by: Eric S. Raymond <esr@thyrsus.com>
1704 commit fc6b7f564df88ca773ae245b1b4e278b47dffd59
1705 Author: Chris Wilson <chris@chris-wilson.co.uk>
1706 Date: Thu Aug 23 15:13:14 2012 +0100
1708 sna: Flush the batch if it contains any DRI pixmaps
1710 This fixes a regression from
1712 commit 02963f489b177d0085006753e91e240545933387
1713 Author: Chris Wilson <chris@chris-wilson.co.uk>
1714 Date: Sun Aug 19 15:45:35 2012 +0100
1716 sna: Only submit the batch if flushing a DRI client bo
1718 which made the presumption that we called sna_add_flush_pixmap() for
1719 every DRI pixmap that we used. However, that is only called for the
1720 dirty pixmaps, any native exported pixmap only marks the batch as
1721 requiring a flush. So in those cases we always need to submit the batch
1722 if it contains an exported DRI pixmap.
1724 Reported-by: chr.ohm@gmx.net
1725 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53967
1726 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1728 commit fd38f45ec5421802b426867a050c978a2feec937
1729 Author: Chris Wilson <chris@chris-wilson.co.uk>
1730 Date: Thu Aug 23 12:11:48 2012 +0100
1732 sna: Mark the CPU damage as needing flushing for DRI buffers
1734 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1736 commit ac6cb667546a82b865c959a1be32f52b2da7bf7d
1737 Author: Chris Wilson <chris@chris-wilson.co.uk>
1738 Date: Wed Aug 22 12:00:49 2012 +0100
1740 sna: Flush before adding any SHM pixmap into the batch
1742 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1744 commit 785c1046b94fd9ca6f22b2a6d73639408a411cea
1745 Author: Chris Wilson <chris@chris-wilson.co.uk>
1746 Date: Wed Aug 22 10:56:12 2012 +0100
1748 sna: Only use the GPU for an active CPU bo unless forced
1750 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1752 commit 8032f51859ce1928922edf6892f493a84d9c39f0
1753 Author: Chris Wilson <chris@chris-wilson.co.uk>
1754 Date: Wed Aug 22 10:40:53 2012 +0100
1756 sna: If we cannot use the CPU bo along a render pathway, promote to GPU
1758 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1760 commit cc967507bbf357e1d5ec3cd0f3c0c5ecfa8b867a
1761 Author: Chris Wilson <chris@chris-wilson.co.uk>
1762 Date: Wed Aug 22 10:36:04 2012 +0100
1764 sna: Convert to using IGNORE_CPU flag rather than complicating the CPU damage
1766 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1768 commit 5aa59ce7c012eb309c4f5a362ccf531c065bd7ff
1769 Author: Chris Wilson <chris@chris-wilson.co.uk>
1770 Date: Wed Aug 22 10:10:33 2012 +0100
1772 sna: Assert that the CPU bo is not used if the GPU is clear
1774 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1776 commit 273765033223024ff6a662195e0e4b96f8960463
1777 Author: Chris Wilson <chris@chris-wilson.co.uk>
1778 Date: Wed Aug 22 00:48:08 2012 +0100
1780 sna: Make sure the opposite damage is destroyed after reducing to all
1782 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1784 commit 31e398c9dbd8539e6fd2cc7398e97ee1df2f7f23
1785 Author: Chris Wilson <chris@chris-wilson.co.uk>
1786 Date: Wed Aug 22 00:38:17 2012 +0100
1788 sna: Discard a no-longer-used GPU bo after moving to the CPU domain
1790 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1792 commit 8be15c37df9d9b34e0f21700673212bd19c772b2
1793 Author: Chris Wilson <chris@chris-wilson.co.uk>
1794 Date: Wed Aug 22 00:27:00 2012 +0100
1796 sna: Balance CPU bo accounting for SHM pixmaps
1798 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1800 commit 262d57a5aaac46508d8e29860c7a567bcd5bc4d8
1801 Author: Chris Wilson <chris@chris-wilson.co.uk>
1802 Date: Tue Aug 21 23:36:12 2012 +0100
1804 sna: Display still resident memory in inactive/snoop caches under DEBUG_MEMORY
1806 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1808 commit 304581cc02adeb865a24edde934b5b9ceba68d96
1809 Author: Chris Wilson <chris@chris-wilson.co.uk>
1810 Date: Tue Aug 21 23:29:12 2012 +0100
1812 sna: Add a DBG to log pixmap destruction
1814 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1816 commit 946d54238d90f5e31772f0df336ac9a5e7f2d62b
1817 Author: Chris Wilson <chris@chris-wilson.co.uk>
1818 Date: Tue Aug 21 21:02:00 2012 +0100
1820 sna: Fix the assertion for tracking proxies in the batch
1822 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1824 commit 4ee2e227ddf61c87f08f55d4922d2562b563ca87
1825 Author: Chris Wilson <chris@chris-wilson.co.uk>
1826 Date: Tue Aug 21 00:28:40 2012 +0100
1828 sna: Mark all levels of a proxy as dirty
1830 So that if we write to a surface through one view, we make sure that the
1831 sample cache is invalidated for all future views.
1833 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1835 commit eee181e9d71273d94fe735805ed84e8f7b7b2180
1836 Author: Chris Wilson <chris@chris-wilson.co.uk>
1837 Date: Tue Aug 21 00:19:10 2012 +0100
1839 sna/gen6+: Only mark the dst as dirty again if it already is in the batch
1841 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1843 commit c86df17c1455a53cb52f33a25c8c362e5331621e
1844 Author: Chris Wilson <chris@chris-wilson.co.uk>
1845 Date: Mon Aug 20 22:54:06 2012 +0100
1847 sna/gen3: Fix assertion to check the freshly allocated vertex bo
1849 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1851 commit 6aabe90587f4916a01a1cd2bbc577a1e7fa20eca
1852 Author: Chris Wilson <chris@chris-wilson.co.uk>
1853 Date: Mon Aug 20 22:09:54 2012 +0100
1855 sna: Allow target bo promotion to GPU even on old architectures
1857 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1859 commit 1a4b6fea7b1516de35e6800efa5b85f8401a5b2a
1860 Author: Chris Wilson <chris@chris-wilson.co.uk>
1861 Date: Mon Aug 20 22:08:23 2012 +0100
1863 sna: Assign a unique id to snoopable CPU bo
1865 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1867 commit 892b1a1e431e8f27133825f8a27dde4955da0054
1868 Author: Chris Wilson <chris@chris-wilson.co.uk>
1869 Date: Mon Aug 20 22:07:05 2012 +0100
1871 sna/gen3: Convert to sna_drawable_use_bo()
1873 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1875 commit 3ca1bfb51ba522454433d58131e7dab7fcbe7e34
1876 Author: Chris Wilson <chris@chris-wilson.co.uk>
1877 Date: Mon Aug 20 21:28:47 2012 +0100
1879 sna: Trim a parameter from kgem_bo_mark_dirty() and add some assertions
1881 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1883 commit 16f3d3a9ae145a3af51d2c0c42c6c585d676a863
1884 Author: Chris Wilson <chris@chris-wilson.co.uk>
1885 Date: Mon Aug 20 19:42:22 2012 +0100
1887 sna: Keep a stash of the most recently allocated requests
1889 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1891 commit fb349ced91e15ecaa025321bd37d1fe3cfdd2f44
1892 Author: Chris Wilson <chris@chris-wilson.co.uk>
1893 Date: Mon Aug 20 17:35:44 2012 +0100
1895 sna: A few more buffer cache management assertions
1897 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1899 commit ae6b3c7508d570af94f172d6a91d62d2987d654b
1900 Author: Chris Wilson <chris@chris-wilson.co.uk>
1901 Date: Mon Aug 20 17:03:16 2012 +0100
1903 Check that the module that indeed i915 before using custom ioctls
1905 Thanks to Adam Jackson for pointing me towards drmGetVersion() and
1906 Julien Cristau for saying "Yuck!"
1908 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1910 commit 3c611087e8ae09fc3fe4271f16d912ac8ae89f6b
1911 Author: Chris Wilson <chris@chris-wilson.co.uk>
1912 Date: Mon Aug 20 16:09:25 2012 +0100
1914 Only open the matching BusID and not the first named
1916 If you pass a name to drmOpen() it will attempt to open any device
1917 corresponding to that name if it first fails to open the device
1918 corresponding to the BusID. Obviously we want the failure from opening
1919 the specified device in order to prevent wrongly opening the first found
1920 device multiple times.
1922 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1924 commit dd6c67b32f726b6ad7f12f3b83f6d8c868ff4dc1
1925 Author: Chris Wilson <chris@chris-wilson.co.uk>
1926 Date: Mon Aug 20 15:58:03 2012 +0100
1928 sna: Add a couple of buffer cache management assertions
1930 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1932 commit 01ebdb4d7a8bb751167153554f9122d996e4ea91
1933 Author: Chris Wilson <chris@chris-wilson.co.uk>
1934 Date: Mon Aug 20 15:33:26 2012 +0100
1936 sna: Remove confusing is_cpu()
1938 The only real user now has its own heuristics, so convert the remaining
1939 users over to !is_gpu().
1941 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1943 commit 410316d20299b9ed3447d1d897f904af786ed097
1944 Author: Chris Wilson <chris@chris-wilson.co.uk>
1945 Date: Mon Aug 20 14:56:12 2012 +0100
1947 sna: Refine decision making for maybe-inplace trapezoids
1949 In particular, we want to avoid preferentially taking the CPU paths
1950 when it may force any migration (including clear).
1952 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1954 commit 91f1bf971f9cdc6498f513a5ddec1ad7a4e24b3d
1955 Author: Chris Wilson <chris@chris-wilson.co.uk>
1956 Date: Mon Aug 20 15:27:27 2012 +0100
1958 sna: Correct ordering of calls to memcpy for BLT cpu composite paths
1960 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1962 commit 027569bf83fc6d67dca7cfd65fdfa37ef6b47204
1963 Author: Chris Wilson <chris@chris-wilson.co.uk>
1964 Date: Mon Aug 20 12:55:20 2012 +0100
1966 Missing includes for b5b76ad849b
1968 The warnings of implicit function declarations were lost amongst the
1971 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1973 commit f1b259f627814c765ea93d8c839ee7533249974f
1974 Author: Chris Wilson <chris@chris-wilson.co.uk>
1975 Date: Mon Aug 20 12:51:51 2012 +0100
1977 sna: Add a modicum of DBG for PolyFillRect
1979 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1981 commit b5b76ad849bfda1e75192d1cb3c6c0fcc623bb91
1982 Author: Chris Wilson <chris@chris-wilson.co.uk>
1983 Date: Mon Aug 20 12:01:39 2012 +0100
1985 Sanity check that the driver is an i915.ko GEM device before claiming it
1987 This fixes an issue with us claiming Poulsbo and friends even though we
1988 do not speak their language.
1990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
1992 commit 74f998136bac441d782faf779616af6f7bdcadad
1993 Author: Chris Wilson <chris@chris-wilson.co.uk>
1994 Date: Mon Aug 20 11:07:55 2012 +0100
1996 sna: Avoid migrating the BLT composite src to the GPU if the dst is not
1998 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2000 commit a05643eb670e91ab102465df799301def88faaf9
2001 Author: Chris Wilson <chris@chris-wilson.co.uk>
2002 Date: Mon Aug 20 10:46:14 2012 +0100
2004 sna: Remove unneeded source bo unref after __sna_render_pixmap_bo()
2006 As __sna_render_pixmap_bo() deliberately does not reference its returned
2007 bo, we need to avoid unreferencing it else we cause explosions later.
2009 Fixes regression from commit a13781d19defc97af6a279c11a85e33ef825020e
2010 Author: Chris Wilson <chris@chris-wilson.co.uk>
2011 Date: Sun Aug 19 09:45:12 2012 +0100
2013 sna: Enable BLT composite functions to target CPU buffers
2015 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2017 commit c80db9bc2ef0a37a4abb78c9ef667c8b36ab6fba
2018 Author: Chris Wilson <chris@chris-wilson.co.uk>
2019 Date: Mon Aug 20 10:45:36 2012 +0100
2021 sna: Flush the batch before preparing for a FlushCallback
2023 As we end up submitting and maybe synchronising upon a batch within
2024 FlushCallback (or our client will) it is important that we start that
2025 serialized batch as early as possible to minimise the impending stalls.
2027 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2029 commit c39fe0253847f5a86e16b47ba420c8ba819c9110
2030 Author: Chris Wilson <chris@chris-wilson.co.uk>
2031 Date: Sun Aug 19 20:42:55 2012 +0100
2033 sna: Do not use the GPU to migrate to the CPU whilst wedged!
2035 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2037 commit fe05268d70088c8cad5f4b5ef756e1ffe2069fca
2038 Author: Chris Wilson <chris@chris-wilson.co.uk>
2039 Date: Sun Aug 19 19:28:42 2012 +0100
2041 sna: Experiment with flushing the batch prior to rendering to a ShmPixmap
2043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2045 commit bbd7a825810cc9772e6d613df449cb5ecb0be3f6
2046 Author: Chris Wilson <chris@chris-wilson.co.uk>
2047 Date: Sun Aug 19 19:28:20 2012 +0100
2049 sna: Don't promote a ShmPixmap to GPU for a CopyArea
2051 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2053 commit 7bf7a5ad1057f1aeb5b261da6dc501323c022287
2054 Author: Chris Wilson <chris@chris-wilson.co.uk>
2055 Date: Sun Aug 19 19:17:35 2012 +0100
2057 sna/gen3: Tidy vbo discard
2059 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2061 commit d1b808fd72b477bde96f7c6737a993bd1a20baf2
2062 Author: Chris Wilson <chris@chris-wilson.co.uk>
2063 Date: Sun Aug 19 17:53:24 2012 +0100
2065 sna: Tweak is_cpu/is_gpu heuristics
2067 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2069 commit c682c1a37692021cbd9bfc1e3f1ccf2b648c73f9
2070 Author: Chris Wilson <chris@chris-wilson.co.uk>
2071 Date: Sun Aug 19 17:28:25 2012 +0100
2073 sna: Discard GPU (and damage) after applying clear on migration to CPU
2075 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2077 commit dc83ef49f9e8ff94ab77aa9d54fe29e32b9d7a3d
2078 Author: Chris Wilson <chris@chris-wilson.co.uk>
2079 Date: Sun Aug 19 17:20:45 2012 +0100
2081 sna/trapezoids: Accept more operators for maybe-inplace
2083 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2085 commit 47d948f5df0cd0e975cfe5183b6ce79cd3fd27ee
2086 Author: Chris Wilson <chris@chris-wilson.co.uk>
2087 Date: Sun Aug 19 17:09:21 2012 +0100
2089 sna: Update maybe_inplace to recognise more types of handled pixel formats
2091 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2093 commit 02963f489b177d0085006753e91e240545933387
2094 Author: Chris Wilson <chris@chris-wilson.co.uk>
2095 Date: Sun Aug 19 15:45:35 2012 +0100
2097 sna: Only submit the batch if flushing a DRI client bo
2099 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2101 commit 13d1a105159222518800d3c5ad5660725864ec6b
2102 Author: Chris Wilson <chris@chris-wilson.co.uk>
2103 Date: Sun Aug 19 09:48:05 2012 +0100
2105 sna: compare the correct trailing dword when skipping identical bitmap lines
2107 Fixes regression in 2.20.4 from
2109 commit 85192f00e345830541e3715e211b1f98154bbef4
2110 Author: Chris Wilson <chris@chris-wilson.co.uk>
2111 Date: Wed Aug 8 12:11:50 2012 +0100
2113 sna: Ignore trailing bits when comparing lines inside the bitmap
2115 Reported-by: Edward Sheldrake <ejsheldrake@gmail.com>
2116 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53699
2117 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2119 commit a13781d19defc97af6a279c11a85e33ef825020e
2120 Author: Chris Wilson <chris@chris-wilson.co.uk>
2121 Date: Sun Aug 19 09:45:12 2012 +0100
2123 sna: Enable BLT composite functions to target CPU buffers
2125 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2127 commit be940856c74fbedd27997dd61e2a85959b321193
2128 Author: Chris Wilson <chris@chris-wilson.co.uk>
2129 Date: Sat Aug 18 21:23:03 2012 +0100
2131 sna: Consider sample wraparound in each direction independently
2133 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2135 commit 110c7ef7f6c31929affa038918e6ce087bccddc6
2136 Author: Chris Wilson <chris@chris-wilson.co.uk>
2137 Date: Sat Aug 18 17:42:51 2012 +0100
2139 sna/damage: Replace the damage with a larger box if subsumed
2141 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2143 commit 75a2fab766d8aed180ef795919e503db22c0e0fd
2144 Author: Chris Wilson <chris@chris-wilson.co.uk>
2145 Date: Sat Aug 18 17:28:08 2012 +0100
2147 sna: Avoid forcing an upload for an unblittable bo unless on a fallback path
2149 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2151 commit 110d5519f3523b1e2c50db637cdc4c5bc44c960a
2152 Author: Chris Wilson <chris@chris-wilson.co.uk>
2153 Date: Sat Aug 18 16:39:52 2012 +0100
2155 sna: Reduce subtracted damage earlier
2157 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2159 commit 8812e8b6e89e6432a6a768a0566ce4c153e9b256
2160 Author: Chris Wilson <chris@chris-wilson.co.uk>
2161 Date: Sat Aug 18 16:21:13 2012 +0100
2163 sna: Reduce damage after a large composite operation
2165 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2167 commit e361627b90ea6bf2f9a8c46cf8debe562fdf4f09
2168 Author: Chris Wilson <chris@chris-wilson.co.uk>
2169 Date: Sat Aug 18 16:16:17 2012 +0100
2171 sna/damage: Add some more sanity checks for creating empty regions
2173 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2175 commit 10f334872e9dd190e18c768219e60815acabe4d3
2176 Author: Chris Wilson <chris@chris-wilson.co.uk>
2177 Date: Sat Aug 18 15:49:02 2012 +0100
2179 sna: Add damage for the whole unaligned trapezoid not per component
2181 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2183 commit 5ff6198c9346d84717bac28980329b048f4406e8
2184 Author: Chris Wilson <chris@chris-wilson.co.uk>
2185 Date: Sat Aug 18 11:59:56 2012 +0100
2189 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2191 commit bc77a466531fdbdd21e9354af567e5215a66edf9
2192 Author: Chris Wilson <chris@chris-wilson.co.uk>
2193 Date: Sat Aug 18 11:09:23 2012 +0100
2195 sna: Silence a couple of recent compiler warnings
2197 sna_accel.c: In function 'sna_poly_fill_rect_extents':
2198 sna_accel.c:11438:11: warning: unused variable 'v' [-Wunused-variable]
2200 sna_blt.c: In function 'sna_blt_composite__convert':
2201 sna_blt.c:1712:3: warning: format '%s' expects a matching 'char *' argument [-Wformat]
2202 sna_blt.c:1738:3: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'CARD32' [-Wformat]
2204 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2206 commit d1da2c71461798bb027a260ff8480b71ebffde26
2207 Author: Chris Wilson <chris@chris-wilson.co.uk>
2208 Date: Thu Aug 16 21:15:54 2012 +0100
2210 sna: Correct assertion for StoreColors
2212 Fixing the assertion reveals that it was bogus!
2214 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2216 commit 31ddaa0cfbde083aef42d9fdb5ed13bb4edc36c8
2217 Author: Chris Wilson <chris@chris-wilson.co.uk>
2218 Date: Thu Aug 16 21:13:52 2012 +0100
2220 sna: Fixup DBG after 7f5a9e3
2222 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2224 commit b8bc56c12c6fdc559354fbbf96b6e043a491aa8f
2225 Author: Chris Wilson <chris@chris-wilson.co.uk>
2226 Date: Thu Aug 16 21:00:57 2012 +0100
2228 uxa/gen3: Simply simple fill shader generation
2230 Suggested-by: Zdenek Kabelac <zkabelac@redhat.com>
2231 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2233 commit 182581edde356a21a0681de526f83884e1a0de5d
2234 Author: Chris Wilson <chris@chris-wilson.co.uk>
2235 Date: Thu Aug 16 20:59:09 2012 +0100
2237 uxa/glamor: Check for failed prepares
2239 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2240 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2242 commit c289b607d1526d8d3625fa84c093552a9f3ea168
2243 Author: Chris Wilson <chris@chris-wilson.co.uk>
2244 Date: Thu Aug 16 20:54:40 2012 +0100
2246 uxa: Check for failed mmapping of the scanout
2248 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2249 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2251 commit 32ec8b979bc8cf7d8ce351ca752b806d42bc1c0f
2252 Author: Chris Wilson <chris@chris-wilson.co.uk>
2253 Date: Thu Aug 16 20:53:08 2012 +0100
2255 test/dri2: Discard error returns from _XReply
2257 This is only test code, so keep the static analyser quiet
2259 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2260 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2262 commit 5675c36e0dca92b8d24235d82cd0e87f82fdc7b0
2263 Author: Chris Wilson <chris@chris-wilson.co.uk>
2264 Date: Thu Aug 16 20:44:32 2012 +0100
2266 sna: Check for failure to pin the front buffer
2268 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2269 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2271 commit c9db2c74ab3b482489bec63b11df80bb35e21aca
2272 Author: Chris Wilson <chris@chris-wilson.co.uk>
2273 Date: Thu Aug 16 20:42:50 2012 +0100
2275 sna/gen2: Add break rather than fallthrough
2277 The fall-through in this instance is irrelevant, and the static
2278 analysers complain for not commenting on the fall-through. Silence the
2279 analyser by removing the fall-through.
2281 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2282 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2284 commit e0abbc3ccafb51d6b2114e1c7d6dc2eda3a8c1f4
2285 Author: Chris Wilson <chris@chris-wilson.co.uk>
2286 Date: Thu Aug 16 20:40:17 2012 +0100
2288 i810: Query PortPriv for the surface properties
2290 Avoid the NULL deferences.
2292 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2293 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2295 commit c09518f98d5c533860b084506d9ee2c819cca52f
2296 Author: Chris Wilson <chris@chris-wilson.co.uk>
2297 Date: Thu Aug 16 20:32:53 2012 +0100
2299 sna: Fix assignment inside assertion (should be an equality check)
2301 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2302 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2304 commit 7d3e4328d4f4421178377933ca98fb80daa85c76
2305 Author: Chris Wilson <chris@chris-wilson.co.uk>
2306 Date: Thu Aug 16 20:32:11 2012 +0100
2308 sna/glyphs: Fix potential leak of glyph extents (for many lists of many formats)
2310 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
2311 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2313 commit 9b00c82df78cb8820c0e32fbecc7c933eed81d52
2314 Author: Chris Wilson <chris@chris-wilson.co.uk>
2315 Date: Thu Aug 16 20:30:52 2012 +0100
2317 sna: Remember alpha-fixup for BLT upload
2319 Still dubious whether this is truly beneficial...
2321 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2323 commit 94871944a0e1351273d6029df7bf0300f31a8571
2324 Author: Chris Wilson <chris@chris-wilson.co.uk>
2325 Date: Thu Aug 16 00:07:58 2012 +0100
2327 sna: Avoid conflating tearfree and randr shadow modes
2329 For ordinary TearFree we do not want to adjust the crtc offsets.
2331 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2333 commit 1a389842347bab9d91240444e161589071eb9a47
2334 Author: Chris Wilson <chris@chris-wilson.co.uk>
2335 Date: Wed Aug 15 23:07:16 2012 +0100
2337 sna: Avoid queuing a pageflip on a DPMS off pipe
2339 If the pipe is not running, attempting to queue a pageflip will result
2340 in an error and us disabling the output in retaliation.
2342 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2344 commit 615739556dd1cc4565eb1c47f93fe8abd697802f
2345 Author: Chris Wilson <chris@chris-wilson.co.uk>
2346 Date: Wed Aug 15 20:10:52 2012 +0100
2348 sna: Only continue to update the shadow buffer if the flips succeeded
2350 If the flip fail, we disable the crtc and may end up disabling the
2351 output, removing the shadow (and old scanout bo) in the process.
2353 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2355 commit 2554d0d76e0fcae6e324938c28bb50deeb8814dc
2356 Author: Chris Wilson <chris@chris-wilson.co.uk>
2357 Date: Wed Aug 15 18:04:18 2012 +0100
2359 sna: Submit the batch upon destroying a ShmPixmap
2361 The midlayer has already detached the pixmap from the segment (possibly
2362 destroying that segment in the process, thanks midlayer!) so we need to
2363 submit the batch asap before the segment disappears.
2365 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2367 commit 4e604d721be5a1c0f8eaf91e3a5cf8c01a609a69
2368 Author: Chris Wilson <chris@chris-wilson.co.uk>
2369 Date: Wed Aug 15 16:53:20 2012 +0100
2371 sna: Flush shm pixmaps after upload
2373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2375 commit 7f5a9e30c953c174a4d6bc83e0b0a5e068cbda52
2376 Author: Chris Wilson <chris@chris-wilson.co.uk>
2377 Date: Wed Aug 15 09:34:24 2012 +0100
2379 sna: Remove zero-size rectangles fro PolyFillRectangle
2381 Some applications like to waste precious bw by attempting to fill
2382 zero-sized rectangles; remove them upfront rather than process them.
2384 Reported-by: Jiri Slaby <jirislaby@gmail.com>
2385 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
2386 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2388 commit 0c882f4fa991ef379e416628eb41cd2178818ad1
2389 Author: Chris Wilson <chris@chris-wilson.co.uk>
2390 Date: Tue Aug 14 16:02:38 2012 +0100
2392 sna/gen2,3: Prevent dereferencing a NULL bo with solid render fills
2394 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2396 commit 48f436d3705059711097a222aac65f862b4a5cfd
2397 Author: Chris Wilson <chris@chris-wilson.co.uk>
2398 Date: Tue Aug 14 15:05:33 2012 +0100
2400 sna: Defer the release of a busy SHM pixmap until the next flush callback
2402 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2404 commit 9f07d1fc0181f22da36837337241f228626e0d88
2405 Author: Chris Wilson <chris@chris-wilson.co.uk>
2406 Date: Tue Aug 14 12:03:24 2012 +0100
2408 sna: Enable use of shm pixmap for readback
2410 A little too overzealous in the degradation of writing to the CPU bo
2411 if we need to read from a GPU bo.
2413 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2415 commit 9b016d2e466f342cc6649504fa48ab6a810c7e94
2416 Author: Chris Wilson <chris@chris-wilson.co.uk>
2417 Date: Sun Aug 12 21:36:30 2012 +0100
2419 sna: do not bother attaching to a miniscule ShmPixmap
2421 If the pixmap is less than a page in size we are unlikely to recover the
2422 cost of tracking it via the GPU.
2424 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2426 commit 2f4de90709264ad19a3e3f5f0f79f4bba78a760a
2427 Author: Chris Wilson <chris@chris-wilson.co.uk>
2428 Date: Sun Aug 12 13:01:24 2012 +0100
2430 uxa: Only emit a warning if we fail to tile the frontbuffer and the hw allows
2432 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2434 commit 4d3cafc99ab49da6744046bc2004981144cb064d
2435 Author: Chris Wilson <chris@chris-wilson.co.uk>
2436 Date: Sun Aug 12 13:02:43 2012 +0100
2438 sna: Fix compilation without pixman glyphs
2440 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2442 commit dc18eaa585c36c8e5f5b4ec405a976a835fd2ac3
2443 Author: Chris Wilson <chris@chris-wilson.co.uk>
2444 Date: Sun Aug 12 10:34:10 2012 +0100
2446 sna: Make the failure to create render caches non-fatal
2448 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2450 commit b580abdfa68108f3e63ee1f897b6ea83b9c47935
2451 Author: Dave Airlie <airlied@gmail.com>
2452 Date: Sun Aug 12 20:52:12 2012 +1000
2454 intel: fix video xvPipe range check
2456 Pointed out by Russell King on irc.
2458 Signed-off-by: Dave Airlie <airlied@redhat.com>
2460 commit 44f848f9b2f2a2dcd9087210ea46bc4fdb63c057
2461 Author: Chris Wilson <chris@chris-wilson.co.uk>
2462 Date: Sat Aug 11 19:44:15 2012 +0100
2464 sna: Fix typo in computation of texel offsets for tiled 8x8 blts
2466 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
2467 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53353
2468 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2470 commit b335d369bbcbfaef878795a5c8e72c2d40d3a536
2471 Author: Chris Wilson <chris@chris-wilson.co.uk>
2472 Date: Sat Aug 11 12:59:49 2012 +0100
2474 sna: Discard the GPU bo after syncing to CPU bo for writing
2476 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2478 commit 029934662e0bad6cf47baf4d7202656dd3e0ba08
2479 Author: Chris Wilson <chris@chris-wilson.co.uk>
2480 Date: Sat Aug 11 10:47:44 2012 +0100
2482 sna: Add a little DBG for promotion of CPU source bo
2484 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2486 commit f20f8556bfc8b4df6ba5050cca656a4ff0fdf18f
2487 Author: Chris Wilson <chris@chris-wilson.co.uk>
2488 Date: Sat Aug 11 10:46:40 2012 +0100
2490 sna: Treat ShmPixmap as ordinary
2492 In theory the code handles the automatic promotion and demotion of the
2495 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2497 commit 16c751a06c503b85c5ab6588bae277af4348487d
2498 Author: Chris Wilson <chris@chris-wilson.co.uk>
2499 Date: Sat Aug 11 10:44:08 2012 +0100
2501 sna: Only mark userptr bo as snooped on non-LLC architectures
2503 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2505 commit 59359dba6496229eff7c60dd04536c5add69cec8
2506 Author: Chris Wilson <chris@chris-wilson.co.uk>
2507 Date: Sat Aug 11 10:25:40 2012 +0100
2509 sna: Mark SHM userptr bo as unreusable
2511 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2513 commit 64488010504a6e76008bb3b3c1e61caeb025913c
2514 Author: Chris Wilson <chris@chris-wilson.co.uk>
2515 Date: Sat Aug 11 10:21:39 2012 +0100
2517 sna/gen2-5: Substitute an equivalent BLT composite operation
2519 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2521 commit f464d508c870293699616626d64bd64f16051467
2522 Author: Chris Wilson <chris@chris-wilson.co.uk>
2523 Date: Sat Aug 11 10:10:32 2012 +0100
2525 sna/gen6+: Try to use the BLT to avoid TLB misses
2527 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2529 commit e9c0e54e69a5ce93bb4e79d56da4a83fddf49c4e
2530 Author: Chris Wilson <chris@chris-wilson.co.uk>
2531 Date: Sat Aug 11 08:58:14 2012 +0100
2533 sna: Handle userptr failures more gracefully
2535 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2537 commit 5d6d9231cd2003fda1c6f2dd3174014317a45704
2538 Author: Chris Wilson <chris@chris-wilson.co.uk>
2539 Date: Fri Aug 10 23:07:07 2012 +0100
2541 sna: Reset BLT state after copy-boxes
2543 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2545 commit 50bfc172f1f89c023bb3a47418b845e31eba7126
2546 Author: Chris Wilson <chris@chris-wilson.co.uk>
2547 Date: Fri Aug 10 22:49:26 2012 +0100
2549 sna: Improve a DBG statement
2551 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2553 commit b33f6754a99f6d11e423d6a03739fa2c04eeed88
2554 Author: Chris Wilson <chris@chris-wilson.co.uk>
2555 Date: Fri Aug 10 21:59:36 2012 +0100
2557 sna: Add assertions to 8x8 tiled BLTs and reset BLT state afterwards
2559 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
2560 References: https://bugs.freedesktop.org/show_bug.cgi?id=53353
2561 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2563 commit ff54d4af39edcddd55f597c404ab6dd548a24aa9
2564 Author: Chris Wilson <chris@chris-wilson.co.uk>
2565 Date: Fri Aug 10 21:41:38 2012 +0100
2567 sna: Add a couple of debug options to control use of BLT patterns
2569 References: https://bugs.freedesktop.org/show_bug.cgi?id=53353
2570 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2572 commit 22306144030b5d37df6d46321555bced6e33c50c
2573 Author: Chris Wilson <chris@chris-wilson.co.uk>
2574 Date: Wed Aug 8 19:45:09 2012 +0100
2576 sna: Use the HAS_SEMAPHORES parameter when available
2578 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2580 commit 465ae1840a883c50d5a5aab7380b30527f00e17b
2581 Author: Chris Wilson <chris@chris-wilson.co.uk>
2582 Date: Wed Aug 8 12:23:25 2012 +0100
2584 sna: Use a compile time constant for determine the zeroth bit
2586 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2588 commit 85192f00e345830541e3715e211b1f98154bbef4
2589 Author: Chris Wilson <chris@chris-wilson.co.uk>
2590 Date: Wed Aug 8 12:11:50 2012 +0100
2592 sna: Ignore trailing bits when comparing lines inside the bitmap
2594 References: https://bugs.freedesktop.org/show_bug.cgi?id=51422
2595 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2597 commit edc1427f3dcddb73acdb5b5e03756ecb30cb3797
2598 Author: Chris Wilson <chris@chris-wilson.co.uk>
2599 Date: Wed Aug 8 10:53:40 2012 +0100
2601 sna/gen5: Add DBG for context switching
2603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2605 commit fec78ddc60b1f03097b334672546c2c7ef21ab4f
2606 Author: Chris Wilson <chris@chris-wilson.co.uk>
2607 Date: Wed Aug 8 10:08:52 2012 +0100
2609 sna: Add DBG option to disable discarding of glyph masks
2611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2613 commit b0cd430082d219500729d20c4740440e58773892
2614 Author: Chris Wilson <chris@chris-wilson.co.uk>
2615 Date: Wed Aug 8 10:07:06 2012 +0100
2617 sna/gen5: Fix use of uninitialised rgba value for DBG
2619 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2621 commit 9ad923f95c59cdf42363f95c7dd52188c0987296
2622 Author: Chris Wilson <chris@chris-wilson.co.uk>
2623 Date: Wed Aug 8 09:59:50 2012 +0100
2625 sna: Minor DBG for fallback glyphs and discarding the mask
2627 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2629 commit 1a0590d133ea6991e0939d1f170f9c10df6856a0
2630 Author: Chris Wilson <chris@chris-wilson.co.uk>
2631 Date: Tue Aug 7 17:16:35 2012 +0100
2633 sna: Check the composite extents against the clip
2635 When computing the composite extents (as opposed to the composite
2636 region) also check if the resultant box overlaps the destination clip
2637 region (we know it already fits into the extents). This helps in cases
2638 with small roi against clipped drawables, such as drawing text onto
2641 References: https://bugs.freedesktop.org/show_bug.cgi?id=51422
2642 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2644 commit 5f5a10ef04a8c01b22da2284583851d84273dc2a
2645 Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
2646 Date: Mon Aug 6 18:48:09 2012 -0300
2650 Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
2651 Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
2652 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2654 commit d8f7f933bc2d30e529730521d5628c10df0361de
2655 Author: Chris Wilson <chris@chris-wilson.co.uk>
2656 Date: Sun Aug 5 17:18:54 2012 +0100
2658 sna: Format markup to suppress compiler warning
2660 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2662 commit 4d0a259988f67f5c569c5d8000c010a7b662efd3
2663 Author: Chris Wilson <chris@chris-wilson.co.uk>
2664 Date: Sat Aug 4 15:54:19 2012 +0100
2666 sna/gen6: Compile fix for DBG
2668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2670 commit d917a276bff4e933590e78404526b929b6b3b497
2671 Author: Chris Wilson <chris@chris-wilson.co.uk>
2672 Date: Sat Aug 4 18:05:03 2012 +0100
2674 sna/gen4: Restore the w/a flush for fill/copy as well
2676 So far, it looks like that the only one we can indeed drop is the
2677 composite with mask.
2679 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2681 commit d754bc3435889b5d19085588c9c888dc08f9522e
2682 Author: Cyril Brulebois <kibi@debian.org>
2683 Date: Mon Nov 7 23:29:29 2011 +0100
2685 Fix build failure when passing --with-builderstring.
2687 If --with-builderstring="foo" is passed to configure (and sna enabled):
2689 src/sna/sna_driver.c: In function 'sna_init_scrn':
2690 src/sna/sna_driver.c:1023:7: error: token ""foo"" is not valid in preprocessor expressions
2692 Fix the missing defined() around the BUILDER_DESCRIPTION variable.
2694 Signed-off-by: Cyril Brulebois <kibi@debian.org>
2696 commit 4bad76b4510d8cf4ccf3f64bcd13ebb80fae4d26
2697 Author: Chris Wilson <chris@chris-wilson.co.uk>
2698 Date: Sat Aug 4 14:28:41 2012 +0100
2700 sna/dri: Don't force ring selection if we have semaphores
2702 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2704 commit fb7b584ec9d0ea96175ba3a621699378c6724940
2705 Author: Chris Wilson <chris@chris-wilson.co.uk>
2706 Date: Sat Aug 4 14:27:46 2012 +0100
2708 sna/dri: Correct ring selection for a busy bo
2710 Confused the RENDER ring with the BLT, limiting swap performance.
2712 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2714 commit 20e58077fe071fd03d44a97194aa163376dc46a4
2715 Author: Chris Wilson <chris@chris-wilson.co.uk>
2716 Date: Sat Aug 4 11:21:10 2012 +0100
2718 sna/gen6: Reduce ring switching for overlapping copy boxes
2720 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2722 commit 2345227663f443baa3c30bdf8b8c33827b7f4947
2723 Author: Chris Wilson <chris@chris-wilson.co.uk>
2724 Date: Sat Aug 4 11:12:12 2012 +0100
2726 sna/gen7: Use a temporary to avoid switching rings for overlapping copies
2728 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2730 commit 61ec2999afd5dad22580e024421f13afe7c82f3a
2731 Author: Chris Wilson <chris@chris-wilson.co.uk>
2732 Date: Sat Aug 4 10:33:05 2012 +0100
2734 sna/gen4: Restore w/a flush for video
2736 One flush removal too many, keep those fingers crossed that the others
2737 do not make an unwanted return.
2739 Reported-by: Roman Jarosz <kedgedev@gmail.com>
2740 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53119
2741 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2743 commit 5833ef173a01afb710acf10e806b83c5ca6efc09
2744 Author: Chris Wilson <chris@chris-wilson.co.uk>
2745 Date: Sat Aug 4 09:31:41 2012 +0100
2749 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2751 commit 036b90f099af21e60fb4c3684616daf1927f705e
2752 Author: Chris Wilson <chris@chris-wilson.co.uk>
2753 Date: Fri Aug 3 21:41:59 2012 +0100
2755 sna/gen7: Correct number of texture coordinates used for video
2757 Fixes regresion from
2759 commit 33c028f8be829caa4fdb9416ff177dc71f24b68e
2760 Author: Chris Wilson <chris@chris-wilson.co.uk>
2761 Date: Wed Aug 1 01:17:50 2012 +0100
2763 sna/gen6+: Reduce floats-per-vertex for spans
2765 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2767 commit 05dcc5f1699ba90fc14c50882e8d4be89bc4a4f9
2768 Author: Chris Wilson <chris@chris-wilson.co.uk>
2769 Date: Fri Aug 3 15:08:45 2012 +0100
2771 Pass the chipset info through driverPrivate rather than a global pointer
2773 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2775 commit 2b3f4ca33a00440a7005fef69099f8dbaddbbad1
2776 Author: Chris Wilson <chris@chris-wilson.co.uk>
2777 Date: Fri Aug 3 14:27:51 2012 +0100
2779 Unexport intel_chipsets
2781 Only used by the core module code, so make it static.
2783 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2785 commit 5ff749727d3590368806508ac0d0fa8efd1d1d51
2786 Author: Chris Wilson <chris@chris-wilson.co.uk>
2787 Date: Wed Jul 25 22:21:29 2012 +0100
2789 sna/gen7: Add constant variations and hookup a basic GT descriptor for Haswell
2791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2793 commit cd028cad3dc9b059a3d83b818d581f86e16ec317
2794 Author: Chris Wilson <chris@chris-wilson.co.uk>
2795 Date: Thu Jul 26 13:17:11 2012 +0100
2797 sna: Limit the batch size on all gen7 variants
2799 Seems the limit on the surface state size is common across the family
2801 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2803 commit 4cd9ec9d404d934268952a1058afa07741b09efe
2804 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2805 Date: Fri May 4 18:26:46 2012 +0200
2807 uxa: fix 3DSTATE_PS to fill in number of samples for Haswell
2809 The sample mask value must match what is set for 3DSTATE_SAMPLE_MASK,
2810 through gen6_upload_invariant_states().
2812 Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2813 Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2815 commit 412668464cf9505629eac20001701af3402dc6e8
2816 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2817 Date: Fri May 4 17:55:10 2012 +0200
2819 uxa: set "Shader Channel Select" fields in surface state for Haswell
2821 For normal behaviour, each Shader Channel Select should be set to the
2822 value indicating that same channel. i.e. Shader Channel Select Red is
2823 set to SCS_RED, Shader Channel Select Green is set to SCS_GREEN, etc.
2825 Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2826 Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2828 commit a47ba68996f117fabcb601d35bcc5f99cbcd6122
2829 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2830 Date: Fri May 4 17:17:22 2012 +0200
2832 uxa: fix max PS threads shift value for Haswell
2834 The maximum number of threads is now a 9-bit value. Thus, one more bit
2835 towards LSB was re-used. i.e. bit position is now 23 instead of 24 on
2838 Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2839 Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2841 commit ce4421e175ceb9259208c7c223af8d66282c3db3
2842 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2843 Date: Fri May 4 17:09:19 2012 +0200
2845 uxa: use at least 64 URB entries for Haswell
2847 Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2848 Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2850 commit 8c880aa34c522b0d67cbb932771f00c947d00dec
2851 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2852 Date: Fri May 4 17:43:19 2012 +0200
2854 uxa: add IS_HSW() macro to distinguish Haswell from Ivybridge
2856 Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2858 commit 0c0d1d956a8ba37d9e6f4a5e4f52018c8ce498e5
2859 Author: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2860 Date: Fri Aug 3 12:03:00 2012 +0100
2862 Introduce a chipset identifier for Haswell (Ivybridge successor)
2864 Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2866 commit 146959dd5ef28384a3db4fce4bf7840f2b3ec58c
2867 Author: Chris Wilson <chris@chris-wilson.co.uk>
2868 Date: Wed Aug 1 23:43:15 2012 +0100
2870 sna: Drop the clear flag as we discard the GPU damage
2872 Hopefully only to keep the sanity checks happy...
2874 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2876 commit 7404e3085b2ee36fa24f77a02d156b4b1d2dff60
2877 Author: Chris Wilson <chris@chris-wilson.co.uk>
2878 Date: Wed Aug 1 23:37:35 2012 +0100
2880 sna: Ensure we only mark a clear for a fill on the GPU bo
2882 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2884 commit ca46d1c7a18596ea9fe2b0577ccf1d110e3e42ac
2885 Author: Chris Wilson <chris@chris-wilson.co.uk>
2886 Date: Wed Aug 1 20:20:29 2012 +0100
2888 sna/gen7: Prefer the BLT for self-copies
2890 Looking at the test results for a third time, gives the edge to the BLT
2893 commit e4a3cd3d16447b5d83d1c8c63c342f1240935267
2894 Author: Chris Wilson <chris@chris-wilson.co.uk>
2895 Date: Wed Aug 1 17:37:33 2012 +0100
2897 sna: Add validation of the clear flag to pixmap debugging
2899 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2901 commit eaeda34bef711cc566f51dee092a19a3c4ac1a16
2902 Author: Chris Wilson <chris@chris-wilson.co.uk>
2903 Date: Wed Aug 1 16:23:05 2012 +0100
2905 sna: Fix computation of st values for SIMD8 dispatch
2907 Fixes regression with enabling 8-pixels.
2909 Reported-by: Mehran Kholdi <semekh.dev@gmail.com>
2910 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53044
2911 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2913 commit 55231eca818c82620c0146384b19b5d5659f6cd6
2914 Author: Chris Wilson <chris@chris-wilson.co.uk>
2915 Date: Wed Aug 1 16:22:23 2012 +0100
2917 sna/gen6: Install a fallback 16-pixel shader
2919 In case the DBG options leave no shader compiled, make sure we always
2922 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2924 commit 4e79c1fef064ce68914eb644edd7f588be3d7300
2925 Author: Chris Wilson <chris@chris-wilson.co.uk>
2926 Date: Wed Aug 1 15:58:43 2012 +0100
2928 Revert "sna/gen7: Prefer the BLT for self-copies"
2930 This reverts commit 89e75dbcb6749bde7587ecc08abed276c255e7f9.
2932 Having removed the forced stall for a RENDER self-copy there is no
2933 longer a need to encourage ring switching.
2935 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2937 commit 85cef78a40c6a7a0254f8fba685f224eac6038e1
2938 Author: Chris Wilson <chris@chris-wilson.co.uk>
2939 Date: Wed Aug 1 13:39:36 2012 +0100
2941 sna/gen7: Simplify the force-stall detection
2943 After reducing the number of conditions where we think we need to force
2944 the stall on the results, we can then simplify the code to detect
2945 that remaining case.
2947 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2949 commit 9391a2c71f020541a2a62ae68eadd486216a38df
2950 Author: Chris Wilson <chris@chris-wilson.co.uk>
2951 Date: Wed Aug 1 12:50:32 2012 +0100
2953 sna/gen7: Only force a stall for a dirty target if also used as a blend source
2955 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2957 commit 0a4bb8663b9fa9b39d13bfb49aea30f2aaecce78
2958 Author: Chris Wilson <chris@chris-wilson.co.uk>
2959 Date: Wed Aug 1 11:19:27 2012 +0100
2961 sna/gen4: Flush not required between fill vertices, only nomaskcomposite
2963 A small breakthrough... Still need to flush the primitive between state
2966 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2968 commit 33c028f8be829caa4fdb9416ff177dc71f24b68e
2969 Author: Chris Wilson <chris@chris-wilson.co.uk>
2970 Date: Wed Aug 1 01:17:50 2012 +0100
2972 sna/gen6+: Reduce floats-per-vertex for spans
2974 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2976 commit 9b2873d3d97b6780d878bd9b821fba0302470f9f
2977 Author: Chris Wilson <chris@chris-wilson.co.uk>
2978 Date: Wed Aug 1 00:01:15 2012 +0100
2980 sna/gen4+: Implement an opacity shader
2982 Avoid the cumbersome lookup through the alpha gradient texture and
2983 simply multiply the incoming opacity value. The next step will be to
2984 reduce the number of floats required per vertex.
2986 Now that we have removed the primary user of the alpha solid cache, it
2987 may be time to retire that as well.
2989 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2991 commit fd3a1236051265fab700aad689a171de47d7338f
2992 Author: Chris Wilson <chris@chris-wilson.co.uk>
2993 Date: Tue Jul 31 10:20:51 2012 +0100
2995 sna/gen6: Enable 8 pixel dispatch
2997 This gives a small performance increase when operating with rectangles,
2998 which is reasonably frequent.
3000 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3002 commit 8922b804bc9ed27957c81f7cda4812ab4ecbd4de
3003 Author: Chris Wilson <chris@chris-wilson.co.uk>
3004 Date: Tue Jul 31 10:20:51 2012 +0100
3006 sna/gen7: Enable 8 pixel dispatch
3008 This gives a small performance increase when operating with rectangles,
3009 which is reasonably frequent.
3011 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3013 commit 492093d04b1486dd34aafe2f109a77ddeb836f18
3014 Author: Chris Wilson <chris@chris-wilson.co.uk>
3015 Date: Tue Jul 31 18:39:17 2012 +0100
3017 sna: Generate shaders for SNB+ 8-pixel dispatch
3019 Not ideal yet, sampling an alpha-only surface using SIMD8 only seems to
3022 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3024 commit 6a5ed88f9fab654c9c11c566b841d42150d26c5d
3025 Author: Chris Wilson <chris@chris-wilson.co.uk>
3026 Date: Mon Jul 30 17:08:20 2012 +0100
3028 sna/gen4: Tidy debugging code
3030 Cluster the ifdefs together in the initialisation code.
3032 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3034 commit 46ec9b0ed55d0fcade40f92206e59c02e402d870
3035 Author: Chris Wilson <chris@chris-wilson.co.uk>
3036 Date: Tue Jul 31 17:41:34 2012 +0100
3038 sna: Update DPMS mode on CRTC after forcing the outputs on
3040 If we forcibly update the outputs to be on, then the core will not issue
3041 its on DPMS event and we miss out on updating the CRTC bookkeeping in
3042 sna_crtc_dpms(). So we need to update the flag on the CRTC as we
3043 manipulate the outputs during modesetting.
3045 References: https://bugs.freedesktop.org/show_bug.cgi?id=52142
3046 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3048 commit 8f166d26b8a93592939068c5a8d160981c724cfd
3049 Author: Chris Wilson <chris@chris-wilson.co.uk>
3050 Date: Tue Jul 31 11:58:24 2012 +0100
3052 sna: Be more careful with damage reduction during CompositeRectangles
3054 We actually need to force DAMAGE_ALL in case we are promoting the GPU
3057 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3059 commit e6cb5d93eaa01e7f4763f797bba341f3cc481d98
3060 Author: Chris Wilson <chris@chris-wilson.co.uk>
3061 Date: Mon Jul 30 11:14:58 2012 +0100
3063 sna: Avoid overlapping gpu/cpu damage with IGNORE_CPU
3065 We cannot simply ignore the presence of CPU damage with IGNORE_CPU but
3066 must remember to discard it.
3068 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3070 commit c9805ba98775bb1e969ff59c7044fe1a49673ca8
3071 Author: Chris Wilson <chris@chris-wilson.co.uk>
3072 Date: Sun Jul 29 18:19:37 2012 +0100
3074 sna: Export sna_drawable_use_bo() to select target for FillRectangles
3076 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3078 commit 89e75dbcb6749bde7587ecc08abed276c255e7f9
3079 Author: Chris Wilson <chris@chris-wilson.co.uk>
3080 Date: Sun Jul 29 15:36:44 2012 +0100
3082 sna/gen7: Prefer the BLT for self-copies
3084 If we are copying to ourselves, we have to regularly flush the render
3085 cache at which point the RENDER pipeline is slower than the BLT
3088 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3090 commit 33d6afda6cec124494f49b74152768da8a3fbdb5
3091 Author: Chris Wilson <chris@chris-wilson.co.uk>
3092 Date: Sun Jul 29 09:51:16 2012 +0100
3094 sna/gen7: Compile basic kernels at runtime
3096 commit eba8d3b3e14a5a16cea6cb8a89f12d3feb8f3d99
3097 Author: Chris Wilson <chris@chris-wilson.co.uk>
3098 Date: Sun Jul 29 09:51:16 2012 +0100
3100 sna/gen6: Compile basic kernels at runtime
3102 commit 8515ec90405912b3d776defcd6e81b1b5f699f1e
3103 Author: Chris Wilson <chris@chris-wilson.co.uk>
3104 Date: Sun Jul 29 09:51:16 2012 +0100
3106 sna/gen5: Compile basic kernels at runtime
3108 commit 00c08b1842c9493ca918a868202946b2e7150de0
3109 Author: Chris Wilson <chris@chris-wilson.co.uk>
3110 Date: Sun Jul 29 09:51:16 2012 +0100
3112 sna/gen4: Compile basic kernels at runtime
3114 commit 7c9dbc980b760e0053d83ca2d7cb147613285680
3115 Author: Chris Wilson <chris@chris-wilson.co.uk>
3116 Date: Sun Jul 29 09:50:39 2012 +0100
3118 sna: Assemble SF and WM kernels using brw
3120 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3122 commit 8ebafa0493c0fa08ab9d80eeb1191b7560dc0863
3123 Author: Chris Wilson <chris@chris-wilson.co.uk>
3124 Date: Wed Sep 21 19:06:07 2011 +0100
3126 sna: Add the brw assembler
3128 In order to construct programs on the fly to cater for the combinatorial
3129 number of possible shaders, we need an assembler, whilst also taking the
3130 opportunity to remove some of the inefficiencies and mistakes from the
3133 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3135 commit ca9d9c02a260bf7930e04bf64e93cc051893c04e
3136 Author: Chris Wilson <chris@chris-wilson.co.uk>
3137 Date: Sat Jul 28 23:08:07 2012 +0100
3139 sna: Prefer not to create a GPU bo without RENDER acceleration
3141 Unless that bo happens to be used on a render chain to the scanout.
3143 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3145 commit eefbe5b6038424566faf7333bb09764b050dd6b4
3146 Author: Chris Wilson <chris@chris-wilson.co.uk>
3147 Date: Mon Jul 30 10:56:29 2012 +0100
3149 sna: Debug option to test migration of inactive pixmaps
3151 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3153 commit d3499cacb59f19b5a3439a630ffbc3e105a27b75
3154 Author: Chris Wilson <chris@chris-wilson.co.uk>
3155 Date: Sun Jul 29 23:00:15 2012 +0100
3157 sna: Assert that we never attempt to submit a batch whilst wedged
3159 We should be asserting at the point that we insert the invalid operation
3160 into the batch, but asserting upon submitting the batch is a useful
3163 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3165 commit cb4d789f98e6e05ec29e5242887018c7450ddf10
3166 Author: Chris Wilson <chris@chris-wilson.co.uk>
3167 Date: Sun Jul 29 22:55:44 2012 +0100
3169 sna: Disable the warning for a hung GPU is we manually set wedged
3171 Only warn about a hung GPU if we encounter an EIO during operation, but
3172 don't warn if we set wedged during initialisation based on unsupported
3175 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3177 commit 0fd680ff52f7ce0a101c617dfb8997c4e228e3ad
3178 Author: Chris Wilson <chris@chris-wilson.co.uk>
3179 Date: Sun Jul 29 22:51:26 2012 +0100
3181 Don't disable acceleration on 830/845g by default
3183 Run the risk of a GPU hang (it shouldn't endanger the entire machine
3184 normally) and let the user elect to disable it through
3186 Option "NoAccel" "true"
3188 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3190 commit 3d45f0affe263985f440e144203ed7cbb3803696
3191 Author: Chris Wilson <chris@chris-wilson.co.uk>
3192 Date: Sat Jul 28 18:21:08 2012 +0100
3194 sna: Honour the Option "DRI"
3196 References: https://bugs.freedesktop.org/show_bug.cgi?id=52624
3197 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3199 commit fb385745a2347f8966765567e78229d67ddc8d60
3200 Author: Chris Wilson <chris@chris-wilson.co.uk>
3201 Date: Fri Jul 27 13:24:04 2012 +0100
3203 sna/gen4: Move the common vertex_offset==0 check into the flush()
3205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3207 commit 73ddd8b0decee444a57f10a11f05deebba686649
3208 Author: Chris Wilson <chris@chris-wilson.co.uk>
3209 Date: Fri Jul 27 12:43:00 2012 +0100
3211 sna/gen4: Further refinement to the GT allocation
3213 Still hunting for why gen4 fails utterly.
3215 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3217 commit 717823a9f77cfedc50e03c5c31c1da44a396b561
3218 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
3219 Date: Fri Jul 27 18:12:26 2012 +0800
3221 uxa/dri (glamor): Use exchange buffer in glamor fixup.
3223 The previous implementation is to create a new textured
3224 pixmap based on the newly created pixmap's buffer object.
3226 This is not efficient, as we already created it when we
3227 call CreatePixmap. We can just exchange the underlying
3228 texture/image buffers by calling intel_glamor_exchange_buffers().
3230 And this commit seems also fix a weird rendering problem
3231 when working with compiz/mutter.
3233 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
3234 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3236 commit a8ee1406244d8b8399bf933d6b61bfd14374b5f9
3237 Author: Chris Wilson <chris@chris-wilson.co.uk>
3238 Date: Fri Jul 27 09:07:16 2012 +0100
3242 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3244 commit bef73cd9279be3438e467981db39c67bc13104f5
3245 Author: Chris Wilson <chris@chris-wilson.co.uk>
3246 Date: Thu Jul 26 21:54:33 2012 +0100
3248 sna/dri: Select the engine before emitting the wait
3250 So that if we have a flexible WAIT_FOR_EVENT that can go on either
3251 pipeline, we can choose our preferred pipeline for DRI.
3253 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3255 commit 1ced4f1ddcf30b518e1760c7aa4a5ed4f934b9f5
3256 Author: Chris Wilson <chris@chris-wilson.co.uk>
3257 Date: Thu Jul 26 10:50:31 2012 +0100
3259 Reduce maximum thread count for IVB GT1 to avoid spontaneous combustion
3261 Somewhere along the way it seems that IVB GT1 was reduced to only allow
3262 a maximum of 48 threads, as revealed in the lastest bspecs.
3264 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52473
3265 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3267 commit 8f8f8759111f791ee99adfd87296443fb0e6acad
3268 Author: Chris Wilson <chris@chris-wilson.co.uk>
3269 Date: Thu Jul 26 17:39:34 2012 +0100
3271 sna/gen4: Tweak heuristics for render/blt usage
3273 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3275 commit c9dd1401615f0ed9492a0c0f547fb37150e013d1
3276 Author: Chris Wilson <chris@chris-wilson.co.uk>
3277 Date: Thu Jul 26 16:31:16 2012 +0100
3279 sna/gen4: Bump thread counts
3281 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3283 commit 7f3fdef98c1ab2fa27439c3be9810b7a934017ce
3284 Author: Chris Wilson <chris@chris-wilson.co.uk>
3285 Date: Thu Jul 26 15:39:05 2012 +0100
3287 sna/gen7: IVB requires a complete pipeline stall when changing blend modes
3289 Similar to how SandyBridge behaves, I had hoped that with IvyBridge they
3290 would have made the pipelined operation actually pipelined, but alas.
3292 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52473
3293 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3295 commit 0938b3df8c25178c8ea0012e1ead1061d03a4e7c
3296 Author: Chris Wilson <chris@chris-wilson.co.uk>
3297 Date: Thu Jul 26 15:21:59 2012 +0100
3299 sna/dri: Add an explanatory assertion
3301 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3303 commit c621183466aa55a5938027b702069e792df2272d
3304 Author: Chris Wilson <chris@chris-wilson.co.uk>
3305 Date: Thu Jul 26 15:18:56 2012 +0100
3307 sna/dri: Tidy fallback/normal CopyRegion
3309 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3311 commit 52d2491a1bafb979d79bb970027c55788f199acb
3312 Author: Chris Wilson <chris@chris-wilson.co.uk>
3313 Date: Thu Jul 26 15:19:14 2012 +0100
3315 sna/video: Protect against attempting to use TexturedVideo whilst wedged
3317 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3319 commit 8c0e3bbb0c301d0fa4652aa38edd84a9fd6b555e
3320 Author: Chris Wilson <chris@chris-wilson.co.uk>
3321 Date: Thu Jul 26 14:58:04 2012 +0100
3323 sna: Force the damage on the DRI pixmap to be flushed to the GPU
3325 Otherwise nothing will happen if we consider ourselves wedged.
3327 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3329 commit f50d7b03b5d0b81b24f0acb4ae555545bbdaa179
3330 Author: Chris Wilson <chris@chris-wilson.co.uk>
3331 Date: Thu Jul 26 14:47:44 2012 +0100
3333 sna: Only recommend not to create bo for a pixmap whilst wedged
3335 This allows us to continue to map a GPU bo and operate inplace if we are
3336 force to create a GPU bo for a compositor.
3338 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3340 commit 49403ddd1bd9a23e1b32a10e7d0757ae2897a579
3341 Author: Chris Wilson <chris@chris-wilson.co.uk>
3342 Date: Thu Jul 26 14:44:14 2012 +0100
3344 sna/dri: Mark a flush required for any new DRI already on exec/dirty lists
3346 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3348 commit efe3c8ff48738f2a274e1d4514d32499fc4aadbc
3349 Author: Chris Wilson <chris@chris-wilson.co.uk>
3350 Date: Thu Jul 26 14:43:33 2012 +0100
3352 sna: Allow DRI to force allocation of a GPU bo and backing pixmap
3354 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3356 commit 2069384f9f06d3ef9dbb0c3f2c64cac4b24e10fc
3357 Author: Chris Wilson <chris@chris-wilson.co.uk>
3358 Date: Thu Jul 26 14:15:45 2012 +0100
3360 sna/dri: Implement fallback handling for CopyRegion whilst wedged
3362 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3364 commit b18143de47e060b67a46d9c68590a2d35df9fca6
3365 Author: Paul Menzel <paulepanter@users.sourceforge.net>
3366 Date: Thu Jul 26 12:51:57 2012 +0200
3368 NEWS: Correct release version: s/2.12.0/2.20.0/
3370 Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
3371 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3373 commit c262d02fb5defe9100df54cf83cc00e11e335745
3374 Author: Chris Wilson <chris@chris-wilson.co.uk>
3375 Date: Thu Jul 26 01:12:11 2012 +0100
3377 Limit PCI matching to VGA devices
3381 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3383 commit ad6355311b8b80777bc0fec95f6bf6cd1b4969d9
3384 Author: Chris Wilson <chris@chris-wilson.co.uk>
3385 Date: Thu Jul 26 01:02:53 2012 +0100
3387 sna: Compile against xorg-1.10
3389 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3391 commit b6c7c490d76c683b7b5c89d20f902603b85bd3bc
3392 Author: Chris Wilson <chris@chris-wilson.co.uk>
3393 Date: Thu Jul 26 00:48:59 2012 +0100
3395 sna: Check for a NULL scanout after DPMS off with shadow enabled
3397 We may mark the scanout as detached when all outputs are turned off (for
3398 example during rotation) and so in the subsequent block handler we need
3399 to be careful in case we are handling a detached shadow.
3401 Reported-by: chr.ohm@gmx.net
3402 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52514
3403 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3405 commit f4acc01bb09f68edbad4c6cc7e04c271c92661c1
3406 Author: Chris Wilson <chris@chris-wilson.co.uk>
3407 Date: Wed Jul 25 22:43:32 2012 +0100
3409 sna: Restore inplace upload for DRI compositors
3411 With a DRI compositor we have to flush the output after every request,
3412 which leads to major inefficiencies. This can be mitigated slightly if
3413 we know we will have to upload shortly, which we track using the pixmap
3416 In particular PutImage is meant to upload inplace to an active DRI
3417 buffer, however this was accidentally dropped in
3419 commit a253c95ec63b2b075e66ae7380fed6a73469eba5
3420 Author: Chris Wilson <chris@chris-wilson.co.uk>
3421 Date: Sun Jul 15 13:32:35 2012 +0100
3423 sna: Prefer uploads to be staged in snoopable bo
3425 Performace of putimage500 on ivb i7-3720qm:
3430 Reported-by: Michael Larabel <Michael@phoronix.com>
3431 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3433 commit 5c969a05ef815b261e157fe8d1172aebfd7f5841
3434 Author: Chris Wilson <chris@chris-wilson.co.uk>
3435 Date: Wed Jul 25 20:28:41 2012 +0100
3437 sna/gen7: Move the PS thread count definition into the constant struct
3439 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3441 commit 901bb618215d65747eb0a8d481c77c1201f69362
3442 Author: Chris Wilson <chris@chris-wilson.co.uk>
3443 Date: Wed Jul 25 20:13:15 2012 +0100
3445 sna/gen7: Remove duplicated constants
3447 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3449 commit 8c5077e4ed055a97bf9deda59c0e9a45e42317ca
3450 Author: Chris Wilson <chris@chris-wilson.co.uk>
3451 Date: Wed Jul 25 16:59:11 2012 +0100
3453 Assume all unknown chipsets are future gen
3455 I think the likelihood of a new product being launched based on a 8xx
3456 design is remote enough not to worry about.
3458 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3460 commit 954d9c6aca792098cb16a46b4ec8d5fc104b15a9
3461 Author: Chris Wilson <chris@chris-wilson.co.uk>
3462 Date: Wed Jul 25 15:41:29 2012 +0100
3464 sna: Initialise single-shot tile offsets before use
3466 As noted by the compiler amidst all the noise.
3468 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3470 commit 484b072072c8297a87940c9e32097923f0a77c8f
3471 Author: Chris Wilson <chris@chris-wilson.co.uk>
3472 Date: Wed Jul 25 12:22:47 2012 +0100
3474 sna: Add 'gmux_backlight' to list of known devices for backwards compatability
3476 Reported-by: Austin Lund <austin.lund@gmail.com>
3477 References: https://bugs.freedesktop.org/show_bug.cgi?id=52423
3478 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3480 commit 0ab6d7a50d37cf4454577cb8c333d4b8683aa054
3481 Author: Chris Wilson <chris@chris-wilson.co.uk>
3482 Date: Wed Jul 25 12:20:36 2012 +0100
3484 sna: Prefer platform backlight devices over firmware
3486 This is in contrast to libbacklight but closer to our original code as
3487 we prefer a known custom backlight controller over the ACPI interfaces.
3488 As only the ACPI interfaces are marked as firmware, and the custom
3489 backlight controllers as platform, we therefore need to prefer platform
3492 Reported-by: Austin Lund <austin.lund@gmail.com>
3493 References: https://bugs.freedesktop.org/show_bug.cgi?id=52423
3494 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3496 commit b6d82ab07661aba98963f239f9501b50c3910962
3497 Author: Chris Wilson <chris@chris-wilson.co.uk>
3498 Date: Wed Jul 25 10:40:07 2012 +0100
3500 sna: Reduce ping-pong for Composite with render disabled
3502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3504 commit de707b7dc853a3b315ead9789d5ca541412c99bc
3505 Author: Chris Wilson <chris@chris-wilson.co.uk>
3506 Date: Wed Jul 25 09:05:46 2012 +0100
3508 uxa: Add Apple's gmux to the list of known preferred backlights
3510 Reported-by: Austin Lund <austin.lund@gmail.com>
3511 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52423
3512 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3514 commit 3d254e4010d0753f433cfe62c6a7546b02482847
3515 Author: Chris Wilson <chris@chris-wilson.co.uk>
3516 Date: Mon Jul 23 23:45:33 2012 +0100
3518 sna: Use SETUP_BLT to reduce overheads for tiled BLT
3520 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3522 commit 2ed44149eaa27b9632ec83a776f3ee67b0eec7b0
3523 Author: Chris Wilson <chris@chris-wilson.co.uk>
3524 Date: Mon Jul 23 23:45:33 2012 +0100
3526 sna: Use SETUP_BLT to reduce overheads for stippled BLT
3528 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3530 commit c1785aaca2c9347114d28f114ee59ef8206d829b
3531 Author: Chris Wilson <chris@chris-wilson.co.uk>
3532 Date: Mon Jul 23 23:00:25 2012 +0100
3534 sna: Remove a surplus function
3536 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3538 commit 554fce8a65b2518cae032f1eadf58830559299c8
3539 Author: Chris Wilson <chris@chris-wilson.co.uk>
3540 Date: Mon Jul 23 22:59:16 2012 +0100
3542 uxa: Pass the correct parameters to intel_detect_chipset()
3544 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3546 commit 40d90dfd8674c255a45b46bbdc09d497af5b3f50
3547 Author: Chris Wilson <chris@chris-wilson.co.uk>
3548 Date: Mon Jul 23 21:55:46 2012 +0100
3550 intel: Refactor the common chipset detection/override
3552 Reduce the duplicate messages for which type of chip we by
3553 amalgamating the common code.
3555 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3557 commit e3f6c48d18c316899c71b6fc34971039c6f9e5f8
3558 Author: Chris Wilson <chris@chris-wilson.co.uk>
3559 Date: Mon Jul 23 14:43:23 2012 +0100
3561 sna: Refactor PutImage to avoid calling drawable_gc_flags() too early
3563 drawable_gc_flags() asserts that the gc has been moved to the CPU prior
3564 to its calls so that it can read the reduced raster operation.
3566 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
3567 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3569 commit 1af26ea4228a9d7768b475b4f9164d2c7620d4fd
3570 Author: Chris Wilson <chris@chris-wilson.co.uk>
3571 Date: Mon Jul 23 14:39:44 2012 +0100
3573 sna: Flesh out tiled operations using the BLT
3575 Before enabling the RENDER pipeline for this operation, let's just see
3576 what is required to fully use the BLT pipeline as well.
3578 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3580 commit ac182a006732525a921a9c539e5ebfb537ad3b52
3581 Author: Chris Wilson <chris@chris-wilson.co.uk>
3582 Date: Mon Jul 23 11:05:05 2012 +0100
3584 sna: Hold a reference to the full stipple pattern for repeated tiles
3586 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3588 commit 83f683b47063eab8cfb5037d02133dd977c3fc25
3589 Author: Chris Wilson <chris@chris-wilson.co.uk>
3590 Date: Sun Jul 22 23:20:23 2012 +0100
3594 A good brown paper bag bug release for SNA.
3596 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3598 commit 9402bdcc13f7e96dfe527ff4a3da8d13a7870a02
3599 Author: Chris Wilson <chris@chris-wilson.co.uk>
3600 Date: Sun Jul 22 22:14:52 2012 +0100
3602 sna/glyphs: Also discard the glyph mask for bitmaps with an opaque source
3604 Though I expect all such glyphs to be caught by the non-overlapping
3607 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3609 commit b315e0ebb75d8391ebef7ebe53741a5e33c968bb
3610 Author: Chris Wilson <chris@chris-wilson.co.uk>
3611 Date: Sun Jul 22 20:07:38 2012 +0100
3613 sna: Tweak the fallback hints for XYPixmap PutImage
3615 As the fallback uses a multiple-pass algorithm updating one plane at a
3616 time, we wish to prepare the fallback surface for reads.
3618 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3620 commit 8acaf2693e176a92993a498683f121cfe0343fd4
3621 Author: Chris Wilson <chris@chris-wilson.co.uk>
3622 Date: Sun Jul 22 12:28:34 2012 +0100
3624 sna: Promote tiled operations to the GPU if the tile is already on the GPU
3626 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3628 commit 7d4a3e371beea65bf66e54ae13789d6d5ca91f8b
3629 Author: Chris Wilson <chris@chris-wilson.co.uk>
3630 Date: Sun Jul 22 11:19:13 2012 +0100
3632 sna: Use an upload buffer for large stipples
3634 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3636 commit 40e0cf32a25e43e16184b2af87a1e1abeb8e4052
3637 Author: Chris Wilson <chris@chris-wilson.co.uk>
3638 Date: Sat Jul 21 19:43:23 2012 +0100
3640 sna/dri: We fail at predicting the flip frame
3642 Simply report the values from the kernel, and transfer the blame...
3644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3646 commit c6e316eeba3008b351f2cd63829154f4672c5417
3647 Author: Chris Wilson <chris@chris-wilson.co.uk>
3648 Date: Sat Jul 21 18:18:32 2012 +0100
3650 sna: Adjust hints to prefer rendering convex polygon with the GPU
3652 Keep the general polygons as only using the GPU if necessary, until the
3653 cost of the routines is analysed.
3655 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3657 commit f1e7248cb353d634f27d297059911168ce1a0762
3658 Author: Chris Wilson <chris@chris-wilson.co.uk>
3659 Date: Sat Jul 21 16:08:31 2012 +0100
3661 sna: Expand the heuristic for predicting when to use CPU bo for readback
3663 For tiny transfers, the cost of setting up the GPU operation outweighs
3664 the actual savings through increased throughput. So we try to guess when
3665 it will be preferrable to simply read from the GPU bo directly.
3667 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3669 commit 06db69c2c7023f702f9773be90144fdf7a1159e4
3670 Author: Chris Wilson <chris@chris-wilson.co.uk>
3671 Date: Sat Jul 21 15:28:10 2012 +0100
3673 sna: Update assertion for cached io buffers
3675 As kgem_buffers may be reused and repurposed through the snoop cache it
3676 is no longer true that only proxies will have the io flag set.
3678 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3680 commit d715e1e01437049e167462281d51b5e214594361
3681 Author: Chris Wilson <chris@chris-wilson.co.uk>
3682 Date: Sat Jul 21 15:04:31 2012 +0100
3684 sna: Also discard the last-was-cpu flag when overwriting cpu damage
3686 We interpret a FillRect that erradicates the existing damage as a
3687 clear-event and an opportunity to see if it is worth migrating the
3688 render commands to the GPU. This is undermined if we leave the
3689 'prefer-cpu' flag intact.
3691 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3693 commit e95825d17ce65ad8173a5e6518a98969e236a4f8
3694 Author: Chris Wilson <chris@chris-wilson.co.uk>
3695 Date: Sat Jul 21 14:38:17 2012 +0100
3697 sna: Enable snooping on the reused linear buffer
3699 This explains why suddenly I was seeing clflush again on a couple of
3702 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3704 commit c5e6b5874f334b9124a17f017c6eb175cf88f115
3705 Author: Chris Wilson <chris@chris-wilson.co.uk>
3706 Date: Sat Jul 21 14:26:52 2012 +0100
3708 sna: Fix the reversed not SHM assertion
3710 Should be double checking that we are not about to free a CPU bo pinned
3713 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3715 commit f36b656ab2bc16ec8849cadb0afb574bb742c3a3
3716 Author: Chris Wilson <chris@chris-wilson.co.uk>
3717 Date: Sat Jul 21 13:18:40 2012 +0100
3719 sna: Avoid marking io/snoop buffers as unreusable unnecessarily
3721 As they are kept in special caches, we can reserve the unreusable flags
3722 for exceptional buffers.
3724 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3726 commit 37dfdb0e9e86effc3ca8b590c98aa2382e8f0cea
3727 Author: Chris Wilson <chris@chris-wilson.co.uk>
3728 Date: Sat Jul 21 13:01:09 2012 +0100
3730 sna: Correct assertion for __kgem_bo_size()
3732 Only proxies are measured in bytes not pages.
3734 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3736 commit 83ad661bc73e9d0094b669c5203e25afc3937bb7
3737 Author: Chris Wilson <chris@chris-wilson.co.uk>
3738 Date: Sat Jul 21 12:51:41 2012 +0100
3740 sna: Change the vmap interface name to userptr
3742 This is in common with the other drivers and avoids the conflict with
3743 'vmalloc/vmap' used by the kernel for allocation of contiguous virtual
3746 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3748 commit 8dcccd308222bcf1b96f2ee15842b4558ea5f29e
3749 Author: Chris Wilson <chris@chris-wilson.co.uk>
3750 Date: Sat Jul 21 12:47:21 2012 +0100
3752 sna: s/vmap/snoop/ since we use the flag more generically
3754 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3756 commit 6acc9e6a6e1de2a11597c810e02f793774cef2dd
3757 Author: Chris Wilson <chris@chris-wilson.co.uk>
3758 Date: Sat Jul 21 12:07:46 2012 +0100
3760 sna: Fix role reversal of __kgem_bo_size() and kgem_bo_size()!
3762 Reported-by: Jiri Slaby <jirislaby@gmail.com>
3763 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
3764 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3766 commit 286b0e1a48cab85191dfbb112c8dd14aeaa70956
3767 Author: Chris Wilson <chris@chris-wilson.co.uk>
3768 Date: Fri Jul 20 16:04:37 2012 +0100
3770 sna: Refresh experimental userptr vmap support
3772 Bring the code uptodate with both kernel interface changes and internal
3773 adjustments following the creation of CPU buffers with set-cacheing.
3775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3777 commit 93c794eb3f80bef64f1619986a7c950229dc7a47
3778 Author: Chris Wilson <chris@chris-wilson.co.uk>
3779 Date: Fri Jul 20 20:34:53 2012 +0100
3781 sna: Micro-optimise copying boxes with the blitter
3783 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3785 commit a0d95a9c2d3a27eafbe459e2aefe772c006e596f
3786 Author: Chris Wilson <chris@chris-wilson.co.uk>
3787 Date: Fri Jul 20 20:34:23 2012 +0100
3789 sna: Only update a buffer when it becomes dirty
3791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3793 commit c52d265b83b033fb2a275fcc9a8a8d146e3afdf6
3794 Author: Chris Wilson <chris@chris-wilson.co.uk>
3795 Date: Fri Jul 20 19:38:38 2012 +0100
3797 sna: Tweak CPU bo promotion rules for CopyArea
3799 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3801 commit f92a64dd9162731210b14368b6ee408356d7fefc
3802 Author: Chris Wilson <chris@chris-wilson.co.uk>
3803 Date: Fri Jul 20 15:54:35 2012 +0100
3805 sna: Only set the vmap flag after we make the bo snoopable
3807 Otherwise if we fail then we incorrectly add the handle to the vmap
3810 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3812 commit 8b4cf24f1403bf3d929cc0725de66b3d0e08ebaf
3813 Author: Chris Wilson <chris@chris-wilson.co.uk>
3814 Date: Fri Jul 20 14:46:32 2012 +0100
3816 sna: Also check whether the first upload box can use the BLT
3818 No point checking boxes 1..n if box 0 is the troublemaker!
3820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3822 commit df14b285be44f0c40a718bb8ae09a9558b1eb2c7
3823 Author: Chris Wilson <chris@chris-wilson.co.uk>
3824 Date: Fri Jul 20 14:35:28 2012 +0100
3826 sna/gen6: Prefer the more flexible render ring for large surfaces
3828 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3830 commit 578ff11c3753ede2c81afc47302991e3d3b316f2
3831 Author: Chris Wilson <chris@chris-wilson.co.uk>
3832 Date: Fri Jul 20 14:24:06 2012 +0100
3834 sna: Just use composite.box() when we only have one box
3836 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3838 commit fb7987fc0b51cf3b83dcf78bcefe65ec3af32ccf
3839 Author: Chris Wilson <chris@chris-wilson.co.uk>
3840 Date: Fri Jul 20 13:12:27 2012 +0100
3842 sna/dri: Cleanup ring selection for SNB+ CopyRegion
3844 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3846 commit 3b56588fbaa2c4ccdfb2f2a8f5656d2cda9dacd7
3847 Author: Chris Wilson <chris@chris-wilson.co.uk>
3848 Date: Fri Jul 20 10:19:25 2012 +0100
3850 sna: Update WIP userptr example usage
3852 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3854 commit 473a1dfb683ed576d86b37aba36aaa0e379f4606
3855 Author: Chris Wilson <chris@chris-wilson.co.uk>
3856 Date: Fri Jul 20 09:56:13 2012 +0100
3858 sna: Rename kgem_partial_bo to kgem_buffer
3860 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3862 commit 8e6e8a2fa8adda9ae9be8a88fbb14851e9d2df2e
3863 Author: Chris Wilson <chris@chris-wilson.co.uk>
3864 Date: Fri Jul 20 09:51:46 2012 +0100
3866 sna: Allow the snoopable upload buffer to take pages from the CPU vma cache
3868 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3870 commit 979035bb9ce04db5fe30efa4f6daab0a40f6af57
3871 Author: Chris Wilson <chris@chris-wilson.co.uk>
3872 Date: Thu Jul 19 20:06:28 2012 +0100
3874 sna: Remove topmost unused 'flush' attribute
3876 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3878 commit b83011909aaf185f05fc2df743882c2410eff46d
3879 Author: Chris Wilson <chris@chris-wilson.co.uk>
3880 Date: Thu Jul 19 19:51:46 2012 +0100
3882 sna: Replace 'sync' flag with equivalent 'flush'
3884 The only difference is in semantics. Currently 'sync' was only used on
3885 CPU buffers for shared memory segments with 2D clients, and 'flush' on GPU
3886 buffers shared with DRI clients.
3888 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3890 commit 88bee3caeaacbbb1b4d789ea3db9a3802a62b59d
3891 Author: Chris Wilson <chris@chris-wilson.co.uk>
3892 Date: Thu Jul 19 19:40:34 2012 +0100
3894 sna: Remove unused scanout-is-dirty? flag
3896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3898 commit 6f60f89588caa70e7d8ed53ba453bbe8c2094a95
3899 Author: Chris Wilson <chris@chris-wilson.co.uk>
3900 Date: Thu Jul 19 16:58:34 2012 +0100
3902 sna/gen6: Bump the WM thread count to 80
3904 Note that we should only do this when "WiZ Hashing" is disabled. So we
3905 should be checking the GT_MODE register (bring on i915_read!) to be sure
3906 that is safe to do so. However, it gives a big boost to performance of
3907 render copies... It also causes perf benchmarks to hit thermal limits
3910 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3912 commit fc39d4b5cb105d269c5349e479daf112f5d93580
3913 Author: Chris Wilson <chris@chris-wilson.co.uk>
3914 Date: Thu Jul 19 16:25:54 2012 +0100
3916 sna/gen6: Add a simple DBG option to limit usage of either BLT/RENDER
3918 We can force the code to either select only BLT or RENDER operations -
3919 for those that we have a choice for at least!
3921 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3923 commit 15d3eea7004822e5cbd48d676692e1b6a2b26d3e
3924 Author: Chris Wilson <chris@chris-wilson.co.uk>
3925 Date: Thu Jul 19 16:22:20 2012 +0100
3927 sna: Handle mixed bo/buffers in assertions
3929 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3931 commit e4fce3b7801038e4f64d848a0995f4b441b4d2aa
3932 Author: Chris Wilson <chris@chris-wilson.co.uk>
3933 Date: Thu Jul 19 10:50:09 2012 +0100
3935 sna/gen4: Hookup composite spans
3937 Due to the unresolved flushing bug it is no faster (so only enable when
3938 we definitely can't do the operation inplace), however it does eliminate
3939 a chunk of CPU overhead.
3941 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3943 commit 5f138176bf15682324d2e8cfa9fac3e49604bf8f
3944 Author: Chris Wilson <chris@chris-wilson.co.uk>
3945 Date: Thu Jul 19 10:32:33 2012 +0100
3947 sna: Tweak order of screen re-initialisation
3949 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3951 commit 9bd0f8f3e7783d7a6bab707fc08ec96830cd5809
3952 Author: Chris Wilson <chris@chris-wilson.co.uk>
3953 Date: Thu Jul 19 09:40:07 2012 +0100
3955 i810: Correct the double negative and enable XAA when available
3957 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3959 commit d145d0e1459f578eea621e6944814642e5dd431f
3960 Author: Chris Wilson <chris@chris-wilson.co.uk>
3961 Date: Thu Jul 19 09:23:10 2012 +0100
3963 i810: Handle initialisation without the XAA module present at runtime
3965 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3967 commit 7a3b98e05b706548527e73b2008600391c601a62
3968 Author: Chris Wilson <chris@chris-wilson.co.uk>
3969 Date: Thu Jul 19 08:42:05 2012 +0100
3971 sna: Re-register the SHM funcs every server generation
3973 As the SHM layer hooks into the CloseScreen chain to free its privates,
3974 we then need to call the registration function again on the next
3975 generation to ensure that the private is reallocated before use.
3977 Reported-by: Pawel Sikora <pluto@agmk.net>
3978 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52255
3979 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3981 commit 4bcab83bbddf8a698aa83f5038f9ab019a404bd5
3982 Author: Chris Wilson <chris@chris-wilson.co.uk>
3983 Date: Wed Jul 18 22:19:45 2012 +0100
3985 i810: DRI is not dependent upon XAA
3987 The blit routines is uses are independent of the XAA driver interface
3988 and can be used separately.
3990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
3992 commit 558c8251299b786cab1ac83dbd35f077224b5950
3993 Author: Chris Wilson <chris@chris-wilson.co.uk>
3994 Date: Wed Jul 18 22:00:30 2012 +0100
3996 sna/gen4+: Drop unsupported source formats
3998 Once again I've confused existence of the enum with the ability of the
3999 sampler to read that format.
4001 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4003 commit 9f3b3098c9f870d303a9de2b9c0db119eff5a865
4004 Author: Chris Wilson <chris@chris-wilson.co.uk>
4005 Date: Wed Jul 18 21:39:47 2012 +0100
4007 sna/dri: Allow DRI2 to be loaded even if we are wedged
4009 Just because the GPU is spitting EIO at us does not necessarily imply
4010 that a DRI client will also suffer. Spit out a warning for later bug
4011 reporting and let them find out for themselves!
4013 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4015 commit 15b7191fd363e9e6083844a218e25419695d55f1
4016 Author: Chris Wilson <chris@chris-wilson.co.uk>
4017 Date: Wed Jul 18 12:59:41 2012 +0100
4019 sna/gen6: Micro-optimise render copy emission
4021 Backport of the changes made for IVB.
4023 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4025 commit 4eea9ac0035dd72f3c637adc39eeaeda46472e9e
4026 Author: Chris Wilson <chris@chris-wilson.co.uk>
4027 Date: Wed Jul 18 12:59:41 2012 +0100
4029 sna/gen7: Micro-optimise render copy emission
4031 The goal is bring the overhead down to that of using the blitter. Tricky
4032 given the number of steps to using the 3D pipeline compared to the
4035 A stretch goal would be to make IVB GPU bound for -copywinpix10!
4037 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4039 commit 267429bbb146449ee4d3b88fa8e23c5b1d53470a
4040 Author: Chris Wilson <chris@chris-wilson.co.uk>
4041 Date: Wed Jul 18 19:24:44 2012 +0100
4043 sna: Enable runtime detection of set-cacheing ioctl
4045 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4047 commit c0b3674d042ff55d64ad1fd0d64926e1967be323
4048 Author: Chris Wilson <chris@chris-wilson.co.uk>
4049 Date: Wed Jul 18 20:48:27 2012 +0100
4051 sna/trapezoids: Only reduce bounded operators to a single pass
4053 Only for a few operators can we replace the opacity mask by
4054 premultiplying into the source.
4056 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4058 commit bb0303677c38076db14dfbceec3636197a971e8c
4059 Author: Chris Wilson <chris@chris-wilson.co.uk>
4060 Date: Wed Jul 18 10:40:50 2012 +0100
4062 sna/trapezoids: Use pixman from within the spans to reduce two-pass operations
4064 Reduce the two pass CompositeTrapezoids if we can perform the operation
4065 inplace by calling pixman_image_composite from the span. This step
4066 enables this for xrgb32.
4068 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4070 commit bee1a14618797b3d3a1c1a20eb72644fa907c048
4071 Author: Chris Wilson <chris@chris-wilson.co.uk>
4072 Date: Wed Jul 18 09:38:32 2012 +0100
4074 sna: Fix processing of the last fallback box
4076 The evil typo caused us to misalign the clip boxes and run over a
4077 garbage array on 64-bit builds.
4079 Reported-by: Edward Sheldrake <ejsheldrake@gmail.com>
4080 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
4081 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52163
4082 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4084 commit 88cb1968b6dbf3edfa885da9503e91124af46007
4085 Author: Chris Wilson <chris@chris-wilson.co.uk>
4086 Date: Wed Jul 18 09:38:03 2012 +0100
4088 sna: Add more DBG for fallback processing
4090 Hunting the lost box...
4092 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4094 commit 36f2e46619598e9bca4fe1207aa2f157bfa1ecf4
4095 Author: Chris Wilson <chris@chris-wilson.co.uk>
4096 Date: Wed Jul 18 00:45:54 2012 +0100
4098 sna: Reuse the snoopable cache more frequently for upload buffers
4100 Now that we are keeping a small cache of snoopable buffers, experiment
4101 with using them for uploads more frequently.
4103 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4105 commit 73f07abbd2d78418e5a66262f293b5ed80b7ccb4
4106 Author: Chris Wilson <chris@chris-wilson.co.uk>
4107 Date: Wed Jul 18 00:19:49 2012 +0100
4109 sna: Maintain a short-lived cache of snoopable CPU bo for older gen
4111 Once again, we find that frequent buffer creation and manipulation of the
4112 GTT is a painful experience leading to noticeable and frequent application
4113 stalls. So mitigate the need for fresh pages by keeping a small stash of
4114 recently freed and inactive bo.
4116 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4118 commit 77520641a332a622c0b5378bd254ed5cb46a5f0a
4119 Author: Chris Wilson <chris@chris-wilson.co.uk>
4120 Date: Tue Jul 17 22:09:33 2012 +0100
4122 i810: Replace XAAGet.*ROP() with local tables
4124 The XAAGetPatternROP() and XAAGetCopyROP() functions were removed along
4125 with the rest of XAA so we need to implement those tables locally.
4127 Reported-by: Knut Petersen <Knut_Petersen@t-online.de>
4128 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4130 commit caef63e0268e59e439b030a9a338e81d5cf8e311
4131 Author: Chris Wilson <chris@chris-wilson.co.uk>
4132 Date: Tue Jul 17 21:22:57 2012 +0100
4134 i810: Split xaa routines from common acceleration methods
4136 Some of the routines in i810_accel.c are specific to XAA whilst others
4137 are used elsewhere, for example in i810_dri.c. Therefore we have to be
4138 selective over which ones we compile out without xaa.
4140 Reported-by: Knut Petersen <Knut_Petersen@t-online.de>
4141 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4143 commit 53ff19f45a3cc4863845c23e8d3c2c2b95e03fd9
4144 Author: Chris Wilson <chris@chris-wilson.co.uk>
4145 Date: Tue Jul 17 19:40:16 2012 +0100
4147 sna: Allow wedged CopyPlane to operate inplace on the destination
4149 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4151 commit d4fa4d5494db45b227c9ae7f7a90cd5dfd940027
4152 Author: Chris Wilson <chris@chris-wilson.co.uk>
4153 Date: Tue Jul 17 19:29:32 2012 +0100
4155 sna: Allow inplace copies for wedged CopyArea
4157 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4159 commit 217eeadf81a8cbb43e495e1e937acdd95c703377
4160 Author: Chris Wilson <chris@chris-wilson.co.uk>
4161 Date: Tue Jul 17 19:35:06 2012 +0100
4163 sna: Allow operation inplace to scanout whilst wedged
4165 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4167 commit 40ff29480a0dbf458adf1a1b0d3275ad1361530e
4168 Author: Chris Wilson <chris@chris-wilson.co.uk>
4169 Date: Tue Jul 17 18:38:49 2012 +0100
4171 sna: Tweak fast blt path
4173 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4175 commit fce69c79c4840e7863d7c382da0d22be90a9f19a
4176 Author: Chris Wilson <chris@chris-wilson.co.uk>
4177 Date: Tue Jul 17 18:28:24 2012 +0100
4179 sna: prefer fbBlt over pixman_blt
4181 It is currently much better optimised through memcpy.
4183 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4185 commit c29f96d50839388377ad57c6366f9bc7ad8b9d0a
4186 Author: Chris Wilson <chris@chris-wilson.co.uk>
4187 Date: Tue Jul 17 13:02:51 2012 +0100
4189 sna/gen7: Bump the number of pixel shader threads for IVB GT2
4191 Spotted-by: Kilarski, Bernard R" <bernard.r.kilarski@intel.com>
4192 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4194 commit 799bae9e8ff53fb1b5c74f3278d530a58d66de9a
4195 Author: Chris Wilson <chris@chris-wilson.co.uk>
4196 Date: Tue Jul 17 17:07:10 2012 +0100
4198 sna/dri: Do not allow an exchange to take place on invalid buffers
4200 If the SwapBuffers is called after we have resized a Window but before
4201 the client has processed the Invalidate notification, then the
4202 SwapBuffers will be referring to a pair of stale buffers. As the buffers
4203 are no longer attached to the Pixmap, we can not simply exchange them.
4205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4207 commit 067aeaddb8047f01ae3a20b26ba0acf5ba2d035f
4208 Author: Chris Wilson <chris@chris-wilson.co.uk>
4209 Date: Tue Jul 17 15:33:19 2012 +0100
4211 sna: Rebalance choice of GPU vs CPU bo
4213 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4215 commit 7ebeea3f5c71959773478de44b08a967fe5acc8b
4216 Author: Chris Wilson <chris@chris-wilson.co.uk>
4217 Date: Tue Jul 17 15:10:43 2012 +0100
4219 sna: Avoid the CPU bo readback for render paths
4221 As we exclude using the CPU bo if there is overlapping GPU damage, we
4222 can forgo the call to keep the transfer the damage.
4224 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4226 commit ed8c729ed02705fd03be1ab22a94b5aae13567c8
4227 Author: Chris Wilson <chris@chris-wilson.co.uk>
4228 Date: Tue Jul 17 11:21:30 2012 +0100
4230 sna: Catch the short-circuit path for clearing clear on move-to-gpu as well
4232 I thought the short-circuit path was only taken when already clear, I
4235 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4237 commit 359b9cc82de13b0ac89692896ac6104ff3be308b
4238 Author: Chris Wilson <chris@chris-wilson.co.uk>
4239 Date: Tue Jul 17 10:26:27 2012 +0100
4241 sna: Limit the use of snoopable buffers to read/write uploads
4243 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4245 commit 4f21dba6ee505217d63edd84611622e05aeb4593
4246 Author: Chris Wilson <chris@chris-wilson.co.uk>
4247 Date: Tue Jul 17 09:26:46 2012 +0100
4249 sna: Only drop the clear flag when writing to the GPU pixmap
4251 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4253 commit fbfbbee8288aba1e4754fd2dbc02e71f5e118cda
4254 Author: Chris Wilson <chris@chris-wilson.co.uk>
4255 Date: Tue Jul 17 09:20:21 2012 +0100
4257 sna: Fix glyph DBG to include clip extents and actual glyph origin
4259 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4261 commit f0ed0ca234a4bed986824845ff70e8554c0e579f
4262 Author: Chris Wilson <chris@chris-wilson.co.uk>
4263 Date: Tue Jul 17 08:35:20 2012 +0100
4265 sna: Promote an undamaged pixmap to use the full GPU
4267 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4269 commit 1f79e877fb6602bd0f9dd14ac9c3511f3b7044fb
4270 Author: Chris Wilson <chris@chris-wilson.co.uk>
4271 Date: Mon Jul 16 21:18:24 2012 +0100
4273 sna: Share the pixmap migration decision with the BLT composite routines
4275 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4277 commit d141a2d59007866c9eaad020c744be446e70c346
4278 Author: Chris Wilson <chris@chris-wilson.co.uk>
4279 Date: Mon Jul 16 23:20:58 2012 +0100
4281 sna: Disable snoopable bo for gen4
4283 Further inspection reveals that whilst it may not hang the GPU, the
4284 results are not pleasant or complete.
4286 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4288 commit 107feed2a4ca044313c70f83a62909187ff1f905
4289 Author: Chris Wilson <chris@chris-wilson.co.uk>
4290 Date: Mon Jul 16 18:58:30 2012 +0100
4292 sna: Disable snoopable uplaod buffers for gen4
4294 The sampler really does not like using snoopable buffers...
4296 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4298 commit 818c21165c746b7b410a6e6e23b1675d88db685d
4299 Author: Chris Wilson <chris@chris-wilson.co.uk>
4300 Date: Mon Jul 16 16:28:00 2012 +0100
4302 sna: Fixup pixmap validation for sna_copy_area()
4304 Remember to offset the box by the drawable deltas in order to
4305 compensate for compositing.
4307 Reported-by: Jiri Slaby <jirislaby@gmail.com>
4308 References: https://bugs.freedesktop.org/show_bug.cgi?id=52142
4309 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4311 commit 623d84bed7c47ac39348775ce35eec54196f6dac
4312 Author: Chris Wilson <chris@chris-wilson.co.uk>
4313 Date: Mon Jul 16 16:07:37 2012 +0100
4315 Wrap defines to avoid redefinition warnings
4317 Currently this only catches out ARRAY_SIZE, but wrap the other common
4318 defines for consistency.
4320 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4322 commit 907a2a7c97514d3f7610648ed87c7042a857f786
4323 Author: Chris Wilson <chris@chris-wilson.co.uk>
4324 Date: Mon Jul 16 14:47:03 2012 +0100
4326 sna/trapezoids: Fix inplace unaligned fills (on gen4)
4328 Reported-by: Sergio Callegari <sergio.callegari@gmail.com>
4329 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52150
4330 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4332 commit 6ce2f40249231f57cf464361ea5329cee1932ccf
4333 Author: Chris Wilson <chris@chris-wilson.co.uk>
4334 Date: Mon Jul 16 14:46:39 2012 +0100
4336 sna/trapezoids: Add some DBG to unaligned fills
4338 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4340 commit 2721214868685123c973970a8ce0d93346ae0ee2
4341 Author: Chris Wilson <chris@chris-wilson.co.uk>
4342 Date: Mon Jul 16 13:52:39 2012 +0100
4344 sna: Move the disabling of CPU bo for gen4 to the render unit
4346 They appear to work fine with the BLT and only seem to cause issues when
4347 used with the sammpler. So enable them for accelerated uploads and
4350 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4352 commit 0777b146bf1a63c99e4d4af141e676a47b1f2dc9
4353 Author: Chris Wilson <chris@chris-wilson.co.uk>
4354 Date: Mon Jul 16 13:11:07 2012 +0100
4356 sna: Use set-cache-level to allocate snoopable upload buffers
4358 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4360 commit 33443f7ee48fa54b6f4d09c93cddac0e32314b9c
4361 Author: Chris Wilson <chris@chris-wilson.co.uk>
4362 Date: Mon Jul 16 12:51:54 2012 +0100
4364 sna: Add a couple of DBG options to control accelerated up/downloads
4366 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4368 commit 924060293826a1cc0d9d7bc26e913e46c6b2d054
4369 Author: Chris Wilson <chris@chris-wilson.co.uk>
4370 Date: Mon Jul 16 12:03:47 2012 +0100
4372 sna: Discard and recreate the CPU buffer when busy during move-to-cpu
4374 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4376 commit 7024ef771ff170e61e788b5216c86b46e0f8ae6a
4377 Author: Chris Wilson <chris@chris-wilson.co.uk>
4378 Date: Mon Jul 16 12:15:54 2012 +0100
4380 sna: Add a few DBG to show when CPU bos are being used for xfer
4382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4384 commit c564414157e27417f0de1c0542dafd9b47e01eda
4385 Author: Chris Wilson <chris@chris-wilson.co.uk>
4386 Date: Mon Jul 16 09:34:30 2012 +0100
4388 sna: Disable the scanout flush when switch off via DPMS
4390 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4392 commit 536e7ab756d6821db79e4cd79a250af1c0f7d5a3
4393 Author: Chris Wilson <chris@chris-wilson.co.uk>
4394 Date: Mon Jul 16 12:16:26 2012 +0100
4396 intel: Don't use stdbool without declaring it
4398 Reported-by: Fabio Pedretti <fabio.ped@libero.it>
4399 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4401 commit 0c32be15b06ad63c1fc1371de879f2d879080f6b
4402 Author: Chris Wilson <chris@chris-wilson.co.uk>
4403 Date: Thu Jul 5 19:26:48 2012 +0100
4405 uxa: Remove Shadow hack
4407 This was an incomplete hack so deprecate in favour of Shadow-on-Steriods,
4410 References: https://bugs.freedesktop.org/show_bug.cgi?id=47324
4411 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4413 commit 6a18a0936eafc45ab920ab0eecf2fc2a601c41a7
4414 Author: Chris Wilson <chris@chris-wilson.co.uk>
4415 Date: Sun Jul 15 20:26:00 2012 +0100
4417 configure: version bump for 2.20.0 release
4419 The day SNA hits mainstream...
4421 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4423 commit 6fa059330decd1437eef4928d732ec91fd4e17e7
4424 Author: Chris Wilson <chris@chris-wilson.co.uk>
4425 Date: Sun Jul 15 14:04:48 2012 +0100
4427 sna: Avoid creating a CPU bo for uploads if we already have a large GPU bo
4429 And vice-versa if already have a large CPU bo.
4431 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4433 commit a253c95ec63b2b075e66ae7380fed6a73469eba5
4434 Author: Chris Wilson <chris@chris-wilson.co.uk>
4435 Date: Sun Jul 15 13:32:35 2012 +0100
4437 sna: Prefer uploads to be staged in snoopable bo
4439 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4441 commit ef6d94a8444927941db108811e1a26357dc3f18e
4442 Author: Chris Wilson <chris@chris-wilson.co.uk>
4443 Date: Sun Jul 15 11:46:53 2012 +0100
4445 sna: Simply reverse all the boxes if dx <= 0 and dy <= 0
4447 In this fairly common case, avoid both the double pass and use a simpler
4448 algorithm as we can simply reverse the order of the boxes.
4450 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4452 commit 6601a943ff968ac39ba198351c50dc883cb4232e
4453 Author: Chris Wilson <chris@chris-wilson.co.uk>
4454 Date: Sun Jul 15 11:39:56 2012 +0100
4456 sna: Keep track of the base pointer for the reordered boxes
4458 So that we avoid freeing an invalid pointer.
4460 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4462 commit ef34d5cf415ad7459ab44b0ec2e70b14150735fc
4463 Author: Chris Wilson <chris@chris-wilson.co.uk>
4464 Date: Sat Jul 14 23:25:17 2012 +0100
4466 sna: Make sure we check for a busy CPU bo before declaring is-cpu
4468 Even if the pixmap is entirely damaged on the CPU, we still may be in
4469 the process of transferring it and so cause an unwanted stall.
4471 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4473 commit 0e397e4a1dc23ed07089c967612d705584f3b376
4474 Author: Chris Wilson <chris@chris-wilson.co.uk>
4475 Date: Sat Jul 14 21:35:23 2012 +0100
4477 sna/glyphs: Perform the fallback mask reduce before moving the glyph pointers
4479 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4481 commit db1ee13a53b0c1348b7566ee60ee1b7b384ef59a
4482 Author: Chris Wilson <chris@chris-wilson.co.uk>
4483 Date: Sat Jul 14 21:26:01 2012 +0100
4485 sna/gen7: Check for gradient allocation failure within spans
4487 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4489 commit 86479e97460da798a3804cbb4ae39e62de881af1
4490 Author: Chris Wilson <chris@chris-wilson.co.uk>
4491 Date: Sat Jul 14 21:24:04 2012 +0100
4493 sna/gen7: Uses EXTEND_PAD for its alpha-gradient with spans
4495 Fixes regression from 2b94f9a043.
4497 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4499 commit f17037275c05198c3c3f456964fd42032f9085b6
4500 Author: Chris Wilson <chris@chris-wilson.co.uk>
4501 Date: Sat Jul 14 20:11:17 2012 +0100
4503 sna: Reorder overlapping boxes for CopyArea/Window
4505 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4507 commit 86e09d14bd00344d378b86a19ebb44f7d946926c
4508 Author: Chris Wilson <chris@chris-wilson.co.uk>
4509 Date: Sat Jul 14 17:51:31 2012 +0100
4511 sna: Tidy sna_copy_boxes
4513 So there appears to be a bug hidden here. But only when we scroll
4514 upwards in a GTK+ application. Hmm.
4516 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4518 commit 924f596463555db27214fd8227218c2e21ecddc8
4519 Author: Chris Wilson <chris@chris-wilson.co.uk>
4520 Date: Sat Jul 14 19:03:08 2012 +0100
4522 sna: Avoid BLT to snoopable bo on older gen
4524 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4526 commit 44e226b1d9fca8cb95b0864adf8708b03ee8472c
4527 Author: Chris Wilson <chris@chris-wilson.co.uk>
4528 Date: Sat Jul 14 18:36:29 2012 +0100
4530 sna/glyphs: A repeat of the earlier typo for pixman glyphs
4532 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4534 commit 797ebf937fcfcc87502727c70e6b52f89fecc799
4535 Author: Chris Wilson <chris@chris-wilson.co.uk>
4536 Date: Sat Jul 14 16:58:18 2012 +0100
4538 sna/glyphs: Fix typo and render glyphs to the small mask
4540 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4542 commit 2b94f9a043372ffede01339eea99377a71169fbc
4543 Author: Chris Wilson <chris@chris-wilson.co.uk>
4544 Date: Sat Jul 14 16:38:21 2012 +0100
4546 sna/gen7: Preselect sampler
4548 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4550 commit 50d61c38bdb720c17ba5bfcaf97032338e466d9b
4551 Author: Chris Wilson <chris@chris-wilson.co.uk>
4552 Date: Sat Jul 14 16:07:32 2012 +0100
4554 sna/gen4: Fix typo introduced in checking dst format
4556 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4558 commit fc91b1f157f9e6e4097e662e2890574e530344ac
4559 Author: Chris Wilson <chris@chris-wilson.co.uk>
4560 Date: Sat Jul 14 15:48:50 2012 +0100
4562 sna/gen7: Only emit the render flushes between operations
4564 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4566 commit d30dc59651f156e9d4572141e2c54af89d007a37
4567 Author: Chris Wilson <chris@chris-wilson.co.uk>
4568 Date: Sat Jul 14 15:42:19 2012 +0100
4570 sna/gen7: Only fallback to BLT for a redirected target
4572 As we can always use a partial surface for the source.
4574 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4576 commit 9f05b0c03b1fdab474b436431c430028fca1937e
4577 Author: Chris Wilson <chris@chris-wilson.co.uk>
4578 Date: Sat Jul 14 15:03:32 2012 +0100
4580 sna/gen4+: Fix up card/dest format confusion in previous commit
4582 That would have been much more successful had I not supplied the wrong
4583 opaque formats to the sampler.
4585 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4587 commit fffbc34e4621898eee9b80bf8b6b3699bcade52a
4588 Author: Chris Wilson <chris@chris-wilson.co.uk>
4589 Date: Sat Jul 14 14:54:27 2012 +0100
4591 sna/gen4+: Translate to card format using a switch
4593 GCC produces faster code than a walk over the format tables.
4595 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4597 commit 95fdd9af5c8a8360d02debc400e75869c36f05ca
4598 Author: Chris Wilson <chris@chris-wilson.co.uk>
4599 Date: Sat Jul 14 13:05:27 2012 +0100
4601 sna: Enable pixman_glyphs if available
4603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4605 commit 3f764ee4c50567cfb831495d42cb6c2bb94055ad
4606 Author: Chris Wilson <chris@chris-wilson.co.uk>
4607 Date: Sat Jul 14 12:29:47 2012 +0100
4609 sna/dri: Presume an unknown chipset will want i965_dri.so
4611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4613 commit 39845280b5a9d0ed60c6a158c3d0df9fb8756d40
4614 Author: Chris Wilson <chris@chris-wilson.co.uk>
4615 Date: Sat Jul 14 12:14:28 2012 +0100
4617 sna: Allow BLT support on future unknown hardware
4619 Assume that if the kernel recognises the chipset and declares it has a
4622 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4624 commit b260ca44b3aaba2c8cd25640ad7ac9ca6478b0f2
4625 Author: Chris Wilson <chris@chris-wilson.co.uk>
4626 Date: Sat Jul 14 09:59:07 2012 +0100
4628 Drop some unused includes
4630 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4632 commit f517bdb12b909ef9d0897166bdabc537551a065b
4633 Author: Chris Wilson <chris@chris-wilson.co.uk>
4634 Date: Sat Jul 14 08:56:21 2012 +0100
4636 sna: Do not perform batch compaction on old kernels
4638 As they will reject the delta pointing outside of the target batch.
4640 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4642 commit ce69a1e8686889f1eebb4cb3a39c41f473e58b93
4643 Author: Chris Wilson <chris@chris-wilson.co.uk>
4644 Date: Fri Jul 13 23:25:02 2012 +0100
4646 sna: Add some DBG for selecting glyph path
4648 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4650 commit b68b2d90915f7a97e43a39e3117927544886fba0
4651 Author: Chris Wilson <chris@chris-wilson.co.uk>
4652 Date: Fri Jul 13 23:24:32 2012 +0100
4654 sna: Add some DBG for short-circuiting choice of source bo
4656 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4658 commit ad4786b285074b5bd70b3ad0bea6ec1b77ad6740
4659 Author: Chris Wilson <chris@chris-wilson.co.uk>
4660 Date: Fri Jul 13 22:24:51 2012 +0100
4662 sna: Aim for consistency and use stdbool except for core X APIs
4664 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4666 commit a05c3547bba52288bae872ea672ffe2f4dab2ffa
4667 Author: Chris Wilson <chris@chris-wilson.co.uk>
4668 Date: Fri Jul 13 15:06:15 2012 +0100
4670 sna/gen4: Simplify comparing the pipeline-pointers against the previous
4672 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4674 commit ea9ec18505645dfec85ab96bbbbbc1793830c737
4675 Author: Chris Wilson <chris@chris-wilson.co.uk>
4676 Date: Fri Jul 13 11:39:01 2012 +0100
4678 sna: Split CloseScreen into two phases
4680 In order to get the ordering correct we need to free the xf86_cursors
4681 before calling the miPointerCloseScreen. This requires us to insert a
4682 hook at the top of the CloseScreen chain. However we still require the
4683 final CloseScreen hook in order to do the fundamental clean up, hence
4684 split the CloseScreen callback into two phases.
4686 Reported-by: Jiri Slaby <jirislaby@gmail.com>
4687 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
4688 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4690 commit 0f03f7b4359fcbcde651bc1554ddff4fe10bc53b
4691 Author: Chris Wilson <chris@chris-wilson.co.uk>
4692 Date: Fri Jul 13 10:35:27 2012 +0100
4694 sna: Only try to enable DRI2 if the module is available at runtime
4696 Blatantly cribbed from Michel Danzer's patch for ati...
4698 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4700 commit b5d6a57f12025aef9850c7d9baa6905f776be971
4701 Author: Chris Wilson <chris@chris-wilson.co.uk>
4702 Date: Thu Jul 12 22:16:27 2012 +0100
4704 Enable compilation of SNA by default
4706 But only if we meet the required versions of Xorg and leave UXA as the
4707 default AccelMethod for the time being.
4709 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4711 commit 6c2975ab2943478b3a246b5fb231f9f3df2d8475
4712 Author: Chris Wilson <chris@chris-wilson.co.uk>
4713 Date: Thu Jul 12 22:14:39 2012 +0100
4715 Fix the reporting of whether SNA is configured
4717 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4719 commit fd15ce65ab8ce5cf571e37daa7db1ee245616cd4
4720 Author: Chris Wilson <chris@chris-wilson.co.uk>
4721 Date: Thu Jul 12 20:32:14 2012 +0100
4723 sna: Fix build without DRI2
4725 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4727 commit 22be9988b933f33fc5247a9abc3b00a7f2e4a202
4728 Author: Chris Wilson <chris@chris-wilson.co.uk>
4729 Date: Thu Jul 12 19:49:34 2012 +0100
4731 sna: Check for failure to initialize the sprite pointers
4733 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4735 commit 32e7f4ee64867779b2def6fcd882708d7b0e2cf5
4736 Author: Chris Wilson <chris@chris-wilson.co.uk>
4737 Date: Thu Jul 12 19:39:10 2012 +0100
4739 sna/glyphs: Fix array allocation for list_extents
4741 Originally I intended to skip assigning the box on the last list.
4742 However, loop simplicity failed and now we run the risk of writing
4743 beyond the end of stack_extents, and overwriting the list_extents
4746 Reported-by: Jiri Slaby <jirislaby@gmail.com>
4747 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
4748 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4750 commit 0477b5fb6f040f3bad86bb314a24df1bcd660aed
4751 Author: Chris Wilson <chris@chris-wilson.co.uk>
4752 Date: Thu Jul 12 19:37:02 2012 +0100
4754 sna/glyphs: Apply mask reduction along fallback paths as well
4756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4758 commit 16aaa51b5d326f44974489f5b29716c7ff5ab48e
4759 Author: Chris Wilson <chris@chris-wilson.co.uk>
4760 Date: Thu Jul 12 15:26:54 2012 +0100
4762 sna: Rearrange the tests for dropping the glyph mask
4764 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4766 commit 600746f923b046187a66b60b5e5d2b187475730e
4767 Author: Chris Wilson <chris@chris-wilson.co.uk>
4768 Date: Thu Jul 12 14:16:31 2012 +0100
4770 i810: Remove vestigial reference to xaa.h
4772 i810_hwmc.c can pull it in via i810.h like everybody else. As for
4773 xaalocal.h, I have no idea what that is... Both appear to be cut'n'paste
4776 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4778 commit 1d9ab2e7101167075112a472ee82530dc0365183
4779 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
4780 Date: Thu Jul 12 18:47:50 2012 +0800
4782 uxa/dri: Fix a buffer leak on pageflipping when enabling glamor.
4784 We need to put current front_buffer to back buffer thus we
4785 don't need to create a new back buffer next time. This behaviou
4786 should be the same with or without glamor. Previous code
4787 incorrectly discard the previous front_buffer and cause a
4788 big buffer leak problem.
4790 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
4791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4793 commit fa89e0614563cbe0cc45e6d7e7584cb333244c27
4794 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
4795 Date: Thu Jul 12 18:47:51 2012 +0800
4797 uxa/glyphs: Bypass uxa glyphs operations if using glamor.
4799 glamor_glyphs will never fallback. We don't need to keep a
4800 uxa glyphs cache picture here. Thus simply bypass the
4801 corresponding operations.
4803 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
4804 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4806 commit d25000e1f31e78aff2ab43adb12aec0aac36f56f
4807 Author: Chris Wilson <chris@chris-wilson.co.uk>
4808 Date: Wed Jul 11 22:37:25 2012 +0100
4810 i810: Tidy configure detection for XAA/DGA
4812 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4814 commit 44a1528c15eec9b3f93651e779013137864d4d2b
4815 Author: Chris Wilson <chris@chris-wilson.co.uk>
4816 Date: Wed Jul 11 15:05:32 2012 +0100
4818 sna: Remove unused cached partials list
4820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4822 commit 94d489ae43a2c4d4d9ddc9ce30ff1a9142b77d4a
4823 Author: Chris Wilson <chris@chris-wilson.co.uk>
4824 Date: Wed Jul 11 14:25:20 2012 +0100
4826 sna: Minor tweak to upload in place if the CPU bo is busy
4828 Since we have to pay the price of the stall anyway...
4830 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4832 commit b5db90aa52f10897ad2d7795df94c0e3d2878aea
4833 Author: Chris Wilson <chris@chris-wilson.co.uk>
4834 Date: Wed Jul 11 12:52:24 2012 +0100
4836 sna: Discard the mask for "non-overlapping" glyphs
4838 If we can acheive the same rasterisation results without the mask,
4839 rendering the glyphs-to-dst is so much faster that it outweighs the cost
4840 of checking for overlapping glyphs.
4842 The penalty is then for code that correctly declared that it required
4843 a mask, who now have an extra ~10% overhead in the processing of their
4846 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4848 commit 665c9cbdcf2c493cac29d316eaffa2abe197a183
4849 Author: Chris Wilson <chris@chris-wilson.co.uk>
4850 Date: Tue Jul 10 19:01:12 2012 +0100
4852 sna: Disable periodic scanout refresh if all outputs are disconnected
4854 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4856 commit 85fdc3143b157fd7ba3453efc86da8238d9de316
4857 Author: Chris Wilson <chris@chris-wilson.co.uk>
4858 Date: Tue Jul 10 12:43:46 2012 +0100
4860 sna: Avoid the expensive recomputation of the region when copying
4862 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4864 commit 747ed0ca28bb46fc45ff20acd1f82212fa2b46bd
4865 Author: Chris Wilson <chris@chris-wilson.co.uk>
4866 Date: Tue Jul 10 17:05:20 2012 +0100
4868 sna: Only promote to full GPU if we already have a GPU bo
4870 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4872 commit 232217eef8f99f5678d65bf9aa5b898ef6d3b3c6
4873 Author: Daniel Stone <daniel@fooishbar.org>
4874 Date: Tue Jul 10 00:32:13 2012 +0100
4876 i810: Make DGA optional
4878 Don't build DGA when it's not available, or when we don't want it.
4880 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4882 commit 78dc0c04745ad4485b994f67833f4a155749f01d
4883 Author: Daniel Stone <daniel@fooishbar.org>
4884 Date: Tue Jul 10 00:32:13 2012 +0100
4886 i810: Make XAA optional
4888 Don't build XAA when it's not available, or when we don't want it.
4890 Signed-off-by: Daniel Stone <daniel@fooishbar.org>
4892 commit 45ab003a5860fd4290df24739d2520fddfe27a8f
4893 Author: Chris Wilson <chris@chris-wilson.co.uk>
4894 Date: Mon Jul 9 21:28:40 2012 +0100
4896 sna: Check for overlapping glyphs within each list, then overlapping lists
4898 Spotted by Zhigang Gong is this optimisation to avoid the problem with
4899 multiple lines passed in a single request (using multiple lists). As the
4900 start of line will overlap with the previous line when we use the simple
4901 bbox comparison, we always declare those runs as overlapping and so we
4902 cannot substitute a glyph mask. However, we can reduce the problem to
4903 only checking for overlapping glyphs within a list and then checking for
4904 overlapping lists. Very, very clever.
4906 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4908 commit 8066bc33d78e78ce7c13833b08a7daaea2f3ed22
4909 Author: Chris Wilson <chris@chris-wilson.co.uk>
4910 Date: Mon Jul 9 11:25:11 2012 +0100
4912 sna: Fix reversal of inside logic for BitmapToRegion
4914 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4916 commit 8a9a17cd092d156272e8953974119f527b2b77eb
4917 Author: Chris Wilson <chris@chris-wilson.co.uk>
4918 Date: Mon Jul 9 11:12:12 2012 +0100
4920 sna: Improve the check for assertions
4922 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4924 commit b11bc37684181390fc1400afb44054785104ca15
4925 Author: Chris Wilson <chris@chris-wilson.co.uk>
4926 Date: Mon Jul 9 11:09:16 2012 +0100
4928 sna: Simplify the DBG incarnation
4930 It was only ever used in conjunction with HAS_DEBUG_FULL. For debug
4931 purposes it is as easy to redefine DBG locally. By simplifying the DBG
4932 macro we can create it consistently and so reduce the number of compiler
4935 Long term, this has to be dynamic. Sigh.
4937 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4939 commit 21798a88676e91049917fafd3196dd4374b94226
4940 Author: Chris Wilson <chris@chris-wilson.co.uk>
4941 Date: Mon Jul 9 09:23:20 2012 +0100
4943 sna: Promote large operations to use the whole GPU
4945 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4947 commit 1b6ad7a6ae6820c8f66d1c80613885ee84e7d316
4948 Author: Chris Wilson <chris@chris-wilson.co.uk>
4949 Date: Mon Jul 9 08:54:11 2012 +0100
4951 sna: Only consider large clears as candidates for GPU migration
4953 If we only operating on a small region of the pixmap and have require
4954 damage migration in the past, we are likely to require migration again
4955 at some point. So keep track of small damage areas.
4957 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4959 commit d8a75538ea1d2a79d6282b0e2dfd73cfdea1a480
4960 Author: Chris Wilson <chris@chris-wilson.co.uk>
4961 Date: Mon Jul 9 08:38:39 2012 +0100
4963 sna: PadPixmap only writes to the out-of-bounds bits
4965 So we only need to delcare it as reading the source pixmap and not mark
4968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4970 commit eafb454edf188e7dada1ddf886d1e46f0151968d
4971 Author: Chris Wilson <chris@chris-wilson.co.uk>
4972 Date: Mon Jul 9 08:38:08 2012 +0100
4974 sna: Rename conflicting symbols with uxa
4976 Reported-by: Christoph Reiter <reiter.christoph@gmail.com>
4977 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=51887
4978 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4980 commit 0af29175a087cc2e509962f8828790c8f7232611
4981 Author: Chris Wilson <chris@chris-wilson.co.uk>
4982 Date: Mon Jul 9 01:24:23 2012 +0100
4984 sna: Just use a linear scan to find the terminating clip box
4986 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
4988 commit 2941a5fe15626730869a48a63bb088e8ae2c0549
4989 Author: Chris Wilson <chris@chris-wilson.co.uk>
4990 Date: Mon Jul 9 00:20:57 2012 +0100
4992 sna: Remove the consideration of CPU damage for overwriting FillRect
4994 We consideer a singular FillRect to be a sequence point in the rendering
4995 commands, that is it is usually used to clear the background as the first
4996 operation in a drawing sequence. So it is useful to ask if we can move
4997 the sequence to the GPU at that point.
4999 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5001 commit 8be00b6d4767ffc09328d246d02ee75312ad5842
5002 Author: Chris Wilson <chris@chris-wilson.co.uk>
5003 Date: Sun Jul 8 20:48:40 2012 +0100
5005 sna: Substitute the reduce clip region for fallback ops
5007 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5009 commit 0457935a70f6c1ae06f61f9ad0dd1bdc825465fa
5010 Author: Chris Wilson <chris@chris-wilson.co.uk>
5011 Date: Sun Jul 8 20:38:03 2012 +0100
5013 sna: Review placement hints for fallback operations
5015 Look for those operations that maybe better via the GTT and those that
5016 are preferred to be in CPU cache. The wonders of multiple layers of
5019 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5021 commit b7f0b0e7e3f66165b87c46f897de536cf74a9daf
5022 Author: Chris Wilson <chris@chris-wilson.co.uk>
5023 Date: Sun Jul 8 19:59:34 2012 +0100
5025 sna: Remove function for force-to-gpu
5027 This is now enitrely done in the core move-to-gpu as a special case.
5029 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5031 commit bb8770158c08394c2de79e0ca1c1b3112e17dd23
5032 Author: Chris Wilson <chris@chris-wilson.co.uk>
5033 Date: Sun Jul 8 18:53:25 2012 +0100
5035 sna: Apply the clear color when resizing the front buffer
5037 If the existing front buffer is clear, just apply the clear color to
5038 then new buffer rather than copy the old one across.
5040 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5042 commit 8dd14855d75240501aa2d089bcdfa46e7badbf18
5043 Author: Chris Wilson <chris@chris-wilson.co.uk>
5044 Date: Sun Jul 8 15:46:41 2012 +0100
5046 sna/dri: Review stale comments
5048 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5050 commit 3bb7a530e77a5c6e8bae0fb14c570feafe37c2bd
5051 Author: Chris Wilson <chris@chris-wilson.co.uk>
5052 Date: Sat Jul 7 20:53:27 2012 +0100
5054 sna/dri: Fix cross-chaining of pageflip vs vblank
5056 And double-check that the drawable is still flippable before completing
5059 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5061 commit 1454df8caa51fff9810a9b792d17aa82b247a4ae
5062 Author: Chris Wilson <chris@chris-wilson.co.uk>
5063 Date: Sat Jul 7 19:59:50 2012 +0100
5065 sna/dri: Use draw ref directly
5067 As we hook into the DestroyWindow notification, we can reliably use the
5068 original Drawable reference and avoid the secondary object lookups.
5070 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5072 commit 9abb6c09bdb9831db5f0626eed70663a2dc39fc6
5073 Author: Chris Wilson <chris@chris-wilson.co.uk>
5074 Date: Sat Jul 7 18:43:06 2012 +0100
5076 sna/dri: Remove dead code for 'old_fb'
5078 The member still exists but is never set and is unused.
5080 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5082 commit ad877abdc70b842afd202b2f97892d97a0b8d151
5083 Author: Chris Wilson <chris@chris-wilson.co.uk>
5084 Date: Sat Jul 7 17:48:52 2012 +0100
5086 sna/dri: Attach the chain of frame events to the Window, not pixmap
5088 So that we can have multiple clients swapping in separate windows
5091 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5093 commit 81cd9aa80091b9bb08b50062f117d678a3bc7a91
5094 Author: Chris Wilson <chris@chris-wilson.co.uk>
5095 Date: Sat Jul 7 16:12:49 2012 +0100
5097 sna: Tweak start/stop of the deferred flush
5099 As we now emit work whenever we wakeup and find the GPU idle, we rarely
5100 actually have pending work in the deferred flush queue, so try to avoid
5101 installing a timer if we are not accumulating work.
5103 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5105 commit 6cb0c631e4eafc09f1677c73906de9108d735de4
5106 Author: Chris Wilson <chris@chris-wilson.co.uk>
5107 Date: Sat Jul 7 15:10:18 2012 +0100
5109 sna/dri: Clarify the message for one failure case
5111 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5113 commit 9a314d18cef1e08b23f9dca861ad2ba396bb1080
5114 Author: Chris Wilson <chris@chris-wilson.co.uk>
5115 Date: Sat Jul 7 14:56:05 2012 +0100
5117 sna/dri: Add a couple of missing VG_CLEAR on vblanks
5119 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5121 commit 66a53c15cb5ee729fb43ea9713fd8538a3f982ad
5122 Author: Chris Wilson <chris@chris-wilson.co.uk>
5123 Date: Sat Jul 7 14:46:08 2012 +0100
5125 sna/dri: Couple the frame events into DestroyWindow
5127 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5129 commit 975a566bed72ddc79853b329307ed72a82df24b0
5130 Author: Chris Wilson <chris@chris-wilson.co.uk>
5131 Date: Sat Jul 7 14:12:51 2012 +0100
5133 sna/dri: Replace the DRI2 drawable type with a devPrivate
5135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5137 commit 0da1c98f660269806408af5fd08c1ab5e538082e
5138 Author: Chris Wilson <chris@chris-wilson.co.uk>
5139 Date: Sat Jul 7 12:15:28 2012 +0100
5141 test: Add missing header for distcheck
5143 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5145 commit e3e58123d36924c760ab6f58a7155a040422e91d
5146 Author: Chris Wilson <chris@chris-wilson.co.uk>
5147 Date: Fri Jul 6 15:22:26 2012 +0100
5149 sna: Fixup fb wrapper
5151 To accommodate changes in the Xserver and avoid breakage; would have been
5152 much easier had the fb been exported in the first place.
5154 commit 5d2f88fd9972c62c87098ddc7fee7b6f0cea0fdb
5155 Author: Chris Wilson <chris@chris-wilson.co.uk>
5156 Date: Sat Jul 7 10:07:51 2012 +0100
5158 sna: Add a couple of DBG flags to control upload buffers
5160 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5162 commit 18b3090381ac0fb61df96a7c6b2f12aacf180ee8
5163 Author: Chris Wilson <chris@chris-wilson.co.uk>
5164 Date: Sat Jul 7 00:06:36 2012 +0100
5166 sna: Fixup USE_INPLACE debug hint
5168 After we enabled inplace readback, we can just apply the INPLACE hint to
5169 all fallbacks, but instead apply it to the automatic testing for a
5170 potential inplace operation.
5172 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5174 commit 24b59a8955eb15522ad334a541530aca937fcf07
5175 Author: Chris Wilson <chris@chris-wilson.co.uk>
5176 Date: Fri Jul 6 14:16:36 2012 +0100
5178 sna: Refactor a common function: is_clipped()
5180 Had I done this earlier, I would not have the bug in the open-coded
5183 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5185 commit 68b357454af705f1ed3a9599435a402f2611a180
5186 Author: Chris Wilson <chris@chris-wilson.co.uk>
5187 Date: Fri Jul 6 10:58:06 2012 +0100
5189 sna: Clear garbage from the new front buffer when resizing
5191 Avoid displaying a buffer filled with random junk when resizing (and
5192 thereby creating a new) framebuffer.
5194 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5196 commit 5784e0f21dc91f33c99a507105a0695cc53d6574
5197 Author: Chris Wilson <chris@chris-wilson.co.uk>
5198 Date: Fri Jul 6 09:42:37 2012 +0100
5200 Allow matching against any device supported by drm/i915
5202 However we cannot enable acceleration if we do not recognise its
5203 hardware layout or instruction set.
5205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5207 commit 1ee10cc3b2aa0888753eeb25c7fde7296a3c92eb
5208 Author: Chris Wilson <chris@chris-wilson.co.uk>
5209 Date: Fri Jul 6 00:01:58 2012 +0100
5211 Make the detection of broken pre-production hardware verbose
5213 These SDV should have been returned to the manufacturer long ago and
5214 replaced with real systems. So if they are still in use, add a gentle
5217 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5219 commit cd2dd3016e0834d1636aa96511608022a4cdbcd1
5220 Author: Chris Wilson <chris@chris-wilson.co.uk>
5221 Date: Thu Jul 5 19:58:54 2012 +0100
5223 sna: Fix clipping of glyphs-to-dst for partially obscurred windows
5225 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5227 commit 0f086acb259d7732560c5d0d642308de028a4445
5228 Author: Chris Wilson <chris@chris-wilson.co.uk>
5229 Date: Wed Jul 4 12:26:18 2012 +0100
5231 sna: Fallback to attaching a shadow fb if we fail to setup the crtc
5233 As we've chosen to fix the kernel to handle CRTC offsets > 4096, drop
5234 the automatic workaround. However, allow the user to force creation of
5235 PerCrtcPixmaps for the purpose of debugging (and to workaround the bug
5236 in older kernels) and to fallback to trying a shadow fb if the setcrtc
5237 fails with the composite fb.
5239 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5241 commit 7e8060f837475c85cc061ba4a5388140cd227613
5242 Author: Chris Wilson <chris@chris-wilson.co.uk>
5243 Date: Thu Jul 5 03:31:12 2012 +0100
5245 sna: Do not force GPU allocation if CPU bo is already busy
5247 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5249 commit c32bb286dc9a489232030f6abe9076411fbcecfd
5250 Author: Chris Wilson <chris@chris-wilson.co.uk>
5251 Date: Thu Jul 5 03:18:12 2012 +0100
5253 sna: Make sure damage is flushed to the CPU bo before use
5255 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5257 commit d46cc00b3cd903bfaf37ad7d4a60676c4b346983
5258 Author: Chris Wilson <chris@chris-wilson.co.uk>
5259 Date: Thu Jul 5 02:19:26 2012 +0100
5261 sna/dri: Assert that our pixmaps sizes are invariant
5263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5265 commit 3c1f58fe45e8f4f18c45733f862ca1f290dde84d
5266 Author: Chris Wilson <chris@chris-wilson.co.uk>
5267 Date: Wed Jul 4 23:03:44 2012 +0100
5269 sna: drop an unused ret var
5271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5273 commit a69a47f94037698a360f8b8611f6f3e58a20ca87
5274 Author: Chris Wilson <chris@chris-wilson.co.uk>
5275 Date: Wed Jul 4 22:15:45 2012 +0100
5277 sna: Fix detection of EIO through throttle
5279 When using drmIoctl, one needs to check the errno for the actual error
5282 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5284 commit f2fdb0da83d6208c884ffde2bb30d0a67fc75217
5285 Author: Chris Wilson <chris@chris-wilson.co.uk>
5286 Date: Wed Jul 4 21:43:22 2012 +0100
5288 sna: Use throttle() to confirm a GPU hang
5290 This has two effects. The first is that we confirm that the EIO was
5291 indeed a GPU hang and not some other form of I/O failure (for example
5292 swapin). And the second is that it means we also print the warning to
5293 look for an i915_error_state when we detect a GPU hang.
5295 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5297 commit b8c2b34e39c7f2ee2597f8727b9b98c6aac93fb1
5298 Author: Chris Wilson <chris@chris-wilson.co.uk>
5299 Date: Wed Jul 4 19:53:27 2012 +0100
5301 sna: Simplify timer execution
5303 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5305 commit 844ab84c64fde38db315787e16a9f7f674d0f88c
5306 Author: Chris Wilson <chris@chris-wilson.co.uk>
5307 Date: Wed Jul 4 17:58:27 2012 +0100
5309 sna: Fix leak around ENOMEM error path in sna_copy_boxes
5311 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5313 commit 54ee41bf05e328ed2f2cd2a64c21afe3efe2556a
5314 Author: Chris Wilson <chris@chris-wilson.co.uk>
5315 Date: Wed Jul 4 17:40:58 2012 +0100
5317 sna: Pass a hint to the backends when using copy_boxes for readback
5319 If we expect to only emit this set of copy_boxes() and then submit the
5320 batch, we would prefer to use the BLT for its lower latency.
5322 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5324 commit d36623aaebdc454bd281f34332edcad5d94f34c0
5325 Author: Chris Wilson <chris@chris-wilson.co.uk>
5326 Date: Wed Jul 4 13:55:25 2012 +0100
5328 sna/dri: Remove broken code for buffer xchg from AsyncSwap
5330 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5332 commit 0bdb4d0d3693df007a6cfc9a75bb0deddd812d53
5333 Author: Chris Wilson <chris@chris-wilson.co.uk>
5334 Date: Wed Jul 4 13:12:43 2012 +0100
5336 sna/dri: Fixup blit fallback path to use xchg when possible
5338 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5340 commit aeeed323f9081875a804ae19fb356bee2a61e7b0
5341 Author: Chris Wilson <chris@chris-wilson.co.uk>
5342 Date: Wed Jul 4 11:52:27 2012 +0100
5344 sna/dri: Select the appropriate copy engine based on the current src ring
5346 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5348 commit 5026b6a147f411582af24ffd17c87a81da5eea0b
5349 Author: Chris Wilson <chris@chris-wilson.co.uk>
5350 Date: Wed Jul 4 09:39:59 2012 +0100
5352 uxa: Install the drm_wakeup_handler for vblank events in !use_page_flipping
5354 Even if page-flipping itself is disabled, we still want to allow the
5355 client to schedule wakeups for some future vblank which requires
5356 listening to the kernel vblank notifications.
5358 Reported-by: Eric Anholt <eric@anholt.net>
5359 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=51699
5360 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5362 commit 79309dd55f20098e12ead5427c811f237d5592fa
5363 Author: Chris Wilson <chris@chris-wilson.co.uk>
5364 Date: Tue Jul 3 22:27:39 2012 +0100
5366 sna: check for failure to change cache level on the bo
5368 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5370 commit 9e2ac8a413c05aafd1c69775c5b0fa5f1f37ac23
5371 Author: Chris Wilson <chris@chris-wilson.co.uk>
5372 Date: Tue Jul 3 19:33:55 2012 +0100
5374 sna: Use set-cache-level ioctl to create CPU bo
5376 As an alternative to vmap, we can use the kernel for all memory
5377 management through bo, which is much preferred for its simplicity (i.e.
5378 avoiding introducing even more vm complexity).
5380 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5382 commit f26163e916e21f6b64a19c147165f2f2c9c5466d
5383 Author: Chris Wilson <chris@chris-wilson.co.uk>
5384 Date: Tue Jul 3 18:51:46 2012 +0100
5386 sna: Guess when the client is attempting to read back the whole pixmap
5388 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5390 commit c597e6cd4cb4406878eae07c04ad420d8185d99e
5391 Author: Chris Wilson <chris@chris-wilson.co.uk>
5392 Date: Tue Jul 3 13:26:33 2012 +0100
5394 sna: Fix iterator typo
5396 Increment the iterator, not the end-stop.
5398 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5400 commit 2d087eadd9cd3aa2e2ccd73a568286d21702d29a
5401 Author: Chris Wilson <chris@chris-wilson.co.uk>
5402 Date: Tue Jul 3 11:54:02 2012 +0100
5404 sna: Also prefer to use the dst GPU bo if CopyArea reduces to an inplace upload
5406 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5408 commit dbe3a5ca1e6558bcee29d893aba8cc352220a36d
5409 Author: Chris Wilson <chris@chris-wilson.co.uk>
5410 Date: Tue Jul 3 11:48:15 2012 +0100
5412 sna: Attempt to reduce all-damage on the CPU
5414 As we now treat CPU all-damaged as a special case for deciding when and
5415 where to migrate, look out for that condition after adding damage.
5417 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5419 commit 3a41248195e8b327a5d970726450bd2077cdaf0f
5420 Author: Chris Wilson <chris@chris-wilson.co.uk>
5421 Date: Tue Jul 3 11:22:54 2012 +0100
5423 sna: Operate on the original boxes for CopyArea
5425 Be consistent and avoid the confusion when mixing operations on the
5426 region boxes and the original boxes.
5428 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5430 commit e3bc91842bf186ec149c852f48993235ef7ad27f
5431 Author: Chris Wilson <chris@chris-wilson.co.uk>
5432 Date: Tue Jul 3 10:21:06 2012 +0100
5434 sna: Allow booting on older kernels by disabling HW acceleration
5436 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5438 commit affdebcb2d09cd8fdc5aadb2d8df3193587e4a06
5439 Author: Chris Wilson <chris@chris-wilson.co.uk>
5440 Date: Mon Jul 2 18:50:26 2012 +0100
5442 sna: And free the DIR after use
5444 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5446 commit e7b31b6d0a32f76db4a8aef64c77d4afe808fb6c
5447 Author: Chris Wilson <chris@chris-wilson.co.uk>
5448 Date: Mon Jul 2 14:01:36 2012 +0100
5450 sna: Consolidate CopyArea with the aim of reducing migration ping-pong
5452 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5454 commit ce27a81ac508368d54f1237893a9b1214cf3e3d0
5455 Author: Chris Wilson <chris@chris-wilson.co.uk>
5456 Date: Mon Jul 2 17:52:24 2012 +0100
5458 sna: Clear cpu flag after deciding to use gpu bo
5460 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5462 commit 2f1b7e8a23ac3086dda0025ecf09dd1feac94837
5463 Author: Chris Wilson <chris@chris-wilson.co.uk>
5464 Date: Mon Jul 2 17:53:01 2012 +0100
5466 sna: Check for non-existent /sys/class/backlight directory
5468 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5470 commit d12d50d107d403c3cf4dfe24bb63ce1006d0e025
5471 Author: Chris Wilson <chris@chris-wilson.co.uk>
5472 Date: Mon Jul 2 14:09:21 2012 +0100
5474 sna: Skip hidden and special entries inside /sys/class/backlight
5476 Just to avoid the warnings after fallback to comparing the d_name to
5477 the list of known interfaces.
5479 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5481 commit e80f9c4670a0e84521907b1baa059322784b1558
5482 Author: Chris Wilson <chris@chris-wilson.co.uk>
5483 Date: Mon Jul 2 10:21:42 2012 +0100
5485 sna: Prefer backlight iface based on /sys/class/backlight/*/type
5487 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5489 commit 61e16dc5673a1ac96b2ecee072cc3e80971be5d9
5490 Author: Tom Hughes <tom@compton.nu>
5491 Date: Mon Jul 2 10:23:56 2012 +0100
5493 Add asus-nb-wmi backlight control
5495 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=51593
5497 commit 543816011d05f0cf40b05ec58f87a5954397a224
5498 Author: Chris Wilson <chris@chris-wilson.co.uk>
5499 Date: Sun Jul 1 18:36:55 2012 +0100
5501 sna/gen7: Correct two minor misues of equivalent GEN6 constants
5503 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5505 commit 839fea7f779ca63c32817ee4ba695dfce3344980
5506 Author: Chris Wilson <chris@chris-wilson.co.uk>
5507 Date: Sun Jul 1 18:23:38 2012 +0100
5509 sna/gen6: Boost VS entries to maximum
5511 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5513 commit fbd114507d9bf2e2b1d1e52c5e42dc6cdbd8c9a0
5514 Author: Chris Wilson <chris@chris-wilson.co.uk>
5515 Date: Sun Jul 1 14:03:31 2012 +0100
5517 sna/dri: Assert that the replacement bo is large enough for the pixmap
5519 Just another paranoid sanity check.
5521 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5523 commit 675cbd5fade91fd6a6bf533a31b0a211237af6e8
5524 Author: Chris Wilson <chris@chris-wilson.co.uk>
5525 Date: Sun Jul 1 12:39:13 2012 +0100
5527 sna/trapezoids: Skip the division when converting coverage to floating point
5529 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5531 commit 182c3637cc5d3a6ce52127087aa2f19ca2b42719
5532 Author: Chris Wilson <chris@chris-wilson.co.uk>
5533 Date: Sun Jul 1 13:23:44 2012 +0100
5535 sna: If the pixmap is marked for flushing, prefer to use the GPU
5537 Again, to avoid the forced ping-pong as we upload the damage after
5538 nearly every operation, simply prefer to use the GPU in such cases.
5540 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5542 commit 35b1ac138002c206a6d6b866d49a0d73705dd3ac
5543 Author: Chris Wilson <chris@chris-wilson.co.uk>
5544 Date: Sun Jul 1 10:53:07 2012 +0100
5546 sna: After an operation on the CPU, prefer not to use the GPU
5548 A nasty habit of applications is to fill an area, only to read it back
5549 with GetImage, render locally and replace with PutImage. This causes a
5550 readback of an active bo everytime, so let's try to mitigate that by
5551 preferring not to use the GPU after a forced readback.
5553 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5555 commit e625c02e6266403fcd8a72ccce2c6c6291e2e5fc
5556 Author: Chris Wilson <chris@chris-wilson.co.uk>
5557 Date: Sat Jun 30 09:34:21 2012 +0100
5559 sna/damage: Early check for contains-box? if subtract and box outside region
5561 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5563 commit abd7be1cee6f2f494a11cd9d2e7888c3043ffc02
5564 Author: Chris Wilson <chris@chris-wilson.co.uk>
5565 Date: Fri Jun 29 15:32:10 2012 +0100
5567 sna/dri: Prefer GPU rendering if no more CPU damage on a DRI bo
5569 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5571 commit 67b87e4f7cf6c3ab9cfccc9fe43a824bfe84f393
5572 Author: Chris Wilson <chris@chris-wilson.co.uk>
5573 Date: Fri Jun 29 15:03:12 2012 +0100
5575 sna/dri: Optimise clip reduction with copy-to-front to an unclipped Window
5577 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5579 commit eae5e1275cd11703de2bca67dacd1d57129b561a
5580 Author: Chris Wilson <chris@chris-wilson.co.uk>
5581 Date: Fri Jun 29 13:12:52 2012 +0100
5583 sna: Install the ModeSet handler as the base handler
5585 This way we can safely ignore it across server regen.
5587 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5589 commit 15a0761cad862a5d73bbc2af81bc5267e66c307e
5590 Author: Chris Wilson <chris@chris-wilson.co.uk>
5591 Date: Fri Jun 29 12:22:55 2012 +0100
5593 sna: Only consider the request list when deciding whether the GPU is busy
5595 Micro-optimisation to overhead extra checks and to make sure an
5596 unflushed bo doesn't prevent us from submitting more work before
5599 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5601 commit 4061f05dd61a7200d91d30b9b4a2ab0c61306870
5602 Author: Chris Wilson <chris@chris-wilson.co.uk>
5603 Date: Fri Jun 29 11:40:18 2012 +0100
5605 sna/trapezoids: Write unaligned fallback boxes inplace
5607 As this is a pure write operation (though we will write the edge pixels
5608 twice) we can perform this operation inplace and incur a slightly slower
5609 trap creation at the benefit of avoiding the later copy.
5611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5613 commit 44e41536b7728c03a4899a97242960ae7ed35624
5614 Author: Chris Wilson <chris@chris-wilson.co.uk>
5615 Date: Fri Jun 29 11:39:47 2012 +0100
5617 sna/trapezoids: Render the partial left-edge of fallback unaligned boxes
5619 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5621 commit e6f9bfe1e20c4c76822d77598f4f1c7e10754929
5622 Author: Chris Wilson <chris@chris-wilson.co.uk>
5623 Date: Fri Jun 29 11:06:33 2012 +0100
5625 sna: Use currentTime rather than GetTimeInMillis()
5627 The overhead of reading the hpet() on every block handler (more or less)
5628 is exorbitant, so trust that we update currentTime frequently enough to
5629 be a good approximation - the side effect is that we will wakeup
5630 slightly to earlier from using an old value for the current time.
5632 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5634 commit c6c4f30e194eb3b8a2421134bb005f1542b3ef9f
5635 Author: Chris Wilson <chris@chris-wilson.co.uk>
5636 Date: Fri Jun 29 10:14:52 2012 +0100
5638 sna: Add assertions to check that we do install the timers
5640 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5642 commit 87c8f5a47e3abb7ff887aab45b6389cf9b8d8008
5643 Author: Chris Wilson <chris@chris-wilson.co.uk>
5644 Date: Fri Jun 29 09:33:09 2012 +0100
5646 sna: Make the post-flip delay explictit
5648 As the kernel is inconsistent in enforcing this across generations,
5649 handle the synchronisation of the pageflip explicity. Ultimately this
5650 should be replaced with a tripple buffer mechanism.
5652 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5654 commit 31caa43a21a68174386682fd558c911c7c03d76f
5655 Author: Chris Wilson <chris@chris-wilson.co.uk>
5656 Date: Thu Jun 28 17:05:46 2012 +0100
5658 sna/gen5: Check harder for need_upload() fallbacks
5660 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5662 commit 7c3eb1fda9e567a300a8138acdac9ec11e010b11
5663 Author: Chris Wilson <chris@chris-wilson.co.uk>
5664 Date: Thu Jun 28 17:04:26 2012 +0100
5666 sna: Correct inverted logic for checking xrgb drawables
5668 Reported-by: Christoph Reiter <reiter.christoph@gmail.com>
5669 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=51472
5670 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5672 commit c3e2c1332d8d5a3944df99cc11aa66c586add3e8
5673 Author: Chris Wilson <chris@chris-wilson.co.uk>
5674 Date: Thu Jun 28 12:34:36 2012 +0100
5676 sna: Fix the application of the crtc offset for posting damage
5678 The damage boxes are in framebuffer (source) space, so we need to apply
5679 the offset for the boxes in crtc (destination) space.
5681 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5683 commit 47e6bfa4f40cf7efcfe7eee24d2512d737fd7e89
5684 Author: Chris Wilson <chris@chris-wilson.co.uk>
5685 Date: Thu Jun 28 12:24:27 2012 +0100
5687 sna: Force use of per-crtc scanout if the offset is too large
5689 On gen4+, the scanout offset into a tiled surface is specified through
5690 the DSPTILEOFF register and limited to 12bits of precision. So if we
5691 have a CRTC positioned in that nether-region, we need to allocate a
5692 separate per-crtc pixmap for it and perform shadowing.
5694 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5696 commit 93e77ee019248fe77483e83f2210d584bb5d1be2
5697 Author: Chris Wilson <chris@chris-wilson.co.uk>
5698 Date: Thu Jun 28 12:02:32 2012 +0100
5700 sna: Quieten kernel debug complaints when disabling crtc
5702 Even if we are obviously turning the crtc off, it still complains if the
5703 number of connectors is non-zero. So make it so.
5705 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5707 commit 85e4f48a87ddbc227af8f4af5ea46ae17902b111
5708 Author: Chris Wilson <chris@chris-wilson.co.uk>
5709 Date: Thu Jun 28 11:44:45 2012 +0100
5711 sna: Add a DBG to the periodic flush mechanism
5713 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5715 commit 87dd6408a5c29e4808283df78a981de0a3c0a79c
5716 Author: Chris Wilson <chris@chris-wilson.co.uk>
5717 Date: Thu Jun 28 10:53:59 2012 +0100
5719 sna: Correct the reversal of the periodic flushing semantics
5721 Regression from 1e9319d (sna: extend RandR to support super sized
5722 monitor configurations) which tried to take into account the need to
5723 flush the shadow CRTC bo in addition to the normal scanout bo. In the
5724 refactoring of the need_flush(), the double negative was missed.
5726 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
5727 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5729 commit 05f486f64bc7ea4a8a71f5d792fa586ac0843414
5730 Author: Chris Wilson <chris@chris-wilson.co.uk>
5731 Date: Thu Jun 28 10:42:21 2012 +0100
5733 sna: Flush the per-crtc render caches for rotated scanouts
5735 We need to manually flush the render cache in order for results to be
5736 visible on the scanout.
5738 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5740 commit db7979981032d5647069f213a063dcbf78e57890
5741 Author: Chris Wilson <chris@chris-wilson.co.uk>
5742 Date: Wed Jun 27 09:36:58 2012 +0100
5744 sna: s/width/height/ cut'n'paste typo
5746 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
5747 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5749 commit fcbbe1664ae62ae275068cc040f385d2320d5213
5750 Author: Chris Wilson <chris@chris-wilson.co.uk>
5751 Date: Wed Jun 27 09:30:44 2012 +0100
5753 sna: Remove a trailing ';'
5755 The unwanted ';' caused is_cpu() to always return false if a GPU bo was
5756 attached. Not necessary a bad thing, just misses the potential
5757 optimisation where having chosen to prefer to use the CPU path we then
5758 have to migrate to the GPU even though the bo is undamaged or idle.
5760 Spotted-by: Zdenek Kabelac <zkabelac@redhat.com>
5761 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5763 commit a072ab506569ecff5b4c57fa90f7a417db69f33b
5764 Author: Chris Wilson <chris@chris-wilson.co.uk>
5765 Date: Tue Jun 26 17:24:37 2012 +0100
5767 test: Add client side copy to FakeFront for emulating CopyBuffer correctly
5769 The server manages FakeFront following a flip, but it the client
5770 optimises a swap by replacing it with a CopyRegion, it is expected to
5771 also update the FakeFront itself. Replicate that behaviour so that the
5772 timings for the test case are consistent with mesa.
5774 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5776 commit 96804c74f8f0d19fc1b9db467cf050f4934fe83b
5777 Author: Chris Wilson <chris@chris-wilson.co.uk>
5778 Date: Tue Jun 26 17:09:45 2012 +0100
5780 test: FakeFront rules
5782 Oh my, I just once again rediscovered the copy on every flip due to the
5783 requirement for keeping FakeFront uptodate for reads after a SwapBuffers.
5785 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5787 commit f306cd557ef263ff5057e413c335cc75bc0f7895
5788 Author: Chris Wilson <chris@chris-wilson.co.uk>
5789 Date: Tue Jun 26 14:59:37 2012 +0100
5791 sna/dri: Hold a reference to the cached DRI2 buffer on the front buffer
5793 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5795 commit a87f2b9325bfad2bb3f93226706b6f9a09598945
5796 Author: Chris Wilson <chris@chris-wilson.co.uk>
5797 Date: Sun Jun 24 13:48:38 2012 +0100
5799 sna/gen4: Check for peculiar initial values for the surface offset
5801 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5803 commit 8f4221a2520b7ed1f67b41185c2e4842e83e53ef
5804 Author: Chris Wilson <chris@chris-wilson.co.uk>
5805 Date: Tue Jun 26 15:56:50 2012 +0100
5807 test: Add a simple exercise for DRI2 swap paths
5809 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5811 commit a505015a254d6c6e24f0542bc141cde873dc6f34
5812 Author: Chris Wilson <chris@chris-wilson.co.uk>
5813 Date: Mon Jun 25 22:50:47 2012 +0100
5815 sna: Force DPMS to be on following a modeset
5817 Similarly to UXA, this papers over inconsistent behaviour in the kernel
5818 in handling the DPMS upon a modeswitch.
5820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5822 commit b7a8c94cdb9cf42a31f8ce128d70e23458ba2042
5823 Author: Chris Wilson <chris@chris-wilson.co.uk>
5824 Date: Mon Jun 25 20:49:59 2012 +0100
5826 sna: remove the assert(0)s along error paths
5828 This were there as a debugging aide to see if we ever hit unreachable
5829 code paths - mainly along corruption inducing GPU wedged recovery paths.
5830 They are superfluous and just scare the reader.
5832 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5834 commit 15c0ee445f603033c82f357fedfc7737d198d7b3
5835 Author: Chris Wilson <chris@chris-wilson.co.uk>
5836 Date: Sun Jun 24 19:28:03 2012 +0100
5838 sna/gen5: Tweak thread allocations
5840 Bump the alloted number of threads to their max. Using more threads than
5841 cores helps hide the stalls due to sampler fetch, math functions and urb
5842 write. Specifying too many threads seems to not incur a performance
5843 regression, suggesting that the hardware scheduler is sane enough not to
5844 overpopulate the EU.
5846 A small but significant boost, peak x11perf -aa10text on an i3-330m is
5847 raised from 1.93Mglyphs/s to 2.35Mglyphs/s.
5849 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5851 commit fa10005ce31483827547b7f71eae066899f0026c
5852 Author: Chris Wilson <chris@chris-wilson.co.uk>
5853 Date: Sat Jun 23 17:28:00 2012 +0100
5855 sna/dri: Perform an exchange for a composited windowed SwapBuffers
5857 If the front buffer is not attached to the scanout and has not been
5858 reparented, we can simply exchange the underlying bo between the
5859 front/back attachments and inform the compositor of the damage.
5861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5863 commit 53d735ddb16b0204662b8584aa22998ba53deec1
5864 Author: Chris Wilson <chris@chris-wilson.co.uk>
5865 Date: Sat Jun 23 10:16:44 2012 +0100
5867 sna/dri: Queue windowed swaps
5869 Implement "tripple-buffering" for windowed SwapBuffers by allowing the
5870 client to submit one extra frame before throttling. That is we emit the
5871 vsync'ed blit and immediately unblock the client so that it renders to
5872 the GPU (which is guaranteed to be executed after the blit so that its
5873 Front/Back buffers are still correct) and requests another SwapBuffers.
5874 The subsequent swapbuffers are appended to the vsync chain with the
5875 blit/unblock then executed on the vblank following the original blit.
5876 That is both the client and xserver render concurrently.
5878 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5880 commit 1e9319d5f56583be99f573f208cebb0ee3b5cc26
5881 Author: Chris Wilson <chris@chris-wilson.co.uk>
5882 Date: Fri Jun 22 11:22:16 2012 +0100
5884 sna: extend RandR to support super sized monitor configurations
5886 With the introduction of the third pipe on IvyBridge it is possible to
5887 encounter situations where the combination of the three monitors exceed
5888 the limits of the scanout engine and so prevent them being used at their
5889 native resolutions. (It is conceivable to hit similar issues on earlier
5890 generation, especially gen2/3.) One workaround, this patch, is to extend
5891 the RandR shadow support to break the extended framebuffer into per-crtc
5894 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5896 commit e8b090902e788257610374deae659f01a91888f3
5897 Author: Chris Wilson <chris@chris-wilson.co.uk>
5898 Date: Fri Jun 22 22:01:37 2012 +0100
5900 sna/gen3+: Remove stale assertions for cached vbo
5902 Following the previous commit, we reset the vbo when it becomes idle
5903 rather than discard it. As such, the assertions to check that we are
5904 discarding the vbo are now bogus.
5906 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5908 commit 565297e6bd3457a150036af9c62fe0dc67b794ac
5909 Author: Chris Wilson <chris@chris-wilson.co.uk>
5910 Date: Thu Jun 21 13:53:02 2012 +0100
5912 sna/gen3+: Keep vbo cached
5914 Once we switch to using a vbo, keep it cached (resetting everytime it is
5915 idle) until we expire our caches.
5917 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5919 commit d806973e21cd46e605b3cd405323ae7a64c12798
5920 Author: Chris Wilson <chris@chris-wilson.co.uk>
5921 Date: Thu Jun 21 12:57:13 2012 +0100
5923 sna: Micro-optimise search_inactive_cache
5925 Discard the unneeded next parameter to drop a memory reference in a hot
5926 path, and don't wait for a retirement if we are looking in a larger
5929 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5931 commit d39fef0a7f3daf5c07686b44e4dea01c0f06c77a
5932 Author: Chris Wilson <chris@chris-wilson.co.uk>
5933 Date: Thu Jun 21 12:25:35 2012 +0100
5935 sna: Tiles are only 128 bytes wide on gen2
5937 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5939 commit 4f2dde1fa3b04b27bae8fc0bca9c824bd362d23b
5940 Author: Chris Wilson <chris@chris-wilson.co.uk>
5941 Date: Thu Jun 21 10:31:24 2012 +0100
5943 sna/gen7: Eliminate the pipeline stall after a non-pipelined operation
5945 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5947 commit 3ef05a8d0833203e265aff392f225a11a11c2d01
5948 Author: Chris Wilson <chris@chris-wilson.co.uk>
5949 Date: Thu Jun 21 09:36:42 2012 +0100
5951 sna/gen7: Do not emit a pipeline stall after a non-pipelined command
5953 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5955 commit 4501e131e6b737cb8f2581c8b1f7ea9d29a8e912
5956 Author: Chris Wilson <chris@chris-wilson.co.uk>
5957 Date: Thu Jun 21 16:38:32 2012 +0100
5959 sna/gen7: prefer using RENDER copy
5961 Further testing and the balance of doubt swings in favour of using the
5962 3D pipeline for copies.
5964 For small copies the BLT unit is faster,
5965 2.14M/sec vs 1.71M/sec for comppixwin10
5967 And for large copies the RENDER pipeline is faster,
5968 13000/sec vs 8000/sec for comppixwin500
5970 I think the implication is that we are not efficiently utilising the EU
5971 for small primitives - i.e. something that we might be able to improve.
5973 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5975 commit 3da56c48b7820ec77d704c5a16670eb86a6f673f
5976 Author: Chris Wilson <chris@chris-wilson.co.uk>
5977 Date: Wed Jun 20 15:14:23 2012 +0100
5979 sna/gen7: Prefer using BLT rather than redirect for copies
5981 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5983 commit b1f8386db6e9b3eea1bdbf8cde90f33792640ce8
5984 Author: Chris Wilson <chris@chris-wilson.co.uk>
5985 Date: Wed Jun 20 23:28:14 2012 +0100
5987 sna/gen7: Emit a pipeline flush after every render operation
5989 For whatever reason, this produces a 30% improvement with the fish-demo
5990 (500 -> 660 fps on i7-3730qm at 1024x768). However, it does cause about
5991 a 5% regression in aa10text. We can appear to alleviate that by only
5992 doing the flush when the composite op != PictOpSrc.
5994 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
5996 commit d02e6d81420a114c9622bbdaf90fc3ae5d4b15a7
5997 Author: Chris Wilson <chris@chris-wilson.co.uk>
5998 Date: Thu Jun 21 16:10:02 2012 +0100
6000 Encode the third pipe using the HIGH_CRTC shift for vblanks
6002 The original vblank interface only understood 2 pipes (primary and
6003 secondary) and so selecting the third pipe (introduced with IvyBridge)
6004 requires use of the HIGH_CRTC. Using the second pipe where we meant the
6005 third pipe could result in some spurious timings when waiting on the
6008 Reported-by: Adam Jackson <ajax@redhat.com>
6009 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6011 commit f8b67be8d3b5d5cbe730ba5dba3edd2d30a99d9f
6012 Author: Chris Wilson <chris@chris-wilson.co.uk>
6013 Date: Wed Jun 20 12:39:19 2012 +0100
6015 sna: Don't clear the needs_flush flag after emitting a flush on the busy bo
6017 We use that flag to check whether we need to check whether the bo is
6018 still busy upon destruction, so only clear it if the bo is marked as
6021 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6023 commit 5419bbb483af595d7021e49debef7700c599217e
6024 Author: Chris Wilson <chris@chris-wilson.co.uk>
6025 Date: Wed Jun 20 11:21:52 2012 +0100
6027 sna/gen7: Prefer BLT for copies
6029 It's faster for where the cost of the extra batches and ring switching
6032 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6034 commit 1c0bb8c4c93a43e2932429a9f6c23d91f1be2060
6035 Author: Chris Wilson <chris@chris-wilson.co.uk>
6036 Date: Wed Jun 20 10:57:40 2012 +0100
6038 sna/gen7: Keep using RENDER paths for large pixmaps
6040 As the 3D pipeline is quite versatile and we only need to force BLT if
6041 we cannot extract the subregion.
6043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6045 commit b238f64e8a53883cfe7c568e37ef18bbee77996e
6046 Author: Chris Wilson <chris@chris-wilson.co.uk>
6047 Date: Wed Jun 20 10:46:59 2012 +0100
6049 sna/gen[67]: Prefer to not force BLT paths for large pixmaps
6051 The sampler can in fact handler subregions of large pixmaps quite well,
6052 and so we prefer to keep using the 3D pipeline so long as the operation
6053 fits in. If not, then switch to the BLT in order to avoid the temporary
6056 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6058 commit 38f06a351fdd35fa5c6136e9a25e034a812a94c2
6059 Author: Chris Wilson <chris@chris-wilson.co.uk>
6060 Date: Tue Jun 19 22:22:12 2012 +0100
6062 uxa: Fix second regression in glyph fallback from 64a4bc
6064 To complete my show of incompetence for the evening, not only do we have
6065 to restore the original source when compositing the mask onto the
6066 destination, we also need to restore the original dst (rather than
6067 composite the mask onto the mask!).
6069 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6071 commit fda9faee755cb35906ca1179a568332ef2de35a5
6072 Author: Chris Wilson <chris@chris-wilson.co.uk>
6073 Date: Tue Jun 19 21:01:47 2012 +0100
6075 uxa: Use the original src for fallback glyph compositing
6077 In 64a4bcb8cefff, I introduced a WHITE source for the purposes of
6078 accumulating the glyph mask correctly. Unfortunately I neglected to
6079 restore the original source picture for compositing the glyph mask on
6080 the destination, resulting in a use-after-free and then corruption.
6082 Reported-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
6083 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6085 commit 8141e290b1ac7c4d1524bb389a84a8f375df4634
6086 Author: Chris Wilson <chris@chris-wilson.co.uk>
6087 Date: Tue Jun 19 20:24:33 2012 +0100
6089 sna: Explain why we ignore the busy status result during kgem_bo_flush()
6091 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6093 commit eb1d07624e5aeb7b5db7ceb46975091e92185d63
6094 Author: Chris Wilson <chris@chris-wilson.co.uk>
6095 Date: Tue Jun 19 16:00:13 2012 +0100
6097 sna: Ensure extents is initialised if short-circuit use-cpu-bo
6099 As we may attempt to end up using the GPU bo is the CPU bo is busy, we
6100 need to make sure we have initialised the damage extents first.
6102 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6103 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6105 commit 9f216e159bd05ddd269eb8ddf3ca7a407c2901e2
6106 Author: Chris Wilson <chris@chris-wilson.co.uk>
6107 Date: Tue Jun 19 15:57:31 2012 +0100
6109 sna: Assert expected return values
6111 Keep the semantic analyser happy by consuming the expected return value
6114 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6115 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6117 commit 2dc93b2a6c832ce8b972de90d09080e860dcd40f
6118 Author: Chris Wilson <chris@chris-wilson.co.uk>
6119 Date: Tue Jun 19 15:34:09 2012 +0100
6121 sna: Check results from syscalls
6123 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6124 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6126 commit 06634604abf15fdd27dd007fcf81595da994146b
6127 Author: Chris Wilson <chris@chris-wilson.co.uk>
6128 Date: Tue Jun 19 15:28:43 2012 +0100
6130 Initialise adaptors to 0 in case xf86XVListGenericAdaptors does not
6132 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6133 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6135 commit 8bfea58dbc634cadc399d3132030c591e086880c
6136 Author: Chris Wilson <chris@chris-wilson.co.uk>
6137 Date: Tue Jun 19 15:26:18 2012 +0100
6139 sna: Minor cleanups from sematic analyser in DBG
6141 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6142 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6144 commit 0a43d425670b883b04565296c0510e7ba03ba6de
6145 Author: Chris Wilson <chris@chris-wilson.co.uk>
6146 Date: Tue Jun 19 14:46:58 2012 +0100
6148 uxa: Implement glyphs-to-dst to avoid fallbacks
6150 An earlier version was buggy and introduced corruption as it failed to
6151 fallback gracefully with ComponentAlpha glpyhs. This is a much simpler
6152 implementation that composites each glyph individually, leaving it to the
6153 backend to optimise away state changes. It should still be many times
6154 faster than incurring the fallback...
6156 Reported-by: Oleksandr Natalenko <pfactum@gmail.com>
6157 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50508
6158 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6160 commit 64a4bcb8ceffff8e4ee448d8a467620596703509
6161 Author: Chris Wilson <chris@chris-wilson.co.uk>
6162 Date: Tue Jun 19 13:42:12 2012 +0100
6164 uxa: Use (white IN glyph) ADD mask to compose the glyph mask
6166 As pointed out by Soren Sandmann and Behdad Esfahbod, it is essential to
6167 use white IN glyph when adding to the mask so that the channel expansion
6168 is correctly performed when adding to an incompatible mask format.
6170 For example, loading alpha as the source results in the value 000a being
6171 added to the rgba glyph mask (for mixed subpixel rendering with
6172 grayscale glyphs), whereas the desired value is aaaa.
6174 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6176 commit 99845dcb3ba862269b29aec782e2bcef31c0403e
6177 Author: Chris Wilson <chris@chris-wilson.co.uk>
6178 Date: Tue Jun 19 10:37:10 2012 +0100
6180 Post Damage on the Screen Pixmap after a pageflip
6182 This issue was raised by Dave Airlie as he is trying to integrate
6183 multiple GPUs into the xserver, and a particular setup has a slave
6184 rendering device that copies the contents from the GPU over a
6185 DisplayLink USB adaptor. As such the slave device is listening for
6186 Damage on the Screen Pixmap and needs the update following pageflips.
6187 Since we already are posting damage for all the SwapBuffers paths other
6188 than pageflip, for consistency we should post damage along the pageflip
6191 Reported-by: Dave Airlie <airlied@redhat.com>
6192 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6194 commit 4acf727941a37651f672db3774281a437940c540
6195 Author: Chris Wilson <chris@chris-wilson.co.uk>
6196 Date: Tue Jun 19 10:20:54 2012 +0100
6198 sna: Initialize the color value for fallback unaligned boxes
6200 Reported-by:Zdenek Kabelac <zkabelac@redhat.com>
6201 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=5047
6202 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6204 commit b0b2d3c9663c29e9844aef1608416ae40c194b55
6205 Author: Chris Wilson <chris@chris-wilson.co.uk>
6206 Date: Tue Jun 19 00:41:35 2012 +0100
6208 sna: Avoid copying unintialised data during source picture upload
6210 If we have never written to a pixmap, then there will be neither a GPU
6211 or shadow pointer and we would attempt to copy a NULL pointer. In this
6212 case as the user is expecting to copy unintialised data we are at
6213 liberty to replace those undefined values with the clear color.
6215 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6217 commit 38472fcc53c5dceb98b96458183e6729b8311a43
6218 Author: Chris Wilson <chris@chris-wilson.co.uk>
6219 Date: Tue Jun 19 00:37:12 2012 +0100
6221 sna: Double check that the source is busy before performing indirect reads
6223 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6225 commit 8cdfb8c24c8b49c88451714d80293c66d63e8c01
6226 Author: Chris Wilson <chris@chris-wilson.co.uk>
6227 Date: Tue Jun 19 00:36:41 2012 +0100
6229 sna: Fix up the shadow pointer on the source when copying
6231 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6233 commit 17f3a83fdc8c0ef5c12fb4be34d86021c0c865e5
6234 Author: Chris Wilson <chris@chris-wilson.co.uk>
6235 Date: Mon Jun 18 23:48:16 2012 +0100
6237 sna: Review sna_copy_boxes
6239 A couple of ordering issue and more assertions.
6241 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6243 commit a9045699b9cd66d0b0d96bfc964458c96845f97f
6244 Author: Chris Wilson <chris@chris-wilson.co.uk>
6245 Date: Mon Jun 18 23:47:21 2012 +0100
6247 sna: Reset region after transferring to cpu
6249 If we adjust the region for the pixmap offset, be sure that we reset it
6250 before returning it back to the caller.
6252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6254 commit 9f51311a7d95bf4bc23926b8e6bf8ee52afd180c
6255 Author: Chris Wilson <chris@chris-wilson.co.uk>
6256 Date: Mon Jun 18 23:45:50 2012 +0100
6258 sna: Check if the busy is truly busy before commiting to an indirect upload
6260 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6262 commit 291b3c4367b455bfc5a772f8caaeee73f36d826a
6263 Author: Chris Wilson <chris@chris-wilson.co.uk>
6264 Date: Mon Jun 18 23:43:00 2012 +0100
6266 sna: Align upload buffers to 128
6268 This seems to be a restriction (observed on 965gm at least) that we
6269 have incoherent sampler cache if we write within 128 bytes of a busy
6270 buffer. This is either due to a restriction on neighbouring cachelines
6271 (like the earlier BLT limitations) or an effect of sampler prefetch.
6273 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6274 References: https://bugs.freedesktop.org/show_bug.cgi?id=50477
6275 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6277 commit 39e5c7491535999643c1761bb1602ad757ab486c
6278 Author: Chris Wilson <chris@chris-wilson.co.uk>
6279 Date: Mon Jun 18 21:58:27 2012 +0100
6281 sna: Assert damage is valid after every addition
6283 Even more paranoia than just checking upon migration.
6285 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6287 commit 92e1693e5fb3a1dd89fca5e5ecc660e2de78f9cd
6288 Author: Chris Wilson <chris@chris-wilson.co.uk>
6289 Date: Mon Jun 18 21:26:58 2012 +0100
6291 sna: Validate cpu/gpu damage never overlaps
6293 References: https://bugs.freedesktop.org/show_bug.cgi?id=50477
6294 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6296 commit d2312c8f958002e54ddcb834f37916f4b46ac291
6297 Author: Chris Wilson <chris@chris-wilson.co.uk>
6298 Date: Mon Jun 18 21:29:29 2012 +0100
6300 sna: Fixup tracking of vmap upload buffers
6302 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6304 commit 75e9eeca7e998b1ee3f8b0df780adae1b9e5b408
6305 Author: Chris Wilson <chris@chris-wilson.co.uk>
6306 Date: Mon Jun 18 16:39:20 2012 +0100
6308 sna: Remove overlapping CPU damage when operating inplace on the GPU
6310 Otherwise we gradually introduce garbage into the picture.
6312 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6313 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50477
6314 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6316 commit a936466dd41c4bb8b0956c4da5a5838f991b332a
6317 Author: Chris Wilson <chris@chris-wilson.co.uk>
6318 Date: Mon Jun 18 11:36:53 2012 +0100
6320 sna: Prefer to attempt a Composite operation rather than use pixman composite
6322 As pixman composite performance is atrocious for anything other than
6323 solids, prefer to upload the mask and attempt a composite operation on
6324 the GPU unless we are forcing the fallback.
6326 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6328 commit 4b325d6e2b99ec3f63a7c5175f6f4746f6773550
6329 Author: Chris Wilson <chris@chris-wilson.co.uk>
6330 Date: Mon Jun 18 11:24:55 2012 +0100
6332 sna: Fix rendering of unaligned boxes through pixman
6334 Not only do we need to make sure the source is available to the CPU, we
6335 need to actually check the right conditions for clipping the box.
6337 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6339 commit caef27492b5a70bd46a09cd5094b2c93e38e39ea
6340 Author: Chris Wilson <chris@chris-wilson.co.uk>
6341 Date: Sun Jun 17 20:56:46 2012 +0100
6343 sna: convert another instance of applying the clear to the CPU pixmap
6345 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6347 commit 8695c4c77666cd07eab51efcbc7c4f11c85250fd
6348 Author: Chris Wilson <chris@chris-wilson.co.uk>
6349 Date: Sun Jun 17 16:53:53 2012 +0100
6351 sna: Fix the blt composite op with no-ops
6353 When returning early because the operation is a no-op, we still need to
6354 fill in the function pointers to prevent a later NULL dereference.
6356 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6358 commit 7905ddae1dbc8805d0fadbd6d21c7a5df7e715fc
6359 Author: Chris Wilson <chris@chris-wilson.co.uk>
6360 Date: Sun Jun 17 17:01:12 2012 +0100
6362 sna: Further refine choice of placement when uploading source data.
6364 The goal is cheaply spot a simple copy operation that can be performed
6365 on the CPU without having to load both parties onto the GPU.
6367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6369 commit 5a675b61f27273f7ef344d4056dbba1f8cd97737
6370 Author: Chris Wilson <chris@chris-wilson.co.uk>
6371 Date: Sun Jun 17 15:42:17 2012 +0100
6373 sna: Correct typo forcing everything to be clear to 0!
6375 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6377 commit b55bf1abbe71281e3d9ebde1c4005d40902d5e7f
6378 Author: Chris Wilson <chris@chris-wilson.co.uk>
6379 Date: Sun Jun 17 14:14:44 2012 +0100
6381 sna: Fix cut'n'paste errors in tiling debug
6383 Rename for different variables
6385 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6387 commit 9756c60b4ad15281d025b9c27f19d19e8a630958
6388 Author: Chris Wilson <chris@chris-wilson.co.uk>
6389 Date: Sun Jun 17 11:39:33 2012 +0100
6391 sna/gen7: Enable non-rectilinear spans
6393 Seems we have enough GPU power to overcome the clumsy shaders. Just
6394 imagine the possibilities when we have a true shader for spans...
6396 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6398 commit 41aff56a1f452e409c7a49512a1d2824b74f3838
6399 Author: Chris Wilson <chris@chris-wilson.co.uk>
6400 Date: Sun Jun 17 10:50:31 2012 +0100
6402 sna: Add tiling for spans
6404 Semmingly only advisable when already committed to using the GPU. This
6405 first pass is still a little naive as it makes no attempt to avoid empty
6406 tiles, nor aims to be efficient.
6408 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6410 commit 222e6ff43ef683e82101fb360911fc01fbe00597
6411 Author: Chris Wilson <chris@chris-wilson.co.uk>
6412 Date: Sat Jun 16 23:40:30 2012 +0100
6414 sna: Read inplace for fallback copies
6416 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6418 commit 79d468925bb012806e097337e4e5930818c6ab46
6419 Author: Chris Wilson <chris@chris-wilson.co.uk>
6420 Date: Sat Jun 16 17:54:21 2012 +0100
6422 sna: Decrease latency for 1x1 GetImage by using an inplace mapping
6424 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6426 commit 2c2a8d3780f1d8de3f13bee8e068fdaf608ff9e9
6427 Author: Chris Wilson <chris@chris-wilson.co.uk>
6428 Date: Sat Jun 16 23:19:49 2012 +0100
6430 sna: Allow reads to be performed inplace
6432 If we can guess that we will only readback the data once, then we can
6433 skip the copy into the shadow.
6435 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6437 commit bc6997f6f751d3ba352dfc20c6717ec12b8fac47
6438 Author: Chris Wilson <chris@chris-wilson.co.uk>
6439 Date: Sat Jun 16 17:54:15 2012 +0100
6441 sna: Cleanup damage processing after operating inplace
6443 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6445 commit 937ca8a5d8a0f70a0724db1519bb7b5fc0857425
6446 Author: Chris Wilson <chris@chris-wilson.co.uk>
6447 Date: Sat Jun 16 17:53:58 2012 +0100
6449 sna: Use memset for simple clears
6451 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6453 commit de4572b0b52e2fcfcad04660ee2f81ee88d500a5
6454 Author: Chris Wilson <chris@chris-wilson.co.uk>
6455 Date: Sat Jun 16 13:40:26 2012 +0100
6457 sna: Inspect CPU damaged state when deciding upon Composite placement
6459 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6461 commit b689cd924c500373e1e293dd9eb54a238e400381
6462 Author: Chris Wilson <chris@chris-wilson.co.uk>
6463 Date: Fri Jun 15 16:25:51 2012 +0100
6465 sna: Composite traps inplace if the CPU is already all-damaged
6467 One outcome is that inspecting the usage patterns afterwards indicated
6468 that we were missing an opportunity to reduce unaligned boxes to an
6471 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6473 commit ae3c0963790cfb6f984ed4ad3ecbaae492775e1b
6474 Author: Chris Wilson <chris@chris-wilson.co.uk>
6475 Date: Fri Jun 15 16:25:51 2012 +0100
6477 sna: Composite glyphs inplace if the CPU is already all-damaged
6479 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6481 commit eaed58b2baf30eaea37be06cfc1d9d81059aba27
6482 Author: Chris Wilson <chris@chris-wilson.co.uk>
6483 Date: Sat Jun 16 12:55:54 2012 +0100
6485 sna: Tweak placement of operations
6487 Take in account busyness of the damaged GPU bo for considering placement
6488 of the subsequent operations. In particular, note that is_cpu is only
6489 used for when we feel like the following operation would be better on
6490 the CPU and just want to confirm that doing so will not stall.
6492 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6494 commit 8eac098962891a5deb7c53d36c6dec57c7f2b972
6495 Author: Chris Wilson <chris@chris-wilson.co.uk>
6496 Date: Sun Jun 17 09:51:12 2012 +0100
6498 sna/gen3: Add another DBG for dropping vbo
6500 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6502 commit 515c8b19d638d4a811b159ef0dc7cf4059e30217
6503 Author: Chris Wilson <chris@chris-wilson.co.uk>
6504 Date: Fri Jun 15 15:41:14 2012 +0100
6506 sna: Prefer to operate inplace if already mapped of the GPU is wholly dirty
6508 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6510 commit d1713941e9db3e7a6d83466be1b253978fb4bf01
6511 Author: Chris Wilson <chris@chris-wilson.co.uk>
6512 Date: Fri Jun 15 15:29:59 2012 +0100
6514 sna: Tweaks for DBG missing glyphs through fallbacks
6516 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6518 commit 2b23605efba009fb340ec10b37d54caae159b9b1
6519 Author: Chris Wilson <chris@chris-wilson.co.uk>
6520 Date: Fri Jun 15 15:28:31 2012 +0100
6522 sna: Don't trim prepare for glyphs_via_mask
6524 If we pass the expected width/height without passing the per-glyph
6525 offset into the preparation function, we make the erroneous mistake of
6526 analysing the glyph cache only for the mask extents and so will miss
6527 glyphs that we need to upload for the operation.
6529 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6531 commit 9f66b27114fcc457fa5cb2d5889e875384f89e75
6532 Author: Chris Wilson <chris@chris-wilson.co.uk>
6533 Date: Fri Jun 15 13:47:33 2012 +0100
6535 sna: Remove mark-as-cpu after gem_pread
6537 The kernel no longer moves the read bo into the CPU domain, so remove
6538 the last vestiges of that tracking.
6540 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6542 commit bfd88d5429528b0162eafcc9496bfd1c708b60cd
6543 Author: Chris Wilson <chris@chris-wilson.co.uk>
6544 Date: Fri Jun 15 11:21:29 2012 +0100
6546 sna: Silence a few compiler warnings
6548 Nothing harmful, just noise.
6550 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6552 commit b3659beec8c4126de5da4df5cb2cc077b7af6c0b
6553 Author: Chris Wilson <chris@chris-wilson.co.uk>
6554 Date: Fri Jun 15 11:08:18 2012 +0100
6556 sna: Check for zero-sized damage regions after trimming Windows to their pixmaps
6558 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6559 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6561 commit 293a867ea55d3004f5be21b1d0ad765a89c28a5a
6562 Author: Chris Wilson <chris@chris-wilson.co.uk>
6563 Date: Thu Jun 14 23:02:10 2012 +0100
6565 sna: Specialise the self-copy blitter to handle vertically overlapping copies
6567 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6569 commit 3b76ab995bb70da5f76afead7c696cb7fd72b158
6570 Author: Chris Wilson <chris@chris-wilson.co.uk>
6571 Date: Thu Jun 14 22:33:11 2012 +0100
6573 sna: Add a little bit more DBG to migration IO paths
6575 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6577 commit 49543522e789a0a8213c164b37a18c0642e60ac4
6578 Author: Chris Wilson <chris@chris-wilson.co.uk>
6579 Date: Thu Jun 14 20:25:21 2012 +0100
6581 sna: Fix reversed assertion
6583 Darn the double negative! After adding damage we want to assert that
6584 the region is not empty, not !not empty.
6586 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6588 commit 4e50467b4f50b522a07bedae7ff68b9c37362af8
6589 Author: Chris Wilson <chris@chris-wilson.co.uk>
6590 Date: Thu Jun 14 17:40:09 2012 +0100
6592 uxa: Force the outputs to off for consistency with xf86DisableUnusedFunctions()
6594 Upon a VT switch, we set the desired modes and turn off the DPMS on any
6595 unused output. Make this explicit so that we always maintain consistency
6596 between the kernel and X's list of enabled CRTCs.
6598 References: https://bugs.freedesktop.org/show_bug.cgi?id=50772
6599 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6601 commit 3a56a0b10eb3ab07efd28efdc79d0fd3a6fa20ff
6602 Author: Chris Wilson <chris@chris-wilson.co.uk>
6603 Date: Thu Jun 14 19:26:42 2012 +0100
6605 sna: Make the disable-unused after vt switch distinct from DPMS off
6607 So that we do not lose control over dpms on/off!
6609 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6611 commit b415ca05c2e1c4f09f85d8769d39e5369661ed3a
6612 Author: Chris Wilson <chris@chris-wilson.co.uk>
6613 Date: Thu Jun 14 17:06:19 2012 +0100
6615 sna: Reset damage extents after reduction goes to zero
6617 If the reduction of the damage clears all of the boxes, we need to reset
6618 the -infinite extents so that we continue to accumulate further damage.
6620 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
6621 References: https://bugs.freedesktop.org/show_bug.cgi?id=50744
6622 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6624 commit 0df7c488640d3590d2a88dc353b72167b6644eaa
6625 Author: Chris Wilson <chris@chris-wilson.co.uk>
6626 Date: Thu Jun 14 16:09:38 2012 +0100
6628 sna: Supporting overlapping copies for fallback blits
6630 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6631 References: https://bugs.freedesktop.org/show_bug.cgi?id=50393
6632 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6634 commit d5b273b20b7d917d08af120815c28b1618d3342c
6635 Author: Chris Wilson <chris@chris-wilson.co.uk>
6636 Date: Thu Jun 14 15:35:24 2012 +0100
6638 sna: Explicitly disable the CRTC so that xserver/kernel agree after VT switch
6640 References: https://bugs.freedesktop.org/show_bug.cgi?id=50772
6641 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6643 commit 221534abe2dc04fae8b8fc332104bca275d4863b
6644 Author: Chris Wilson <chris@chris-wilson.co.uk>
6645 Date: Wed Jun 13 22:48:43 2012 +0100
6647 sna: Only reduce damage addition to a region operation if clean
6649 Some paths bypass operating upon the region as they do not have an
6650 YX-banded set of boxes and so prefer to defer the costly construction of
6651 the region till later. As a result, we have to be careful not to
6652 overwrite any existing information if we do operate on the region after
6653 setting the dirty boxes.
6655 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50744
6656 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
6657 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6659 commit 0db789e180b6b01fb8aff547879387058f52a0b9
6660 Author: Zdenek Kabelac <zkabelac@redhat.com>
6661 Date: Wed Jun 13 14:26:37 2012 +0200
6665 Adding preserving const modifiers to decrease amount of const warnings
6667 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
6668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6670 commit 33998a7080aa7f50ba922c764c6a93fe951c5b64
6671 Author: Zdenek Kabelac <zkabelac@redhat.com>
6672 Date: Wed Jun 13 14:26:36 2012 +0200
6674 sna: Fix typo and compare y1 to y2
6676 It seems 'y1' was meant to be compared here
6678 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
6679 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6681 commit 0ade32fbad4014efcc8aa8ed4cd0f8a39f9ab107
6682 Author: Chris Wilson <chris@chris-wilson.co.uk>
6683 Date: Wed Jun 13 14:28:43 2012 +0100
6685 sna: Fix operator preference around a | b & c | d
6687 Tell the compiler what we really mean is a | (b & (c | d))
6689 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6690 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6692 commit 20f3114ccd63052b4fd06ef1d87efaaabbbac7ac
6693 Author: Chris Wilson <chris@chris-wilson.co.uk>
6694 Date: Wed Jun 13 14:27:10 2012 +0100
6696 sna: Use long for simplicity when printing size_t values for debugging
6698 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6699 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6701 commit 6f75b8da6a468188dd0c00015395424598b3f502
6702 Author: Zdenek Kabelac <zkabelac@redhat.com>
6703 Date: Wed Jun 13 14:26:33 2012 +0200
6705 sna: Valgrind updates
6707 Add some mising initialization for unknown ioctl
6709 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
6710 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6712 commit 70cf637ca47c9e67758e09c7e1d10eef51db1f98
6713 Author: Zdenek Kabelac <zkabelac@redhat.com>
6714 Date: Wed Jun 13 14:26:32 2012 +0200
6716 legacy/i810: Fix declaration to match definition
6718 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
6719 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6721 commit 5fea2478cd5ab4156c182210d28c5e27e5f67403
6722 Author: Chris Wilson <chris@chris-wilson.co.uk>
6723 Date: Wed Jun 13 14:10:01 2012 +0100
6725 Lots of trivial semantic fixes
6727 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6728 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6730 commit e499f207c161d1b3cd75f065dc89021ff5f40b63
6731 Author: Chris Wilson <chris@chris-wilson.co.uk>
6732 Date: Wed Jun 13 13:11:17 2012 +0100
6734 sna: Fix memleak from sna_crtc_find_plane()
6736 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6738 commit 49da55da518348fc6b88e09d5132dd1b1d751304
6739 Author: Chris Wilson <chris@chris-wilson.co.uk>
6740 Date: Wed Jun 13 11:20:41 2012 +0100
6742 sna: Free the reversed glyph bits along with the font
6744 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6746 commit 9f68c7fe6aa34ab17d82489fca5f63a3ce335444
6747 Author: Chris Wilson <chris@chris-wilson.co.uk>
6748 Date: Wed Jun 13 11:15:42 2012 +0100
6750 sna: Free clip rectangles after intersection with PolyRectangle extents
6752 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6754 commit ab3b7fe31b5a9d7924e959f21d29c4f7352ec8cb
6755 Author: Chris Wilson <chris@chris-wilson.co.uk>
6756 Date: Wed Jun 13 10:35:14 2012 +0100
6758 sna: Only reuse a write buffer if all external references have been dropped
6760 This avoids the unhappy situation of overwriting an upload buffer that
6761 we intend to use for a fallback.
6763 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6765 commit e8cb656dc4bab1489df9dfb4cb64704b2a9fa34b
6766 Author: Chris Wilson <chris@chris-wilson.co.uk>
6767 Date: Wed Jun 13 08:10:10 2012 +0100
6769 sna: Free clip intersection with line extents after use
6771 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6773 commit 572745eb243a96819c8fa248f4ffdd703ed13b0f
6774 Author: Chris Wilson <chris@chris-wilson.co.uk>
6775 Date: Tue Jun 12 22:49:05 2012 +0100
6777 sna/damage: Remove unused return value
6779 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6781 commit aefc0417dde8b2137a787459a69d91c14902f22b
6782 Author: Chris Wilson <chris@chris-wilson.co.uk>
6783 Date: Tue Jun 12 22:47:41 2012 +0100
6785 sna/damage: Simplify initialisation of damage extents
6787 Just use the already computed region->extents when adding the first box.
6789 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6791 commit 78a96e812b912c5b25a02670f603f455f93e9d00
6792 Author: Chris Wilson <chris@chris-wilson.co.uk>
6793 Date: Tue Jun 12 21:43:31 2012 +0100
6795 sna: Refresh the damage self-tests
6797 They had bitrotted with the revamp and needed some massaging to update
6798 to the new interfaces.
6800 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6802 commit 6138f7434a6fb014299a7e9a8392ef962c8cba30
6803 Author: Cyril Brulebois <kibi@debian.org>
6804 Date: Tue Jun 12 21:19:14 2012 +0100
6806 Fix up braindamage in previous commit.
6808 ickle: Fixing up my idiotic change, obviously too much birthday cake.
6810 commit 224d631a232d3536e0f1f2780bfbe1218ee7590f
6811 Author: Cyril Brulebois <kibi@debian.org>
6812 Date: Tue Jun 12 21:14:53 2012 +0100
6814 Avoid calling xf86nameCompare() with a NULL string
6816 Device sections without a Driver property would lead to a server
6817 segfault because of a NULL pointer's being passed as the second
6818 argument of xf86nameCompare().
6820 Debian bug #677206 <http://bugs.debian.org/677206>
6822 Signed-off-by: Cyril Brulebois <kibi@debian.org>
6824 commit 3b9b64c7c9b5b0bfaafb97c9a9fe5849bbb412da
6825 Author: Dave Airlie <airlied@gmail.com>
6826 Date: Tue Jun 12 10:26:34 2012 +0100
6828 uxa: do copy fb at startup.
6830 Copy the current framebuffer for smooth wayland->gdm handoff.
6832 This has been hanging around in Fedora for too long now, and we've
6833 dropped the feature a few times, and yes I know the Simpsons did it^W^W^W
6836 I've updated the code to have some of the better fixes from nouveau.
6838 I've no idea who wrote this code either, krh or ajax. [ickle: The
6839 earliest version I've found had krh's fingerprints on it, though it may
6840 still have been a joint effort.]
6842 Signed-off-by: Dave Airlie <airlied@redhat.com>
6843 [ickle: improve error handling, only copy the fb during initial takeover]
6845 commit b6525702b9ffd21beb8ea6bb10a8ad5ce7f9de14
6846 Author: Chris Wilson <chris@chris-wilson.co.uk>
6847 Date: Tue Jun 12 11:59:31 2012 +0100
6849 sna: Hook up AdjustFrame()
6851 Looks to be unused by the core, just a solitary invocation in an obscure
6852 extension it seems. However the implementation looks trivial so
6853 incorporate it until it is finally removed, just in case.
6855 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6857 commit 7b281d1300127d65392aaadbbe2299fa9e1749a9
6858 Author: Chris Wilson <chris@chris-wilson.co.uk>
6859 Date: Tue Jun 12 10:37:04 2012 +0100
6861 sna: Remove a pair of empty stub functions
6863 The export wrappers for the glyph cache constructor/destructor existed
6864 in case there was a need to add more routines. Since that never
6865 happened, remove the extra step of indirection.
6867 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6869 commit 392e33a62d729c64c57699505220b4029e015470
6870 Author: Chris Wilson <chris@chris-wilson.co.uk>
6871 Date: Mon Jun 11 09:51:41 2012 +0100
6873 sna: Instrument memory/bo allocations for monitoring over time
6875 Hide it behind --enable-debug=memory to avoid incurring the cost for
6878 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6880 commit 4e984b79cce038ba5026d8bdd93dec9b5a6ced8a
6881 Author: Chris Wilson <chris@chris-wilson.co.uk>
6882 Date: Mon Jun 11 09:02:03 2012 +0100
6884 configure: Correct the help string for --with-default-accel
6886 Prior to finalizing the defaults I changed my mind and realised that the
6887 default had to reflect the current behaviour of someone enabling SNA for
6888 the first time, and not the previous behaviour of --enable-sna to
6889 override UXA. This is so that distro's could offer an SNA enabled DDX
6890 for the brave whilst not affecting their typical no-xorg.conf users.
6892 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6894 commit 7614a541886dd9eb5c9d441e2618785e9cc53a65
6895 Author: Chris Wilson <chris@chris-wilson.co.uk>
6896 Date: Sun Jun 10 09:36:38 2012 +0100
6898 configure: Harden --with-default-accel against simple mistakes
6900 If the user specifies no options, assume automatic selection. Then
6901 double check we found a valid backend and so avoid later breaking the
6904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6906 commit ea0209fbc12fca282bd8ef72223791bf163801ce
6907 Author: Chris Wilson <chris@chris-wilson.co.uk>
6908 Date: Fri Jun 8 22:44:23 2012 +0100
6910 sna: Remove option to disable vmap
6912 It works everywhere, so when it is finally enabled, let it be.
6914 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6916 commit 18baa1428b1c4c0a49927d4c32db091db9e4e97b
6917 Author: Chris Wilson <chris@chris-wilson.co.uk>
6918 Date: Fri Jun 8 22:43:01 2012 +0100
6920 Make the enum list of options consistent with the actual table.
6922 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6924 commit adc872a9654dc18d778323ca0721704878ad3851
6925 Author: Chris Wilson <chris@chris-wilson.co.uk>
6926 Date: Fri Jun 8 06:44:20 2012 +0100
6928 sna: Add DBG spew to flink()
6930 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6932 commit a62ad4e80722db187766c83a16fa84ec236cc5eb
6933 Author: Chris Wilson <chris@chris-wilson.co.uk>
6934 Date: Thu Jun 7 23:37:59 2012 +0100
6936 sna: Skip flushing the active queue if there is not a suitable bo pending
6938 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6940 commit 318982566bbc7145847bd03601087150eef7a8d8
6941 Author: Chris Wilson <chris@chris-wilson.co.uk>
6942 Date: Thu Jun 7 21:56:04 2012 +0100
6944 sna/dri: Disable experimental code by default
6946 Since these require non-upstream patches to other components, we don't
6947 want it enabled by default and randomly breaking builds.
6949 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6951 commit ebf84b8e572b5cb1a509000d412dfa5be3d1aca3
6952 Author: Chris Wilson <chris@chris-wilson.co.uk>
6953 Date: Thu Jun 7 20:15:53 2012 +0100
6955 sna/trapezoids: Micro-optimise cell allocation
6957 The pool is a fixed size so we can delete the generic code to handle
6960 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6962 commit 0a25fc68c5cd82cad4b99b0f2357f430c8783c3f
6963 Author: Chris Wilson <chris@chris-wilson.co.uk>
6964 Date: Thu Jun 7 19:13:09 2012 +0100
6966 sna: Actually create Y-tiled source pixmaps
6968 An inconsistency highlighted by 7c51cabaecac revealed that we had a
6969 mismatch between the check in move_to_gpu() and how we created the
6970 pixmap. This mismatch resulted in us creating and uploading tiled
6971 pixmaps for single shot textures, and the increase aperture pressure was
6972 causing a regression in firefox-fishbowl on pnv, for example.
6974 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6976 commit c58d137d3eeb0e97bfd53e68404e04d9012b5697
6977 Author: Chris Wilson <chris@chris-wilson.co.uk>
6978 Date: Thu Jun 7 11:18:27 2012 +0100
6980 sna: Suppress DBG prints along potential SIGIO paths
6982 As ErrorF/fprintf is not re-entrant due to its mutex we can not use DBG
6983 from code that could be called by a signal handler. X's SIGIO handler
6984 attempts to move the cursor from within the handler (eek!) and so we
6985 need to be careful not to take any locks, such in as the aforementioned
6986 fprintf, along the cursor paths.
6988 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
6989 References: https://bugs.freedesktop.org/show_bug.cgi?id=50744
6990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
6992 commit 3f5b94f3d1625b06840c6441a0b175604ee3d2f9
6993 Author: Chris Wilson <chris@chris-wilson.co.uk>
6994 Date: Wed Jun 6 17:24:07 2012 +0100
6996 sna: Check against integer overflows when computing cache size
6998 Even with a 1nm process, I doubt we will see 4+GiB cache sizes ;-)
7000 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7001 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7003 commit 902391bd798775e1a7a53503d4dd1756162f737f
7004 Author: Chris Wilson <chris@chris-wilson.co.uk>
7005 Date: Wed Jun 6 17:20:39 2012 +0100
7007 uxa: Remove dead-code for SourcePictures
7009 All SourcePictures are now converted into Drawables, which had been
7010 assumed by the driver backend. However, the code still existed to
7011 attempt to pass procedural pictures onwards and so set pSrcPix to NULL
7012 which was being flagged by the static analyser as a potential NULL
7015 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7016 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7018 commit 3d8a1f7176877975a31ad7a6548fa8309065f617
7019 Author: Chris Wilson <chris@chris-wilson.co.uk>
7020 Date: Wed Jun 6 17:15:42 2012 +0100
7022 sna/gen4+: Add missing "fall through" comments
7024 Reported-by: <zkabelac@redhat.com>
7025 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7027 commit 8ae4407c43e6a8d26784508f61b416138f908132
7028 Author: Chris Wilson <chris@chris-wilson.co.uk>
7029 Date: Wed Jun 6 17:13:05 2012 +0100
7031 sna: Silence static analyser complaining about potential NULL pointer
7033 Add an assert to prove that is not.
7035 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7036 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7038 commit 08010b23a3e2bc37d202251923fac814b18fa501
7039 Author: Chris Wilson <chris@chris-wilson.co.uk>
7040 Date: Wed Jun 6 17:10:35 2012 +0100
7042 uxa/dri: Silence static analyser for potential NULL intel_pixmap
7044 If the intel_pixmap was NULL we should have failed to create the DRI2
7045 buffer, so we can safely assert here to keep the analyser quiet.
7047 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7048 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7050 commit 06b1b875ba13227ddaf7f28dbdcdaa3eb49f0857
7051 Author: Chris Wilson <chris@chris-wilson.co.uk>
7052 Date: Wed Jun 6 17:05:13 2012 +0100
7054 uxa/i915: check for failure to allocate temporary destination
7056 If the target drawable is too large for the render pipeline, we need to
7057 create a temporary surface. This may fail, so abort if it does.
7059 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7060 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7062 commit c553dcae2dd714cac413ffc7c7779cd78c9a3e61
7063 Author: Chris Wilson <chris@chris-wilson.co.uk>
7064 Date: Wed Jun 6 17:04:01 2012 +0100
7066 sna: Silence a few unused function warnings
7068 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7069 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7071 commit 536033b4edb40db18210e5272eb8ce9703e5e364
7072 Author: Chris Wilson <chris@chris-wilson.co.uk>
7073 Date: Wed Jun 6 16:02:47 2012 +0100
7075 legacy/i810/dri: Check for malloc failure for BusIdString
7077 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7078 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7080 commit b2da80c1d1058dca5d3d85693918e53d9fa61dd3
7081 Author: Chris Wilson <chris@chris-wilson.co.uk>
7082 Date: Wed Jun 6 15:59:54 2012 +0100
7084 sna: Add some missing 'fall through' comments
7086 Static analysers are dumb and presume readers are too.
7088 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7089 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7091 commit c433fb4521e2fb65a555489bd09ba6bb9448b72f
7092 Author: Chris Wilson <chris@chris-wilson.co.uk>
7093 Date: Wed Jun 6 15:17:30 2012 +0100
7095 uxa/dri: Fix up typo from f2513cb0f
7097 s/true/FALSE/? Wrong.
7099 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7100 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7102 commit 1675f441049d559d891fe378e437c12db21df9e7
7103 Author: Chris Wilson <chris@chris-wilson.co.uk>
7104 Date: Wed Jun 6 12:55:13 2012 +0100
7106 uxa/dri: Transfer reference of the new_back buffer to the flip_chain
7108 In order to prevent a leak of the bo when the chain is eventually
7109 torn-down when the client exits.
7111 Reported-by: Andreas Lampersperger <lampersperger.andreas@heidenhain.de>
7112 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50670
7113 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7115 commit 18726a4975ab2ddf85eaa6eb1602dcbe599217f7
7116 Author: Chris Wilson <chris@chris-wilson.co.uk>
7117 Date: Wed Jun 6 10:58:49 2012 +0100
7119 sna: Perform CopyArea directly onto a CPU bo if available
7121 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7123 commit 57d7d5de78bcf01d75d7a7de03fe50a2a9bd1b7e
7124 Author: Chris Wilson <chris@chris-wilson.co.uk>
7125 Date: Wed Jun 6 00:08:17 2012 +0100
7127 sna: Use GPU for readback onto CPU bo
7129 Time to blt from GTT to LLC 16384 bytes: 125.000µs (snb)
7130 Time to blt from GTT to LLC 16384 bytes: 71.000µs (ivb)
7131 Time to blt from GTT to LLC 1048576 bytes: 1400.000µs (snb)
7132 Time to blt from GTT to LLC 1048576 bytes: 938.000µs (ivb)
7134 Time to copy from GTT to LLC 16384 bytes: 118.000µs (snb)
7135 Time to copy from GTT to LLC 16384 bytes: 134.000µs (ivb)
7136 Time to copy from GTT to LLC 1048576 bytes: 6723.000µs (snb)
7137 Time to copy from GTT to LLC 1048576 bytes: 7424.000µs (ivb)
7141 Time to blt from LLC to GTT 16384 bytes: 10.000µs (snb)
7142 Time to blt from LLC to GTT 16384 bytes: 8.000µs (ivb)
7143 Time to blt from LLC to GTT 1048576 bytes: 217.000µs (snb)
7144 Time to blt from LLC to GTT 1048576 bytes: 135.000µs (ivb)
7146 Time to copy from LLC to GTT 16384 bytes: 4.000µs (snb)
7147 Time to copy from LLC to GTT 16384 bytes: 4.000µs (ivb)
7148 Time to copy from LLC to GTT 1048576 bytes: 270.000µs (snb)
7149 Time to copy from LLC to GTT 1048576 bytes: 179.500µs (ivb)
7151 It seems clear then that even with the extra synchronisation cost
7152 copying from the GTT is much preferable with the GPU than using the
7153 uncached reads by the CPU. Streaming write-combines from the CPU into
7154 the GTT seem about as efficient as we can manage, so continue to use the
7155 mapping unless busy.
7157 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7159 commit f2513cb0fdb0d1214854fd4e4dcd477ba8583862
7160 Author: Chris Wilson <chris@chris-wilson.co.uk>
7161 Date: Wed Jun 6 10:41:35 2012 +0100
7163 uxa/dri: Do not use undeclared stdbool features
7165 The header isn't pulled in, so stop using the undefined values of
7168 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7170 commit 1dafb4777f8378c87f34feae667582498220204c
7171 Author: Chris Wilson <chris@chris-wilson.co.uk>
7172 Date: Tue Jun 5 21:25:40 2012 +0100
7174 sna/damage: Add some assertions to validate that each damage box is non-empty
7176 References: https://bugs.freedesktop.org/show_bug.cgi?id=50744
7177 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7179 commit c4eb5528a456b65c673f7c984d14a622ac67cdca
7180 Author: Chris Wilson <chris@chris-wilson.co.uk>
7181 Date: Tue Jun 5 16:04:16 2012 +0100
7183 uxa: Check for DPMS off before scheduling a WAIT_ON_EVENT
7185 Regression from commit 3f3bde4f0c72f6f31aae322bcdc20b95eade6631
7186 Author: Chris Wilson <chris@chris-wilson.co.uk>
7187 Date: Thu May 24 11:58:46 2012 +0100
7189 uxa: Only consider an output valid if the kernel reports it attached
7191 When backporting from SNA, a key difference that UXA does not track DPMS
7192 state in its enabled flag and that a DPMS off CRTC is still bound to the
7193 fb. So we do need to rescan the outputs and check that we have a
7194 connector enabled *and* the pipe is running prior to emitting a scanline
7197 References: https://bugs.freedesktop.org/show_bug.cgi?id=50668
7198 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7200 commit 7c51cabaecac52348766c622e80ed14b9854e54d
7201 Author: Chris Wilson <chris@chris-wilson.co.uk>
7202 Date: Tue Jun 5 12:59:46 2012 +0100
7204 sna: Try to create Y-tiled pixmaps for initial source bo
7206 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7208 commit a26c05cc8936cab28d83c6beeff906a910353338
7209 Author: Chris Wilson <chris@chris-wilson.co.uk>
7210 Date: Tue Jun 5 12:48:33 2012 +0100
7212 sna/gen2+: Tweak placement of operations for CPU-bound large pixmaps
7214 Try to avoid uncessary migration to the GPU of large pixmaps that are
7215 wholly bound to the CPU.
7217 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7219 commit a21bdbe3e312b40b936c5c68c84f5c1bc0f9fb88
7220 Author: Dave Airlie <airlied@redhat.com>
7221 Date: Tue May 22 15:44:28 2012 +0100
7223 sna: port to compat api
7225 This ports SNA to the new compat-api.
7227 Signed-off-by: Dave Airlie <airlied@redhat.com>
7229 commit 8f5001493e18b6b3b97ca5f0c923678a51975bd0
7230 Author: Dave Airlie <airlied@redhat.com>
7231 Date: Tue May 22 15:35:01 2012 +0100
7233 intel: port legacy and uxa to new API.
7235 This ports the legacy and uxa driver to the new server API.
7237 Signed-off-by: Dave Airlie <airlied@redhat.com>
7239 commit c9824827422f8ec4f46c9cba42d871c98dc54761
7240 Author: Dave Airlie <airlied@redhat.com>
7241 Date: Tue Jun 5 10:38:21 2012 +0100
7243 sna: drop using block/wakeup data.
7245 These went away in the new server API, and really if this
7246 made any measurable difference, I'd be impressed.
7248 Signed-off-by: Dave Airlie <airlied@redhat.com>
7250 commit d9850a05003008d86d0b226bcd775aebdd43bccd
7251 Author: Dave Airlie <airlied@redhat.com>
7252 Date: Tue Jun 5 10:04:48 2012 +0100
7254 uxa: drop enable disable access hook
7256 This looks to be unused.
7258 Signed-off-by: Dave Airlie <airlied@redhat.com>
7260 commit 99df720216787c52d16a350dd25469e2dcea2c47
7261 Author: Chris Wilson <chris@chris-wilson.co.uk>
7262 Date: Tue Jun 5 10:24:34 2012 +0100
7264 sna: Add inline keyword in conjunction with attribute(always_inline)
7266 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7268 commit 47d7da5a986a9f133cd3dc8314ff243001586a36
7269 Author: Chris Wilson <chris@chris-wilson.co.uk>
7270 Date: Tue Jun 5 10:18:40 2012 +0100
7272 sna: Maintain the operation on the CPU if not completely replacing all-damaged
7274 We may as well continue the operation on the CPU if already bound to try to
7275 reduce needless migrations (where the cost of the upload will outweigh
7278 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7280 commit 81f09347f2ab59cf0a3eaca7be83ded555655e93
7281 Author: Chris Wilson <chris@chris-wilson.co.uk>
7282 Date: Mon Jun 4 16:45:05 2012 +0100
7284 uxa/i965: Silence static analyser by asserting the bo exists for the video
7286 This is already checked at the beginning of PutImageTextured, so this
7287 check upon the return value of intel_get_pixmap_bo() should only be
7288 required to keep static analysers happy.
7290 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
7291 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7293 commit 1f43de322b400dcd64eb4545a978ad9b1c7de185
7294 Author: Chris Wilson <chris@chris-wilson.co.uk>
7295 Date: Mon Jun 4 15:39:39 2012 +0100
7297 sna: Exclude consideration of tiling flags from overwriting BLT commands
7299 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7301 commit ebb1c9d5f82e8822f7400ff11a887ab047a0d78e
7302 Author: Chris Wilson <chris@chris-wilson.co.uk>
7303 Date: Mon Jun 4 15:28:31 2012 +0100
7305 sna: Make the bo-is-busy DBG more useful by saying which bo it is
7307 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7309 commit 6dc5ddfe807b307ca10af971c4f84498b2fb82a2
7310 Author: Chris Wilson <chris@chris-wilson.co.uk>
7311 Date: Mon Jun 4 13:49:35 2012 +0100
7313 sna: Remove bogus check for color == 0
7315 This was written when the RGBA color value was being passed around and
7316 not the pointer to the xRenderColor. As such, the NULL deref check
7317 doesn't gain much and the check for rgba==0 irrelevant in this scenario.
7319 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7320 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7322 commit 9b3937228c692598899fb39bfe448a9b457315db
7323 Author: Chris Wilson <chris@chris-wilson.co.uk>
7324 Date: Mon Jun 4 13:47:58 2012 +0100
7326 sna: Simplify selecting default tiling for framebuffers
7328 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7330 commit ff91fd4a2749f1ea8722189aaec938e51bce4222
7331 Author: Chris Wilson <chris@chris-wilson.co.uk>
7332 Date: Mon Jun 4 12:41:05 2012 +0100
7334 uxa: Check for failure from drmModeGetConnector()
7336 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7337 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7339 commit 4712a4008aaff7b3c57a8dd0a5e639992c9b6d30
7340 Author: Chris Wilson <chris@chris-wilson.co.uk>
7341 Date: Mon Jun 4 12:40:56 2012 +0100
7345 commit 63a55e9db97b647cbf7385403a0f30ee4d24dc3b
7346 Author: Chris Wilson <chris@chris-wilson.co.uk>
7347 Date: Mon Jun 4 12:36:11 2012 +0100
7349 uxa/shadow: Free the buffer along the unexpected failure to attach to the pixmap
7351 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7352 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7354 commit c9bd4ed7d7d5c9ba6924e911652e03abc9267865
7355 Author: Chris Wilson <chris@chris-wilson.co.uk>
7356 Date: Mon Jun 4 12:29:58 2012 +0100
7358 legacy/i810: Silence a compiler warning for missing 'const'
7360 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7362 commit 6f8b411b8a833cca6fce8888ce515db6413d57ca
7363 Author: Chris Wilson <chris@chris-wilson.co.uk>
7364 Date: Mon Jun 4 12:21:27 2012 +0100
7366 uxa: Preferred order for calloc is (count, size)
7368 And assert that the allocation succeeds for good measure. A great big
7369 dollop of graceful failure is missing...
7371 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7372 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7374 commit 72a29aa74a1488269c2184300e573d060c49e232
7375 Author: Chris Wilson <chris@chris-wilson.co.uk>
7376 Date: Mon Jun 4 12:21:27 2012 +0100
7378 sna: Check that the buffer was allocated prior to caching it
7380 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7381 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7383 commit 9bff89b1892cee90a80909255876240126b700b2
7384 Author: Chris Wilson <chris@chris-wilson.co.uk>
7385 Date: Mon Jun 4 12:21:27 2012 +0100
7387 legacy/i810: Be paranoid and guard against xf86GetPciInfoForEntity
7389 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7390 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7392 commit ed6004f7ad09329bd4097331aaff38edc83107b6
7393 Author: Chris Wilson <chris@chris-wilson.co.uk>
7394 Date: Mon Jun 4 12:21:27 2012 +0100
7396 uxa: Be paranoid and guard against xf86GetPciInfoForEntity
7398 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7399 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7401 commit abb8d893742f4ba9567330da9706adda4e636e75
7402 Author: Chris Wilson <chris@chris-wilson.co.uk>
7403 Date: Mon Jun 4 12:21:27 2012 +0100
7405 sna: Be paranoid and guard against xf86GetPciInfoForEntity
7407 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7408 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7410 commit 782cd6eafa9917a5bdb181d66b7f4c05392950ba
7411 Author: Chris Wilson <chris@chris-wilson.co.uk>
7412 Date: Mon Jun 4 12:20:17 2012 +0100
7414 uxa: check for failure to allocate drmModeCrtc
7416 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7417 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7419 commit a99bf0125f8f166ff2d65449ecd376c99c3024b8
7420 Author: Chris Wilson <chris@chris-wilson.co.uk>
7421 Date: Mon Jun 4 12:15:51 2012 +0100
7423 uxa/i965: Drop superfluous 'state' variable
7425 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7426 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7428 commit 15a00ba047fdb7d388a322f2e33894d20787af2f
7429 Author: Chris Wilson <chris@chris-wilson.co.uk>
7430 Date: Mon Jun 4 12:09:52 2012 +0100
7432 sna/gen2: Remove double 'const'
7434 So constant, we const'ed it twice.
7436 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7437 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7439 commit 99129c369c16d9ff7588a0e59f9285cc339b7eb3
7440 Author: Chris Wilson <chris@chris-wilson.co.uk>
7441 Date: Mon Jun 4 12:08:11 2012 +0100
7443 sna/gen4+: Use an explicit invalid value for alu
7445 Since the static analyser also doesn't like comparing a uint8_t against
7448 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7449 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7451 commit fae9c054d47203b68a6c09647945d23074ea4df9
7452 Author: Chris Wilson <chris@chris-wilson.co.uk>
7453 Date: Mon Jun 4 12:04:30 2012 +0100
7455 legacy/i810: Remove unused variable
7457 In order to reduce the volumes of output from static analysers.
7459 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7460 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7462 commit 880ec2c9a56d3de33c7fbbfa4c043082fbcf7a78
7463 Author: Chris Wilson <chris@chris-wilson.co.uk>
7464 Date: Mon Jun 4 12:04:30 2012 +0100
7466 legacy/i810: Reorder DRI teardown code to avoid potential NULL derefs
7468 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7469 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7471 commit 7424ea7dd08e304baa4efa09e887b833737ac9f6
7472 Author: Chris Wilson <chris@chris-wilson.co.uk>
7473 Date: Mon Jun 4 11:59:24 2012 +0100
7475 uxa/dri: Free wait_info along DRI2WaitMSC error paths
7477 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7478 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7480 commit ebd6dea009479e612c67d78416b8680bb23dba94
7481 Author: Chris Wilson <chris@chris-wilson.co.uk>
7482 Date: Mon Jun 4 11:56:11 2012 +0100
7484 uxa/dri: Make sure is_glamor_pixmap is always initialised
7486 The code paths are too twisty to be sure otherwise.
7488 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7489 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7491 commit 33f0b4b96662d953052232b1c0477ced0c326a62
7492 Author: Chris Wilson <chris@chris-wilson.co.uk>
7493 Date: Mon Jun 4 11:54:28 2012 +0100
7495 legacy/i810: Free offscreen image info struct on failure
7497 Impossible with the current code, the server aborts on failure. However,
7498 this looks to be the simple answer to keep static analysers quiet.
7500 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7501 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7503 commit d24340747389db971c04349d1ee517f195c2b28e
7504 Author: Chris Wilson <chris@chris-wilson.co.uk>
7505 Date: Mon Jun 4 11:53:26 2012 +0100
7507 sna: Free clip boxes (if allocated) along error path
7509 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7510 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7512 commit 1215abc5c30cc3e183bde59b1523c09b59c484e3
7513 Author: Chris Wilson <chris@chris-wilson.co.uk>
7514 Date: Mon Jun 4 11:49:15 2012 +0100
7516 sna: Make the ignored return value explicit
7518 The return from __kgem_throttle_retire() is just a hint as to whether
7519 any forward progress was made. In the case of the error path, though it
7520 is a last ditch effort before aborting, so we do not really care whether
7521 or not it succeeds, we must try again.
7523 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7524 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7526 commit e230b460515a043a1b7353d153e864754ca5e064
7527 Author: Chris Wilson <chris@chris-wilson.co.uk>
7528 Date: Mon Jun 4 11:48:23 2012 +0100
7530 ux/i965: create_sampler_state_bo() expects enums, so feed it the right types
7532 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7533 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7535 commit 072d7a8b42f04600c9d8054f3648642a1aaff57a
7536 Author: Chris Wilson <chris@chris-wilson.co.uk>
7537 Date: Mon Jun 4 11:40:01 2012 +0100
7539 sna/debug: Assert the non-existence of the relocation handle at source
7541 This should help the static analyzer pinpoint the blame and make it
7544 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7545 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7547 commit 4a3c355e056339aed68b70470556633dea899b1b
7548 Author: Chris Wilson <chris@chris-wilson.co.uk>
7549 Date: Mon Jun 4 11:32:42 2012 +0100
7551 sna: Silence a compiler warning for loss of 'const' qualifier
7553 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7555 commit 791029cc16bca36b8dec82297ff7e07a972c51ab
7556 Author: Chris Wilson <chris@chris-wilson.co.uk>
7557 Date: Sat Jun 2 13:28:43 2012 +0100
7559 sna/trapezoids: Implement trapezoidal opaque fills inplace
7561 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7563 commit 1f78a934a423911e18d340f0585e31941f6e8663
7564 Author: Chris Wilson <chris@chris-wilson.co.uk>
7565 Date: Sat Jun 2 13:27:52 2012 +0100
7567 sna: Clear the counters prior to querying the property
7569 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7571 commit a1953f1d4abc6e158a5e3ca53d3207548842254c
7572 Author: Chris Wilson <chris@chris-wilson.co.uk>
7573 Date: Sat Jun 2 12:02:55 2012 +0100
7575 sna: Prefer to use memset() for extremely large clears
7577 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7579 commit ce85cd1a36e31795a966ea8983c2d6f803a4eccd
7580 Author: Chris Wilson <chris@chris-wilson.co.uk>
7581 Date: Sat Jun 2 11:46:06 2012 +0100
7583 sna: Add some DBG to retreiving EDID
7585 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7587 commit e8eb273bd6153c232a9ffc558e3b7fd4beaab01b
7588 Author: Chris Wilson <chris@chris-wilson.co.uk>
7589 Date: Sat Jun 2 10:43:03 2012 +0100
7591 sna/gen7: Add DBG for ring switching
7593 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7595 commit 722afa6bc910a2ccfbb0442872a878d1b6b78315
7596 Author: Chris Wilson <chris@chris-wilson.co.uk>
7597 Date: Sat Jun 2 10:04:02 2012 +0100
7599 sna: Remove some unused members from the KMS state tracking
7601 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7603 commit 26e7bb3f25f0c83d39ff505fa7e05bfcb976e39c
7604 Author: Chris Wilson <chris@chris-wilson.co.uk>
7605 Date: Sat Jun 2 09:16:51 2012 +0100
7607 sna: Add a DBG message to indicate flushing for GPU idle
7609 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7611 commit 58fc03b8c36688e9fa7925aa82b83d36c9decb7e
7612 Author: Chris Wilson <chris@chris-wilson.co.uk>
7613 Date: Sat Jun 2 08:53:24 2012 +0100
7615 sna: Prevent NULL deref with early termination and DBG enabled
7617 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7619 commit 1be2afb81df887ed664ac31d0641da808cfb8888
7620 Author: Chris Wilson <chris@chris-wilson.co.uk>
7621 Date: Sat Jun 2 08:04:02 2012 +0100
7623 legacy/i810/video: Release memory after creating adaptors
7625 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7627 commit 261d086265fff6c9b28a67d2fcceed8f107d5cb0
7628 Author: Chris Wilson <chris@chris-wilson.co.uk>
7629 Date: Fri Jun 1 23:22:31 2012 +0100
7631 legacy/i810/dri: Propagate failure from allocating texture memory
7633 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7634 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7636 commit 2ea12b56d42686f75deb33fc126d09acee7e2899
7637 Author: Chris Wilson <chris@chris-wilson.co.uk>
7638 Date: Fri Jun 1 23:22:31 2012 +0100
7640 legacy/i810: Fix memset(sizeof(*ptr)) rather than memset(sizeof(ptr))
7642 Clear the entire structure and not the first 4 bytes...
7644 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7645 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7647 commit d3e15d1460faad192781d74fdc2b5092e9c148ad
7648 Author: Chris Wilson <chris@chris-wilson.co.uk>
7649 Date: Sat Jun 2 08:14:23 2012 +0100
7651 uxa/i965: Make the unhandled allocation failures explicit
7653 Add assertions to the point of allocation and write failures.
7655 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7657 commit 46fbb3d31bf73df4bd0259a146e9929f62a9c488
7658 Author: Chris Wilson <chris@chris-wilson.co.uk>
7659 Date: Fri Jun 1 23:22:31 2012 +0100
7661 uxa: NameForAtom may return NULL
7663 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7664 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7666 commit 8451d5dd3c992bcfad9f71aa890b6b225884b978
7667 Author: Chris Wilson <chris@chris-wilson.co.uk>
7668 Date: Fri Jun 1 23:22:31 2012 +0100
7670 uxa: Use unsigned bitfields for single bits
7672 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7673 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7675 commit d66ca06865973833398560a04186e0067d9cb206
7676 Author: Chris Wilson <chris@chris-wilson.co.uk>
7677 Date: Fri Jun 1 23:22:31 2012 +0100
7679 uxa: Intialize is_glamor_pixmap
7681 It is not clear whether is_glamor_pixmap is always initialized prior to
7682 use, so set it to a safe^W likely value.
7684 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7685 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7687 commit 25a17b2ca5e3945150151c4308761891c0e20bab
7688 Author: Chris Wilson <chris@chris-wilson.co.uk>
7689 Date: Fri Jun 1 23:22:31 2012 +0100
7691 uxa: Remove 'render_source_is_solid' dead code
7693 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7694 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7696 commit ba0eb230836fa5f94a2f50da5880fdd3b9dfd72e
7697 Author: Chris Wilson <chris@chris-wilson.co.uk>
7698 Date: Fri Jun 1 23:22:31 2012 +0100
7700 uxa: Fix reallocation of XVAdaptors array
7702 Prevent the leak and remove some unsightly code in the process.
7704 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7705 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7707 commit affb1f82c1401bac3f54654a342e3b16b8bd374b
7708 Author: Chris Wilson <chris@chris-wilson.co.uk>
7709 Date: Sat Jun 2 08:27:59 2012 +0100
7711 sna/gradient: Reuse old gradient bo if allocation of new fails
7713 Prefer a stall to a crash.
7715 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7716 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7718 commit e51f984a4c15de42c5bee3aeacee553e2dd4eca0
7719 Author: Chris Wilson <chris@chris-wilson.co.uk>
7720 Date: Sat Jun 2 08:23:35 2012 +0100
7722 sna: Check gradient allocations during init
7724 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7725 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7727 commit 2c5647a18e7856b604bce70270463fc128c2cfcd
7728 Author: Chris Wilson <chris@chris-wilson.co.uk>
7729 Date: Fri Jun 1 23:22:31 2012 +0100
7731 sna: NameForAtom may return NULL
7733 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7734 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7736 commit 7867bff00caeff9e1fbe569ca3d37b94f704a4cd
7737 Author: Chris Wilson <chris@chris-wilson.co.uk>
7738 Date: Fri Jun 1 23:22:31 2012 +0100
7740 sna: Replace invalid signed value with unsigned ~0 for uint8_t
7742 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7743 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7745 commit ca72b372237287af0f4475801a9b64efca975832
7746 Author: Chris Wilson <chris@chris-wilson.co.uk>
7747 Date: Fri Jun 1 23:22:31 2012 +0100
7749 sna: Drop return value from 3D point transform
7751 And just fixup the computed coordinates in the face of an invalid
7754 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7755 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7757 commit b83e2a79f2a082630e0d3147abec8d3599f3f4b4
7758 Author: Chris Wilson <chris@chris-wilson.co.uk>
7759 Date: Fri Jun 1 17:17:49 2012 +0100
7761 uxa: Static analysis warning fixes
7763 A smattering of bugs and confusing code.
7765 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7766 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7768 commit 67e8ac365c4b7dd9c4eb2abe46aecc8cc04d4278
7769 Author: Chris Wilson <chris@chris-wilson.co.uk>
7770 Date: Fri Jun 1 17:20:51 2012 +0100
7772 sna/gen3+: Guard against a kgem_bo_destroy(NULL)
7774 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7777 commit e91b27ae14a5366175a4fe9c87ee15a0a25b5a42
7778 Author: Chris Wilson <chris@chris-wilson.co.uk>
7779 Date: Fri Jun 1 17:12:36 2012 +0100
7781 sna: Fix potential deference of NULL cpu_bo inside assertion
7783 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7784 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7786 commit 048161e69612324d6c077ec487a2a64aaaa7c1f5
7787 Author: Chris Wilson <chris@chris-wilson.co.uk>
7788 Date: Fri Jun 1 17:09:21 2012 +0100
7790 sna/trapezoids: Check for malloc failure
7792 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7793 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7795 commit 2896aa7b46931d3b95567121ccb07844725ab45c
7796 Author: Chris Wilson <chris@chris-wilson.co.uk>
7797 Date: Fri Jun 1 17:08:01 2012 +0100
7799 sna/io: Free the heap clip boxes on error paths
7801 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7802 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7804 commit 1ef00ff720226c211444806d9fdf67ce2046685b
7805 Author: Chris Wilson <chris@chris-wilson.co.uk>
7806 Date: Fri Jun 1 17:05:55 2012 +0100
7808 sna/dri: Fix typo, check for NULL after allocations to spot failure!
7810 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7811 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7813 commit dc0fe44bebfdee63bcda4d185b85fe9e1cb16a71
7814 Author: Chris Wilson <chris@chris-wilson.co.uk>
7815 Date: Fri Jun 1 16:54:41 2012 +0100
7817 sna: Fix a smattering of static analysis warnings
7819 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7822 commit 8eed569fb386a9af48a8beb28666d72c6678e48c
7823 Author: Chris Wilson <chris@chris-wilson.co.uk>
7824 Date: Fri Jun 1 13:02:50 2012 +0100
7826 sna/trapezoids: Correct extents declaration for fallback
7828 Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
7829 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7831 commit 91419576eef562378cccf90968c4f0277139b03d
7832 Author: Chris Wilson <chris@chris-wilson.co.uk>
7833 Date: Fri Jun 1 12:09:19 2012 +0100
7835 sna: Tiny DBG message tweak
7837 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7839 commit 4bdecc5b07a184ba136129e75a7fef914ac3b8d2
7840 Author: Chris Wilson <chris@chris-wilson.co.uk>
7841 Date: Fri Jun 1 11:06:04 2012 +0100
7843 test: Add a very basic test to exercise BLT text drawing
7845 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7847 commit ea0f326d80f21f89c3cff3595a6e1afc42b52efd
7848 Author: Chris Wilson <chris@chris-wilson.co.uk>
7849 Date: Fri Jun 1 09:04:23 2012 +0100
7851 sna: Debug XDrawRectangle
7853 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7855 commit 6a7efa991a705bdb5a3b9d5300d824a1ce8759af
7856 Author: Chris Wilson <chris@chris-wilson.co.uk>
7857 Date: Fri Jun 1 08:47:20 2012 +0100
7859 test: Exercise basic rectangle drawing code
7861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7863 commit bc4323558bebd53e474fbc5404e1c41ab16d02e6
7864 Author: Chris Wilson <chris@chris-wilson.co.uk>
7865 Date: Thu May 31 21:19:26 2012 +0100
7867 sna: Handle negative values when computing the stipple modulus
7869 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7871 commit aca994e03e6e0e16f55841418b0061b175e91e5d
7872 Author: Chris Wilson <chris@chris-wilson.co.uk>
7873 Date: Thu May 31 21:17:24 2012 +0100
7875 sna: Fill early break for clip process of spans
7877 When on the same Y-band as the span, as soon as the clip boxes are too
7878 far to the right, we can stop searching for more intersections.
7880 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7882 commit 0fe150f898120ba9a00e1e6b9d66bec10d7e8a29
7883 Author: Chris Wilson <chris@chris-wilson.co.uk>
7884 Date: Thu May 31 20:02:47 2012 +0100
7886 sna: Fix computation of box for clipped stippled rectangles
7888 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7890 commit a65c3b7b45df46ba993af6755b2460aa80fde930
7891 Author: Chris Wilson <chris@chris-wilson.co.uk>
7892 Date: Thu May 31 19:15:05 2012 +0100
7894 sna: Add some more DBG for stippled blts
7896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7898 commit b2cc64253217a90cf2ade9a8d6f1ed21b23fe875
7899 Author: Chris Wilson <chris@chris-wilson.co.uk>
7900 Date: Thu May 31 19:45:56 2012 +0100
7902 sna: And combine the two migrate to CPU clauses
7904 Just a small tidy up to remove a duplicated predicate.
7906 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7908 commit bd1dd4cf663ca230724dc6beb329b7eb270ea9d5
7909 Author: Chris Wilson <chris@chris-wilson.co.uk>
7910 Date: Thu May 31 17:57:56 2012 +0100
7912 sna: If the CPU damage already contains the migration region, we are done
7914 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7916 commit a2d92ade3911e62f950c85b5650fb13da1279ced
7917 Author: Chris Wilson <chris@chris-wilson.co.uk>
7918 Date: Thu May 31 17:49:01 2012 +0100
7920 sna: Prevent readback of 1x1 if it lies inside CPU damage
7922 If the pixel exists only in the CPU damage, attempting to read it back
7923 from the GPU only results in garbage.
7925 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7927 commit a34a06c4ba36b21e386b1f6371a4bd3f03ff35a1
7928 Author: Chris Wilson <chris@chris-wilson.co.uk>
7929 Date: Thu May 31 20:14:38 2012 +0100
7931 sna/glyphs: Clip the damage to the drawable for unclipped glyphs-to-dst
7933 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7935 commit dcc7ba8ccf95db1c265bd5492f041f94ce368b7a
7936 Author: Chris Wilson <chris@chris-wilson.co.uk>
7937 Date: Thu May 31 17:48:40 2012 +0100
7939 sna/composite: Add debugging modes to disable acceleration
7941 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7943 commit 984400b95a879c7c020de14545ac2e1d9e759e05
7944 Author: Chris Wilson <chris@chris-wilson.co.uk>
7945 Date: Thu May 31 16:56:21 2012 +0100
7947 sna/glyphs: Don't apply drawable offsets twice
7949 The offsets from image_from_pict() already include the drawable offset
7950 so we need not apply them twice.
7952 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7954 commit 7bbd445c8cbeb7bcea2c56c4deb9bf6ccf6f2497
7955 Author: Chris Wilson <chris@chris-wilson.co.uk>
7956 Date: Thu May 31 16:45:03 2012 +0100
7958 sna/trapezoids: Apply the drawable offset for the CPU migration
7960 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7962 commit c5313620a2f66bc8daf211fc77a9c480a79ea271
7963 Author: Chris Wilson <chris@chris-wilson.co.uk>
7964 Date: Thu May 31 15:06:16 2012 +0100
7966 sna: Suppress flushes once we have cleared the kernel dirtiness
7968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7970 commit 0347c8abec9740d706d87dd2f719ac68f3d1c32e
7971 Author: Chris Wilson <chris@chris-wilson.co.uk>
7972 Date: Thu May 31 13:50:31 2012 +0100
7974 sna/glyphs: Fix upconverting of glyphs via gpu
7976 When using the mask instead of the src for adding the glyphs we need to
7977 pass the glyph coordinates through the mask coordinates rather than the
7980 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7982 commit 90ae4f853222ee33206134f4efdc4accfb2f2c38
7983 Author: Chris Wilson <chris@chris-wilson.co.uk>
7984 Date: Thu May 31 14:17:40 2012 +0100
7986 sna: Avoid mixing signed/unsigned int/int16 arithmetric
7988 Life becomes unpleasant with sign extension.
7990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
7992 commit 660c89e9742bac5ce7cbd480e08b4667e37dee8c
7993 Author: Chris Wilson <chris@chris-wilson.co.uk>
7994 Date: Thu May 31 13:18:21 2012 +0100
7996 sna: Use full 16-bit unsigned values for absolute differences
7998 Beware the overflow implicit in:
7999 adx = x2 >= x1 ? x2 - x1 : x1 - x2;
8000 when both x2 and x1 may be large signed 16-bit values
8002 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
8003 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50532
8004 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8006 commit 035c2953751f58225bd6b1fcb4c1275ccb5526cb
8007 Author: Chris Wilson <chris@chris-wilson.co.uk>
8008 Date: Thu May 31 09:29:40 2012 +0100
8010 sna/glyphs: use add(WHITE, glyph, mask) for channel expansion
8012 If the glyph format does not match the mask format we can not simply add
8013 the two together, but must first perform a channel expansion (or
8014 contraction) by multiplying the glyph against a WHITE source.
8016 Normally the glyph and the mask are equivalent formats and so we hit the
8019 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8021 commit cf5b3e2ebf4ee0330f5421b9377bb512a94ec284
8022 Author: Chris Wilson <chris@chris-wilson.co.uk>
8023 Date: Wed May 30 14:44:59 2012 +0100
8025 sna: Don't attempt to flush scanout if unattached
8027 This was lost in the midst of the OSTimer overhaul.
8029 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
8030 References: https://bugs.freedesktop.org/show_bug.cgi?id=50393
8031 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8033 commit 3a64ba0895035f6567e9a15357396b06c0f2d6dd
8034 Author: Chris Wilson <chris@chris-wilson.co.uk>
8035 Date: Wed May 30 13:51:07 2012 +0100
8037 sna: Convert remaining drmCommands to drmIoctl
8039 This helps to silence valgrind
8041 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8043 commit 47e2db0ba8ae3fbcdf58cba26c58ec0932b4c90b
8044 Author: Chris Wilson <chris@chris-wilson.co.uk>
8045 Date: Wed May 30 10:37:22 2012 +0100
8047 sna: Replace FREELIKE_BLOCK with MAKE_MEM_NOACCESS
8049 And similarly for MALLOCLIKE_BLOCK. The issue being that valgrind may
8050 overwrite such blocks with alloc-fill and free-fill values, but when in
8051 fact they are defined and still in use by the GPU. This can lead to the
8052 GPU processing garbage, and GPU hangs.
8054 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
8055 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50393
8056 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8058 commit 08a630dc5ef87e551865e558fe4fc45ea66457b4
8059 Author: Chris Wilson <chris@chris-wilson.co.uk>
8060 Date: Wed May 30 08:50:44 2012 +0100
8062 sna: Ensure we flush scanout even when otherwise idle.
8064 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
8065 References: https://bugs.freedesktop.org/show_bug.cgi?id=50477
8066 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8068 commit 10ca765ad346931087be146f3e6d1f4069c95f26
8069 Author: Chris Wilson <chris@chris-wilson.co.uk>
8070 Date: Wed May 30 08:32:34 2012 +0100
8072 Mark a few more options as being UXA specific
8074 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8076 commit 740368c4c6eb547adad247ff529d16e594d6459b
8077 Author: Nick Bowler <nbowler@draconx.ca>
8078 Date: Mon May 28 18:28:28 2012 -0400
8080 Include config.h in intel_options.c
8082 Commit 8a9a585341e2 ("Only create a single instance of the intel_options
8083 array") moved the definition of intel_options into a separate .c file.
8084 Several of the defined options are #ifdef'd based on the configuration,
8085 but since config.h is never included, the macros being tested are never
8086 defined. Therefore, none of the configuration-specific options will
8087 ever be available at runtime, even if they should be.
8089 Add an inclusion of config.h so that such configuration-specific options
8092 Signed-off-by: Nick Bowler <nbowler@draconx.ca>
8093 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8095 commit e1b720c40cf511b4ed6d387891d715522d0d54b6
8096 Author: Chris Wilson <chris@chris-wilson.co.uk>
8097 Date: Wed May 30 07:47:53 2012 +0100
8099 sna/sprite: Restore another xf86drm.h
8101 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8103 commit e986b06162abf286b3f8f97b791fd66dcf8f9aec
8104 Author: Chris Wilson <chris@chris-wilson.co.uk>
8105 Date: Tue May 29 22:21:05 2012 +0100
8107 sna/overlay: Handle return from GETPARAM query correctly
8109 drmIoctl() returns 0 on success, not a positive value.
8111 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8113 commit 68874e2d694cc68f5f051ba5fe4a79a71fdb2dbc
8114 Author: Chris Wilson <chris@chris-wilson.co.uk>
8115 Date: Tue May 29 21:28:39 2012 +0100
8117 sna: Fix chunking for large stipples
8119 Reported-and-tested-by: Matti Hamalainen <ccr@tnsp.org>
8120 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49510
8121 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8123 commit 12af82581118470f59817f626d61df655950e243
8124 Author: Chris Wilson <chris@chris-wilson.co.uk>
8125 Date: Tue May 29 21:45:32 2012 +0100
8127 sna/overlay: Replace drmCommand with direct invocations of drmIoctl
8129 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8131 commit 3a1b6e57fbbbb4c8e5ac9588537840f4ec176ae6
8132 Author: Chris Wilson <chris@chris-wilson.co.uk>
8133 Date: Tue May 29 21:37:16 2012 +0100
8135 sna/overlay: Restore one xf86drm.h include
8137 This was removed in haste, when it is required for drmIoctl and friends.
8139 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8141 commit 264d1b1e1fb368f7845a8112bfa2ad2224acabd9
8142 Author: Chris Wilson <chris@chris-wilson.co.uk>
8143 Date: Tue May 29 13:03:56 2012 +0100
8145 sna: Also check for a free exec slot for the upload buffers
8147 As we check before allocating the upload buffer, we can not be certain
8148 that we will allocate an already attached bo or that we have a free exec
8149 slot. So always check that we have an extra exec slot available - the
8150 false positive rate is going to be negligible.
8152 Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
8153 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50457
8154 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8156 commit c7bcca9a3fb6c669d8b093e6399914693770b0b3
8157 Author: Chris Wilson <chris@chris-wilson.co.uk>
8158 Date: Tue May 29 12:49:12 2012 +0100
8160 sna: Make finding DPMS property valgrind clean
8162 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8164 commit 3c074797094dc7675bb6ae549630fdf061ab5258
8165 Author: Chris Wilson <chris@chris-wilson.co.uk>
8166 Date: Tue May 29 12:45:20 2012 +0100
8168 sna: Make GetEDID valgrind clean.
8170 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8172 commit 560e50c0a5afa0824086557581ca6aa5696b44a2
8173 Author: Chris Wilson <chris@chris-wilson.co.uk>
8174 Date: Tue May 29 12:36:31 2012 +0100
8176 sna: Make GetEncoder() valgrind clean
8178 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8180 commit b4b6fa795201379694a03d537064dbca446cfdfd
8181 Author: Chris Wilson <chris@chris-wilson.co.uk>
8182 Date: Tue May 29 11:32:28 2012 +0100
8184 sna: Trim the set of includes
8186 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8188 commit 9d1a9e666763a3e1270c8f000b1b37a6b926441f
8189 Author: Chris Wilson <chris@chris-wilson.co.uk>
8190 Date: Mon May 28 20:06:35 2012 +0100
8192 sna: Bump experimental ioctl number for vmap
8196 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8198 commit 55e6f5f220401318529e81f7c96fe0af3b893a0c
8199 Author: Chris Wilson <chris@chris-wilson.co.uk>
8200 Date: Mon May 28 16:21:09 2012 +0100
8202 sna: Use the local function for turning the cursor off prior to release
8204 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8206 commit 77dd429222922aa1ba7f283553e11e60e4d5c496
8207 Author: Chris Wilson <chris@chris-wilson.co.uk>
8208 Date: Mon May 28 14:40:23 2012 +0100
8210 sna/dri: Make WAIT_VBLANK valgrind clean
8212 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8214 commit 8ebe84818ab70e662a9bb3f232b4664ff40ac375
8215 Author: Chris Wilson <chris@chris-wilson.co.uk>
8216 Date: Mon May 28 14:27:46 2012 +0100
8218 sna: Make sna_copy_fbcon() valgrind clean
8220 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8222 commit f37dae734ebac5c0ed2d6f50fc92bc107f4e4a33
8223 Author: Chris Wilson <chris@chris-wilson.co.uk>
8224 Date: Mon May 28 14:20:54 2012 +0100
8226 sna: Just use the CRTC id from the array
8228 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8230 commit b1f24a0eae4bb0081ff7469a2aee63a1f32140f7
8231 Author: Chris Wilson <chris@chris-wilson.co.uk>
8232 Date: Mon May 28 14:16:03 2012 +0100
8234 sna: Make sna_crtc_apply() valgrind clean
8236 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8238 commit 0ebfcdbb80245f98c0c7ef9f090a2be6900fb457
8239 Author: Chris Wilson <chris@chris-wilson.co.uk>
8240 Date: Mon May 28 14:10:04 2012 +0100
8242 sna: Make AddFB valgrind clean
8244 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8246 commit 32ca3c7b65516b0bd48b0d0979b086ee1580167a
8247 Author: Chris Wilson <chris@chris-wilson.co.uk>
8248 Date: Mon May 28 14:02:11 2012 +0100
8250 sna: Make sna_crtc_is_bound() valgrind clean
8252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8254 commit 7973f6751c9bf565dee4c89aa3e1badbcc45018f
8255 Author: Chris Wilson <chris@chris-wilson.co.uk>
8256 Date: Mon May 28 14:00:55 2012 +0100
8258 sna: Add a little bit more verbosity to cursor routines for valgrind
8260 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8262 commit 19c463e52df919fc75de7e420fd3565f0e9a0576
8263 Author: Chris Wilson <chris@chris-wilson.co.uk>
8264 Date: Mon May 28 12:54:35 2012 +0100
8266 sna: Only wait if there is a suitable active buffer
8268 There is not point repeating the search after retiring if we know that
8269 there is no outstanding suitable active buffer.
8271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8273 commit 5b99c7cd340f782d3057d4257865c5feb96b71f0
8274 Author: Chris Wilson <chris@chris-wilson.co.uk>
8275 Date: Mon May 28 10:55:49 2012 +0100
8277 uxa/dri: Enable vblank scheduling even with pageflipping disabled
8279 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8281 commit 810357ad65d551ec5d35dbf228f1b62fe235801f
8282 Author: Chris Wilson <chris@chris-wilson.co.uk>
8283 Date: Mon May 28 10:12:38 2012 +0100
8285 uxa/dri: Fix ordering of setting vs using swap members
8287 Trivial readibility fix, as the actual ordering is serialised through
8288 there being only a single thread.
8290 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8292 commit 2b56a188ef2c48f82572eca2201a3a0ecf6b4f45
8293 Author: Chris Wilson <chris@chris-wilson.co.uk>
8294 Date: Mon May 28 10:01:09 2012 +0100
8296 Mark another couple of options as being UXA only.
8298 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8300 commit b025f1a604dbbbd456a23f330ceed5f97ecdffcf
8301 Author: Chris Wilson <chris@chris-wilson.co.uk>
8302 Date: Mon May 28 09:48:49 2012 +0100
8304 sna/dri: Tweak vblank_mode=n swapbuffers to account for throttle delay
8306 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8308 commit 45148a714f1461fff898aeb393ad225f081b9bd5
8309 Author: Chris Wilson <chris@chris-wilson.co.uk>
8310 Date: Mon May 28 09:39:51 2012 +0100
8312 sna/dri: Enable handling for Option "SwapBuffersWait"
8314 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8316 commit c709f2447dfc6dc36c50ff741d5d9bbdc7c03b58
8317 Author: Chris Wilson <chris@chris-wilson.co.uk>
8318 Date: Mon May 28 08:17:21 2012 +0100
8320 sna/dri: Requeue vblank throttling until the vsync'ed copy completes
8322 If the GPU is busy, then we may not actually schedule our copy for
8323 several vblanks, resulting in us falsely reporting that the work
8324 completed too early and allowing the client to continue scheduling more
8325 work and racing ahead of the queued copies.
8327 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8329 commit 7e73fa02ed361a9c0c08f61d00421671bf10ce9f
8330 Author: Chris Wilson <chris@chris-wilson.co.uk>
8331 Date: Mon May 28 08:16:43 2012 +0100
8333 sna: Add some debugging to show count of outstanding requests during retire
8335 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8337 commit 62b557065edc0555f2bf83b0eed9169329a2f2ba
8338 Author: Chris Wilson <chris@chris-wilson.co.uk>
8339 Date: Sun May 27 16:52:51 2012 +0100
8341 sna: Use magic upload buffers for video textures
8343 So that we may benefit from the caching of buffers and the automatic
8344 selection of the preferred upload method.
8346 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8348 commit 35291d2db813f75fedcdca9920a40592acd3cca3
8349 Author: Chris Wilson <chris@chris-wilson.co.uk>
8350 Date: Sun May 27 16:54:29 2012 +0100
8352 sna: Search the inactive bo cache for a mappable upload buffer
8354 See if we have a bo that we can cheaply map to an inplace upload, rather
8355 than rely on an existing GTT map.
8357 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8359 commit 36d53ff52b055b2b5fc03aff7b2cab83037d9f42
8360 Author: Chris Wilson <chris@chris-wilson.co.uk>
8361 Date: Sun May 27 16:46:42 2012 +0100
8363 sna: Validate all CRTCs after updating one
8365 Updating one CRTC may cause the kernel to turn off another, so be
8366 paranoid and run the check in a loop after applying the CRTC set.
8368 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8370 commit 7fdd8bd6d2ba7bf113d4109e6bb6f750f909565f
8371 Author: Johannes Obermayr <johannesobermayr@gmx.de>
8372 Date: Sat May 26 23:52:28 2012 +0200
8374 glamor: Fix misspelled xf86GetOptValString and OPTION_ACCEL_METHOD introduced by commit e456291.
8376 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8378 commit 8ea4ba081de0206351394481f54dcbe6922a085b
8379 Author: Chris Wilson <chris@chris-wilson.co.uk>
8380 Date: Sat May 26 21:48:16 2012 +0100
8382 sna: Fix typo for debug compilation
8386 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8388 commit 317bf05196086eb5dedb6436c07f253f01c9bf63
8389 Author: Chris Wilson <chris@chris-wilson.co.uk>
8390 Date: Fri May 25 14:34:16 2012 +0100
8392 sna: Perform the selftest after probing
8394 We do no want to slow down the detection phase by performing our
8395 self-tests, so only running those before initialising the driver.
8397 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8399 commit a5fe863e3f1a6750edaffb518c4e768e9bc11521
8400 Author: Chris Wilson <chris@chris-wilson.co.uk>
8401 Date: Fri May 25 14:28:56 2012 +0100
8403 sna: Check the bus type before declaring probe success
8405 This should never fail, but still better to fail during detection rather
8406 than pretend it works.
8408 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8410 commit 4094826aee76ef24dad13bc5a8a723bfe4a69162
8411 Author: Chris Wilson <chris@chris-wilson.co.uk>
8412 Date: Fri May 25 08:55:13 2012 +0100
8414 sna: Trust the crtc-is-bound determination after modeset and hotplug
8416 As these should be the only time that they change and we now have the
8417 checks in place, we can drop the workaround of doing the check just
8418 before emitting the wait.
8420 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8422 commit 8a9a585341e2dd43c649204fcf6d92a867671ba3
8423 Author: Chris Wilson <chris@chris-wilson.co.uk>
8424 Date: Fri May 25 08:50:33 2012 +0100
8426 Only create a single instance of the intel_options array
8428 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8430 commit e45629135065d0cc73c285f8df35ab4e1d07c6dc
8431 Author: Chris Wilson <chris@chris-wilson.co.uk>
8432 Date: Thu May 24 18:22:45 2012 +0100
8434 Allow runtime switching of AccelMethod between uxa/sna and even glamor
8437 Option "AccelMethod" "uxa/glamor/sna"
8440 The appropriate backend must also be enabled at compile time for the
8441 runtime option to be available (i.e. --enable-uxa (default) --enable-sna
8444 Demanded-by: Adam Jackson <ajax@redhat.com>
8445 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50290
8446 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8448 commit df6ab02c3690eea8393ecc8c113e2f2891856cc6
8449 Author: Eugeni Dodonov <eugeni.dodonov@intel.com>
8450 Date: Mon Oct 31 14:43:22 2011 -0200
8452 Unify options handling between UXA and SNA
8454 Unifies available options for both UXA and SNA drivers, and
8455 moves them into a common header file, intel_opts.h.
8457 Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
8459 commit ae2be7e25bda46551381c19a673b321b4382e1f9
8460 Author: Chris Wilson <chris@chris-wilson.co.uk>
8461 Date: Thu May 24 16:33:27 2012 +0100
8463 sna/trapezoids: Correct rounding for downsampling onto sample grid
8465 Reported-by: S. Christian Collins <s_chriscollins@hotmail.com>
8466 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49446
8467 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8469 commit 0ab226e27e7920bdb9f7eb62c5174cd097ac7f7f
8470 Author: Chris Wilson <chris@chris-wilson.co.uk>
8471 Date: Thu May 24 12:51:46 2012 +0100
8473 sna: Query CRTC states following a hotplug event
8475 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8477 commit 3f3bde4f0c72f6f31aae322bcdc20b95eade6631
8478 Author: Chris Wilson <chris@chris-wilson.co.uk>
8479 Date: Thu May 24 11:58:46 2012 +0100
8481 uxa: Only consider an output valid if the kernel reports it attached
8483 Reported-by: Kyle Hill <kyle.hill@tacomafia.net>
8484 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50078
8485 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8487 commit 11db66fedf96f158cbbac8011a8ba0b29a20ba3a
8488 Author: Chris Wilson <chris@chris-wilson.co.uk>
8489 Date: Thu May 24 11:07:19 2012 +0100
8491 sna: Add some more DBG tracepoints around modesetting
8493 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8495 commit fb9da4cb9e12ca733c31879b8b37906f361a8c35
8496 Author: Chris Wilson <chris@chris-wilson.co.uk>
8497 Date: Wed May 23 13:53:37 2012 +0100
8499 sna: Only mark an unattached output as inactive
8501 So that a latter attempt to set the DesiredMode may succeed and we do
8502 not modify the configuration without notifying clients.
8504 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8506 commit 5ae032e22b127c7c95753197e0914a8028a3b22e
8507 Author: Chris Wilson <chris@chris-wilson.co.uk>
8508 Date: Wed May 23 13:27:07 2012 +0100
8510 sna: Suppress modesetting errors on return from VT switch
8512 If we presume that userspace will set the correct mode shortly
8513 afterwards, we can ignore the failure of the automatic restore.
8515 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8517 commit 34882a979d9817d33bd6a8ae73a9f7083556578c
8518 Author: Chris Wilson <chris@chris-wilson.co.uk>
8519 Date: Wed May 23 10:27:44 2012 +0100
8521 sna: Keep checking retirement for flushing list
8523 Even after all outstanding requests have been completed we may still
8524 have buffers on the flushing list that need to become idle.
8525 Once such consequence would be to prevent the vblank flush from going
8526 idle, causing spurious wakeups every vrefresh when otherwise idle.
8528 References: https://bugs.freedesktop.org/show_bug.cgi?id=50078
8529 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8531 commit 54fde5bf05d43f26e7f1893584af736a138d245f
8532 Author: Dave Airlie <airlied@redhat.com>
8533 Date: Wed May 16 15:34:08 2012 +0100
8535 intel: convert sna/uxa to using new glyph picture accessors.
8537 The compat-api.h takes care of old/new servers.
8539 Signed-off-by: Dave Airlie <airlied@redhat.com>
8541 commit 43a34186d13d29c671431832469ca5301751b3cf
8542 Author: Dave Airlie <airlied@redhat.com>
8543 Date: Wed May 16 15:26:55 2012 +0100
8545 intel: convert to new screen conversion APIs
8547 The compat header takes care of the old server vs new server.
8549 this commit was autogenerated from util/modular/x-driver-screen-scrn-conv.sh
8551 Signed-off-by: Dave Airlie <airlied@redhat.com>
8553 commit 9aabc04d76fa976abb4a74f4e9676c7556e88df3
8554 Author: Dave Airlie <airlied@redhat.com>
8555 Date: Wed May 16 15:14:32 2012 +0100
8557 intel: add compat header file for conversion fns
8559 Signed-off-by: Dave Airlie <airlied@redhat.com>
8561 commit c9ce6ae8a6fc7b1a91aef1e59640c39024d0426d
8562 Author: Dave Airlie <airlied@redhat.com>
8563 Date: Wed May 23 08:33:20 2012 +0100
8565 legacy/i810: rename pScreen->screen
8567 For consistency before moving to new APIs.
8569 This just changes the files where the API changes will touch.
8571 Signed-off-by: Dave Airlie <airlied@redhat.com>
8573 commit 70ee0f84f5b9c6c13a0338f8b2757e587c849c14
8574 Author: Dave Airlie <airlied@redhat.com>
8575 Date: Wed May 23 08:31:52 2012 +0100
8577 legacy/i810: rename pScrn->scrn
8579 This is need to make the compat api stuff easier to implement.
8581 Signed-off-by: Dave Airlie <airlied@redhat.com>
8583 commit 3410db008c6c20dc6590a89c05b40a0c6de6744a
8584 Author: Chris Wilson <chris@chris-wilson.co.uk>
8585 Date: Tue May 22 15:54:52 2012 +0100
8587 sna/dri: Avoid confusion of attempting to exchange pixmaps for DPMS-off flips
8589 Not only do we confuse ourselves, but we end up confusing the damage
8590 tracking on the root window whenever we update the Screen Pixmap. So for
8591 the time being, don't.
8593 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8595 commit 28c4fb0b70e098972af972d30a660d167a8973ad
8596 Author: Chris Wilson <chris@chris-wilson.co.uk>
8597 Date: Tue May 22 09:49:53 2012 +0100
8599 sna: Swallow disconnection event upon mode restoration
8601 So that we can VT switch back to X even though the output configuration
8604 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8606 commit b17c44c8d234cb2c651e88c7892bcd5b33139c1c
8607 Author: Chris Wilson <chris@chris-wilson.co.uk>
8608 Date: Tue May 22 09:43:51 2012 +0100
8610 sna: Remove incorrect assertion for mode restoration
8612 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8614 commit 49b7742999ee11b0c36754ea10bc5609ebe8c609
8615 Author: Chris Wilson <chris@chris-wilson.co.uk>
8616 Date: Tue May 22 09:40:25 2012 +0100
8618 sna: Force config restoration after VT switch
8620 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8622 commit 952e3dcba3a651d22d44dbb039ad97e111d9bc63
8623 Author: Chris Wilson <chris@chris-wilson.co.uk>
8624 Date: Mon May 21 19:48:52 2012 +0100
8626 sna: Only override active on DPMSModeOff
8628 Along the enable path we then only want to change the value if we
8629 actually call sna_crtc_apply.
8631 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8633 commit afdaf184594bfe3633305969eb1166c28e1006bf
8634 Author: Chris Wilson <chris@chris-wilson.co.uk>
8635 Date: Mon May 21 14:32:39 2012 +0100
8637 sna: Add a log message for the change of CRTC mode
8639 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8641 commit 9fb18462ec1428a486f998585e7b4caf4dd08f53
8642 Author: Chris Wilson <chris@chris-wilson.co.uk>
8643 Date: Mon May 21 14:12:23 2012 +0100
8645 sna: Confirm that the modeset takes
8647 If we attempt to change to a mode with a disabled connector the kernel
8648 will silently switch off that connector (and crtc) and report that the
8649 modeswitch is successful.
8651 Reported-by: Kyle Hill <kyle.hill@tacomafia.net>
8652 References: https://bugs.freedesktop.org/show_bug.cgi?id=50078
8653 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8655 commit cd3b48854403ee63723ef570c1a7f9f8447e7105
8656 Author: Chris Wilson <chris@chris-wilson.co.uk>
8657 Date: Sat May 19 12:56:45 2012 +0100
8659 sna/gen7: Trim a dead assignment
8661 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8663 commit cf79cd3e27b7077f67f202b16ed39eb1cc70ac65
8664 Author: Chris Wilson <chris@chris-wilson.co.uk>
8665 Date: Sat May 19 12:44:11 2012 +0100
8667 sna: Emit a new batch if we need to clip after PolyText without space
8669 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8671 commit f91dcc44dcc15850f82666b1bcdd27182400e7dc
8672 Author: Chris Wilson <chris@chris-wilson.co.uk>
8673 Date: Fri May 18 20:09:41 2012 +0100
8675 sna: Give the proxy a unique name
8677 So that if we cache the current destination bo (for example, gen3) then
8678 a new proxy (or even just a new batchbuffer) will indeed cause the
8679 destination buffer to be updated.
8681 Reported-and-tested-by: Clemens Eisserer <linuxhippy@gmail.com>
8682 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48636
8683 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8685 commit ee073d613bba38f90951405d5ecddfcf3ac5e043
8686 Author: Chris Wilson <chris@chris-wilson.co.uk>
8687 Date: Fri May 18 15:19:00 2012 +0100
8689 sna/traps: Fix processing of full-steps
8691 A missing factor of Y-height caused the computation of coverage for the
8692 spans to be completely wrong. This affects the vertical segments of
8693 rounded rectangles, for instance.
8695 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8697 commit 8ba800c63906fb29d34f40b9437092a665bffb14
8698 Author: Chris Wilson <chris@chris-wilson.co.uk>
8699 Date: Thu May 17 22:46:18 2012 +0100
8701 sna: Don't consider uploading inplace if the dst bo is unmappable
8703 Handle (and take advantage of) the fallback at the high level rather
8704 than masquerading an inplace write.
8706 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8708 commit dad24721a13ce3a357e8ddae3c2dea61045f6fc2
8709 Author: Chris Wilson <chris@chris-wilson.co.uk>
8710 Date: Thu May 17 21:44:22 2012 +0100
8712 Revert "sna: Always try to operate inplace if we an LLC gpu bo"
8714 This reverts commit 10b4a9bb5f46ab9d9c8b165084ce4174b54a8d39 as it
8715 causes a regression for pixel data uploads to active buffers.
8717 commit 681c6e72412fff96b203a09be6ac8d393f3489a5
8718 Author: Chris Wilson <chris@chris-wilson.co.uk>
8719 Date: Thu May 17 19:43:02 2012 +0100
8721 sna: Track flushing status of live bo
8723 Currently we only move a bo with an outstanding kernel flush onto the
8724 flushing list if it is no longer in use. This leaves us potentially
8725 stalling on a flush if we try then to write to the object believing it
8726 to be retired and idle.
8728 Reported-by: Jiri Slaby <jirislaby@gmail.com>
8729 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
8730 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8732 commit d99502a33d5bdbad010b7a036c1aee989fe29947
8733 Author: Chris Wilson <chris@chris-wilson.co.uk>
8734 Date: Thu May 17 17:56:43 2012 +0100
8736 sna/glyphs: Pass the extents to the backend for preparing to composite glyphs
8738 This information should not be required, but it might come in handy, so
8741 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8743 commit fb21c2df1a10d66f115a5978b0db508058fdc412
8744 Author: Chris Wilson <chris@chris-wilson.co.uk>
8745 Date: Thu May 17 16:30:31 2012 +0100
8747 sna/io: Assert that we do not replace a bo->flush object
8749 These should be pinned by the higher layers and so we should never be
8750 attempting to replace them. If we do replace a bo->flush, then we will
8751 end up miscounting outstanding flush bo.
8753 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8755 commit 596c0a68709a93bb376647c1b566e9df4f23b35d
8756 Author: Chris Wilson <chris@chris-wilson.co.uk>
8757 Date: Tue May 15 18:53:39 2012 +0100
8759 sna: Assign GCops after checking for fallback to pass sanity checks
8761 We assert that prior to installing the fallback GCops the current ops
8762 are the default set. This is broken if we point GCops to our GPU ops,
8763 but then fallback. So check for the fallback first.
8765 Reported-by: Jiri Slaby <jirislaby@gmail.com>
8766 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
8767 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8769 commit 3c9759ef2ad755bbe720d4aa031ec67dbc3b7734
8770 Author: Chris Wilson <chris@chris-wilson.co.uk>
8771 Date: Mon May 14 14:12:06 2012 +0100
8773 sna/gen3: Fix pre-multiplication of opacity masks
8775 The return type for the multipliation was only 8-bits wide dating back
8776 to the time when the function just computed a single channel and didn't
8777 try to blend the output back into the argb pixel value. Inlining the
8778 shift into the function means that we need the output to be 32-bits wide
8781 Reported-by: Jiri Slaby <jirislaby@gmail.com>
8782 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49887
8783 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8785 commit 053bd5bf2425aed44e1c2bb981d98ee5171211a0
8786 Author: Chris Wilson <chris@chris-wilson.co.uk>
8787 Date: Mon May 14 12:08:09 2012 +0100
8789 sna: Use the correct storage for box_from_seg
8791 Fixes regression from 3aa98289e3a2 with clipped segments. As we clipped
8792 the empty box rather than the segment, we never drew any outlines.
8794 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8796 commit b654b8794db7b00666ce5c59535a9302932c483b
8797 Author: Chris Wilson <chris@chris-wilson.co.uk>
8798 Date: Mon May 14 11:36:15 2012 +0100
8800 sna/trapezoids: Fix picture creation for fallback trapezoids
8802 Being a little lax in not updating the format after upconversion to
8803 PICT_a8, meant we were trying to composite with a depth 1, 8 bpp a8
8804 image and thoroughly confusing everybody when creating the upload
8807 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8809 commit 80567f61afe77a003e663b17c1fc6b6c3ed04042
8810 Author: Chris Wilson <chris@chris-wilson.co.uk>
8811 Date: Mon May 14 10:40:38 2012 +0100
8813 sna/trapezoids: Do not reduce SRC to a clear pixmap to unbounded
8815 As we instruct the migration code to drop the clear when copying from
8816 the GPU to the CPU, we then need to emit the zeros during the span
8819 Fixes some occassional corruption behind complex clip masks.
8821 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8823 commit 067e9375809ea6cfa0c0b5e2159b359535c3b362
8824 Author: Chris Wilson <chris@chris-wilson.co.uk>
8825 Date: Mon May 14 10:00:00 2012 +0100
8827 sna: Trim unused partial buffer uploads
8829 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8831 commit effb0b9ec579624e38f0007cd7096e7da04048ca
8832 Author: Chris Wilson <chris@chris-wilson.co.uk>
8833 Date: Sun May 13 19:29:44 2012 +0100
8835 sna: Tweak usage of kgem_bo_can_map vs kgem_bo_mapped
8837 More often than not we only want to consider whether we can map the bo
8838 and decide whether doing so is the better option. Whether the bo is
8839 already mapped is not such an issue any more with the throttling.
8841 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8843 commit 37aa1e1291127ff4d27407fac936af64d0e353a8
8844 Author: Chris Wilson <chris@chris-wilson.co.uk>
8845 Date: Sun May 13 16:59:30 2012 +0100
8847 sna: Simplify partial buffer tracking
8849 As we only want to track partial buffers with asynchronous reuse,
8850 exclude all overs from the active buffers list. Secondly, keep the list
8851 in most-recently-used order rather than by size.
8853 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8855 commit 7048a71fb6ff5e7c37e21a794f7b42d9a1da3473
8856 Author: Chris Wilson <chris@chris-wilson.co.uk>
8857 Date: Sun May 13 15:42:28 2012 +0100
8859 sna: Flush batch if GPU is idle upon wakeup
8861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8863 commit 12e340c4b8efa70f3305f489af30461b7c0e5edd
8864 Author: Chris Wilson <chris@chris-wilson.co.uk>
8865 Date: Sun May 13 11:01:35 2012 +0100
8867 sna: Optimise kgem_clear_dirty() by keeping dirty buffers at the front
8869 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8871 commit b0dfd8223adee87e28b391a019effa7bb3630533
8872 Author: Chris Wilson <chris@chris-wilson.co.uk>
8873 Date: Sun May 13 08:33:23 2012 +0100
8875 sna: Flush the batch before blocking if the GPU is idle
8877 If we have work pending for the GPU and we believe it is idle, just
8878 submit the batch in order to improve GPU/CPU concurrency and reduce
8881 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8883 commit 6924fc525d6bc82901cfed769c176b44c0bce024
8884 Author: Chris Wilson <chris@chris-wilson.co.uk>
8885 Date: Fri May 11 14:04:09 2012 +0100
8887 sna: Fix off-by-one in computation of width for inplace trapezoids
8889 This lead to the dropping of the last pixel for small area trapezoids,
8890 such as the right hand outline of buttons under ClearLooks.
8892 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48320
8893 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8895 commit a58a543e84ed15c41f4fa7644be3ba7865d31b92
8896 Author: Chris Wilson <chris@chris-wilson.co.uk>
8897 Date: Thu May 10 12:59:55 2012 +0100
8899 sna: handle vmap creation failures gracefully
8901 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8903 commit 9af651fdf93f6ccdbcb147cb16989f84c2d0ebdc
8904 Author: Chris Wilson <chris@chris-wilson.co.uk>
8905 Date: Thu May 10 10:57:56 2012 +0100
8907 sna: Include some DBG for segment drawing
8909 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8911 commit 3aa98289e3a2f59633ac5f1a5f64db4ee55609d9
8912 Author: Chris Wilson <chris@chris-wilson.co.uk>
8913 Date: Thu May 10 10:31:03 2012 +0100
8915 sna: Correct adjustment of LineCap for rectilinear segments
8917 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8919 commit ad69316ab1054c5e4b56c19b8eec87b9f86939e5
8920 Author: Chris Wilson <chris@chris-wilson.co.uk>
8921 Date: Thu May 10 10:03:49 2012 +0100
8923 sna: Include a small amount of timer-slack to avoid very short sleeps
8925 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8927 commit 29d1d409b259ba0c85947ef5eef903ea527aa118
8928 Author: Chris Wilson <chris@chris-wilson.co.uk>
8929 Date: Thu May 10 09:54:53 2012 +0100
8931 sna: Avoid doubling the final step of poly lines
8933 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8935 commit a83d90ee61be44a2a36b56ad24bbc6544320448f
8936 Author: Chris Wilson <chris@chris-wilson.co.uk>
8937 Date: Wed May 9 20:34:52 2012 +0100
8939 sna: Avoid randomly changing domains of active bo
8941 After attaching the bo to the scanout, mark it as retired in order to
8942 update its domains so that the assertion during retirement later is
8945 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
8946 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49526
8947 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8949 commit d5200510b8f2a0ccb2e0c45dfebc659f8360780c
8950 Author: Chris Wilson <chris@chris-wilson.co.uk>
8951 Date: Wed May 9 17:58:58 2012 +0100
8953 sna: Respect the error terms when reversing the PolyLine DDA
8955 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8957 commit a3d37fb29f8dffb0e370ad95783994aaa7eccfaf
8958 Author: Chris Wilson <chris@chris-wilson.co.uk>
8959 Date: Tue May 8 15:47:14 2012 +0100
8961 sna: Force remapping for IO transfer
8963 Should fix regression from fcccc5528 (sna: Improve handling of inplace
8964 IO for large transfers) whereby it was aborting the transfer it we need
8965 to remap the buffer for the upload.
8967 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49546
8968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8970 commit 2a9a93e4484e0c616724610f4c8019fcbaa7ad53
8971 Author: Chris Wilson <chris@chris-wilson.co.uk>
8972 Date: Tue May 8 15:12:51 2012 +0100
8974 sna: Only avoid ring switching for indirect uploads
8976 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8978 commit 613902b60e0f2ca2a916e68306a1a37bc236d00d
8979 Author: Chris Wilson <chris@chris-wilson.co.uk>
8980 Date: Tue May 8 14:17:46 2012 +0100
8982 sna: Fix off-by-one in checking available execbuffer slots
8984 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8986 commit 663e387b35c314c4c2bee8137d6b70d27fa9f729
8987 Author: Chris Wilson <chris@chris-wilson.co.uk>
8988 Date: Tue May 8 11:45:34 2012 +0100
8990 sna: Only submit a batch to the scanout if it is not already busy
8992 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
8994 commit cd7a56b7313233190a4c4a735d4a141e99c9b688
8995 Author: Chris Wilson <chris@chris-wilson.co.uk>
8996 Date: Tue May 8 08:53:41 2012 +0100
8998 sna: Throttle independently of scanout updates
9000 As we are now throttling to relieve GTT pressure, it is a benefit to
9001 consistently throttle before blocking.
9003 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9005 commit 2372176f73b7c945f56b7673eab5eccb86366416
9006 Author: Chris Wilson <chris@chris-wilson.co.uk>
9007 Date: Mon May 7 12:57:35 2012 +0100
9009 sna: Throttle execution when searching for inactive buffers
9011 If we have some active buffers that we may reuse and old outstanding
9012 requests, throttling before retiring should prevent the CPU from running
9013 away from the GPU and hogging the entire GTT and RAM.
9015 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9017 commit 4df228749729dd540b639368400fa20118cdf412
9018 Author: Chris Wilson <chris@chris-wilson.co.uk>
9019 Date: Mon May 7 10:42:30 2012 +0100
9021 sna: Rate-limit and shrink bo usage if we hit system resource limits
9023 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9025 commit ca4d2296e6e42e837627756790b262cae0fd3b6c
9026 Author: Chris Wilson <chris@chris-wilson.co.uk>
9027 Date: Mon May 7 10:23:19 2012 +0100
9029 sna: Expand batch buffers
9031 As batch buffers are compacted to fit into the smallest bo, the only
9032 cost is the larger static array allocation (and presumably cache
9035 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9037 commit 9281b80644ce76ad9e0f3f8f812cbae97c10814a
9038 Author: Chris Wilson <chris@chris-wilson.co.uk>
9039 Date: Mon May 7 23:25:22 2012 +0100
9041 sna/gen[345]: Clear used vertices when discarding unmappable vbo
9043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9045 commit a6ee376e93517659391905e6c9018b3bb735135d
9046 Author: Chris Wilson <chris@chris-wilson.co.uk>
9047 Date: Mon May 7 10:24:21 2012 +0100
9049 sna: Use the correct invocation of kgem_bo_destroy() for sanity-checks
9051 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9053 commit c89c7e9a04314e40cee5514a182a8364c4f99374
9054 Author: Chris Wilson <chris@chris-wilson.co.uk>
9055 Date: Mon May 7 11:54:50 2012 +0100
9057 sna: Update select timeout when installing a timer in the block handler
9059 The block handler is run after the timers are queried for their
9060 expiration and so if we install a timer in the block hander, we must
9061 set the timeout ourselves.
9063 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9065 commit 4ba366ead3a133136554579fe9a604da1fc1da68
9066 Author: Chris Wilson <chris@chris-wilson.co.uk>
9067 Date: Mon May 7 08:55:35 2012 +0100
9069 sna: Manually execute the timer as TimerForce does not run an inactive timer
9071 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9073 commit 2c801c45dbad58d18f2673e17723825dda0c4e55
9074 Author: Chris Wilson <chris@chris-wilson.co.uk>
9075 Date: Sun May 6 17:23:11 2012 +0100
9077 sna: Replace timerfd with OsTimer
9079 As timerfd is linux-specific, and OsTimer an OS-agnostic abraction,
9080 replace the former with the later. Arguably this has slightly better
9081 performance characteristics in select-bound loops.
9083 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9085 commit 74d2707de4a0f94d2143f1e9c60762e4167b1ea6
9086 Author: Chris Wilson <chris@chris-wilson.co.uk>
9087 Date: Sun May 6 21:51:27 2012 +0100
9089 sna: Remove short-circuiting for large font fallbacks
9091 Unlike the fallback for an unhandled depth, we need to ensure that the
9092 pixmaps are mapped to the CPU before calling the fb routines.
9094 Reported-by: Toralf Förster <toralf.foerster@gmx.de>
9095 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49558
9096 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9098 commit c5b6741d3729c6867702ab64a6c59cb8052c0ef3
9099 Author: Chris Wilson <chris@chris-wilson.co.uk>
9100 Date: Sun May 6 12:40:54 2012 +0100
9102 sna/gen2+: Fix typo for computing redirected extents for render copy
9104 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9106 commit 771090f25db702d25ebbd3f2b44429cf0acfe8fd
9107 Author: Chris Wilson <chris@chris-wilson.co.uk>
9108 Date: Fri May 4 20:56:37 2012 +0100
9110 sna: Add a pair of asserts to track down a NULL pointer dereference
9112 Looks like the assumption for the location of the data is invalid,
9113 allocation failure, perhaps?
9115 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
9116 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9118 commit d376a960df7a081a5d449f77b81ae13223b98929
9119 Author: Chris Wilson <chris@chris-wilson.co.uk>
9120 Date: Fri May 4 10:09:46 2012 +0100
9122 sna/dri: Only track a single pending flip across all pipes
9124 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9126 commit 450592f989efd0d3bc9ef2de245fce0a180e91a2
9127 Author: Chris Wilson <chris@chris-wilson.co.uk>
9128 Date: Fri May 4 09:50:19 2012 +0100
9130 sna: Cache the framebuffer id
9132 Also fixup a weakness of only tracking scanout with a single bit, as we
9133 used to clear it forcibly after every flip.
9135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9137 commit 29d035279b2fe98d5ba9cf01125faea34d36fb76
9138 Author: Chris Wilson <chris@chris-wilson.co.uk>
9139 Date: Fri May 4 09:11:31 2012 +0100
9141 sna/dri: pageflip unref debugging
9143 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9145 commit 079b491ced2c9c3c73d938ef6025d040016ad3a7
9146 Author: Chris Wilson <chris@chris-wilson.co.uk>
9147 Date: Thu May 3 23:52:17 2012 +0100
9149 sna: Ensure drawables are clipped against the pixmap before migration
9151 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9153 commit 1a5f8599b150064339d62a97c58026e62b49ff27
9154 Author: Chris Wilson <chris@chris-wilson.co.uk>
9155 Date: Thu May 3 23:21:48 2012 +0100
9157 sna: Compile fix for fresh assertion
9159 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9161 commit 61cac5c265279d45677262216a0ba56f548cd898
9162 Author: Chris Wilson <chris@chris-wilson.co.uk>
9163 Date: Thu May 3 22:33:59 2012 +0100
9165 sna: Maintain a reference to the chain of proxies
9167 Rather than attempt to flatten the chain to the last link, we may need
9168 to hold a reference to the intermediate links in case of batch buffer
9171 Fixes http://tnsp.org/~ccr/intel-gfx/test.html
9173 Reported-by: Matti Hamalainen <ccr@tnsp.org>
9174 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=49436
9175 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9177 commit dea5d429f7a52dfc945b17a57ef79744cc796b0e
9178 Author: Chris Wilson <chris@chris-wilson.co.uk>
9179 Date: Thu May 3 17:35:24 2012 +0100
9181 sna: Remove extraneous SCANOUT flags
9183 These are now fixed by obeying a minimum alignment restriction.
9185 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9187 commit 19fd24a4db994bb5c5ce4a73f06d9394a758ea91
9188 Author: Chris Wilson <chris@chris-wilson.co.uk>
9189 Date: Thu May 3 17:35:10 2012 +0100
9191 sna: Fix offset for combining damage
9193 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9195 commit 1376c81dbf3b789e04e6804df1b1fd32bcb2bd1d
9196 Author: Chris Wilson <chris@chris-wilson.co.uk>
9197 Date: Thu May 3 17:34:28 2012 +0100
9199 sna: Debug option to force particular upload/download paths
9201 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9203 commit f0d464d6b18855616fc43d9a25aa6853f86c8e2b
9204 Author: Chris Wilson <chris@chris-wilson.co.uk>
9205 Date: Thu May 3 16:17:35 2012 +0100
9207 sna/dri: Balance flush counting
9209 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9211 commit d47e98dd64c0b9fe2979db42622c5ee8168e8b35
9212 Author: Chris Wilson <chris@chris-wilson.co.uk>
9213 Date: Thu May 3 14:54:17 2012 +0100
9215 sna: Minor glyph fallback fixes
9217 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9219 commit a1f08b8850616952fb0babe2275eb36b13a380ec
9220 Author: Chris Wilson <chris@chris-wilson.co.uk>
9221 Date: Thu May 3 14:53:29 2012 +0100
9223 sna: Don't discard GPU buffer if we only want to read back for the operation
9225 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9227 commit fcccc5528b8696fb4f9b3f9f528673b95d98a907
9228 Author: Chris Wilson <chris@chris-wilson.co.uk>
9229 Date: Thu May 3 11:27:44 2012 +0100
9231 sna: Improve handling of inplace IO for large transfers
9233 If the transfer is large enough to obliterate the caches, then it is
9234 preferrable to do it inplace rather than upload a proxy texture and
9235 queue a blit. This helps prevent an inconsistency where one layer
9236 believes the operation should be done inplace only for the IO layer to
9237 perform an indirect upload.
9239 Testing show no significant impact upon the cairo-traces, but it does
9240 prevent x11perf -shmput from exploding.
9242 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9244 commit 53568e8e49559094ce5b24b8709669f1f76fe2bf
9245 Author: Chris Wilson <chris@chris-wilson.co.uk>
9246 Date: Mon Apr 30 09:24:06 2012 +0100
9248 sna/gen7: Fix debug printing of primitives
9250 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9252 commit 01c26a44fdce761781908be11102e7a6a3db523c
9253 Author: Chris Wilson <chris@chris-wilson.co.uk>
9254 Date: Wed May 2 21:09:27 2012 +0100
9256 sna: Avoid reducing damage for synchronisation
9258 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9260 commit f4c34e9ab32f31669896b8f626195827a85af337
9261 Author: Chris Wilson <chris@chris-wilson.co.uk>
9262 Date: Wed May 2 19:09:10 2012 +0100
9264 sna: Bring back the minimum alignment for G33
9266 The underlying cause is still not fixed. It should be possible to use
9267 the much laxer alignment for single-stream linear. Still no idea how I
9268 fail to convince the GPU to drop the depth buffer.
9270 Reported-by: Matti Hamalainen <ccr@tnsp.org>
9271 References: https://bugs.freedesktop.org/show_bug.cgi?id=49391
9272 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9274 commit 10b4a9bb5f46ab9d9c8b165084ce4174b54a8d39
9275 Author: Chris Wilson <chris@chris-wilson.co.uk>
9276 Date: Wed May 2 12:45:09 2012 +0100
9278 sna: Always try to operate inplace if we an LLC gpu bo
9280 commit dd80fb00bf7acf37dc3b9125431a12b67d7e92d3
9281 Author: Chris Wilson <chris@chris-wilson.co.uk>
9282 Date: Wed May 2 11:31:42 2012 +0100
9284 sna: Fallback for glyphs too large for XY_TEXT_IMMEDIATE
9286 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9288 commit 7e09babb3e9e23882db30ee7d0c22c503962faa9
9289 Author: Chris Wilson <chris@chris-wilson.co.uk>
9290 Date: Tue May 1 12:37:45 2012 +0100
9292 sna: Only attempt to reuse exported scanout buffers
9296 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9298 commit b4b32e7a0172a74372f800e9c74d639a23c5ff34
9299 Author: Chris Wilson <chris@chris-wilson.co.uk>
9300 Date: Tue May 1 11:15:00 2012 +0100
9302 sna: Fast-path unclipped glyphs
9304 Avoid the redundant computation of the glyph intersection with the
9305 drawable bounding box.
9307 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9309 commit ffdf9aca12adcfa1ec7ab7a1706873105a5f0d4a
9310 Author: Chris Wilson <chris@chris-wilson.co.uk>
9311 Date: Sun Apr 29 21:59:52 2012 +0100
9313 legacy/i810: hwmc additionally depends upon building DRI
9315 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9317 commit 444da84c47266bcbbdf5121507901de8eb36f11b
9318 Author: Chris Wilson <chris@chris-wilson.co.uk>
9319 Date: Sun Apr 29 21:49:04 2012 +0100
9321 configure: Version bump for 2.19.0 release
9323 commit a206a1eee1fb799d567d68db564d663cfc66f6cc
9324 Author: Chris Wilson <chris@chris-wilson.co.uk>
9325 Date: Sat Apr 28 01:54:43 2012 +0100
9327 sna: Tune relocation array size
9329 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9331 commit 4f1908c651ef9e2af33d8831466a605234978c46
9332 Author: Chris Wilson <chris@chris-wilson.co.uk>
9333 Date: Fri Apr 27 23:19:56 2012 +0100
9335 sna: PolyPoint only uses the gc->fgPixel
9337 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9339 commit 8453034c7dd893f1d4c32ee87724f3a13137595d
9340 Author: Chris Wilson <chris@chris-wilson.co.uk>
9341 Date: Fri Apr 27 22:08:51 2012 +0100
9343 sna/gen6: Allow ring switching at the start of a batch
9345 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9347 commit 93ad7793894787600c5074917c753fa7c6816134
9348 Author: Chris Wilson <chris@chris-wilson.co.uk>
9349 Date: Fri Apr 27 22:03:55 2012 +0100
9351 sna: Tweak semaphores-enabled heuristic
9353 The kernel module now defaults to -1, confusing the test.
9355 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9357 commit 986dbdda3bf8dcf208e55543d8a3393c4b53f10b
9358 Author: Chris Wilson <chris@chris-wilson.co.uk>
9359 Date: Fri Apr 27 19:21:44 2012 +0100
9361 sna: Tweak placement choice for high-overhead operations
9363 Some operations cost more to setup than to transfer data back and forth!
9365 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9367 commit daac9a1d036d80ccce83438b49115a236a16bfb6
9368 Author: Chris Wilson <chris@chris-wilson.co.uk>
9369 Date: Fri Apr 27 16:50:28 2012 +0100
9371 sna: Micro-optimise common case of checking a single fenced bo
9373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9375 commit 5967d76ca09a257ec9db66ea664158e1dfd083ba
9376 Author: Chris Wilson <chris@chris-wilson.co.uk>
9377 Date: Wed Apr 25 17:15:37 2012 +0100
9379 sna: Fixup broken assertion
9381 It is valid for the cpu_bo to be NULL, as we may be choosing to free the
9382 large shadow pixel buffer instead.
9384 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9386 commit 860d3859b586939cd52e45b944cb6abd2a2ca71b
9387 Author: Chris Wilson <chris@chris-wilson.co.uk>
9388 Date: Wed Apr 25 16:04:33 2012 +0100
9390 sna/gen7: Add CS stall before changing WM binding table
9392 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9394 commit c219283460c0f2dfdb823e0cb139d05075c6afce
9395 Author: Chris Wilson <chris@chris-wilson.co.uk>
9396 Date: Wed Apr 25 15:04:01 2012 +0100
9398 sna/gen7: Apply more recent improvements from SNB perf tuning
9400 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9402 commit a3371613c9bf577a69cdf811ca1bebaea46bbe95
9403 Author: Chris Wilson <chris@chris-wilson.co.uk>
9404 Date: Wed Apr 25 11:09:35 2012 +0100
9406 sna: Do not automagically convert GTT mappings on untiled scanout to CPU
9408 The likelihood of an untiled mapping of the scanout is slim, except for
9409 gen3 with large desktops, and there it should never be in the CPU
9412 The issue is that we may perform an operation "inplace", yet incoherent
9413 with the display engine, and never flush the CPU cache, resulting in
9414 render corruption. In theory at least!
9416 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9418 commit 1abd92cd012ee46d44ed4873a5e750d56ae6668f
9419 Author: Chris Wilson <chris@chris-wilson.co.uk>
9420 Date: Wed Apr 25 11:16:30 2012 +0100
9422 sna: Clear the domain tracking after attaching the bo to scanout
9424 This is basically to make sure we don't continue treating it as CPU
9427 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9429 commit 8c58c840b1ba579a5601804fc710c58e1e00213f
9430 Author: Chris Wilson <chris@chris-wilson.co.uk>
9431 Date: Mon Apr 23 11:09:37 2012 +0100
9433 sna/dri: Always clear the scanout when destroying dri2 buffers
9435 As we may end up holding onto and releasing the Screen pixmap last, we
9436 may also be responsible for flushing the last reference to the scanout.
9438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9440 commit caf9144271a10f90ea580c246b2df3f69a10b7a0
9441 Author: Chris Wilson <chris@chris-wilson.co.uk>
9442 Date: Fri Apr 20 17:15:37 2012 +0100
9444 sna: Remove the assertions that the cached upload buffers are active
9446 These were added to track down some corruption, but the assertions
9447 themselves are incorrect, just very rare. The upload buffer may
9448 genuinely be cached if we abort the render operation after uploading the
9449 source data, leaving the proxy not coupled to any request.
9451 References: https://bugs.freedesktop.org/show_bug.cgi?id=48400
9452 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9454 commit aff3614efd5c12e658fa5723934e5bd50a83a316
9455 Author: Chris Wilson <chris@chris-wilson.co.uk>
9456 Date: Fri Apr 20 13:21:40 2012 +0100
9458 sna: Always clear the mmapped domains when reusing partial upload buffers
9460 As we need to make sure that we do invalidate the caches appropriately
9461 on reuse. Mildly paranoid, but strictly required by the spec.
9463 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9465 commit cb6a3dc2edf3cd612f833bc9a4656166735ee856
9466 Author: Chris Wilson <chris@chris-wilson.co.uk>
9467 Date: Thu Apr 19 10:34:23 2012 +0100
9469 sna: Discard proxy upload buffer if we choose to render to it
9471 Even if we try to avoid treating an upload buffer as a real GPU target,
9472 we may still choose to migrate the buffer to the GPU in order to keep
9473 other buffers on the GPU. In that case, we do want to create a real GPU
9476 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9478 commit 4cf74d409ca63c6a479c1ee2187908c04f3b830b
9479 Author: Chris Wilson <chris@chris-wilson.co.uk>
9480 Date: Thu Apr 19 09:09:32 2012 +0100
9482 sna: Don't consider upload proxies as being on the GPU for render targets
9484 The upload proxy is a fake buffer that we do not want to render to as
9485 then the damage tracking become extremely confused and the buffer it
9486 self is not optimised for persistent rendering. We assert that we do not
9487 use it as a render target, and this patch adds the check so that we
9488 avoid treating the proxy as a valid target when choosing the render
9491 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9493 commit fd81408b978c9b57c046ee43d2d32e1370e83a7d
9494 Author: Chris Wilson <chris@chris-wilson.co.uk>
9495 Date: Wed Apr 18 11:39:43 2012 +0100
9497 sna: Increase the render target alignment to 4 pixels on gen4+ as well
9499 Repoerted-and-tested-by: Toralf Förster <toralf.foerster@gmx.de
9500 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48865
9501 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9503 commit 11599e52b842b5db76798879b0fbb57762fe6002
9504 Author: Chris Wilson <chris@chris-wilson.co.uk>
9505 Date: Tue Apr 17 20:08:14 2012 +0100
9507 sna/dri: Decouple the frame event info after attaching along error paths
9509 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9511 commit b817200371bfe16f44b879a793cf4a75ad17bc5c
9512 Author: Chris Wilson <chris@chris-wilson.co.uk>
9513 Date: Tue Apr 17 17:54:58 2012 +0100
9515 Don't issue a scanline wait while VT switched
9517 Be paranoid and check that we own the VT before emitting a scanline
9518 wait. If we attempt to wait on a fb/pipe that we do not own, we may
9519 issue an illegal command and cause a lockup.
9521 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9523 commit af4a6e8cb52ace594934446e6d8a7aaa1945a9b0
9524 Author: Chris Wilson <chris@chris-wilson.co.uk>
9525 Date: Mon Apr 16 20:57:03 2012 +0100
9527 sna: Don't assert exported buffers are not busy
9529 As we do not fully control these buffers, we cannot truly say when they
9530 are idle, we can only trust that the split between us and the compositor
9531 doesn't lead to much corruption.
9533 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9535 commit a16616209bb2dcb7aaa859b38e154f0a10faa82b
9536 Author: Chris Wilson <chris@chris-wilson.co.uk>
9537 Date: Sat Apr 14 19:03:25 2012 +0100
9539 uxa: Fix leak of glyph mask for unhandled glyph composition
9541 ==1401== 7,344 bytes in 34 blocks are possibly lost in loss record 570 of 587
9542 ==1401== at 0x4027034: calloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
9543 ==1401== by 0x8BE5150: drm_intel_gem_bo_alloc_internal (intel_bufmgr_gem.c:689)
9544 ==1401== by 0x899FC04: intel_uxa_create_pixmap (intel_uxa.c:1077)
9545 ==1401== by 0x89C2C41: uxa_glyphs (uxa-glyphs.c:254)
9546 ==1401== by 0x21F05E: damageGlyphs (damage.c:647)
9547 ==1401== by 0x218E06: ProcRenderCompositeGlyphs (render.c:1434)
9548 ==1401== by 0x15AA40: Dispatch (dispatch.c:439)
9549 ==1401== by 0x1499E9: main (main.c:287)
9551 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9553 commit ae145c21e9fd3a12164f8b4720d059f9c158249e
9554 Author: Chris Wilson <chris@chris-wilson.co.uk>
9555 Date: Sat Apr 14 18:42:23 2012 +0100
9557 sna: Avoid leaking the plane resources when determining sprite planes
9559 Fixes the tiny, one-off leak:
9561 ==1407== 8 bytes in 1 blocks are definitely lost in loss record 48 of 527
9562 ==1407== at 0x402894D: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
9563 ==1407== by 0x8580BE8: drmMalloc (xf86drm.c:147)
9564 ==1407== by 0x8583D54: drmAllocCpy (xf86drmMode.c:73)
9565 ==1407== by 0x8585265: drmModeGetPlaneResources (xf86drmMode.c:955)
9566 ==1407== by 0x8A1BCE9: sna_video_sprite_setup (sna_video_sprite.c:367)
9567 ==1407== by 0x8A1A0A3: sna_video_init (sna_video.c:523)
9568 ==1407== by 0x89FD4E0: sna_screen_init (sna_driver.c:935)
9569 ==1407== by 0x15AD80: AddScreen (dispatch.c:3909)
9570 ==1407== by 0x19A2DB: InitOutput (xf86Init.c:817)
9571 ==1407== by 0x14981C: main (main.c:204)
9573 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9575 commit 69a7737abeded6ee923643bd8a80a5a84e6a979c
9576 Author: Chris Wilson <chris@chris-wilson.co.uk>
9577 Date: Sat Apr 14 12:06:51 2012 +0100
9579 sna: Align texture subsurfaces to 2x2 texture samples
9581 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9583 commit 1d2a46e0902d82b43a5e12af36521a6a7fd6ba39
9584 Author: Chris Wilson <chris@chris-wilson.co.uk>
9585 Date: Sat Apr 14 12:04:23 2012 +0100
9587 sna: Align redirect subsurfaces to 2x2 or 4x4 render spans
9589 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9591 commit 1ce2b65d622797000e0a4db7dc851d5b1da04f85
9592 Author: Chris Wilson <chris@chris-wilson.co.uk>
9593 Date: Sat Apr 14 11:59:31 2012 +0100
9595 sna: Align render target sizes on gen2/3 to 4x4 render spans
9597 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9599 commit 89f2b09b1e5be9842747998ea4fe32a6f1ede4cc
9600 Author: Chris Wilson <chris@chris-wilson.co.uk>
9601 Date: Fri Apr 13 16:37:43 2012 +0100
9603 sna: Avoid using TILING_Y for large objects on gen2/3
9605 References: https://bugs.freedesktop.org/show_bug.cgi?id=48636
9606 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9608 commit eaadbce122059066353743f1653aa16e9d9b747f
9609 Author: Chris Wilson <chris@chris-wilson.co.uk>
9610 Date: Fri Apr 13 15:12:36 2012 +0100
9612 sna: Relax bogus assertion
9614 The bo may be considered unmappable due to being bound to outside the
9615 mappable region, which we are attempting to rectify through mapping into
9618 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9620 commit b478420740d05fa87ddbd92042b1f7f2d002f73e
9621 Author: Chris Wilson <chris@chris-wilson.co.uk>
9622 Date: Fri Apr 13 15:07:13 2012 +0100
9624 sna: Limit the buffer reuse for mappable uploads to only those with mmaps
9626 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9628 commit 90e2740e7e459c56205fa65bab1ae3dbfd5d3945
9629 Author: Chris Wilson <chris@chris-wilson.co.uk>
9630 Date: Fri Apr 13 13:51:57 2012 +0100
9632 sna: Remove the conflicting assertion during GTT map
9634 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
9635 References: https://bugs.freedesktop.org/show_bug.cgi?id=48636
9636 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9638 commit 9e6d55a8d63f10ca6a2b10e44d00c84b07724485
9639 Author: Chris Wilson <chris@chris-wilson.co.uk>
9640 Date: Fri Apr 13 09:56:39 2012 +0100
9642 sna: Don't use miSpan code for wide-spans by default, too expensive
9644 Only use the fall-forward miSpans code when it prevents a readback.
9646 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9648 commit 9becfbbf89f2b170e50f705cabfc7bbf1dcf9846
9649 Author: Chris Wilson <chris@chris-wilson.co.uk>
9650 Date: Fri Apr 13 09:38:57 2012 +0100
9652 sna: Draw dashed PolyLines twice, once for the fgPixel, once for the bgPixel
9654 As the fast paths only setup state upfront, we were missing the state
9655 changes required between dash-on/off. Take advantage of that each pixel
9656 is only drawn once to batch the state changes and run the
9657 miZeroDashLines twice.
9659 A future task would be to use a custom line drawing routine...
9661 Fixes regression from ec1267df746512c2e262ef0bd9e9527bc5efe6f4.
9663 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9665 commit e269ed5d4e1b9d758aeb9a85ed0fa631f0aff0b1
9666 Author: Chris Wilson <chris@chris-wilson.co.uk>
9667 Date: Thu Apr 12 22:46:22 2012 +0100
9669 sna: Restore CPU domain for vmapped buffers when reusing
9671 For a vmapped upload buffer, we need to notify the kernel (and thereby
9672 the GPU) to invalidate the sampler and flush its caches when we reuse an
9675 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9677 commit d29b8650c40c673e6ddddaf52db9247e9836cba8
9678 Author: Chris Wilson <chris@chris-wilson.co.uk>
9679 Date: Thu Apr 12 22:23:12 2012 +0100
9681 sna: Revert use of mmap64()
9683 As this just causes mayhem on a 64-bit platform. Doomed if you, doomed
9686 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9688 commit a78b1d71a39ae29d5f85bd82c09202ebec3e6539
9689 Author: Chris Wilson <chris@chris-wilson.co.uk>
9690 Date: Thu Apr 12 15:37:25 2012 +0100
9692 sna: Declare AC_SYS_LARGEFILE for mmap64
9694 In order to use the full 32-bits of mmap address space on small
9695 platforms we need to use mmap64().
9697 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9699 commit 09deba927daa96be6230b1c3e1b425622512d8a2
9700 Author: Chris Wilson <chris@chris-wilson.co.uk>
9701 Date: Tue Apr 10 20:59:36 2012 +0100
9703 sna: Check ioctl return from set-domain
9705 Let's not assume it succeeds and so avoid altering our bookkeeping along
9708 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9710 commit 0b12f1d8e4d0a4fafac9553f144535efc4ebe0be
9711 Author: Chris Wilson <chris@chris-wilson.co.uk>
9712 Date: Tue Apr 10 15:23:54 2012 +0100
9714 sna: Fix shadowed variable
9716 sna_accel.c: In function 'sna_pixmap_move_area_to_gpu':
9717 sna_accel.c:1751:12: warning: declaration of 'flags' shadows a parameter
9719 sna_accel.c:1731:72: warning: shadowed declaration is here [-Wshadow]
9721 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9723 commit 755a7107aed268d87c5cc0feb1ba388b0cb7fc59
9724 Author: Chris Wilson <chris@chris-wilson.co.uk>
9725 Date: Tue Apr 10 15:19:19 2012 +0100
9727 sna: Fix typo and use the right pointer for kgem_bo_destroy
9729 Useless warnings in xorg headers ftl.
9731 References: https://bugs.freedesktop.org/show_bug.cgi?id=48400
9732 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9734 commit c8502e350cb18f6f5d821d237ffcee453f347eba
9735 Author: Chris Wilson <chris@chris-wilson.co.uk>
9736 Date: Tue Apr 10 14:43:28 2012 +0100
9738 sna: Add missing alloc failure check for creating tile source
9740 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9742 commit 9cc6f7ccc55cc11f47b3b7d626c9f5a7c1327d57
9743 Author: Chris Wilson <chris@chris-wilson.co.uk>
9744 Date: Tue Apr 10 12:06:56 2012 +0100
9746 sna: Release the freed bo cache upon expire
9748 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9750 commit 102d11906a672140bac099e7bd1b35345d13a2fc
9751 Author: Chris Wilson <chris@chris-wilson.co.uk>
9752 Date: Mon Apr 9 22:54:51 2012 +0100
9754 sna: Check for an inactive partial buffer to reuse after retiring requests
9756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9758 commit e2fb2421d8f5c07925d1699673aa4b1dd6c6b22c
9759 Author: Chris Wilson <chris@chris-wilson.co.uk>
9760 Date: Mon Apr 9 22:54:15 2012 +0100
9762 sna: Release partial buffers during cache expiration
9764 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9766 commit 333fdcad8677675a4758223c1a980c90d970ee42
9767 Author: Chris Wilson <chris@chris-wilson.co.uk>
9768 Date: Mon Apr 9 22:53:31 2012 +0100
9770 sna: Repeat expire whilst there remaining outstanding requests
9772 Do not allow the cache expiration to finish if we are still running
9775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9777 commit 778232e3d2fb5340a3092014801dc00a56c56d42
9778 Author: Chris Wilson <chris@chris-wilson.co.uk>
9779 Date: Tue Apr 10 11:14:50 2012 +0100
9781 sna: Only move the bo into the read domain for readback
9783 And mark it as currently in no domain afterwards, so that if we reuse
9784 it, it will be appropriately moved later.
9786 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9788 commit dd093eafb9b94b8e4cd8853d74078c3aa7e72f57
9789 Author: Chris Wilson <chris@chris-wilson.co.uk>
9790 Date: Mon Apr 9 14:09:42 2012 +0100
9792 sna: Add assertions around proxy list handling
9794 References: https://bugs.freedesktop.org/show_bug.cgi?id=48400
9795 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9797 commit 7f0bede3e7e3f92a637d1c886304b16afc0e34f2
9798 Author: Chris Wilson <chris@chris-wilson.co.uk>
9799 Date: Mon Apr 9 10:48:08 2012 +0100
9801 sna/traps: Use a temporary variable for the write pointer
9803 To avoid accumulating the write offset for wide spans, we need to reset
9804 the destination pointer between spans.
9806 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48332
9807 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9809 commit 2e4da00e3e03b873f5cad0cc5b1f6cc791852ca5
9810 Author: Chris Wilson <chris@chris-wilson.co.uk>
9811 Date: Mon Apr 9 10:42:18 2012 +0100
9813 sna/traps: Assert that the inplace row is contained before writing
9815 References: https://bugs.freedesktop.org/show_bug.cgi?id=48332
9816 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9818 commit 0464e93a088a9e8bc29ad8b36b6e12c3dda32ec6
9819 Author: Chris Wilson <chris@chris-wilson.co.uk>
9820 Date: Sun Apr 8 17:16:03 2012 +0100
9822 sna: Add some assertions for misuse of proxies
9824 References: https://bugs.freedesktop.org/show_bug.cgi?id=48400
9825 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9827 commit 479cb6ba71038fe44f66fb31fad90d0d454fea7a
9828 Author: Chris Wilson <chris@chris-wilson.co.uk>
9829 Date: Sun Apr 8 14:25:49 2012 +0100
9831 sna: Compress adjoining spans during FillSpans
9833 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9835 commit c5c01c13badeb7c2ead0c848b746d8d474277a77
9836 Author: Chris Wilson <chris@chris-wilson.co.uk>
9837 Date: Sun Apr 8 13:51:13 2012 +0100
9839 sna: Remove the duplicated check for use-bo? in PolySegments
9841 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9843 commit f5deea4f60433ee2b0c2d02fba682fff8b7829e8
9844 Author: Chris Wilson <chris@chris-wilson.co.uk>
9845 Date: Sun Apr 8 13:39:47 2012 +0100
9847 sna: Correct partial-write flag for PolySegments fallback
9849 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9851 commit c3d7f4c1cf7d052163b7c4e74bb202c618f0eb76
9852 Author: Chris Wilson <chris@chris-wilson.co.uk>
9853 Date: Sun Apr 8 13:18:56 2012 +0100
9855 sna/gen3: Reset accumulated constants for each composite
9857 In particular the glyph routines require the composite setup to
9858 reinitialise state between glyph runs. This affects anything trying to
9859 use glyphs without a mask with a gradient source, for example.
9861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9863 commit 701473d20485a0557b4fb36efcbfbb8656e2f619
9864 Author: Chris Wilson <chris@chris-wilson.co.uk>
9865 Date: Sun Apr 8 10:09:42 2012 +0100
9867 sna: Release cached upload buffers when reusing a write buffer for readback
9869 References: https://bugs.freedesktop.org/show_bug.cgi?id=48400
9870 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9872 commit 1ecf17b2507f95e1fefea15833fa9f57ec256a2e
9873 Author: Chris Wilson <chris@chris-wilson.co.uk>
9874 Date: Sat Apr 7 10:01:01 2012 +0100
9876 sna/gradient: Compute the absolute delta between color stops
9878 Otherwise we do not detect gradients that start from white!
9880 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
9881 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48407
9882 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9884 commit 4356fae72db3a33935b575edf95c84fbb48072a7
9885 Author: Chris Wilson <chris@chris-wilson.co.uk>
9886 Date: Fri Apr 6 21:10:50 2012 +0100
9888 sna/video: Only wait upon the scanout pixmap
9890 Caught by the addition of the assertion.
9892 Reported-by: Jiri Slaby <jirislaby@gmail.com>
9893 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
9894 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9896 commit b790ba2ec9ead51227d85fc8630bc7505eb7d7b3
9897 Author: Chris Wilson <chris@chris-wilson.co.uk>
9898 Date: Fri Apr 6 15:38:02 2012 +0100
9900 sna: Correct the damage offset for redirected rendering
9902 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
9903 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48385
9904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9906 commit 0b81bafb802bb86454739ed46cf45571bccef735
9907 Author: Chris Wilson <chris@chris-wilson.co.uk>
9908 Date: Fri Apr 6 15:14:45 2012 +0100
9910 sna/glyphs: Prefer a temporary upload mask for large glyph masks
9912 If the required temporary mask is larger than the 3D pipeline can
9913 handle, just render to a CPU buffer rather than redirect every glyph
9916 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9918 commit 42a84613e34522af885b4b50d6c68ef77e81ffc3
9919 Author: Chris Wilson <chris@chris-wilson.co.uk>
9920 Date: Fri Apr 6 14:27:15 2012 +0100
9922 sna: Relase the upload cache when overwriting with PutImage
9924 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
9925 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48359
9926 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9928 commit 999aa210ff87919945c673bdd34bae76ac097681
9929 Author: Chris Wilson <chris@chris-wilson.co.uk>
9930 Date: Fri Apr 6 09:24:36 2012 +0100
9932 sna: Use a sentinel value to prevent accessing beyond the end of the y_buckets
9934 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9936 commit 1ae6328c57eb496072f0d0e27440f5d0901633b0
9937 Author: Chris Wilson <chris@chris-wilson.co.uk>
9938 Date: Fri Apr 6 09:12:08 2012 +0100
9940 sna: Remove redundant check from tor_inplace()
9942 We only execute full-steps for vertical edges so we do not need the
9945 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9947 commit 51b9202d27db3d98c6d82ba224bd8eb218533dd9
9948 Author: Chris Wilson <chris@chris-wilson.co.uk>
9949 Date: Wed Apr 4 11:13:27 2012 +0100
9951 sna: Only engage the GPU detiler for multiple rows
9953 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9955 commit 98ad4c3cd8647ba3ec90fb45157773c8e85e886c
9956 Author: Chris Wilson <chris@chris-wilson.co.uk>
9957 Date: Tue Apr 3 19:47:15 2012 +0100
9959 sna/gen3: Don't force use of the render pipeline just for vmap
9961 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9963 commit 0915d414f55a1bff4171981feb87bae212f29f23
9964 Author: Chris Wilson <chris@chris-wilson.co.uk>
9965 Date: Tue Apr 3 19:04:29 2012 +0100
9967 sna/gen3: Fix pre-multiplication of mask value
9969 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9971 commit 04851e4210d2d71542359c14d4b68d0851b36326
9972 Author: Chris Wilson <chris@chris-wilson.co.uk>
9973 Date: Tue Apr 3 12:34:24 2012 +0100
9975 sna/gen3: Convert the clear-color from picture->format to a8r8g8b8
9977 The shaders treat colours as an argb value, however the clear color is
9978 stored in the pixmap's native format (a8, r5g6b5, x8r8g8b8 etc). So
9979 before using the value of the clear color as a solid we need to convert
9980 it into the a8r8g8b8 format.
9982 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
9983 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48204
9984 Reported-by: Paul Neumann <paul104x@yahoo.de>
9985 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47308
9986 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9988 commit 87a672dafd9d6f47f31b77b406b7f0fb2b4030ac
9989 Author: Chris Wilson <chris@chris-wilson.co.uk>
9990 Date: Tue Apr 3 09:24:06 2012 +0100
9992 sna: Apply CoordMode when computing point extents
9994 Reported-by: Patrick Truebe <eko-priv@gmx.net>
9995 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48220
9996 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
9998 commit 0a0ee491ea18dc59748ff4419ae73bd1a369ae79
9999 Author: Chris Wilson <chris@chris-wilson.co.uk>
10000 Date: Tue Apr 3 09:23:49 2012 +0100
10002 sna: Debugging flil spans and their clipping
10004 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10006 commit ec1267df746512c2e262ef0bd9e9527bc5efe6f4
10007 Author: Chris Wilson <chris@chris-wilson.co.uk>
10008 Date: Mon Apr 2 16:16:24 2012 +0100
10010 sna: Use the solid spans fast paths for dashed zero-width lines as well
10012 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10014 commit 2d1f3cb198f2fe9602356a334a076abf3c68a9c7
10015 Author: Chris Wilson <chris@chris-wilson.co.uk>
10016 Date: Mon Apr 2 13:51:37 2012 +0100
10018 sna/gen4: Remove the accidental debugging hack from the last commit
10020 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10022 commit 5c4dc9c5db7b2a5b936bdbc15536c5cf0f7f5f23
10023 Author: Chris Wilson <chris@chris-wilson.co.uk>
10024 Date: Mon Apr 2 13:40:22 2012 +0100
10026 sna/gen3+: Fix sampling of borders around gradients
10028 Incurs a slight loss of precision for the internal gradient, but much
10029 more preferable to the artefacts around the borders with RepeatNone.
10031 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
10032 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45016
10033 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10035 commit 0b2651dc04cef8f9692b2557684f044b4980700f
10036 Author: Chris Wilson <chris@chris-wilson.co.uk>
10037 Date: Sun Apr 1 19:43:15 2012 +0100
10039 sna: Apply composite offset to damage for spans fast paths
10041 Reported-by: Jiri Slaby <jirislaby@gmail.com>
10042 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
10043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10045 commit 4ea9ab9303d21a62683055b75eaed66c97a5f289
10046 Author: Chris Wilson <chris@chris-wilson.co.uk>
10047 Date: Sun Apr 1 16:36:16 2012 +0100
10049 sna: Fix assertion to look at bbox of all boxes/points
10051 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10053 commit 932743bb333e35d5f6529a701137aad4d7490555
10054 Author: Chris Wilson <chris@chris-wilson.co.uk>
10055 Date: Sun Apr 1 09:54:43 2012 +0100
10057 sna: Assert that drawing boxes are within bounds
10059 More sanity checks required.
10061 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10063 commit cc20c45aa0ca15720510668d6918bf3c99104626
10064 Author: Chris Wilson <chris@chris-wilson.co.uk>
10065 Date: Fri Mar 30 22:51:21 2012 +0100
10067 sna: Minimise the risk of hotplug hangs by checking fb before vsync
10069 Everytime we issue a MI_WAIT_FOR_EVENT on a scan-line from userspace we
10070 run the risk of that pipe being disable before we submit a batch. As the
10071 pipe is then disabled or configured differently, we encounter an
10072 indefinite wait and trigger a GPU hang.
10074 To minimise the risk of a hotplug event being detected and submitting a
10075 vsynced batch prior to noticing the removal of the pipe, perform an
10076 explicit query of the current CRTC and delete the wait if we spot that
10077 our framebuffer is no longer attached. This is about as good as we can
10078 achieve without extra help from the kernel.
10080 Reported-by: Francis Leblanc <Francis.Leblanc-Lebeau@verint.com>
10081 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45413 (and others)
10082 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10084 commit 305734ebdf3d51c084cfbee8804b6c60b1f03a98
10085 Author: Chris Wilson <chris@chris-wilson.co.uk>
10086 Date: Fri Mar 30 21:27:29 2012 +0100
10088 sna: Separate out scanline waiting for gen4
10090 So that we do not set a gen4 bit on gen2 and apply the old workaround of
10091 trimming y2 instead.
10093 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10095 commit 6f2814db6f7b89e94e54b8d73c7e176ab7d1c469
10096 Author: Chris Wilson <chris@chris-wilson.co.uk>
10097 Date: Fri Mar 30 20:45:55 2012 +0100
10099 sna/traps: Align the pointer+index
10101 It's the location of the pixels within the row that matter for
10104 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47418
10105 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10106 Tested-by: Magnus Kessler <Magnus.Kessler@gmx.net>
10108 commit ee075ced844350785685a0f93f88f1dc310bcc73
10109 Author: Chris Wilson <chris@chris-wilson.co.uk>
10110 Date: Fri Mar 30 19:09:30 2012 +0100
10112 sna/traps: Align the pointer not the indices
10114 Magnus found that inplace_row was still crashing on his setup when it
10115 tried to perform an 8-byte aligned write to an unaligned pointer. This
10116 time it looks like the row pointer itself was not 8-byte aligned, so
10117 instead of assuming that and fixing up the indices, ensure that the
10118 (index+row) results in an 8-byte aligned value.
10120 Reported-by: Magnus Kessler <Magnus.Kessler@gmx.net>
10121 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47418
10122 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10124 commit fde8a010b3d9406c2f65ee99978360a6ca54e006
10125 Author: Chris Wilson <chris@chris-wilson.co.uk>
10126 Date: Fri Mar 30 12:47:21 2012 +0100
10128 uxa: Remove broken render glyphs-to-dst
10130 Reported-by: Vincent Untz <vuntz@gnome.org>
10131 Reported-by: Robert Bradford <robert.bradford@intel.com>
10132 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48045
10133 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10135 commit 451489b49916cf5a9d27844196f9656e590d9124
10136 Author: Chris Wilson <chris@chris-wilson.co.uk>
10137 Date: Fri Mar 30 10:21:26 2012 +0100
10139 sna/gen7: Allow per-device specific maxima
10141 As the maximum thread count and urb size differs between different
10142 incarnations of the GT units, be a little more flexible in programming
10145 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10147 commit ea36f2c4a3fa9afa8184eeaf944af9924c080368
10148 Author: Eugeni Dodonov <eugeni.dodonov@intel.com>
10149 Date: Thu Mar 29 21:08:29 2012 -0300
10151 Add support for Ivy Bridge GT2 Server chipset
10153 Sometimes known as Bromlow.
10155 Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
10156 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10158 commit 6142232fa0feeb39412cda85ca727cc770eaa042
10159 Author: Chris Wilson <chris@chris-wilson.co.uk>
10160 Date: Wed Mar 28 18:59:26 2012 +0100
10162 sna: Add video sprite support for ILK+
10164 Based on the work by Jesse Barnes.
10166 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10168 commit ae8aa172a7330439a8e6dda41f5e33eb257a139b
10169 Author: Chris Wilson <chris@chris-wilson.co.uk>
10170 Date: Wed Mar 28 22:07:10 2012 +0100
10172 sna: Fix up 32-bit overflow for maximum object size calculation
10174 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10176 commit 2e7b5f7eafbf452c781e50eba7dc8323260af59e
10177 Author: Chris Wilson <chris@chris-wilson.co.uk>
10178 Date: Tue Mar 27 10:42:59 2012 +0100
10180 sna/traps: Prefer to try mono spans on the GPU before trying inplace CPU
10182 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10184 commit 522b41b393b979fdccf1ad62194cde8129f98c8c
10185 Author: Chris Wilson <chris@chris-wilson.co.uk>
10186 Date: Tue Mar 27 10:42:21 2012 +0100
10188 sna/traps: Fix the width of the left-hand edge of an unaligned box
10190 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10192 commit 008ad39b72eb86dc2f89789427be269c148feaf4
10193 Author: Chris Wilson <chris@chris-wilson.co.uk>
10194 Date: Tue Mar 27 10:41:37 2012 +0100
10196 sna/gen6: Reduce opaque solid OVER to SRC for render composite
10198 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10200 commit a9e9041934b2486e5661ba6c550e6670009f69f3
10201 Author: Chris Wilson <chris@chris-wilson.co.uk>
10202 Date: Mon Mar 26 16:03:42 2012 +0100
10204 sna: Remove bogus assertion of no references to in-flight upload buffers
10206 As we may hold a cached reference to an upload buffer whilst it is
10207 in-flight, the assertion that there are no such references to a buffer
10208 being reused is no longer true. Those cached references will be released
10209 as soon as we retire the buffer during the readback and so we are free
10210 to reuse such an upload buffer for immediate readback.
10212 Reported-by: Jiri Slaby <jirislaby@gmail.com>
10213 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
10214 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10216 commit 89091dd8a3fba6520a651e3e135c87546b81df9b
10217 Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
10218 Date: Tue Mar 20 11:53:21 2012 -0300
10220 Avoid duplicated code with intel_output_create_ranged_atom
10222 Same change for intel_display.c and sna_display.c.
10224 Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
10225 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10227 commit 62f9833298ea936eaefe973eca04b1dde858fbd8
10228 Author: Chris Wilson <chris@chris-wilson.co.uk>
10229 Date: Sun Mar 25 21:25:15 2012 +0100
10231 sna/gen2+: Approximate expensive gradients when using imprecise rendering
10233 If we lack the ability to use a shader to compute the gradients
10234 per-pixel, we need to use pixman to render a fallback texture. We can
10235 reduce the size of this texture and upsample to reduce the cost with
10236 hopefully imperceptible loss of quality.
10238 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10240 commit 25807f472d051163ed96556a409110fa405c24d1
10241 Author: Chris Wilson <chris@chris-wilson.co.uk>
10242 Date: Fri Mar 23 14:56:06 2012 +0000
10244 uxa: Remove hook for CompositeRectangles
10246 It was broken and not flushing damage correctly. With the
10247 improvements made to the kernel, it is no longer a significant advantage
10248 per se and not worth its additional complexity.
10250 Reported-by: Tilman Sauerbeck <tilman@code-monkey.de>
10251 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32547
10252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10254 commit 4460c6d0d30cf20b76854d47fd9e3aecf1839f15
10255 Author: Chris Wilson <chris@chris-wilson.co.uk>
10256 Date: Fri Mar 23 11:05:55 2012 +0000
10258 configure: Stop the debug build erroring out if it cannot find valgrind
10260 Another case where I passed an empty string believing that would be
10261 sufficient to replace the error path...
10263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10265 commit 168ecd96f23756e185af627f814f81b8a4c0c529
10266 Author: Chris Wilson <chris@chris-wilson.co.uk>
10267 Date: Thu Mar 22 23:47:56 2012 +0000
10269 sna: Make the fallback debugging messages more consistent
10271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10273 commit 8693005b5427f256bcd7c2e7cfc48d56f5b8c5ea
10274 Author: Chris Wilson <chris@chris-wilson.co.uk>
10275 Date: Thu Mar 22 23:43:04 2012 +0000
10277 sna: Update the target bo after changing from Y tiling
10279 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10281 commit 83f98d6e5c303e52c1e0fb95b6237ebf62a8edfe
10282 Author: Chris Wilson <chris@chris-wilson.co.uk>
10283 Date: Thu Mar 22 21:57:14 2012 +0000
10285 sna: Correctly test for clear glyphs when searching for ones to skip
10287 With xterm, it is quite common for it to redraw itself by using lots of
10288 spaces and so it is efficient for us if we ellide those clear glyphs and
10289 only draw the backing boxes. However, we were only checking the first 8
10290 pixels in each line because of a missing pointer increment.
10292 Fixes absent '=' characters when using a compositor and ImageText.
10294 Reported-by: Jiri Slaby <jirislaby@gmail.com
10295 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47735
10296 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10298 commit 86121a3af9a9fc9a2c76d7ac9f3ec17105d20d80
10299 Author: Chris Wilson <chris@chris-wilson.co.uk>
10300 Date: Thu Mar 22 11:21:43 2012 +0000
10302 sna: Adjust the damage region for the composite offset
10304 References: https://bugs.freedesktop.org/show_bug.cgi?id=47597
10305 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10307 commit a6b48dd7f1eeb1a8e3841b8f3326c60b300ee9e9
10308 Author: Chris Wilson <chris@chris-wilson.co.uk>
10309 Date: Thu Mar 22 09:22:52 2012 +0000
10311 sna: Force fallbacks if the destination is unattached
10313 Since the removal of the ability to create a backing pixmap after the
10314 creation of its parent, it no longer becomes practical to attempt
10315 rendering with the GPU to unattached pixmaps. So having made the
10316 decision never to render to that pixmap, perform the test explicitly
10317 along the render paths.
10319 This fixes a segmentation fault introduced in 8a303f195 (sna: Remove
10320 existing damage before overwriting with a composite op) which assumed
10321 the existence of a backing pixmap along a render path.
10323 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47700
10324 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10326 commit 52f39ae1697bef86471b7c5eef8553661f255b67
10327 Author: Chris Wilson <chris@chris-wilson.co.uk>
10328 Date: Wed Mar 21 13:31:03 2012 +0000
10330 sna: Assert that the bo created is large enough
10332 Double check that the maximum access size computed from the bo
10333 parameters is within the allocated size for the bo.
10335 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10337 commit 2f23c660f58e0db2107255f32a3503dae62fe304
10338 Author: Chris Wilson <chris@chris-wilson.co.uk>
10339 Date: Wed Mar 21 13:30:45 2012 +0000
10341 sna: Assert that the tiled blt is correctly clipped
10343 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10345 commit e0cd13e2af3db237c171f8ca1c70ddc67b1fbce4
10346 Author: Chris Wilson <chris@chris-wilson.co.uk>
10347 Date: Wed Mar 21 09:06:10 2012 +0000
10349 sna: Fallback to inplace upload if forced to tile the indirect replacement
10351 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
10352 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47629
10353 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10355 commit 1db1eced5462723811e535a53cb41b108fc6c5a3
10356 Author: Chris Wilson <chris@chris-wilson.co.uk>
10357 Date: Tue Mar 20 22:57:50 2012 +0000
10359 sna: Tidy an assertion when handling tiled copies
10361 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10363 commit 38b0cc24b45afd4ed57187fb066896bf1e467f45
10364 Author: Chris Wilson <chris@chris-wilson.co.uk>
10365 Date: Wed Mar 21 01:01:47 2012 +0000
10367 sna: Remove the short-circuiting of move-to-cpu for read if no damage
10369 The danger now is that we may have either discarded the shadow pixels or
10370 have replaced them with a GTT mapping, either way undesirable and so we
10371 should reconstruct the shadow mapping.
10373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10375 commit f49a6b1fd7e23252b2a306f8bdb9baed5c1e8d4b
10376 Author: Chris Wilson <chris@chris-wilson.co.uk>
10377 Date: Tue Mar 20 21:36:26 2012 +0000
10379 sna: Decouple the private data after UnrealizeFont
10381 As the font is kept around and reused after UnrealizeFont, we need to
10382 nullify the pointer to our private data in order to prevent the later
10385 Reported-by: Peter Jordan
10386 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10388 commit 1bdd6461de18a6812d5af0283f4222cbb80ca2f8
10389 Author: Chris Wilson <chris@chris-wilson.co.uk>
10390 Date: Tue Mar 20 11:08:44 2012 +0000
10392 sna/traps: Remove bogus assertion
10394 As we only enter that path for singular unbounded boxes, we are
10395 guaranteed to fill the entire trapezoid extents and so do not need the
10396 unbounded fixup the assertion was fretting about.
10398 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10400 commit 1c2932e9cb283942567c3dd2695d03b8045da27f
10401 Author: Chris Wilson <chris@chris-wilson.co.uk>
10402 Date: Mon Mar 19 15:51:43 2012 +0000
10404 uxa: Defer the call to EnterVT till after outputs are initialised
10406 We need to do this apparently or else we never perform the VT switch.
10407 However, we can not do it too early, especially not before we have
10408 finished intialising the outputs.
10410 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47395
10411 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10413 commit 972e1108a5606399a679f97af0815ec9730ab5be
10414 Author: Chris Wilson <chris@chris-wilson.co.uk>
10415 Date: Mon Mar 19 14:38:28 2012 +0000
10417 sna: Declare videoRam correctly on gen2 devices
10419 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10421 commit c0bab1df509d1f7fe36173715846c13b931e9be1
10422 Author: Chris Wilson <chris@chris-wilson.co.uk>
10423 Date: Mon Mar 19 08:25:22 2012 +0000
10425 sna/dri: Make the drawable as damaged for the off-screen immediate exchange
10427 In some cases off-screen is stil visible, for example under a rotation.
10428 As such xrandr -o left; glxgears -fullscreen was broken.
10430 Reported-by: Phillip Haddad <phillip.haddad@gmail.com>
10431 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10433 commit 17c19ea8e21e1b20eee446045573dfd94ce6f537
10434 Author: Chris Wilson <chris@chris-wilson.co.uk>
10435 Date: Sun Mar 18 13:23:26 2012 +0000
10437 sna/traps: Remove separate edge->vertical flag
10439 Mark vertical edges with dy==0 to reduce structure size and reduce
10440 memory load during edge walking.
10442 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10444 commit 82023397cd05c05cfcb8939829a95e287d438875
10445 Author: Chris Wilson <chris@chris-wilson.co.uk>
10446 Date: Sun Mar 18 11:10:14 2012 +0000
10448 sna/gen3: Improve clear-to-solid reduction
10450 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10452 commit 7ce5f4a11389b2fe9f27a6f09c4848ac71424d5d
10453 Author: Chris Wilson <chris@chris-wilson.co.uk>
10454 Date: Sun Mar 18 10:55:17 2012 +0000
10456 sna/gen3: Do not force tiling for large pixmaps
10458 As the extraction routine is now smarter and can construction
10459 subsurfaces without copying we do not need to force tiling.
10461 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10463 commit 342dda3fe361c8be2f3af5af1516cdc6a5fdcaa9
10464 Author: Chris Wilson <chris@chris-wilson.co.uk>
10465 Date: Sun Mar 18 10:55:06 2012 +0000
10467 sna/gen3: Prevent copy-fallback if we cannot blit
10469 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10471 commit 28c089781fdf74a9bac2e138e65383748558fcc1
10472 Author: Chris Wilson <chris@chris-wilson.co.uk>
10473 Date: Sun Mar 18 10:55:27 2012 +0000
10475 sna: Fixup the cpu shadow mappings before uploading the box
10477 On the off-chance we arrive here with a pointer to the GTT mapping.
10479 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10481 commit 79258fe560d08dd3dd22b4782360e1597e54948c
10482 Author: Chris Wilson <chris@chris-wilson.co.uk>
10483 Date: Sun Mar 18 09:45:27 2012 +0000
10485 sna/traps: Apply some more operator and unbounded reductions
10487 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10489 commit fe8866d6112c3e187d6682e9e4610325668427a0
10490 Author: Chris Wilson <chris@chris-wilson.co.uk>
10491 Date: Sat Mar 17 21:49:56 2012 +0000
10493 sna/gen[345]: Convert CPU mappings to GTT for vertices on submit
10495 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10497 commit 97cd0c7da51024400e8900e46f51620a5f7ad402
10498 Author: Chris Wilson <chris@chris-wilson.co.uk>
10499 Date: Sat Mar 17 23:57:46 2012 +0000
10501 sna/traps: Upon reducing an ADD to a SRC, we need to apply the pending clear
10503 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47444
10504 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10506 commit e31d9dacafe060dc86de801114b475fdd0142eb6
10507 Author: Chris Wilson <chris@chris-wilson.co.uk>
10508 Date: Sat Mar 17 09:21:00 2012 +0000
10510 sna/traps: Align indices for unrolled memset in row_inplace()
10512 The compiler presumes that the uint64_t write is naturally aligned and
10513 so may emit code that crashes with an unaligned moved. To workaround
10514 this, make sure the write is so aligned.
10516 References: https://bugs.freedesktop.org/show_bug.cgi?id=47418
10517 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10519 commit 2b4e11923d9f683f43acf8053bcec1701df25c1f
10520 Author: Chris Wilson <chris@chris-wilson.co.uk>
10521 Date: Sat Mar 17 00:05:47 2012 +0000
10523 sna/traps: Tune inplace_end_subrows()
10525 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10527 commit d887f209f1210a294afc2c3883a97f7ac9f42bad
10528 Author: Chris Wilson <chris@chris-wilson.co.uk>
10529 Date: Fri Mar 16 17:19:11 2012 +0000
10531 sna/dri: Mesa expects the 8-bit stencil buffer to have 2-bytes per pixel
10533 The seperate stencil buffer is full of lies, why worry about one more?
10535 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10537 commit 63c0d10faee3c7cca050505c2e81c416119e57e9
10538 Author: Chris Wilson <chris@chris-wilson.co.uk>
10539 Date: Fri Mar 16 14:51:53 2012 +0000
10541 sna/dri: Improve handling of non-front attachments for CopyRegion
10543 Confusion reigns between using the backing pixmap for the drawable for
10544 the front buffer, and a fake pixmap for the auxiliary buffers.
10546 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10548 commit 324a1dffdc7dd896224fab265bd6a9cf99d7587f
10549 Author: Chris Wilson <chris@chris-wilson.co.uk>
10550 Date: Fri Mar 16 13:07:00 2012 +0000
10552 sna/gen3: Micro-optimise solid span emission
10554 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10556 commit 47c47b85f6b7ef5cbc2192da9ef0cee91f6744ca
10557 Author: Chris Wilson <chris@chris-wilson.co.uk>
10558 Date: Fri Mar 16 12:59:13 2012 +0000
10560 sna/traps: Micro-optimise span emission
10562 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10564 commit 3a26437f195bf3e5d5913f5552a0b838cafb9b2c
10565 Author: Chris Wilson <chris@chris-wilson.co.uk>
10566 Date: Fri Mar 16 12:37:25 2012 +0000
10568 sna/traps: Tune nonzero_row
10570 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10572 commit cb04cf9f4395c258987faead80de5c3a2c93082e
10573 Author: Chris Wilson <chris@chris-wilson.co.uk>
10574 Date: Fri Mar 16 09:28:24 2012 +0000
10576 sna/traps: Make the inline u8 arithmetic more robust
10578 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10580 commit d55bbdf42c75eb0fdb634e963ce81953583964b9
10581 Author: Chris Wilson <chris@chris-wilson.co.uk>
10582 Date: Fri Mar 16 09:18:29 2012 +0000
10584 sna/traps: Remove the old paths for mono inplace traps
10586 Dead code elimination.
10588 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10590 commit fe4ad66aadcfa0c907ba6460a88c49e65225602c
10591 Author: Chris Wilson <chris@chris-wilson.co.uk>
10592 Date: Thu Mar 15 21:33:00 2012 +0000
10594 sna: Correct the order of screen private teardown
10596 So that we do not keep a stale references of the last cached pixmap
10597 across server regeneration (or shutdown).
10599 Reported-by: Thierry Reding <thierry.reding@avionic-design.de>
10600 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47357
10601 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10603 commit 0a8218a535babb5969a58c3a7da0215912f6fef8
10604 Author: Chris Wilson <chris@chris-wilson.co.uk>
10605 Date: Thu Mar 15 13:20:35 2012 +0000
10607 sna: Prefer the CPU bo for uploads if last access was not through the shadow
10609 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10611 commit 77c9c03b102e6fac2279f0085034557bd68c5112
10612 Author: Chris Wilson <chris@chris-wilson.co.uk>
10613 Date: Thu Mar 15 13:18:43 2012 +0000
10615 sna: Fixup the shadow pixmap pointer for move-to-gpu
10617 If we choose not to use the CPU bo for the upload (because we fear the
10618 subsequent synchronisation cost), we need to fixup the shadow pointer
10619 before dereferencing it.
10621 On the move-to-cpu side, the fixup is already performed as we will need
10622 to access the shadow pixels for the subsequent drawing operation.
10624 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10626 commit 0e2fbb6a1c1ad6b0971d07a91a790b0b3236dad3
10627 Merge: 64ccc66 84d7a82
10628 Author: Chris Wilson <chris@chris-wilson.co.uk>
10629 Date: Thu Mar 15 12:43:47 2012 +0000
10631 uxa: Merge 'remove complicated second level caching'
10633 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47345
10634 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10636 commit 84d7a82122530a62ab9a2f4bba3044c23a0db9b0
10637 Author: Chris Wilson <chris@chris-wilson.co.uk>
10638 Date: Sat Mar 10 10:03:23 2012 +0000
10640 uxa: Simplify allocation of backing pixmap
10642 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10644 commit 1b1016624ad16cf5063e88a1d422e28865a9e87f
10645 Author: Chris Wilson <chris@chris-wilson.co.uk>
10646 Date: Sat Mar 10 09:31:24 2012 +0000
10648 uxa/i915: Remove broken CA pass, fallback to magic 2-pass composite helper
10650 The backend failed to handle all the corner cases, so remove the
10653 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10655 commit 895a56a553e1386323a593063e5b7fbb7a410abe
10656 Author: Chris Wilson <chris@chris-wilson.co.uk>
10657 Date: Sat Mar 10 09:28:27 2012 +0000
10659 uxa: Remove unused render_mask_solid members
10661 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10663 commit 9c6f79440e8af9f264e3f6270384064dc1f44721
10664 Author: Chris Wilson <chris@chris-wilson.co.uk>
10665 Date: Sat Mar 10 09:27:36 2012 +0000
10667 uxa: Remove unused tracking of the current render target
10669 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10671 commit 219467ac8bfab98bca82108b22eae8af3fc0bf36
10672 Author: Chris Wilson <chris@chris-wilson.co.uk>
10673 Date: Sat Mar 10 09:26:10 2012 +0000
10675 uxa: Simplify flush tracking
10677 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10679 commit bd8fafe0c48df7f138459f590a0e9e8d0c3267b7
10680 Author: Chris Wilson <chris@chris-wilson.co.uk>
10681 Date: Sat Mar 10 09:10:16 2012 +0000
10683 uxa: Kill the complicated in-flight tracking
10685 Reference leak hunting.
10687 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10689 commit 64ccc6698def517fc37e9ba3c41715626df0e3ca
10690 Author: Thierry Reding <thierry.reding@avionic-design.de>
10691 Date: Thu Mar 15 13:10:20 2012 +0100
10693 configure: Keep passed-in CFLAGS for DRI tests
10695 When the user passes extra CFLAGS and CPPFLAGS to the configure script,
10696 they should be kept when performing subsequent checks with additional
10697 flags. This is required to properly build in cross-compilation setups
10698 where the user may pass in flags like --sysroot in order to pick up the
10699 cross-built dependencies.
10701 Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
10703 commit c5c61ef378cad3c8259631138ea1e03eb02036af
10704 Author: Chris Wilson <chris@chris-wilson.co.uk>
10705 Date: Thu Mar 15 12:38:22 2012 +0000
10707 sna/traps: dst IN WHITE does not reduce to SRC!
10709 I was getting too carried with my reductions. However, IN over a
10710 clear surface is a no-op, though unlikely!
10712 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10714 commit fba49e1bb8e5b6b0e3ceace2dbddb5796ece954e
10715 Author: Chris Wilson <chris@chris-wilson.co.uk>
10716 Date: Thu Mar 15 10:31:51 2012 +0000
10718 sna/traps: Fix off-by-one for filling vertical segments in tor_inplace
10720 If the last solid portion was exactly 4-pixels wide, we would miss
10721 filling in the mask.
10723 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10725 commit 27f8f121b3d7433c6c3470b048428ca088acd2c9
10726 Author: Chris Wilson <chris@chris-wilson.co.uk>
10727 Date: Wed Mar 14 23:43:20 2012 +0000
10729 sna/damage: Handle a reduced damage rather than assert
10731 As we may reduce a damage to empty along the migration paths and not
10732 detect that reduced damage till later, handle those scenarios rather
10735 References: https://bugs.freedesktop.org/show_bug.cgi?42426
10736 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10738 commit abd104600252f0ebd1eb77eb5419693e1c17deda
10739 Author: Chris Wilson <chris@chris-wilson.co.uk>
10740 Date: Wed Mar 14 23:08:31 2012 +0000
10742 sna: Treat unmapped but CPU-mappable bo as available for mapping
10744 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10746 commit 2c492c0064e23457bffdf4f9ab1a7c9f1be93a87
10747 Author: Chris Wilson <chris@chris-wilson.co.uk>
10748 Date: Wed Mar 14 23:05:58 2012 +0000
10750 sna: Disable tiling for single row pixmaps (unless required for hw limits)
10752 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10754 commit 6890592cd2b2d6f0d06c530f5e770fdc98577d4f
10755 Author: Chris Wilson <chris@chris-wilson.co.uk>
10756 Date: Wed Mar 14 21:30:13 2012 +0000
10758 sna/traps: Explicitly create an unattach pixmap for fallback
10760 References: https://bugs.freedesktop.org/show_bug.cgi?id=42426
10761 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10763 commit 23b5640a0f8d8a944275e574f4c72f2380f7ed7a
10764 Author: Chris Wilson <chris@chris-wilson.co.uk>
10765 Date: Wed Mar 14 20:48:56 2012 +0000
10767 sna: Avoid using kgem_bo_reference() internally
10769 So that we can keep the assertion to track the refcnt elsewhere.
10771 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10773 commit 03731b4e37b0cbd3a1cb8c8ec08fddf0f670a918
10774 Author: Chris Wilson <chris@chris-wilson.co.uk>
10775 Date: Wed Mar 14 20:46:59 2012 +0000
10777 sna: Ellide no-op image glyphs
10779 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10781 commit d0e05b4294b2f150a41dd95d52c2e6ee8479283d
10782 Author: Chris Wilson <chris@chris-wilson.co.uk>
10783 Date: Wed Mar 14 20:19:30 2012 +0000
10785 sna: Don't mark cached upload buffers for inactivity expiration
10787 As these do not follow the normal rules of damage tracking, we have to
10788 be careful not to force migration.
10790 References: https://bugs.freedesktop.org/show_bug.cgi?id=42426
10791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10793 commit 96ead26c4dddab637875c74a9afca6fd6662352b
10794 Author: Chris Wilson <chris@chris-wilson.co.uk>
10795 Date: Wed Mar 14 19:56:42 2012 +0000
10797 sna: Add a couple of asserts for inactive_partial reference counting
10799 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10801 commit 6a4ac18e37521d54dc36b215cbd27c56d20b2922
10802 Author: Chris Wilson <chris@chris-wilson.co.uk>
10803 Date: Wed Mar 14 11:51:39 2012 +0000
10805 sna/traps: Use a more direct fallback path for the CPU
10807 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10809 commit 8580677ee6725144acf63f5cc2f8097376352d6e
10810 Author: Chris Wilson <chris@chris-wilson.co.uk>
10811 Date: Wed Mar 14 11:46:18 2012 +0000
10813 sna/traps: Rasterise using pixman inplace where appropriate
10815 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10817 commit 8a303f195b53a78dcbb2b6c84e7e691bedc86d25
10818 Author: Chris Wilson <chris@chris-wilson.co.uk>
10819 Date: Wed Mar 14 11:26:03 2012 +0000
10821 sna: Remove existing damage before overwriting with a composite op
10823 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10825 commit 63eeb98463a77adadc26d9de1e640b75eefebf1a
10826 Author: Chris Wilson <chris@chris-wilson.co.uk>
10827 Date: Wed Mar 14 10:30:47 2012 +0000
10829 sna/gen3: Look harder to see if we can indeed use the BLT for composite
10831 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10833 commit d23ee0380b61e0dfd3ed56b8b4a15fd0b7956491
10834 Author: Chris Wilson <chris@chris-wilson.co.uk>
10835 Date: Tue Mar 13 22:00:25 2012 +0000
10837 sna: Reuse the cached upload as a source GPU bo
10839 References: https://bugs.freedesktop.org/show_bug.cgi?id=42426
10840 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10842 commit 0acec1685328b1dd9dffa804a27a122fc6e4225a
10843 Author: Chris Wilson <chris@chris-wilson.co.uk>
10844 Date: Tue Mar 13 21:29:57 2012 +0000
10846 sna: Defer the release of the upload buffer cache till retirement
10848 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10850 commit 06b28d541bdf2607edc2eb476919b28e747885d8
10851 Author: Chris Wilson <chris@chris-wilson.co.uk>
10852 Date: Tue Mar 13 20:01:53 2012 +0000
10854 sna: Destroy the cached upload buffer if before copying into it
10856 As we discard the buffer after the next batch, we will lose the contents
10857 of the pixmap. Instead discard the cache, and treat it as a normal
10860 References: https://bugs.freedesktop.org/show_bug.cgi?id=42426
10861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10863 commit 015db20287ec90f81ddb7eb2e7ce23b90aacfb2b
10864 Author: Chris Wilson <chris@chris-wilson.co.uk>
10865 Date: Tue Mar 13 16:29:41 2012 +0000
10867 sna: Refactor source upload-to-gpu
10869 This was originally split between into functions for flexibility that
10870 nothing used, so merge the duplicated code into a common function.
10872 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10874 commit 6dfe28f978b59ab2fc44cdd8ddf3743d6507588b
10875 Author: Chris Wilson <chris@chris-wilson.co.uk>
10876 Date: Tue Mar 13 15:47:27 2012 +0000
10878 sna/gen6: Remove the double application of the render offset
10880 Cut'n'paste error from an older generation.
10882 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10884 commit c751242e245eb168b92616bdcea3421be4e420bc
10885 Author: Chris Wilson <chris@chris-wilson.co.uk>
10886 Date: Tue Mar 13 12:04:04 2012 +0000
10888 sna: Only use the cpu bo for xfer between CPU and GPU if either is busy
10890 The synchronisation costs overwhelm any benefit from offloading the
10891 copy, unless we are currently streaming the updates anyway.
10893 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10895 commit c9668a772f15ab36d643a9b0616b309856146c89
10896 Author: Chris Wilson <chris@chris-wilson.co.uk>
10897 Date: Mon Mar 12 23:28:51 2012 +0000
10899 sna: Reduce OVER with a clear pixmap to a BLT
10901 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10903 commit 2ea58256f5ce6e1c0f48d366ff7a1fbf300600de
10904 Author: Chris Wilson <chris@chris-wilson.co.uk>
10905 Date: Mon Mar 12 23:05:17 2012 +0000
10907 sna: Reuse the same upload buffer for the duration of the batch
10909 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10911 commit 8b9abe2be1f54bd8e8593ed155cc4725ac97627a
10912 Author: Chris Wilson <chris@chris-wilson.co.uk>
10913 Date: Mon Mar 12 21:09:19 2012 +0000
10915 sna: Prefer to render very thin trapezoids inplace
10917 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10919 commit 1e2d6ee31a21267ba27e4bebb883aaab08a12f30
10920 Author: Chris Wilson <chris@chris-wilson.co.uk>
10921 Date: Mon Mar 12 20:34:42 2012 +0000
10923 sna: Always reset the source counter after rendering to with the CPU
10925 The goal is to avoid moving to the GPU too early for a frequently
10926 modified CPU buffer.
10928 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10930 commit 7bde1f55a1075eefab0f083833508265fc0e91e7
10931 Author: Chris Wilson <chris@chris-wilson.co.uk>
10932 Date: Mon Mar 12 20:25:50 2012 +0000
10934 sna: After move-to-gpu signals yes, force the GPU bo creation
10936 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10938 commit 2ee7de1f1711d1058e5812c6aa0d94ecf9af7d8d
10939 Author: Chris Wilson <chris@chris-wilson.co.uk>
10940 Date: Mon Mar 12 20:09:05 2012 +0000
10942 sna/trapezoids: Reduce mono ADD/OVER against a clear background to a SRC
10944 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10946 commit 99c239e380b9f5134afc75bf55bf3f69e5113e38
10947 Author: Chris Wilson <chris@chris-wilson.co.uk>
10948 Date: Mon Mar 12 19:49:30 2012 +0000
10950 sna: Treat backing pixmaps no differently from their forward facing cousins
10952 Another fix for the large buffers overhaul.
10954 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10956 commit 28792be1f31190171644d911d7b5573186fe2df2
10957 Author: Chris Wilson <chris@chris-wilson.co.uk>
10958 Date: Mon Mar 12 19:45:35 2012 +0000
10960 sna/display: Only flush pending output when installing a new scanout
10962 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10964 commit 278c329b9e2aa9a94bd0d0ee73939814ac514e3d
10965 Author: Chris Wilson <chris@chris-wilson.co.uk>
10966 Date: Mon Mar 12 11:50:54 2012 +0000
10968 sna/trapezoids: Further improve the clipping criteria for inplace traps
10970 Not only must we defend against the span starting too far to the right,
10971 we must also defend against the span terminating too far to the left.
10973 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10975 commit 7ae45584327a10b05f7aee99bcb71e9d990a3e9b
10976 Author: Chris Wilson <chris@chris-wilson.co.uk>
10977 Date: Mon Mar 12 10:49:46 2012 +0000
10979 sna/trapezoids: Add paranoia to ensure that the span starts within the clip
10981 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
10982 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47226
10983 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10985 commit 8136bc5e113ae06c30def3c91b1615e5fab8af44
10986 Author: Chris Wilson <chris@chris-wilson.co.uk>
10987 Date: Sun Mar 11 19:45:55 2012 +0000
10989 sna: Make the maximum BLT pitch assertions consistent
10991 The maximum permissibly BLT pitch value is 32767, so make the assertions
10994 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47206
10995 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10997 commit 989615493608525fc252e4e94ac7259cba0741f5
10998 Author: Chris Wilson <chris@chris-wilson.co.uk>
10999 Date: Fri Mar 9 22:44:16 2012 +0000
11001 sna: Feed fallback mono trapezoids through the mono rasteriser
11003 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11005 commit 552e4fbd2c25eb5ab0ae77e11f5f8ba2fdb29daa
11006 Author: Chris Wilson <chris@chris-wilson.co.uk>
11007 Date: Fri Mar 9 20:02:44 2012 +0000
11009 sna/traps: Add a fast path for narrow masks
11011 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11013 commit 494edfaaacaae13adfa5e727c66a83cb2294d330
11014 Author: Chris Wilson <chris@chris-wilson.co.uk>
11015 Date: Fri Mar 9 12:19:33 2012 +0000
11017 sna: Handle partial reads with a pending clear
11019 Skip the filling of the whole pixmap if we have a small read and we
11020 know the GPU bo is clear. Also choose to operate inplace on the GPU bo
11021 if we meet the usual criteria.
11023 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11025 commit bd62dc73dcdbab34aa5c83382e46c7315d554a1a
11026 Author: Chris Wilson <chris@chris-wilson.co.uk>
11027 Date: Fri Mar 9 10:37:34 2012 +0000
11029 sna/traps: Apply somes simple but common operator reduction for clipmasks
11031 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11033 commit c25a3f7f46010660f441070ab7b9d5d1bc39ed0d
11034 Author: Chris Wilson <chris@chris-wilson.co.uk>
11035 Date: Fri Mar 9 09:43:46 2012 +0000
11037 sna/dri: Only delivered a delayed flip if the drawable is still on the root
11039 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11041 commit 93846b468e778440549ef0cae171c7fe9678ed9a
11042 Author: Chris Wilson <chris@chris-wilson.co.uk>
11043 Date: Fri Mar 9 09:43:24 2012 +0000
11045 sna/traps: Remove some dead code
11047 This function was never used in this implementation, remove it.
11049 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11051 commit 90c995736555ce14b08b69a42832d9774ba58304
11052 Author: Chris Wilson <chris@chris-wilson.co.uk>
11053 Date: Fri Mar 9 00:37:32 2012 +0000
11055 sna: Emit a INFO when compiled with debugging enabled
11057 It is useful to know and to receive confirmation that you have
11058 successfully compiled and executed the driver with debugging enabled.
11060 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11062 commit 2e194f33db0437ea2f25c22efdad9552aefcab2f
11063 Author: Chris Wilson <chris@chris-wilson.co.uk>
11064 Date: Thu Mar 8 18:07:22 2012 +0000
11066 sna/traps: Fix the initialisation of the error term for vertical mono edges
11068 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11070 commit 635c604b787625f93763001951f8bdf66482c682
11071 Author: Chris Wilson <chris@chris-wilson.co.uk>
11072 Date: Thu Mar 8 18:06:51 2012 +0000
11074 sna/traps: Unroll insertion sort
11076 As the compiler cannot know the loop is bounded by a sentinel, manually
11079 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11081 commit a087430ad99c06e79249d2cdd019cb8bf7f955d3
11082 Author: Chris Wilson <chris@chris-wilson.co.uk>
11083 Date: Thu Mar 8 17:13:39 2012 +0000
11085 sna/gen6: Replace the memset with explict initialisation
11087 The profiles told me to kill it...
11089 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11091 commit 61226cd41faf320f79ee4bd72dc77163079da853
11092 Author: Chris Wilson <chris@chris-wilson.co.uk>
11093 Date: Thu Mar 8 13:41:58 2012 +0000
11095 sna: Fix handling of large glyphs following large and shared buffer work
11097 Part of the large buffer handling was to move the decision making about
11098 whether to create GPU bo for a pixmap to creation time. The single
11099 instance where we change our minds later is involving large glyphs which
11100 we choose not to cache.
11102 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11104 commit aa7a2bd71275e069a2e4383a26355854b0b8044c
11105 Author: Chris Wilson <chris@chris-wilson.co.uk>
11106 Date: Thu Mar 8 12:27:23 2012 +0000
11108 sna: Fix reversed logic for CREATE_NO_RETIRE
11110 If the flag is set, we cannot retire, not the other way around!
11112 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11114 commit f8d520950edb5cec878999a09d3f7e6f8b15bf5e
11115 Author: Chris Wilson <chris@chris-wilson.co.uk>
11116 Date: Thu Mar 8 12:13:36 2012 +0000
11118 sna: Avoid NULL deference in DBG
11120 Only print out the details of the allocated CPU bo, if we actually
11123 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11125 commit 33af42e6284f104ed3c4cdba4bf3b1c29322ce9c
11126 Author: Chris Wilson <chris@chris-wilson.co.uk>
11127 Date: Thu Mar 8 12:10:24 2012 +0000
11129 sna: Force the creation of a backing pixmap for scanout
11131 Ordinarily if the GPU is wedged, we just want to create a shadow buffer.
11132 Except that we must ensure that we do allow a bo to be created for
11133 attaching to the scanout.
11135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11137 commit 6abfa4c5662e5294edc501953dbfa7aabf0b2f57
11138 Author: Chris Wilson <chris@chris-wilson.co.uk>
11139 Date: Thu Mar 8 11:59:02 2012 +0000
11141 sna: Mark the pixmap for writing when creating the screen resources
11143 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11145 commit 5c5ebd6a3ecf3bd83747003bc272a736b7c333c0
11146 Author: Chris Wilson <chris@chris-wilson.co.uk>
11147 Date: Thu Mar 8 11:13:34 2012 +0000
11149 intel: Fix typo s/asert/assert/
11151 The joy of conditional compiles masked this compilation failure when
11154 Reported-by: Reinhard Karcher <reinhard.karcher@gmx.net>
11155 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11157 commit f6474883d325cff443da9ceaa99ec734e6cdc1d6
11158 Author: Chris Wilson <chris@chris-wilson.co.uk>
11159 Date: Thu Mar 8 11:04:05 2012 +0000
11161 sna/gen2+: Use the reduced operator from CompositeRectangles
11163 Do not attempt to further reduce the operator locally in each backend as
11164 the reduction is already performed in the upper layer.
11166 References: https://bugs.freedesktop.org/show_bug.cgi?id=42606
11167 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11169 commit 2ae3cd36ae98fa43716ef482e70364e7563cf1ea
11170 Author: Chris Wilson <chris@chris-wilson.co.uk>
11171 Date: Thu Mar 8 10:51:05 2012 +0000
11173 sna: Tidy marking pixmap->clear for CompositeRectangles
11175 Reduce the two unsightly checks into one.
11177 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11179 commit 8b21659ef2f9cec799b8215a2e5481667bc744bf
11180 Author: Chris Wilson <chris@chris-wilson.co.uk>
11181 Date: Thu Mar 8 10:50:32 2012 +0000
11183 sna: Add some assertions around pixmap creation for render operations
11185 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11187 commit 7d74300647dd815ae5b4399f84ea339045d12563
11188 Author: Chris Wilson <chris@chris-wilson.co.uk>
11189 Date: Thu Mar 8 09:54:16 2012 +0000
11191 sna: Discard unbound partial buffers
11193 Instead of keeping a virgin partial buffer around on its inactive list,
11194 just transfer it to the global bo cache (in actuality destroy it since
11195 it is just a kmalloc with no pages bound).
11197 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11199 commit 9eb8ba4df2a99342dbb3569c043d84aa82506aba
11200 Author: Chris Wilson <chris@chris-wilson.co.uk>
11201 Date: Thu Mar 8 09:42:58 2012 +0000
11203 sna: Preserve the offset alignment when trimming unused rows from partials
11205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11207 commit 8e6166a0e87a00248d0d4925953c923850f5b840
11208 Author: Chris Wilson <chris@chris-wilson.co.uk>
11209 Date: Thu Mar 8 02:09:25 2012 +0000
11211 sna: Becareful not to reduce operators for superluminal colors
11213 wine-1.4 is such an example of a crazy application.
11215 References: https://bugs.freedesktop.org/show_bug.cgi?id=42606
11216 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11218 commit 3f73cc706ff39cd4c10433791f12b5f829f62e6d
11219 Author: Chris Wilson <chris@chris-wilson.co.uk>
11220 Date: Thu Mar 8 08:54:24 2012 +0000
11222 sna/dri: Use a counter for the number of DRI drawables attached to a pixmap
11224 The root pixmap, for instance, may have unique DRI2Drawables for each
11225 inferior window. We only want to clear the flush flag on the last
11226 release, so we need to keep a count of how many DRI drawables remain
11227 attached rather than a solitary flag.
11229 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11231 commit 49a80ce1ff336fb2fa7d214bd3fddbce5a62b77a
11232 Author: Chris Wilson <chris@chris-wilson.co.uk>
11233 Date: Wed Mar 7 15:52:41 2012 +0000
11235 sna/gen2+: Prefer not to fallback if the source is busy
11237 As if we try to perform the operation with outstanding operations on the
11238 source pixmaps, we will stall waiting for them to complete.
11240 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11242 commit 4899740f6f0f5b1a5b2b3490502ccdeb7b092877
11243 Author: Chris Wilson <chris@chris-wilson.co.uk>
11244 Date: Thu Mar 8 00:57:12 2012 +0000
11246 sna: Do not reset partial buffers if they are not attached to the current batch
11248 As we may be holding on to them as an active mapping whilst they are
11249 executing; reseting the used counter back to zero in this case can cause
11252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11254 commit 72a7538d4e7bcf0bd7455d9e67d8751e17739e6c
11255 Author: Chris Wilson <chris@chris-wilson.co.uk>
11256 Date: Wed Mar 7 15:45:21 2012 +0000
11258 sna: Convolution filter fixes
11260 A couple of typos made the convolution filter explode rather than
11263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11265 commit 34fe3cbb316c36c7022735cf9b03d8b655e04434
11266 Author: Chris Wilson <chris@chris-wilson.co.uk>
11267 Date: Wed Mar 7 17:49:01 2012 +0000
11269 sna: Avoid recursive calls to kgem_retire_partials()
11271 Whilst iterating the partial list and uploading the buffers, we need to
11272 avoid trigger a recursive call into retire should we attempt to shrink a
11273 buffer. Such a recursive call will modify the list beneath us so that we
11274 chase a stale pointer and wreak havoc with memory corruption.
11276 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
11277 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47061
11278 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11280 commit 46c79e4d59ec4f90a1fa97b24a3e7058fdbfa6ba
11281 Author: Chris Wilson <chris@chris-wilson.co.uk>
11282 Date: Wed Mar 7 13:17:21 2012 +0000
11284 sna: Restore checking for all-clipped-out for CompositeRectangles
11286 In the refactoring to avoid repeatedly applying the singular
11287 pCompositeClip, the check for the all-clipped state was lost.
11289 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11291 commit 9944f1834a9c53091e4415f928f917e9247f17c5
11292 Author: Chris Wilson <chris@chris-wilson.co.uk>
11293 Date: Wed Mar 7 10:47:42 2012 +0000
11295 sna/gen2: Fix transformation of linear gradients
11297 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11299 commit df25495eaab5bcd5baf86047a2dd0149eea00d1e
11300 Author: Chris Wilson <chris@chris-wilson.co.uk>
11301 Date: Wed Mar 7 10:43:24 2012 +0000
11303 sna/gen4: Hook in the poor-man's linear gradient
11305 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11307 commit 69d6a0387b8eec8396631714a0564b7e28d187ab
11308 Author: Chris Wilson <chris@chris-wilson.co.uk>
11309 Date: Wed Mar 7 10:43:24 2012 +0000
11311 sna/gen5: Hook in the poor-man's linear gradient
11313 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11315 commit b0d3c4f661bace84b78b194eaeee7bfa009a32cf
11316 Author: Chris Wilson <chris@chris-wilson.co.uk>
11317 Date: Wed Mar 7 10:43:24 2012 +0000
11319 sna/gen7: Hook in the poor-man's linear gradient
11321 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11323 commit dcc364a7b190ce9cec2ff35aa22c10529f84e4c3
11324 Author: Chris Wilson <chris@chris-wilson.co.uk>
11325 Date: Wed Mar 7 10:40:50 2012 +0000
11327 sna/gen6: Add poor-man's linear implementation
11329 Still no JIT, in the meantime we can at least cache the gradient ramps.
11331 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11333 commit 232972c0e5bd833c1d0b33432fa8092c601fd0e5
11334 Author: Chris Wilson <chris@chris-wilson.co.uk>
11335 Date: Tue Mar 6 12:17:03 2012 +0000
11337 sna: Remove the 2-step damage flush
11339 The idea was to reduce the number of unnecessary flushes by checking for
11340 outgoing damage (could be refined further by inspecting the reply/event
11341 callback for a XDamageNotifyEvent). However, it does not flush
11342 sufficiently for the compositors' liking. As it doesn't appear to restore
11343 performance to near uncomposited levels anyway, remove the complication.
11345 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11347 commit eb10ade0fc8e21cd01c0d8acba7891fd7f2aa222
11348 Author: Chris Wilson <chris@chris-wilson.co.uk>
11349 Date: Mon Mar 5 22:55:57 2012 +0000
11351 sna: Defer the FlushCallback removal until after the next flush
11353 Try to reduce the amount of Add/Delete ping-pong, in particular around
11354 the recreation of the DRI2 attachment to the scanout after pageflipping.
11356 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11358 commit 60dacdb127599606db13ad111af8ce26c1141da5
11359 Author: Chris Wilson <chris@chris-wilson.co.uk>
11360 Date: Mon Mar 5 22:46:20 2012 +0000
11362 sna: Only install the flush callback for the duration of the foriegn buffer
11364 After we are no longer sharing the bo with foreign clients, we no longer
11365 need to keep flushing before every X_Reply and so we can remove the
11366 callbacks to remove the overhead of having to check every time.
11368 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11370 commit b39d9f9166547effe066acfc3327dd88a019d273
11371 Author: Chris Wilson <chris@chris-wilson.co.uk>
11372 Date: Mon Mar 5 22:29:38 2012 +0000
11374 sna: Check for flush at the start of every WriteToClient
11376 The goal is to simply avoid the flush before going to sleep when we have
11377 no pending events. That is we only want to flush when we know there will
11378 be at least on X_Reply sent to a Client. (Preferably, it would a Damage
11379 reply!) We can safe assume that every WriteToClient marks the beginning
11380 of a new reply added to the Client output queue and thus know that upon
11381 the next flush event we will emitting a Reply and so need to submit our
11384 Second attempt to fix a438e4ac.
11385 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11387 commit f30b0beea4f5657a60eb5b286f41105298fa451a
11388 Author: Chris Wilson <chris@chris-wilson.co.uk>
11389 Date: Sun Mar 4 22:23:39 2012 +0000
11391 sna/trapezoids: Ellide empty cells
11393 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11395 commit b69c9dfae128ed69a397a066b8fbe62012742bf1
11396 Author: Chris Wilson <chris@chris-wilson.co.uk>
11397 Date: Mon Mar 5 21:05:34 2012 +0000
11399 sna/composite: Skip clipping the rectangle region against the singular clip
11401 As we will already have taken it into account when constructing the
11402 region from the rectangles.
11404 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11406 commit f4846168a60feca7437f0039612ba9986c5f8d77
11407 Author: Chris Wilson <chris@chris-wilson.co.uk>
11408 Date: Mon Mar 5 21:04:25 2012 +0000
11410 sna: Flush dirty CPU damage before notifying the compositor
11412 Fixes regression from a438e4ac (sna: Revamp vmap support)
11413 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11415 commit d7600e4e7726bb8bd6c7256ad2baf195d4427c60
11416 Author: Chris Wilson <chris@chris-wilson.co.uk>
11417 Date: Sun Mar 4 19:12:29 2012 +0000
11419 sna: Add some assertions to partial buffer list tracking
11421 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11423 commit 3b5d556a93ac0afebf6c1dd02e0c92f6eb73f633
11424 Author: Chris Wilson <chris@chris-wilson.co.uk>
11425 Date: Sun Mar 4 15:48:33 2012 +0000
11427 sna: Fix assertion for checking inactive shadow buffers
11429 We may have an ordinary malloc with no CPU bo attached so check before
11432 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11434 commit aaed9e9722aa30a3d6dc9a3f07309655de65b6bd
11435 Author: Chris Wilson <chris@chris-wilson.co.uk>
11436 Date: Fri Mar 2 23:31:24 2012 +0000
11438 sna: Encourage promotion of snooped CPU bo to real GPU bo
11440 This fixes the regression in performance of fishietank on gen2. As
11441 the texture atlas is too large to be tiled, one might presume that it
11442 has the same performance characteristics as the snooped linear CPU
11443 buffer. It does not. Therefore if we attempt to reuse a vmap bo, promote
11444 it to a full GPU bo. This hopefully gains the benefit of avoiding the
11445 copy for single shot sources, but still gives us the benefit of avoiding
11448 On the plus side, it does prove that gen2 handles snoopable memory from
11449 both the blitter and the sampler!
11451 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11453 commit 599cd0e8ef3080fc735860bef4e47107c1c05f9a
11454 Author: Chris Wilson <chris@chris-wilson.co.uk>
11455 Date: Fri Mar 2 20:18:32 2012 +0000
11457 sna: Align allocations with partial buffers to 64 bytes.
11459 A magic number required for so many functions of the GPU. In this
11460 particular case it is likely to be that the offset of a texture in the
11461 GTT has to have a minimum alignment of 64 bytes.
11463 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=46415
11464 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11466 commit 4918e309dfadaf14da6997468f08db03ac0a6327
11467 Author: Chris Wilson <chris@chris-wilson.co.uk>
11468 Date: Fri Mar 2 18:18:48 2012 +0000
11470 sna: Silence an assertion failure during shutdown
11472 Clear the scanout flag on the front buffer during teardown to silence
11475 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11477 commit f890fc25c6c2ca358323be5a0d636c3f2ab4b298
11478 Author: Chris Wilson <chris@chris-wilson.co.uk>
11479 Date: Fri Mar 2 18:11:56 2012 +0000
11481 sna: And fix compilation for last commit
11483 I skipped a GCC warning about the implicit function declaration, which
11484 of course results in a runtime silent death. Oops.
11486 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11488 commit 4f853acfeccb92885f154b03125d5716591a87bd
11489 Author: Chris Wilson <chris@chris-wilson.co.uk>
11490 Date: Fri Mar 2 17:36:50 2012 +0000
11492 sna: Prevent backing pixmaps being created later
11494 We used to allow the backing pixmap to be created later in order to
11495 accommodate ShmPixmaps and ShmPutImage. However, they are now correctly
11496 handled upfront if we choose to accelerate those paths, and so all
11497 choice over whether to attach to a pixmap are made during creation and
11500 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11502 commit 866a61a2590f0c5ae6592a13d4e3de3e68f5e373
11503 Author: Chris Wilson <chris@chris-wilson.co.uk>
11504 Date: Fri Mar 2 16:06:59 2012 +0000
11506 sna: Disable vmap on 965gm
11508 The sampler just dies if it encounters a snoopable page, for no apparent
11509 reason. Whilst I encountered the bug on Crestline, disable it for the
11510 rest of gen4 just to be safe.
11512 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11514 commit 1c653786895fc30be0e88455ce5f9caf9adc835d
11515 Author: Chris Wilson <chris@chris-wilson.co.uk>
11516 Date: Fri Mar 2 14:34:23 2012 +0000
11518 sna: Pass usage hint for creating linear buffers
11520 As we wish to immediate map the vertices buffers, it is beneficial to
11521 search the linear cache for an existing mapping to reuse first.
11523 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11525 commit 29ec36ff063472e0744af99aa81ed5ad8e291a36
11526 Author: Chris Wilson <chris@chris-wilson.co.uk>
11527 Date: Fri Mar 2 10:01:07 2012 +0000
11529 sna: Only discard the inplace flag for LLC partial buffers
11531 KGEM_BUFFER_WRITE_INPLACE is WRITE | INPLACE and so the typo prevented
11532 uploading of partial data through the pwrite paths.
11534 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11536 commit f039ccf9587eb07528034c3247a6e700c87a5500
11537 Author: Chris Wilson <chris@chris-wilson.co.uk>
11538 Date: Fri Mar 2 09:47:10 2012 +0000
11540 sna: Be careful not to discard the clear operation for move-region-to-cpu
11542 When moving only a region to the CPU and we detect a pending clear, we
11543 transform the operation into a move whole pixmap. In such situations, we
11544 only have a partial damage area and so need to or in MOVE_READ to
11545 prevent the pending clear of the whole pixmap from being discarded.
11547 References: https://bugs.freedesktop.org/show_bug.cgi?id=46792
11548 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11550 commit 392593e61dac3ac65ee8e32de492c4439413ee85
11551 Author: Chris Wilson <chris@chris-wilson.co.uk>
11552 Date: Thu Mar 1 17:54:51 2012 +0000
11554 sna/gen5: Help the compiler avoid an uncached read
11556 Debug builds are excruitatingly slow as the compiler doesn't store the
11557 temporary in a register but uses an uncached readback instead. Maybe
11560 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11562 commit 9c0c04cac245db046ef17ff24c32e6ab93535f48
11563 Author: Chris Wilson <chris@chris-wilson.co.uk>
11564 Date: Thu Mar 1 14:52:39 2012 +0000
11566 sna: Split storage of inactive partials
11568 As we now attempt to keep retain partial buffers after execution, we can
11569 end up will lots of inactive buffers sitting on the partial buffer list.
11570 In any one batch, we wish to minimise the number of buffers used, so
11571 keep all the inactive buffers on a seperate list and only pull from them
11574 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11576 commit a438e4ac9ba162e870fb22bc54024d35daa2121e
11577 Author: Chris Wilson <chris@chris-wilson.co.uk>
11578 Date: Tue Feb 28 19:15:34 2012 +0000
11580 sna: Revamp vmap support
11582 Dust off the kernel patches and update to reflect the changes made to
11583 support LLC CPU bo, in particular to support the unsynchronized shadow
11586 However, due to the forced synchronisation required for strict client
11587 coherency we prefer not to use the vmap for shared pixmaps unless we are
11588 already busy (i.e. sync afterwards rather than before in the hope that
11589 we can squash a few operations into one). Being able to block the reply
11590 to the client until the request is actually complete and so avoid the
11591 sync remains a dream.
11593 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11595 commit 272f5d9f8407d8084846b429c1722bddb3e861e9
11596 Author: Chris Wilson <chris@chris-wilson.co.uk>
11597 Date: Thu Mar 1 11:10:03 2012 +0000
11599 sna: Discard use of inplace GTT uploads on LLC architectures
11601 As the buffer is cache-coherent, we can read as well as write to any
11602 partial buffer so the distinction is irrelevant.
11604 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11606 commit 43b1a717bae047c7ebbf99e6fa4c03b7a67896b8
11607 Author: Chris Wilson <chris@chris-wilson.co.uk>
11608 Date: Tue Feb 28 13:37:14 2012 +0000
11610 sna: Sort the partial buffers after stealing a write buffer
11612 It will be decoupled and not used again, but this keeps the sanity
11615 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11617 commit 8198e5872c3771e2aefabe1e3e93afa94d2ea0ec
11618 Author: Chris Wilson <chris@chris-wilson.co.uk>
11619 Date: Tue Feb 28 10:42:19 2012 +0000
11621 sna/gen3: Tweak glyph rendering fast paths
11623 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11625 commit 3c4f29820bca336af2c997bafc7ef288b455813c
11626 Author: Chris Wilson <chris@chris-wilson.co.uk>
11627 Date: Mon Feb 27 16:29:38 2012 +0000
11629 uxa/gen3: Remove special casing of solid pictures
11631 Fixes use of alpha-groups and opacity masks in cairo.
11633 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11635 commit 8f3066f0c70654f04f4acadf140e3c5c8dda051d
11636 Author: Chris Wilson <chris@chris-wilson.co.uk>
11637 Date: Mon Feb 27 13:58:58 2012 +0000
11639 sna/gen2; Initialise channel.is-opaque for fills
11641 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11643 commit 3640a0d4cb9e0f115fda9ea36212670f6ccafb22
11644 Author: Chris Wilson <chris@chris-wilson.co.uk>
11645 Date: Mon Feb 27 12:28:22 2012 +0000
11649 This reverts commit 4adb6967a84af8a04769c2d936a41f4a49ed1428.
11651 Oops, this debugging commit was not intended to be pushed along with the
11654 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11656 commit 6fd8d74a6aedb6484e53e704d9f44e0bf83ae9ab
11657 Author: Chris Wilson <chris@chris-wilson.co.uk>
11658 Date: Mon Feb 27 12:06:50 2012 +0000
11660 sna: Upload the ordinary partial buffers!
11662 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11664 commit 4adb6967a84af8a04769c2d936a41f4a49ed1428
11665 Author: Chris Wilson <chris@chris-wilson.co.uk>
11666 Date: Mon Feb 27 11:36:35 2012 +0000
11670 commit 4fbb0baff525115037f7e5a1689541880a6cdafb
11671 Author: Chris Wilson <chris@chris-wilson.co.uk>
11672 Date: Sun Feb 26 22:34:30 2012 +0000
11674 sna: Avoid reusing mmapped partial write buffers for readback
11676 An artefact of retaining the mmapped partial buffers is that it
11677 magnified the effect of stealing those for readback, causing extra
11678 writes on non-llc platforms.
11680 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11682 commit a3c398a6731874ba47e0a46bbd42bf9378e12ab8
11683 Author: Chris Wilson <chris@chris-wilson.co.uk>
11684 Date: Sat Feb 25 10:59:14 2012 +0000
11686 sna: Retain unfinished partial buffers between batches
11688 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11690 commit 8d773b88f45594f45174dc6f1a264d968690ce84
11691 Author: Chris Wilson <chris@chris-wilson.co.uk>
11692 Date: Sat Feb 25 09:32:20 2012 +0000
11694 sna/gen3+: Keep the vertex buffer resident between batches
11696 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11698 commit 8cb773e7c809e1de23cd64d3db862d1f8e7e955a
11699 Author: Chris Wilson <chris@chris-wilson.co.uk>
11700 Date: Sat Feb 25 11:07:16 2012 +0000
11702 sna: Ensure we trigger a retire for search_linear_cache
11704 Bo used for batch buffers are handled differently and not tracked
11705 through the active cache, so we failed to notice when we might be able
11706 to run retire and recover a suitable buffer for reuse. So simply always
11707 run retire when we might need to create a new linear buffer.
11709 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11711 commit b1b4db8942e69d47aabfad3751165dc2252fa448
11712 Author: Chris Wilson <chris@chris-wilson.co.uk>
11713 Date: Sat Feb 25 00:43:30 2012 +0000
11715 sna: Skip a tiled bo when searching the cache for a linear mmap
11717 If we change tiling on a bo, we are effectively discarding the cached
11718 mmap so it is preferable to look for another.
11720 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11722 commit 85e48d2e5eb029d8c17714cb5e7db39ea06a2455
11723 Author: Chris Wilson <chris@chris-wilson.co.uk>
11724 Date: Fri Feb 24 21:40:44 2012 +0000
11726 legacy: Rename XF86DRI to HAVE_DRI1 to avoid conflicts with xorg-server.h
11728 We use the XF86DRI as a user configurable option to control whether to
11729 build DRI support for i810, but it is also used internally within xorg
11730 and there exists a public define in xorg-server.h which overrides our
11731 configure option. So rename our define to HAVE_DRI1 to avoid the
11734 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=46590
11735 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11737 commit 96db90e819a1990c2d139725e522055e92def959
11738 Author: Chris Wilson <chris@chris-wilson.co.uk>
11739 Date: Fri Feb 24 21:36:30 2012 +0000
11741 legacy: Delete unused XF86DRI_DEVEL #define
11743 References: https://bugs.freedesktop.org/show_bug.cgi?id=46590
11744 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11746 commit b870a3e5cd922ce37050ece73e2469802dd21da2
11747 Author: Chris Wilson <chris@chris-wilson.co.uk>
11748 Date: Fri Feb 24 11:21:49 2012 +0000
11750 configure, NEWS: Bump version to 2.18.0 for release
11752 Another quarter, a bit late as I was debugging a few regressions,
11755 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11757 commit 5b5cd6780ef7cae8f49d71d7c8532597291402d8
11758 Author: Chris Wilson <chris@chris-wilson.co.uk>
11759 Date: Fri Feb 24 11:14:26 2012 +0000
11761 uxa: Add a option to disable the bo cache
11763 If you are suffering from regular X crashes and rendering corruption
11764 with a flood of ENOSPC or even EFILE reported in the Xorg.log, try
11765 adding this snippet to your xorg.conf:
11768 Option "BufferCache" "False"
11771 References: https://bugs.freedesktop.org/show_bug.cgi?id=39552
11772 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11774 commit f8ca50818cd4f82e0ad8eeb245e8ee838b083056
11775 Author: Gaetan Nadon <memsize@videotron.ca>
11776 Date: Sat Feb 18 13:49:02 2012 -0500
11778 Revert "Update autotools configuration"
11780 This reverts commit 9184af921bc2f332fcb6c9b47001414378eab8e2.
11782 All X.Org modules must be able to be configured with autoconf 2.60.
11783 In addition, version 2.63 has GPL licensing issues which prevents
11784 some vendor to release software based on it.
11786 The AM_SILENT_RULES are already handled by XORG_DEFAULT_OPTIONS.
11788 All X.Org modules must be able to be configured with libtool 1.5.
11790 AM_MAINTAINER_MODE default value is "enabled" already.
11792 We use the same autogen script for all x.org modules.
11793 There are proposals for changes which should be reviewed and eventually
11794 applied to all modules together.
11796 The lt*.m4 patterns are already included in the root .gitignore file.
11797 This can be proposed as a change to all modules, but it invloves
11798 changing the topvel .gitignore, the m4/.gitignore, the ACLOCAL_AMFLAGS
11799 and the AC_CONFIG_MACRO_DIR together.
11801 For more information on project wide configuration guidelines,
11802 consult http://www.x.org/wiki/ModularDevelopersGuide
11803 and http://www.x.org/wiki/NewModuleGuidelines.
11805 Acked-by: Matthieu Herrb <matthieu.herrb@laas.fr>
11806 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
11808 commit a647aff5124674b638ef52152a5c16c27466eed7
11809 Author: Chris Wilson <chris@chris-wilson.co.uk>
11810 Date: Thu Feb 23 12:04:09 2012 +0000
11812 sna/gen3: Silence the compiler complaining with DBG enabled
11814 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11816 commit cd3a618f583a546bafbe0c171a267774adc4d72b
11817 Author: Chris Wilson <chris@chris-wilson.co.uk>
11818 Date: Thu Feb 23 10:40:57 2012 +0000
11820 sna/gen4 Refactor get_rectangles() to re-emit state after a flush
11822 Condense the work performed by each caller into the callee.
11824 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11826 commit 6a3fa4d1b6849b666a9232b017ce82329494621b
11827 Author: Chris Wilson <chris@chris-wilson.co.uk>
11828 Date: Thu Feb 23 10:40:57 2012 +0000
11830 sna/gen7 Refactor get_rectangles() to re-emit state after a flush
11832 Condense the work performed by each caller into the callee.
11834 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11836 commit fe914eaca4b9178ab3d14053a703772f93295895
11837 Author: Chris Wilson <chris@chris-wilson.co.uk>
11838 Date: Thu Feb 23 10:40:57 2012 +0000
11840 sna/gen5 Refactor get_rectangles() to re-emit state after a flush
11842 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11844 commit 4ecf882c838c1f044b4ed3add486b6579ea2c431
11845 Author: Chris Wilson <chris@chris-wilson.co.uk>
11846 Date: Thu Feb 23 10:40:57 2012 +0000
11848 sna/gen6: Refactor get_rectangles() to re-emit state after a flush
11850 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11852 commit dfa21713c27eb9135cb5e1745807324e49422fde
11853 Author: Chris Wilson <chris@chris-wilson.co.uk>
11854 Date: Thu Feb 23 10:29:38 2012 +0000
11856 sna/gen3: Refactor get_rectangles() to emit composite state and retry
11858 As gen3 only uses the single state emission block, and uniformly calls
11859 get_rectangles(), we can move that caller protocol into the callee.
11861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11863 commit a48e6e0db970ad07cd7452ba24c362d0c2fcf1bf
11864 Author: Chris Wilson <chris@chris-wilson.co.uk>
11865 Date: Thu Feb 23 10:17:34 2012 +0000
11867 sna/gen3+: Force a batch flush when run out of CA vbo
11869 As we prematurely end the batch if we bail on extending the vbo for CA
11870 glyphs, we need to force the flush.
11872 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11874 commit 57c19b10db2b512c52a593fad98b5ac5db4f1497
11875 Author: Chris Wilson <chris@chris-wilson.co.uk>
11876 Date: Thu Feb 23 00:33:16 2012 +0000
11878 sna: Use a CPU mapping if the bo is already in the CPU domain
11880 The heuristic of using the mapping only before the first use in an
11881 execbuffer was suboptimal and broken by the change in bo initialisation.
11883 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11885 commit 510767e213c2f44563f5c438ad1234113567be90
11886 Author: Chris Wilson <chris@chris-wilson.co.uk>
11887 Date: Wed Feb 22 18:33:09 2012 +0000
11889 sna/gen4: Fix vertex flushing across batch flushing
11891 Due to the w/a for its buggy shaders, gen4 is significantly different
11892 that backporting the simple patch from gen5 was prone to failure. We
11893 need to check that the vertices have not already been flushed prior to
11896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11898 commit f6392048e3e761b35644268ef161045524cfa369
11899 Author: Chris Wilson <chris@chris-wilson.co.uk>
11900 Date: Wed Feb 22 17:00:38 2012 +0000
11902 sna/blt: Avoid clobbering the composite state if we fail to setup the BLT
11904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11906 commit 4e86da64be947fe4b8b5f285d26d200827847a8c
11907 Author: Chris Wilson <chris@chris-wilson.co.uk>
11908 Date: Wed Feb 22 14:36:35 2012 +0000
11910 sna/gen3+: Flush the vertices during vertex-finish
11912 But only when finishing the vbo, which is the tricky part of the recent
11915 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11917 commit 71512b2438e06b2bd256aff3657797152fbef38e
11918 Author: Chris Wilson <chris@chris-wilson.co.uk>
11919 Date: Wed Feb 22 13:44:24 2012 +0000
11921 sna: Clear the flush flag on release of scanout/dri
11923 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=46445
11924 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11926 commit 025d3d46dc28e967b4f42e5748d13c7fb3ad283b
11927 Author: Chris Wilson <chris@chris-wilson.co.uk>
11928 Date: Wed Feb 22 13:17:01 2012 +0000
11930 sna/trapezoids: Only the inplace PictOpIn is unbounded
11932 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11934 commit b02f866d67d60538368619849f2acda4c1706476
11935 Author: Chris Wilson <chris@chris-wilson.co.uk>
11936 Date: Wed Feb 22 09:15:25 2012 +0000
11938 sna: Ensure we restore the shadow pixels before uploading CPU data
11940 Reported-by: Joe Nahmias <joe@nahmias.net>
11941 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=46425
11942 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11944 commit be5df7b5ab487a1282c96dda92f3799e35cd19af
11945 Author: Chris Wilson <chris@chris-wilson.co.uk>
11946 Date: Tue Feb 21 23:27:33 2012 +0000
11948 sna/gen5: Remove CA glyph workaround
11950 The root cause has been found and destroyed, so the w/a is now
11953 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11955 commit 6038cede83e7f360428b4625d288411794f9d052
11956 Author: Chris Wilson <chris@chris-wilson.co.uk>
11957 Date: Tue Feb 21 21:26:29 2012 +0000
11959 sna/gen3+: Re-emit composite state after flushing CA vertices
11961 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
11962 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42891
11963 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11965 commit fe6602cbbc4eed1b88ac731a30b46cc970ea444f
11966 Author: Chris Wilson <chris@chris-wilson.co.uk>
11967 Date: Tue Feb 21 21:26:29 2012 +0000
11969 sna/gen3+: Only flush the vertices after checking for end-of-batch
11971 Or upon actually closing the vertex buffer.
11973 However, the underlying issue remains. That is we are failing to re-emit
11974 the first-pass for CA text after flushing the vertex buffer (and so
11975 emitting the second-pass for the flushed vertices).
11977 Reported-by: lemens Eisserer <linuxhippy@gmail.com>
11978 References: https://bugs.freedesktop.org/show_bug.cgi?id=42891
11979 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11981 commit 168c87a340119e65b1d7ccbbf59da820044ca936
11982 Author: Chris Wilson <chris@chris-wilson.co.uk>
11983 Date: Tue Feb 21 13:16:43 2012 +0000
11985 sna: Clear the scanout flag after releasing the scanout pixmap
11987 In the future, this will be a good place to restore the cache level of
11990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
11992 commit d7415742a5f78958489216f450411603b1eff9a7
11993 Author: Chris Wilson <chris@chris-wilson.co.uk>
11994 Date: Tue Feb 21 16:33:26 2012 +0000
11996 sna/dri: Queue a flush on the back DRI2 when enqueing a flip
11998 As we may wait upon the bo having finished rendering before we can
11999 execute the flip, flushing the render cache as early as possible is
12002 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12004 commit 2715a455f7bfbecd7a6977184dc6180a09d06e1f
12005 Author: Chris Wilson <chris@chris-wilson.co.uk>
12006 Date: Tue Feb 21 14:49:30 2012 +0000
12008 uxa: Prevent laggy applications by throttling after rendering
12010 Before blocking and waiting for further input, we need to make sure that
12011 we have not developed too large a queue of outstanding rendering. As we
12012 rendering to the front-buffer with no natural throttling and allow X
12013 clients to render as fast as they wish, it is entirely possible for a
12014 large queue of outstanding rendering to develop. For such an example,
12015 watch firefox rendering the fishietank demo and notice the delay that
12016 can build up before the tooltips appear.
12018 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12020 commit 26721893cb41cef66db7ef626881d1eba8a5bdea
12021 Author: Chris Wilson <chris@chris-wilson.co.uk>
12022 Date: Tue Feb 21 13:31:16 2012 +0000
12024 sna: Mark the pixmap as active for the force-to-gpu short-circuit
12026 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12028 commit 1541f1afca028453379083e564bc82debf0ba39a
12029 Author: Chris Wilson <chris@chris-wilson.co.uk>
12030 Date: Tue Feb 21 13:17:35 2012 +0000
12032 sna: Skip the CPU synchronization when marking a pixmap as inactive
12034 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12036 commit 74fd55a96bd6c63e9c49a4b7fee7ef91f6b00a2c
12037 Author: Chris Wilson <chris@chris-wilson.co.uk>
12038 Date: Tue Feb 21 13:11:32 2012 +0000
12040 sna/dri: Ensure that we reattach to the DRI2 front buffer after modeswitch
12042 If we change the Screen pixmap due to a change of mode, we lose the
12043 flag that we've attached a DRI2 buffer to it. So the next time we try to
12044 copy from/to it, reassert its DRI2 status.
12046 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12048 commit d051793b9194060f5408503b1fac56958c6e58e4
12049 Author: Chris Wilson <chris@chris-wilson.co.uk>
12050 Date: Tue Feb 21 11:42:31 2012 +0000
12052 sna/dri: Improve error handling of failing to create a DRI2 pixmap
12054 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12056 commit 580ae520cad749fb86a1bddd0fa2bcadfd60abb6
12057 Author: Chris Wilson <chris@chris-wilson.co.uk>
12058 Date: Tue Feb 21 10:55:46 2012 +0000
12060 sna: Short-circuit repeated calls to force-to-gpu
12062 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12064 commit f2aafb98026b5c476b7f84aa2dc4c1f9ba2e573d
12065 Author: Chris Wilson <chris@chris-wilson.co.uk>
12066 Date: Tue Feb 21 10:43:11 2012 +0000
12068 uxa: Silence compiler warning for const arguments
12070 i965_video.c: In function 'gen6_create_cc_state':
12071 i965_video.c:1374:12: warning: passing argument 4 of
12072 'intel_bo_alloc_for_data' discards 'const' qualifier from pointer target
12073 type [enabled by default]
12075 Repeated ad nauseam.
12077 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12079 commit 507f72d6d44963cae5d8d4b9da68165f73c6fd36
12080 Author: Chris Wilson <chris@chris-wilson.co.uk>
12081 Date: Tue Feb 21 10:39:48 2012 +0000
12083 uxa: Remove DPRINTF stubs
12085 It wasn't being used for anything non-trivial and was throwing compiler
12086 warnings, so remove it.
12088 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12090 commit c72a67390ea243bf43e2ee4efe237ab88a4615b7
12091 Author: Chris Wilson <chris@chris-wilson.co.uk>
12092 Date: Tue Feb 21 10:21:56 2012 +0000
12094 sna/dri: Update for AsyncSwap interface changes
12096 We now need to return TRUE/FALSE depending on whether we need to
12097 invalidate the drawable.
12099 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12101 commit 27bc2acf0e6a0e5e071e0d187bdf71577e821af8
12102 Author: Chris Wilson <chris@chris-wilson.co.uk>
12103 Date: Tue Feb 21 09:29:41 2012 +0000
12105 sna: Fix use of RegionInit() for singular regions
12107 For a singular region, we want to use a value for nboxes of 0 not 1,
12108 fortunately if you pass in a box, it ignores the value of nboxes.
12109 RegionInit() is a most peculiar API!
12111 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12113 commit d98b41ce4323b43d2359349a04a4a56559e341a2
12114 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
12115 Date: Fri Feb 17 19:50:52 2012 +0800
12117 uxa/glamor/dri: Should fixup the drawable pixmap.
12119 Two fixes in this commit, first we only need to check the
12120 front left buffer, for other attachment we don't need to
12121 check them. The second is, we should fixup the pixmap's
12122 drawable not the original drawable.
12124 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
12125 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12127 commit 1e0d702c3a77f6db3dfd55b8cafc5fca4d778751
12128 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
12129 Date: Fri Feb 17 19:50:51 2012 +0800
12131 uxa/glamor/dri: Enable the pageflip support on glamor.
12133 To support easy buffer exchange at glamor layer, glamor
12134 added a new API glamor_egl_exchange_buffers() to exchange
12135 two pixmaps' EGL image and fbos and textures without
12136 recreating any of them. But this simple method's requirement
12137 is that there are two pixmaps. A exceptional case is:
12138 If we are using triple buffer when do page flipping, we
12139 will have an extra back_buffer which doesn't have a pixmap
12140 attached to it. Then each time we set that buffer to a
12141 pixmap, we will have to call the create_egl_textured_pixmap
12142 to create the corresponding EGL image and fbo and texture
12143 for it. This is not efficient.
12145 To fix this issue, this commit introduces a new back_pixmap
12146 to intel structure to hold the back buffer and corresponding
12147 glamor resources. Then we will just need to do the light
12148 weight buffer exchanging at both DDX and glamor layer.
12150 As the new back pixmap is similar to the screen pixmap
12151 and need to be handled carefully when close screen. As the
12152 glamor data structure is a per screen data, and will be
12153 released at its close screen method. The glamor's close
12154 screen method must cleanup the screen pixmap and back
12155 pixmap's glamor resources. screen pixmap is easy to get,
12156 but there is no good way to store the back pixmap.
12158 So the glamor add a new API glamor_egl_create_textured_screen_ext
12159 function to pass the back pixmap's pointer to glamor layer.
12161 This commit make us depend on glamor commit: 4e58c4f.
12162 And we increased the required glamor version from 0.3.0 to 0.3.1
12164 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
12165 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12167 commit ce7a57994d662f340b9457a2750e4385e7d669cd
12168 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
12169 Date: Fri Feb 17 19:50:50 2012 +0800
12171 uxa/dri: Refine the pageflip processing.
12173 Add a new element back_name to intel structure to track
12174 the back bo's name then avoid flink every time.
12175 And at function I830DRI2ExchangeBuffers, after finish
12176 the BO exchange between info's front and back pixmap,
12177 it set the new front bo to the screen pixmap. But the
12178 screen pixmap should be the same as front's pixmap,
12179 so this is a duplicate operation and can be removed.
12181 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
12183 commit 1a65e2b8a2ebfb4d736efb7631515babad75faf2
12184 Author: Chris Wilson <chris@chris-wilson.co.uk>
12185 Date: Tue Feb 21 08:09:52 2012 +0000
12187 sna: Split up/down edge walking in order to handle endpoint clipping
12189 In order to prevent walking upwards off the top of the pixmap when
12190 rendering a clipped vertical edge, we need to tweak the boundary
12191 conditions for the vertical edge walker.
12193 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
12194 References: https://bugs.freedesktop.org/show_bug.cgi?id=46261
12195 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12197 commit 7ea44997553ffdf57b346dc9d83742c511c9e5a4
12198 Author: Chris Wilson <chris@chris-wilson.co.uk>
12199 Date: Mon Feb 20 23:50:42 2012 +0000
12201 sna: Restore the shadow pixels when reducing CPU damage to all
12203 Reported-by: Joe Nahmias <joe@nahmias.net>
12204 References: https://bugs.freedesktop.org/show_bug.cgi?id=46346
12205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12207 commit dbe0580e207ad85cb6a659f86c5746a7ecbcd036
12208 Author: Chris Wilson <chris@chris-wilson.co.uk>
12209 Date: Mon Feb 20 16:30:53 2012 +0000
12211 sna: gen4+ suffer no penalty for changing tiling
12213 On gen4, the tiling/fence constraints are fairly lax, only requiring
12214 page alignment of the object and its size, and so we can switch
12215 tiling modes without incurring a GPU stall on active bo.
12217 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12219 commit b68b76cf54a322e80685f1ec93538cd6c5813ea4
12220 Author: Chris Wilson <chris@chris-wilson.co.uk>
12221 Date: Mon Feb 20 13:38:11 2012 +0000
12223 sna: Move sync'ing of CPU bo after allocation to first write
12225 The idea was that we could afford to allocate an active CPU bo for
12226 copying to from using the GPU and later sync just before we need to
12227 write to the shadow pixels. Having the sync inside the allocation
12228 function potentially causes an unwanted stall.
12230 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12232 commit bbb6794a3b97b1fcf72c8712ab0ec591683b128b
12233 Author: Chris Wilson <chris@chris-wilson.co.uk>
12234 Date: Mon Feb 20 12:25:31 2012 +0000
12236 sna: Trim clipped lines to end within bounds
12238 References: https://bugs.freedesktop.org/show_bug.cgi?id=46261
12239 References: https://bugs.freedesktop.org/show_bug.cgi?id=45673
12240 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12242 commit 805bc3310cd0a13eab8e48e7615bdd42638cfa33
12243 Author: Chris Wilson <chris@chris-wilson.co.uk>
12244 Date: Mon Feb 20 12:09:19 2012 +0000
12246 sna: When reversing line-drawing direction, use the clipped endpoint
12248 Make sure we take the clipping into account if we choose to reverse the
12249 draw direction (to maintain left-to-right inside the box emission).
12251 References: https://bugs.freedesktop.org/show_bug.cgi?id=46261
12252 References: https://bugs.freedesktop.org/show_bug.cgi?id=45673
12253 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12255 commit 7f0a4a5f7224003a0cd226137de5a068949a41b4
12256 Author: Chris Wilson <chris@chris-wilson.co.uk>
12257 Date: Mon Feb 20 09:52:37 2012 +0000
12259 sna/dri: Ensure the domain tracking is reset when releasing bo used for swaps
12261 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12263 commit 7d147c7462ba183f5316e7e8dafeca62add6f97c
12264 Author: Chris Wilson <chris@chris-wilson.co.uk>
12265 Date: Sun Feb 19 17:50:56 2012 +0000
12267 sna: Correct tile sizes for Y-tiling on i915g
12269 128-byte Y-tiling wasn't introduced until the 945.
12271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12273 commit 95391b7312147760d8da01fce68b8398aa8e4e2f
12274 Author: Chris Wilson <chris@chris-wilson.co.uk>
12275 Date: Thu Feb 16 22:01:41 2012 +0000
12277 sna/trapezoids: Presume that Precise mono rasterisation adheres to the spec
12279 References: https://bugs.freedesktop.org/show_bug.cgi?id=46156
12280 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12282 commit 52b11f63d7922032caef0f0a5979b080dbddcbfc
12283 Author: Chris Wilson <chris@chris-wilson.co.uk>
12284 Date: Thu Feb 16 11:22:23 2012 +0000
12286 sna: Upconvert fallback trapezoids to a8
12288 Since the hardware only handles a8 without tricky emulation and pixman
12289 insists on using a1 for sharp trapezoids we need to ensure that we
12290 convert the a1 to a8 for our trapezoidal mask.
12292 More worryingly, this path should never be hit...
12294 References: https://bugs.freedesktop.org/show_bug.cgi?id=46156
12295 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12297 commit 8050ced6204f5aca12e6c57f86308b6ad1b98209
12298 Author: Chris Wilson <chris@chris-wilson.co.uk>
12299 Date: Wed Feb 15 17:01:18 2012 +0000
12301 sna/dri: Mark bo as reusable after completion of a flip-event
12303 After the flip chain is completed, any residual buffers are no longer in
12304 use and so available for reuse.
12306 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12308 commit fc046aabde76142fce130773d78d797d7d750ab7
12309 Author: Chris Wilson <chris@chris-wilson.co.uk>
12310 Date: Wed Feb 15 16:08:23 2012 +0000
12312 sna/dri: Don't attempt to change tiling if it is a no-op
12314 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12316 commit 66cc9c69657ac2703f2c7fc3c2c50f06bf5daa99
12317 Author: Chris Wilson <chris@chris-wilson.co.uk>
12318 Date: Wed Feb 15 11:58:42 2012 +0000
12320 Be paranoid about the definition of container_of
12322 Replace any existing definition with a correct version, since there are
12323 broken container_of macros floating around the xorg includes.
12325 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12327 commit c0376b7f7b083ab2e87edc36e56fd8eb99c3cd05
12328 Author: Chris Wilson <chris@chris-wilson.co.uk>
12329 Date: Mon Feb 13 00:48:15 2012 +0000
12331 Add a missing macro for old xorg/list.h
12333 list_last_entry() needs to be defined if we are including the xorg
12334 list.h as opposed to our standalone variant.
12336 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12338 commit 87bed52180cd2abd80ef6b58384f9fd9c9968f7b
12339 Author: Chris Wilson <chris@chris-wilson.co.uk>
12340 Date: Sat Feb 11 20:54:18 2012 +0000
12342 Include a local copy of list.h
12344 In 1.11.903, the list.h was renamed to xorg-list.h with a corresponding
12345 change to all structures. As we carried local fixes to list.h and
12346 extended functionality, just create our own list.h with a bit of
12347 handwaving to protect us for the brief existence of xorg/include/list.h.
12349 Reported-by: Armin K <krejzi@email.com>
12350 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45938
12351 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12353 commit c64ebee5fdccf313cbd3c73850e02e6fa7dd2a65
12354 Author: Chris Wilson <chris@chris-wilson.co.uk>
12355 Date: Sat Feb 11 13:31:12 2012 +0000
12357 sna/gen6: Prefer the render ring for copies
12359 Slower for fills, but on the current stack faster for copies, both large
12360 and small. Hopefully, when we write some good shaders for SNB, we will
12361 not only improve performance for copies but also make fills faster on
12362 the render ring than the blt?
12364 As the BLT copy routine is GPU bound for copywinpix10, and the RENDER
12365 copy routine is CPU bound and faster, I believe that we have reached the
12366 potential of the BLT ring and not yet saturated the GPU using the render
12369 Note that we still do not casually switch rings, so the actual routine
12370 chosen will still be selected by the preceeding operations, so is
12371 unlikely to have any effect in practice during, for example, cairo-traces.
12373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12375 commit 6a9b50177408c919ed5c6c2463f687476af2c698
12376 Author: Chris Wilson <chris@chris-wilson.co.uk>
12377 Date: Sat Feb 11 11:02:53 2012 +0000
12379 sna/gen6: Suppress the CS stall for the first command in the batch
12381 The batch emission serves as a full stall, so we do not need to incur a
12382 second before our first rendering.
12384 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12386 commit cbe8bed83f0b4097958c4541ad7809a05e6c6f43
12387 Author: Chris Wilson <chris@chris-wilson.co.uk>
12388 Date: Sat Feb 11 10:58:05 2012 +0000
12390 sna/gen7: Mention the depth-stall required before changing VS state
12392 Because one day we may actually start using VS! Copied from the addition
12393 of the w/a to Mesa by Kenneth Graunke.
12395 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12397 commit 6193f2f00fa7205f9d736340318c66d116dca53e
12398 Author: Chris Wilson <chris@chris-wilson.co.uk>
12399 Date: Thu Feb 9 14:16:17 2012 +0000
12401 sna: Fix retire after readback
12403 Upon reading, we encounter a serialisation point and so can retire all
12404 requests. However, kgem_bo_retire() wasn't correctly detecting that
12405 barrier and so we continued to using GPU detiling thinking the target
12408 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12410 commit 4d8369f8e60fd4f5a0ef49f3e9866ea5ecb21927
12411 Author: Chris Wilson <chris@chris-wilson.co.uk>
12412 Date: Wed Feb 8 13:15:46 2012 +0000
12414 sna/gen2+: Force upload rather than perform source transformations on the CPU
12416 If both the source and destination is on the CPU, then the thinking was
12417 it would be quicker to operate on those on the CPU rather than copy both
12418 to the GPU and then perform the operation. This turns out to be a false
12419 assumption if transformation is involved -- something to be reconsidered
12420 if pixman should ever be improved.
12422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12424 commit 8634d461bd9e5a3d3f75b0efc11db87b8d3e1245
12425 Author: Chris Wilson <chris@chris-wilson.co.uk>
12426 Date: Wed Feb 8 09:13:27 2012 +0000
12428 sna: Limit max CPU bo size to prevent aperture thrashing on upload
12430 Copying between two objects that consume more than the available GATT
12431 space is a painful experience due to the forced use of an intermediatory
12432 and eviction on every batch. The tiled upload paths are in comparison
12433 remarkably efficient, so favour their use when handling extremely large
12436 This reverses the previous idea in that we now prefer large GPU bo
12437 rather than large CPU bo, as the render pipeline is far more flexible
12438 for handling those than the blitter is for handling the CPU bo (at least
12441 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12443 commit 5b16972d7850b2347efc084311d664e14263cba1
12444 Author: Chris Wilson <chris@chris-wilson.co.uk>
12445 Date: Tue Feb 7 23:45:37 2012 +0000
12447 sna: Check that we successfully retired an active linear buffer
12449 If we go to the trouble of running retire before searching, we may as
12450 well check that we retired something before proceeding to check all the
12453 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12455 commit 207b4d4482a6af4a39472ec20ff04fa0c9322d73
12456 Author: Chris Wilson <chris@chris-wilson.co.uk>
12457 Date: Tue Feb 7 21:56:29 2012 +0000
12459 sna: Relax must-be-blittable rules for gen4+
12461 The render pipeline is actually more flexible than the blitter for
12462 dealing with large surfaces and so the BLT is no longer the limiting
12465 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12467 commit 13c960db9ef876ee99991d97dfc34fef184c0341
12468 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
12469 Date: Thu Feb 2 11:30:57 2012 +0800
12471 uxa/glamor: Use a macro to specify module name.
12473 This depends upon glamor commit b5f8d, just after the 0.3.0 tag.
12475 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
12476 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12478 commit 70092bfbc51ddc5a51c9cae21c6b2852c216a6fc
12479 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
12480 Date: Wed Feb 1 19:47:28 2012 +0800
12482 uxa/glamor: Refine CloseScreen and InitScreen process.
12484 The previous version calls glamor_egl_close_screen and
12485 glamor_egl_free_screen manually which is not align with
12486 standard process. Now glamor change the way to follow
12489 glamor layer and glamor egl layer both have their internal
12490 CloseScreens. The correct sequence is after the I830CloseScreen
12491 is registered, then register glamor_egl_close_screen and
12492 the last one is glamor_close_screen. So we move out the
12493 intel_glamor_init from the intel_uxa_init to I830ScreenInit
12494 and just after the registration of I830CloseScreen.
12496 As the glamor interfaces changed, we need to check the
12497 glamor version when load the glamor egl module to make
12498 sure we are loading the right glamor module. If
12499 failed, it will switch back to UXA path.
12501 This depends upon glamor commit 1bc8bf tagged with version 0.3.0.
12503 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
12504 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12506 commit 798aad6c95a1a95fd587430dc7a6d59497a10ce1
12507 Author: Chris Wilson <chris@chris-wilson.co.uk>
12508 Date: Tue Feb 7 20:16:48 2012 +0000
12510 sna/gen[4-7]: Fix erroneous scale factor for partial large bo render copies
12512 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12514 commit ea6588726107f1ab0ef5a8f69d420b5bff819a76
12515 Author: Chris Wilson <chris@chris-wilson.co.uk>
12516 Date: Tue Feb 7 15:26:50 2012 +0000
12518 sna: Apply offsets correctly for partial src/dst in large copy boxes
12520 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12522 commit 14c91e108464a305fb23b3313b842bfaeb9420fc
12523 Author: Chris Wilson <chris@chris-wilson.co.uk>
12524 Date: Tue Feb 7 15:25:48 2012 +0000
12526 sna/tiling: Request Y-tiles if we know we cannot BLT to either the src or dst
12528 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12530 commit 3131217e3ecbdf6cd40b044bd31e8228ff5c6bff
12531 Author: Chris Wilson <chris@chris-wilson.co.uk>
12532 Date: Tue Feb 7 14:35:59 2012 +0000
12534 sna: Mark up the temporary allocations
12536 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12538 commit ec1ccb6bf6a984cbe3317e636d7ff73887348a46
12539 Author: Chris Wilson <chris@chris-wilson.co.uk>
12540 Date: Tue Feb 7 13:37:52 2012 +0000
12542 sna: Set the damage for render->copy_boxes to NULL before use
12544 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12546 commit 58f634b792c6d109c5f2374a3e24d759e744abb4
12547 Author: Chris Wilson <chris@chris-wilson.co.uk>
12548 Date: Tue Feb 7 13:32:20 2012 +0000
12550 sna: Handle tile alignment for untiled large bo more carefully
12552 We ended up trying to align the upper bound to zero as the integer
12553 divsion of the tile width by pixel was zero.
12555 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12557 commit bf3518ea910a2a3c5b4500dcba7c829576fbfd47
12558 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
12559 Date: Tue Feb 7 09:40:54 2012 +0800
12561 uxa/glamor/dri: Fix a typo bug when fixup glamor pixmap.
12563 Should modify the old pixmap's header not the new one which
12564 was already destroyed.
12566 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
12567 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12569 commit 1467a4ba1a327877026cc76b3eabeb51d1415509
12570 Author: Chris Wilson <chris@chris-wilson.co.uk>
12571 Date: Mon Feb 6 20:41:53 2012 +0000
12573 sna: Use the proper sna_picture_is_solid() test
12575 Rather than the specialised routines that assumed pDrawable was
12576 non-NULL, which was no longer true after f30be6f743.
12578 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12580 commit ef335a65a9d9f7eb96873201b546c25fd03d090c
12581 Author: Chris Wilson <chris@chris-wilson.co.uk>
12582 Date: Mon Feb 6 20:37:54 2012 +0000
12584 sna: Search all active buckets for a temporary allocation
12586 Reduce the need for creating a new object if we only need the allocation
12587 for a single operation.
12589 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12591 commit b7e3aaf773f05ce82405e135c0f99b40b3c2f434
12592 Author: Chris Wilson <chris@chris-wilson.co.uk>
12593 Date: Mon Feb 6 18:08:19 2012 +0000
12595 sna: Use the clipped end-point for recomputing segment length after clipping
12597 References: https://bugs.freedesktop.org/show_bug.cgi?id=45673
12598 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12600 commit f30be6f74392f5687ffe8bc9dd2c6dc024ae06c8
12601 Author: Chris Wilson <chris@chris-wilson.co.uk>
12602 Date: Mon Feb 6 15:59:21 2012 +0000
12604 sna/gen2+: Exclude solids from being classed as requiring an upload
12606 We treat any pixmap that is not attached to either a CPU or GPU bo as
12607 requiring the pixel data to be uploaded to the GPU before we can
12608 composite. Normally this is true, except for the solid cache.
12610 References: https://bugs.freedesktop.org/show_bug.cgi?id=45672
12611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12613 commit f009386de87acccadb1803567e4c494e5e80a2a9
12614 Author: Chris Wilson <chris@chris-wilson.co.uk>
12615 Date: Mon Feb 6 09:50:03 2012 +0000
12617 sna: If we have a CPU bo, do not assert we have shadow pixels
12619 When transferring damage to the GPU, on SNB it is not necessarily true
12620 that we have a shadow pixmap, we may instead have drawn onto an unmapped
12621 CPU bo and now simply need to copy from that bo onto the GPU. Move the
12622 assertion onto the path where it truly matters.
12624 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45672
12625 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12627 commit 22e452ebe01c32a08599411743cf18f9ad0545a7
12628 Author: Chris Wilson <chris@chris-wilson.co.uk>
12629 Date: Mon Feb 6 09:19:56 2012 +0000
12631 sna: Disable use of xvmc for SNB+
12633 Not yet implemented, so don't bother setting it to fail.
12635 References: https://bugs.freedesktop.org/show_bug.cgi?id=44874
12636 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12638 commit a8ed1a02ada1f8d2f910dfefb150d26c840bf9ea
12639 Author: Chris Wilson <chris@chris-wilson.co.uk>
12640 Date: Sat Feb 4 20:13:07 2012 +0000
12642 sna: Discard the redundant clear of the unbounded area if already clear
12644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12646 commit b899a4b69696141ca8b897a7abf52649b09f7b3b
12647 Author: Chris Wilson <chris@chris-wilson.co.uk>
12648 Date: Sat Feb 4 20:07:49 2012 +0000
12650 sna: Always pass the clear colour for PictOpClear
12652 Having made that optimisation for Composite, and then made the
12653 assumption that it is always true in the backends, we failed to clear
12654 the unbounded area outside of a trapezoid since we passed in the
12655 original colour and the operation was optimised as a continuation.
12657 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12659 commit c107b90a44abb45c837ff8924939872be5b490eb
12660 Author: Chris Wilson <chris@chris-wilson.co.uk>
12661 Date: Sat Feb 4 16:33:34 2012 +0000
12663 sna/gen6: Reduce PictOpClear to PictOpSrc (with blending disabled)
12665 The advantage of PictOpSrc is that it writes its results directly to
12666 memory bypassing the blend unit.
12668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12670 commit 4baa2806bc0f51e7576b769ca6750deb3821c4d3
12671 Author: Chris Wilson <chris@chris-wilson.co.uk>
12672 Date: Sat Feb 4 12:06:22 2012 +0000
12674 sna: Check if the damage reduces to all before performing the migration
12676 An assert exposed a situation where we had accumulated an unreduced
12677 damage-all and so we were taking the slow path only to discover later
12678 that it was a damage-all and that we had performed needless checks.
12680 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12682 commit 2653524dffc1fe0dbff7d74bfc9be535d9ececb1
12683 Author: Chris Wilson <chris@chris-wilson.co.uk>
12684 Date: Fri Feb 3 20:06:43 2012 +0000
12686 sna: Reduce the downsample tile size to accommodate alignment
12688 If we need to enlarge the sampled tile due to tiling alignments, the
12689 resulting sample can become larger than we can accommodate through the 3D
12690 pipeline, resulting in FAIL.
12692 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12694 commit 93a0b10f163ee79b6a6a7ea46b0a33b622b1f86e
12695 Author: Chris Wilson <chris@chris-wilson.co.uk>
12696 Date: Fri Feb 3 19:30:24 2012 +0000
12698 sna: Apply redirection for the render copy into large pixmaps
12700 If the pixmap is larger than the pipeline, but the operation extents fit
12701 within the pipeline, we may be able to create a proxy target to
12702 transform the operation into one that fits within the constraints of the
12705 This fixes the infinite recursion hit with partially displayed extremely
12708 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12710 commit 4774c6b8331831e0c9f3b24f5f6e1b6ea399f628
12711 Author: Chris Wilson <chris@chris-wilson.co.uk>
12712 Date: Thu Feb 2 15:23:03 2012 +0000
12714 sna: Add a couple of sanity checks that the CPU drawable is on the CPU
12716 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12718 commit 418cd98db7c4a2886c9e310f3691eb6c77421dd7
12719 Author: Chris Wilson <chris@chris-wilson.co.uk>
12720 Date: Wed Feb 1 19:10:41 2012 +0000
12722 sna/gen6: Ring switching outweighs the benefits for cairo-traces
12724 At the moment, the jury is still out on whether freely switching rings
12725 for fills is a Good Idea. So make it easier to turn it on and off for
12728 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12730 commit 2d0e7c7ecd7371ac7b5fe3f382fc5d04792f7019
12731 Author: Chris Wilson <chris@chris-wilson.co.uk>
12732 Date: Wed Feb 1 14:52:56 2012 +0000
12734 sna: Search again for a just-large-enough mapping for inplace uploads
12736 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12738 commit 55c7088f54655609fbb00106679a566b46ee8dba
12739 Author: Chris Wilson <chris@chris-wilson.co.uk>
12740 Date: Wed Feb 1 01:33:52 2012 +0000
12742 sna: Add debugging code to verify damage extents of fallback paths
12744 After using the CPU, upload the damage and read back the pixels from the
12745 GPU bo and verify that the two are equivalent.
12747 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12749 commit c8fc2cde53ef7aa011ec7c47e7fc5486de0651f5
12750 Author: Chris Wilson <chris@chris-wilson.co.uk>
12751 Date: Wed Feb 1 01:27:43 2012 +0000
12753 sna: Fill extents for ImageGlyphs
12755 The spec says to fill the characters boxes, which is what the hardware
12756 does. The implementation fills the extents instead. rxvt expects the
12757 former, emacs the latter. Overdraw is a nuisance, but less than leaving
12760 Reported-by: walch.martin@web.de
12761 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45438
12762 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12764 commit 13508ab5ea136caca90c846ff1026c0c1acd2ad5
12765 Author: Chris Wilson <chris@chris-wilson.co.uk>
12766 Date: Tue Jan 31 22:38:46 2012 +0000
12768 sna: PolyGlyph supports all of fill/tile/stipple rules
12770 The hw routines only directly supports solid fill so fallback for the
12771 interesting cases. An alternative would be to investigate using the
12772 miPolyGlyph routine to convert the weird fills into spans in order to
12773 fallback. Sounds cheaper to fallback, so wait for an actual use case.
12775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12777 commit df4e1059a4e09998334dde6aa1c8ccfe76e442c0
12778 Author: Chris Wilson <chris@chris-wilson.co.uk>
12779 Date: Tue Jan 31 19:19:13 2012 +0000
12781 sna/gen6: Prefer to do fills using the BLT
12783 Using the BLT is substantially faster than the current shaders for solid
12784 fill. The downside is that it invokes more ring switching.
12786 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12788 commit 8b012de0a1af4ec97c3197af3f1efdcc67bc2118
12789 Author: Chris Wilson <chris@chris-wilson.co.uk>
12790 Date: Tue Jan 31 20:29:58 2012 +0000
12792 sna/gen5: Always prefer to emit solid fills using the BLT
12794 As the BLT is far, far faster than using a shader.
12796 Improves cairo-demos/chart from 6 to 13 fps.
12798 Reported-by: Michael Larabel <Michael@phoronix.com>
12799 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12801 commit 0a748fc49d60dc2bc9494f95c4934592b111831a
12802 Author: Chris Wilson <chris@chris-wilson.co.uk>
12803 Date: Tue Jan 31 10:51:02 2012 +0000
12805 sna: Split the tiling limits between upload and copying
12807 The kernel has a bug that prevents pwriting buffers large than the
12808 aperture. Whilst waiting for the fix, limit the upload where possible to
12809 fit within that constraint.
12811 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12813 commit 9c1f8a768ca1f762c722f63bab2747e4ff1fd773
12814 Author: Chris Wilson <chris@chris-wilson.co.uk>
12815 Date: Tue Jan 31 10:29:02 2012 +0000
12817 sna: Avoid converting requested Y to X tiling for large pitches on gen4+
12819 The only strong requirement is that to utilize large pitches, the object
12820 must be tiled. Having it as X tiling is a pure convenience to facilitate
12821 use of the blitter. A DRI client may want to keep using Y tiling
12824 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12826 commit e872c1011fc7b67683703fd891234f07dd7acd04
12827 Author: Chris Wilson <chris@chris-wilson.co.uk>
12828 Date: Tue Jan 31 10:21:05 2012 +0000
12830 sna/dri: We need to reduce tiling on older gen if we cannot fence
12832 Only apply the architectural limits to enable bo creation for DRI buffers.
12834 Reported-by: Alban Browaeys <prahal@yahoo.com>
12835 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45414
12836 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12838 commit a4caf67d8da37d04f8915d96b10411ba7267937e
12839 Author: Chris Wilson <chris@chris-wilson.co.uk>
12840 Date: Tue Jan 31 00:35:42 2012 +0000
12842 sna: Trim tile sizes to fit into bo cache
12844 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12846 commit 3f7c1646c78d8854c88b214d3699e51839ba9711
12847 Author: Chris Wilson <chris@chris-wilson.co.uk>
12848 Date: Tue Jan 31 00:09:42 2012 +0000
12850 sna: Check that the intermediate IO buffer can also be used for blitting
12852 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12854 commit e504fab6c5354ae9d18ccefb10bd586fa49b924c
12855 Author: Chris Wilson <chris@chris-wilson.co.uk>
12856 Date: Mon Jan 30 23:49:18 2012 +0000
12858 sna: Discard the cleared GPU buffer upon PutImage to the CPU buffer
12860 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12862 commit ed1c1a7468d78e99cb4f9a4a8b8a6b00c3257a75
12863 Author: Chris Wilson <chris@chris-wilson.co.uk>
12864 Date: Mon Jan 30 13:43:59 2012 +0000
12866 sna: Track large objects and limit prefer-gpu hint to small objects
12868 As the GATT is irrespective of actual RAM size, we need to be careful
12869 not to be too generous when allocating GPU bo and their shadows. So
12870 first of all we limit default render targets to those small enough to
12871 fit comfortably in RAM alongside others, and secondly we try to only
12872 keep a single copy of large objects in memory.
12874 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12876 commit d53d93ffa6e133f46c39595294ecf8e2182b5a68
12877 Author: Chris Wilson <chris@chris-wilson.co.uk>
12878 Date: Mon Jan 30 15:11:34 2012 +0000
12880 sna: Update the partial buffer allocation size when reusing an old mapping
12882 Whilst the old mapping is guaranteed to be larger than the requested
12883 allocation size, keep track of the actual size allows for better packing
12884 of future buffers. And the code also performs a sanity check that the
12885 buffer is the size we claim it to be...
12887 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12889 commit 6f99555b6b64a0e1baad1853569f7bf521c327c3
12890 Author: Chris Wilson <chris@chris-wilson.co.uk>
12891 Date: Mon Jan 30 11:48:59 2012 +0000
12893 sna: Allow the creation of render targets larger than the maximum bo cache
12895 Given that we now handle uploads to and from bo that are larger than the
12896 aperture and that usage of such large bo is rare and so unlikely to
12897 benefit from caching, allow them to be created as render targets and
12898 destroy as soon as they become inactive.
12900 In principle, this finally enables GPU acceleration of ocitysmap on gen4+,
12901 but due to the large cost of creating and destroying large bo it is
12902 disabled on systems that require clflushing. It is, however, a
12903 pre-requisite for exploiting the enhanced capabilities of IvyBridge.
12905 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12907 commit c65ec096e79aa6bda7b2b3ef235e3fd9698b4da7
12908 Author: Chris Wilson <chris@chris-wilson.co.uk>
12909 Date: Mon Jan 30 11:41:07 2012 +0000
12911 sna: Decrease tiling step size in case we need to enlarge the box later
12913 We can juggle rendering into large bo on gen4 by redirecting the
12914 rendering through a proxy that is tile aligned, and so the render target
12915 may be slightly larger than the tiling step size. As that is then larger
12916 than the maximum 3D pipeline, the trick fails and we need to resort to a
12917 temporary render target with copies in and out. In this case, check that
12918 the tile is aligned to the most pessimistic tiling width and reduce the
12919 step size to accomodate the enlargement.
12921 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12923 commit 95f3734dd69b82e007095a599cc21f4c63d6ec00
12924 Author: Chris Wilson <chris@chris-wilson.co.uk>
12925 Date: Mon Jan 30 11:40:02 2012 +0000
12927 sna: Allow creation of proxies to proxies
12929 Just update the offset of the new bo by the offset of the existing
12932 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12934 commit 488937edb67a60389380b405f8f8a548f51e64c7
12935 Author: Chris Wilson <chris@chris-wilson.co.uk>
12936 Date: Mon Jan 30 11:38:36 2012 +0000
12938 sna: Base prefer-gpu hint on default tiling choice
12940 As on gen4+, tiling increases the maximum usable pitch we can
12941 accommodate wider pixmaps on the GPU.
12943 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12945 commit ca252e5b51d7b2f5a7b2c2e0d8fdb024b08096db
12946 Author: Chris Wilson <chris@chris-wilson.co.uk>
12947 Date: Sun Jan 29 19:36:55 2012 +0000
12949 sna: Detect batch overflow and fallback rather an risk an ENOSPC
12951 Having noticed that eog was failing to perform a 8k x 8k copy with
12952 compiz running on a 965gm, it was time the checks for batch overflow
12955 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12957 commit 3aee521bf236994628c4d103a2b8f391a4be2aa7
12958 Author: Chris Wilson <chris@chris-wilson.co.uk>
12959 Date: Sun Jan 29 19:21:54 2012 +0000
12961 sna: Add a tiled fallback for large BLT copies
12963 If we are attempting to copy between two large bo, larger than we can
12964 fit into the aperture, break the copy into smaller steps and use an
12967 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12969 commit df148c962108a7f3efead0b80ab4fe77f3f79c8b
12970 Author: Chris Wilson <chris@chris-wilson.co.uk>
12971 Date: Sun Jan 29 18:07:14 2012 +0000
12973 sna: Limit the tile size for uploading into large pixmaps
12975 As we may have a constrained aperture, we need to be careful not to
12976 exceed our resources limits when uploading the pixel data. (For example,
12977 fitting two of the maximum bo into a single batch may fail due to
12978 fragmentation of the GATT.) So be cautious and use more tiles to reduce
12979 the size of each individual batch.
12981 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12983 commit e1e67e8f394480eb4fef1238ccfd49cc36e4b6f2
12984 Author: Chris Wilson <chris@chris-wilson.co.uk>
12985 Date: Sun Jan 29 15:43:42 2012 +0000
12987 sna: Fix the "trivial" fix to improve error handling
12989 The logic was just backwards and we tried to upload a shadowless GPU
12992 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
12994 commit d3fb1e1e89ccf5cefe6add66de4f960ef07cac60
12995 Author: Chris Wilson <chris@chris-wilson.co.uk>
12996 Date: Sun Jan 29 14:20:33 2012 +0000
12998 sna: Handle GPU creation failure when uploading subtexture
13000 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13002 commit 518a99ea34b26aa094f29a4cc1ea5419f63a0e56
13003 Author: Chris Wilson <chris@chris-wilson.co.uk>
13004 Date: Sun Jan 29 14:09:46 2012 +0000
13006 sna: Always create a GPU bo for copying from an existent source GPU bo
13008 Make sure we prevent the readback of an active source GPU bo by always
13009 prefering to do the copy on the GPU if the data is already resisent.
13010 This fixes the second regression from e583af9cc, (sna: Experiment with
13011 creating large objects as CPU bo).
13013 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13015 commit 624d9843abda9ca6bd1b004d70a6fdc082ba9653
13016 Author: Chris Wilson <chris@chris-wilson.co.uk>
13017 Date: Sun Jan 29 13:55:20 2012 +0000
13019 sna: Ignore map status and pick the first inactive bo for reuse
13021 This fixes the performance regression introduced with e583af9cca,
13022 (sna: Experiment with creating large objects as CPU bo), as we ended up
13023 creating fresh bo and incurring setup and thrashing overhead, when we
13024 already had plenty cached.
13026 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13028 commit 5c6255ba2f12f04938fd586ca02562ee3cae05af
13029 Author: Chris Wilson <chris@chris-wilson.co.uk>
13030 Date: Sun Jan 29 11:02:38 2012 +0000
13032 sna: Determine whether to use a partial proxy based on the pitch
13034 On gen4+ devices the maximum render pitch is much larger than is simply
13035 required for the maximum coordinates. This makes it possible to use
13036 proxy textures as a subimage into the oversized texture without having
13037 to blit into a temporary copy for virtually every single bo we use.
13039 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13041 commit 65466f86263b3788b438fe021a12ade371190b01
13042 Author: Chris Wilson <chris@chris-wilson.co.uk>
13043 Date: Sat Jan 28 01:54:47 2012 +0000
13045 sna: Allow ridiculously large bo, up to half the total GATT
13047 Such large bo place extreme stress on the system, for example trying to
13048 mmap a 1GiB into the CPU domain currently fails due to a kernel bug. :(
13049 So if you can avoid the swap thrashing during the upload, the ddx can now
13050 handle 16k x 16k images on gen4+ on the GPU. That is fine until you want
13053 The real complication comes in uploading (and downloading) from such
13054 large textures as they are too large for a single operation with
13055 automatic detiling via either the BLT or the RENDER ring. We could do
13056 manual tiling/switching or, as this patch does, tile the transfer in
13057 chunks small enough to fit into either pipeline.
13059 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13061 commit 03211f4b0b7e32b6d7dc28e60be72db204b8c8d4
13062 Author: Chris Wilson <chris@chris-wilson.co.uk>
13063 Date: Fri Jan 27 23:18:05 2012 +0000
13065 sna: Guard against the upload buffer growing past the maximum bo size
13067 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13069 commit 2afd49a28429cdeb36583cfc31cc9b1742c1fb83
13070 Author: Chris Wilson <chris@chris-wilson.co.uk>
13071 Date: Fri Jan 27 22:18:30 2012 +0000
13073 sna: Limit inplace upload buffers to maximum mappable size
13075 References: https://bugs.freedesktop.org/show_bug.cgi?id=45323
13076 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13078 commit 8f4bae01e39392faa8978090db4cbe28fa00b013
13079 Author: Chris Wilson <chris@chris-wilson.co.uk>
13080 Date: Fri Jan 27 20:28:44 2012 +0000
13082 sna/video: Ensure the video pixmap is on the GPU
13084 The presumption that the pixmap is the scanout and so will always be
13085 pinned is false if there is a shadow or under a compositor. In those
13086 cases, the pixmap may be idle and so the GPU bo reaped. This was
13087 compounded by that the video path did not mark the pixmap as busy. So
13088 whilst watching a video under xfce4 with compositing enabled (has to be
13089 a non-GL compositor) the video would suddenly stall.
13091 Reported-by: Paul Neumann <paul104x@yahoo.de>
13092 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45279
13093 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13095 commit d02bd80b2f9f8ee3840be5d8d6b8d389192c57f5
13096 Author: Chris Wilson <chris@chris-wilson.co.uk>
13097 Date: Fri Jan 27 20:12:49 2012 +0000
13099 sna: Use a proxy rather than a temporary bo for too-tall but thin targets
13101 If the render target is thin enough to fit within the 3D pipeline, but is
13102 too tall, we can fudge the address of the origin and coordinates to fit
13103 within the constaints of the pipeline.
13105 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13107 commit ea433995a37f0a9d1579f74029418f22a63a2bc0
13108 Author: Chris Wilson <chris@chris-wilson.co.uk>
13109 Date: Fri Jan 27 19:34:39 2012 +0000
13111 sna: Experiment with a partial source
13113 If the source is thin enough such that the pitch is within the sampler's
13114 constraints and the sample size is small enough, just fudge the origin
13115 of the bo such that it can be sampled.
13117 This avoids having to create a temporary bo and use the BLT to extract
13118 it and helps, for example, firefox-asteroids which uses an 64x11200
13121 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13123 commit ad910949beb0c42e2d7b864f030b055ca40adacd
13124 Author: Chris Wilson <chris@chris-wilson.co.uk>
13125 Date: Fri Jan 27 18:37:39 2012 +0000
13127 sna: Mark diagonal lines as partial write
13129 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13131 commit b9c83e0b2c42f7657b1854a7cdcbc0407957c382
13132 Author: Chris Wilson <chris@chris-wilson.co.uk>
13133 Date: Fri Jan 27 18:24:01 2012 +0000
13135 sna/video: Add some DBG messages to track the error paths
13137 References: https://bugs.freedesktop.org/show_bug.cgi?id=45279
13138 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13140 commit 45d831c8b11d6f9f38c0cb9b5877541a971f2c67
13141 Author: Chris Wilson <chris@chris-wilson.co.uk>
13142 Date: Fri Jan 27 15:45:17 2012 +0000
13144 sna: Consolidate routines to choice destination bo
13146 Combine the two very similar routines that decided if we should render
13147 into the GPU bo, CPU bo or shadow pixmap into a single function.
13149 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13151 commit 6402e7f1195f60f96dc269a701532c3270d8920f
13152 Author: Chris Wilson <chris@chris-wilson.co.uk>
13153 Date: Fri Jan 27 18:05:51 2012 +0000
13155 sna: Ensure that we have a source bo for tiled fills
13157 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13159 commit 6c5fb84f4de346b06e5a538e683c5a118f2579bc
13160 Author: Chris Wilson <chris@chris-wilson.co.uk>
13161 Date: Fri Jan 27 14:08:57 2012 +0000
13163 sna/glyphs: Check that we attached to the cache pixmaps upon creation
13165 If the hw is wedged, then the pixmap creation routines will return an
13166 ordinary unattached pixmap. The code presumed that it would only return
13167 a pixmap with an attached bo, and so would segfault as it chased the
13168 invalid pointer after a GPU hang and the server was restarted.
13169 Considering that we already checked that the GPU wasn't wedged before we
13170 started, this is just mild paranoia, but on a run-once piece of code.
13172 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13174 commit 86f1ae9164a94323c08e1dc6cb301e5bc1126b10
13175 Author: Chris Wilson <chris@chris-wilson.co.uk>
13176 Date: Fri Jan 27 10:26:11 2012 +0000
13178 sna/video: Add some more DBG breadcrumbs to the textured PutImage
13180 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13182 commit ce1cae7f4715fc8e14327c5b705d9f2cc45a3741
13183 Author: Chris Wilson <chris@chris-wilson.co.uk>
13184 Date: Fri Jan 27 00:37:51 2012 +0000
13186 sna/video: Simplify the gen2/915gm check
13188 And make the later check in put image match.
13190 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13192 commit 541908524f9ee754db3bc45d2e1681d34479c1cc
13193 Author: Chris Wilson <chris@chris-wilson.co.uk>
13194 Date: Thu Jan 26 23:14:14 2012 +0000
13196 sna: Remove extraneous clipping from GetImage
13198 The spec says that they must wholly contained with the valid BorderClip
13199 for a Window or within the Pixmap or else a BadMatch is thrown. Rely on
13200 this behaviour and not perform the clipping ourselves.
13202 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13204 commit 7ff40b572ec5cd860d7c7ff23beca0388f37c31c
13205 Author: Chris Wilson <chris@chris-wilson.co.uk>
13206 Date: Thu Jan 26 22:55:04 2012 +0000
13208 sna: Avoid fbBlt for the easy GetImage cases
13211 60000 trep @ 0.6145 msec ( 1630.0/sec): GetImage 500x500 square
13213 60000 trep @ 0.4949 msec ( 2020.0/sec): GetImage 500x500 square
13215 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13217 commit adb1320bba15a3a3b4fa8e7d0fd0360fa696721d
13218 Author: Chris Wilson <chris@chris-wilson.co.uk>
13219 Date: Thu Jan 26 16:05:48 2012 +0000
13221 sna/gen2+: Include being unattached in the list of source fallbacks
13223 If the source is not attached to a buffer (be it a GPU bo or a CPU bo),
13224 a temporary upload buffer would be required and so it is not worth
13225 forcing the target to the destination in that case (should the target
13226 not be on the GPU already).
13228 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13230 commit b1f9415bf355e9fbd6fc3ce8fa1af59083ca2943
13231 Author: Chris Wilson <chris@chris-wilson.co.uk>
13232 Date: Thu Jan 26 13:59:24 2012 +0000
13234 sna: Allow gen4+ to use larger GPU bo
13236 As the blitter on gen4+ does not require fence registers, it is not
13237 restricted to operating on large objects within the mappable aperture.
13238 As we do not need to operate on such large GPU bo in place, we can relax
13239 the restriction on the maximum bo size for gen4+ to allocate for use
13242 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13244 commit d35b6955dbb5d652d8685d2c1ea82c5e08de55ea
13245 Author: Chris Wilson <chris@chris-wilson.co.uk>
13246 Date: Thu Jan 26 14:34:36 2012 +0000
13248 sna: Prevent mapping through the GTT for large bo
13250 If the bo is larger than a quarter of the aperture, it is unlikely that
13251 we will be able to evict enough contiguous space in the GATT to
13252 accommodate that buffer. So don't attempt to map them and use the
13253 indirect access instead.
13255 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13257 commit 7c81bcd0c425cc0f7ddf2ad8289bb739c8d44289
13258 Author: Chris Wilson <chris@chris-wilson.co.uk>
13259 Date: Thu Jan 26 13:35:15 2012 +0000
13261 sna: Add FORCE_FALLBACK debugging hook for PutImage
13263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13265 commit 35c0ef586bf508c577642d772f18eae0b64cfd44
13266 Author: Chris Wilson <chris@chris-wilson.co.uk>
13267 Date: Thu Jan 26 11:20:03 2012 +0000
13269 sna/gen3: Use cpu bo if already in use
13271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13273 commit b76a6da3fa0148ef32600dd9505e22b90de037df
13274 Author: Chris Wilson <chris@chris-wilson.co.uk>
13275 Date: Thu Jan 26 10:47:01 2012 +0000
13277 sna: Search the buckets above the desired size in the bo cache
13279 It is preferrable to reuse a slightly larger bo, than it is to create a
13280 fresh one and map it into the aperture. So search the bucket above us as
13283 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13285 commit e2b8b1c145932e2254a705905c60f18c200cf2e8
13286 Author: Chris Wilson <chris@chris-wilson.co.uk>
13287 Date: Thu Jan 26 12:42:12 2012 +0000
13289 sna: Apply any previous transformation when downsampling
13291 In order to handle rotations and fractional offsets produced by the act
13292 of downsampling, we need to compute the full affine transformation and
13293 apply it to the vertices rather than attempt to fudge it with an integer
13296 References: https://bugs.freedesktop.org/show_bug.cgi?id=45086
13297 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13299 commit 352828ee59164a9e81093d88dfdd45bc21f0c739
13300 Author: Chris Wilson <chris@chris-wilson.co.uk>
13301 Date: Thu Jan 26 12:41:34 2012 +0000
13303 sna: Tweak aperture thresholds for batch flushing
13305 In order to more easily accommodate operations on large source CPU bo.
13307 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13309 commit cff6a1a2e4648eb211a1789ae9f711e2f16e9d4d
13310 Author: Chris Wilson <chris@chris-wilson.co.uk>
13311 Date: Wed Jan 25 23:21:36 2012 +0000
13313 sna: Use the cpu bo where possible as the source for texture extraction
13315 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13317 commit e583af9cca4ad2e5643317447c6b065d3ee7d11e
13318 Author: Chris Wilson <chris@chris-wilson.co.uk>
13319 Date: Wed Jan 25 23:04:50 2012 +0000
13321 sna: Experiment with creating large objects as CPU bo
13323 Even on non-LLC systems if we can prevent the migration of such
13324 objects, we can still benefit immensely from being able to map them into
13325 the GTT as required.
13327 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13329 commit 55569272f7d4232ef50f7b964dda82f85a190b99
13330 Author: Chris Wilson <chris@chris-wilson.co.uk>
13331 Date: Wed Jan 25 20:13:27 2012 +0000
13333 sna: Apply the same migration flags for the dst alphamap as for the dst pixmap
13335 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13337 commit 4a132ddbf06e5ffc364c25002a1e46ad8bf0e45a
13338 Author: Chris Wilson <chris@chris-wilson.co.uk>
13339 Date: Wed Jan 25 20:12:55 2012 +0000
13341 sna: Correct offset for moving drawable regions to the CPU
13343 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13345 commit 65164d90b7b17ec7eea1e24d4b02ec037b55b1ff
13346 Author: Chris Wilson <chris@chris-wilson.co.uk>
13347 Date: Wed Jan 25 20:12:27 2012 +0000
13349 sna/gen2+: Do not force use of GPU if the target is simply cleared
13351 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13353 commit 307f493d76580687a3cf56106bf296475f1f53e5
13354 Author: Chris Wilson <chris@chris-wilson.co.uk>
13355 Date: Wed Jan 25 20:11:21 2012 +0000
13357 sna: Map freshly created, unbound bo through the CPU
13359 Take advantage that we know we will have to clflush the unbound bo
13360 before use by the GPU and populate it inplace.
13362 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13364 commit d785bb7df054a1f15d59db69b089deb743bbdb40
13365 Author: Chris Wilson <chris@chris-wilson.co.uk>
13366 Date: Wed Jan 25 16:45:43 2012 +0000
13368 sna: GetImage is allowed to read a window's border
13370 We need to adjust the clip to include the border pixels when migrating
13371 damage from the backing pixmap. This also requires relaxing the
13372 constraint that a read must be within the drawable.
13374 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13376 commit 36425ba49ecbd87b1e3bf4340ca2496d8de24a7f
13377 Author: Chris Wilson <chris@chris-wilson.co.uk>
13378 Date: Wed Jan 25 12:49:24 2012 +0000
13380 sna: Round up buffer allocations when downsampling
13382 The pathological case being nx1 or 1xm resulting in an illegal allocation
13383 request of 0 bytes.
13385 One such example is
13386 wolframalpha.com: x = (200 + x) / 100
13387 which generates an approximately 8500x1 image and so needs downscaling
13388 to fit in the render pipeline on all but IvyBridge. Bring on Ivy!
13390 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13392 commit a2e83c6dcba1e911f42a3004b3d0782049e243e2
13393 Author: Chris Wilson <chris@chris-wilson.co.uk>
13394 Date: Wed Jan 25 11:22:38 2012 +0000
13396 sna: Silence compiler warning for a potential uninitialised return on error
13398 sna_accel.c: In function 'sna_copy_plane':
13399 sna_accel.c:5022:21: warning: 'ret' may be used uninitialized in this
13400 function [-Wuninitialized]
13402 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13404 commit 8d22a76506133e0f76424159c0944d29bdf39da9
13405 Author: Chris Wilson <chris@chris-wilson.co.uk>
13406 Date: Wed Jan 25 09:36:20 2012 +0000
13408 sna: Run the miHandleExposures for no-op CopyPlane
13410 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13412 commit 338941eda3c7591a85b83000eafae0407d0d7cd0
13413 Author: Chris Wilson <chris@chris-wilson.co.uk>
13414 Date: Wed Jan 25 02:42:25 2012 +0000
13416 sna: Handle self-copies for CopyPlane
13418 Prepare the source first as this has the dual benefit of letting us
13419 decide how best to proceed with the op (on the CPU or GPU) and prevents
13420 modification of the damage after we have choosen our preferred path.
13422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13424 commit 2e8b398ca383f5292adab8b351b8837dde3e131a
13425 Author: Chris Wilson <chris@chris-wilson.co.uk>
13426 Date: Wed Jan 25 01:42:56 2012 +0000
13428 sna: Only shrink partial buffers that are being written to
13430 Ignore inactive and mmapped buffers.
13432 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13434 commit b79252efaafe2ebc998d6cf6176a425dd897e66f
13435 Author: Chris Wilson <chris@chris-wilson.co.uk>
13436 Date: Wed Jan 25 01:36:27 2012 +0000
13438 sna: Apply source clipping to sna_copy_plane()
13440 Ensure that the migration region is within bounds for both the source
13441 and destination pixmaps.
13443 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13445 commit 46252bc7bcc7e08e47d00cdc87d6c1ed93830fcc
13446 Author: Chris Wilson <chris@chris-wilson.co.uk>
13447 Date: Wed Jan 25 01:31:34 2012 +0000
13449 sna: Set the source clip for CopyArea fallback correctly
13451 The source window is (src->x, src->y)x(src->width, src->height) in
13452 pixmap space. However, we then need to use this to clip against the
13453 desination region, and so we need to translate from the source
13454 coordinate to the destination coordinate.
13456 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13458 commit ae6d3a311783d7e063de0347363331f14bd74d74
13459 Author: Chris Wilson <chris@chris-wilson.co.uk>
13460 Date: Wed Jan 25 01:18:11 2012 +0000
13462 sna: Print source and destination regions for CopyArea fallback for DBG
13464 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13466 commit dd5e90adfc73870cebcb215ad9fb9b5aedd38673
13467 Author: Chris Wilson <chris@chris-wilson.co.uk>
13468 Date: Wed Jan 25 01:17:49 2012 +0000
13470 sna: Clip GetImage to drawable so that damage migration is within bounds
13472 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13474 commit b1fba5e8534da7fe253e21a3501854c04d82a108
13475 Author: Chris Wilson <chris@chris-wilson.co.uk>
13476 Date: Wed Jan 25 01:16:34 2012 +0000
13478 sna: Clear GPU damage first when moving a clear pixmap to the CPU
13480 This allows us to discard any busy GPU or CPU bo when we know we are
13481 going to clear the shadow pixmap afterwards.
13483 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13485 commit 5ad95d66665802bce25e127ae0d06f3e0a9b0e62
13486 Author: Chris Wilson <chris@chris-wilson.co.uk>
13487 Date: Tue Jan 24 22:11:20 2012 +0000
13489 sna: Reduce number of reads required to inspect timers
13491 By using the information provided by select at wakeup.
13493 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13495 commit aae19cbc5d0ddcf247451d06e063b2550a7ff16f
13496 Author: Chris Wilson <chris@chris-wilson.co.uk>
13497 Date: Tue Jan 24 19:58:29 2012 +0000
13499 sna: Only reset devPrivate.ptr if owned by the CPU bo when freeing
13501 If the pixmap is mapped to the GPU bo, we should continue to use the
13502 current mapping rather than revoke it. Otherwise if we write to the GPU
13503 bo inplace, thereby discarding the CPU bo, we set the pointer we are
13504 about to copy to, to NULL.
13506 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13508 commit 5312ee90ad80eda1c6ff72fbef60f77f03094972
13509 Author: Chris Wilson <chris@chris-wilson.co.uk>
13510 Date: Tue Jan 24 19:02:56 2012 +0000
13512 sna: mark the pixmap as no longer clear after rendering video
13514 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13516 commit 69d3fc91f4a397f704b9971f97a905626e86eb3c
13517 Author: Chris Wilson <chris@chris-wilson.co.uk>
13518 Date: Tue Jan 24 18:51:32 2012 +0000
13520 sna: Set up GC for general FillArc to FillSpans callback
13522 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13524 commit 855c45c65a69d34ad32a97a5fc54dd09d40d3e26
13525 Author: Chris Wilson <chris@chris-wilson.co.uk>
13526 Date: Tue Jan 24 18:48:17 2012 +0000
13528 sna: Enable clipping for mi FillSpans gpu callback
13530 The mi routines do not ensure that their output is suitably constrained
13531 to the clip extents, so we must run it through the clipper.
13533 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13535 commit c3fe42cde70a06a9a9a6190934df5cddb41d2f5f
13536 Author: Chris Wilson <chris@chris-wilson.co.uk>
13537 Date: Tue Jan 24 18:36:08 2012 +0000
13539 sna: Don't attempt to create pixmaps for dpeths smaller than 8
13541 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13543 commit 22bab3cd7d0b2bcedfdd86f198d1ed9a21dd82aa
13544 Author: Chris Wilson <chris@chris-wilson.co.uk>
13545 Date: Tue Jan 24 18:33:14 2012 +0000
13547 sna: Handle malloc failures when downsampling
13549 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13551 commit 4aef9e6f50cb96c57b7ba345a18e55c97d788192
13552 Author: Chris Wilson <chris@chris-wilson.co.uk>
13553 Date: Tue Jan 24 17:23:43 2012 +0000
13555 sna: Add more damage-is-contained asserts
13557 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13559 commit 16462e551d575bd5a370a8ac8b450ba72fea7d47
13560 Author: Chris Wilson <chris@chris-wilson.co.uk>
13561 Date: Tue Jan 24 17:47:22 2012 +0000
13563 sna: Install CPU handler for PolyLine for miPolySegment callback
13565 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13567 commit b8b49ae70b3aadf4423bd6830c446c040f87a2f1
13568 Author: Chris Wilson <chris@chris-wilson.co.uk>
13569 Date: Tue Jan 24 16:32:40 2012 +0000
13571 sna: Mutate a temporary copy of the GC->ops so we don't alter new GCs
13573 In order to avoid using the wrong function for a scratch GC created
13574 during the course of a MI function whilst we have a specialised GC in
13575 use, we need to avoid modifying the original function table.
13577 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13579 commit 1a116a15a5d978a1edf413a439f6ef6d3205a7aa
13580 Author: Chris Wilson <chris@chris-wilson.co.uk>
13581 Date: Tue Jan 24 16:53:21 2012 +0000
13583 sna: Reduce the assertion that we would have hinted to operate inplace
13585 As the drawable_gc_flags() operate on lower level information than the
13586 hint, it is able to spot more oportunities to reduce the READ flags and
13587 so the assertion was overly optimistic.
13589 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13591 commit 3d1bba033bc29fdf498dc082f3542c520a5ed39a
13592 Author: Chris Wilson <chris@chris-wilson.co.uk>
13593 Date: Tue Jan 24 18:22:12 2012 +0000
13595 sna/gen3: Apply damage to video pixmap
13597 Reported-by: Paul Neumann <paul104x@yahoo.de>
13598 References: https://bugs.freedesktop.org/show_bug.cgi?id=44504
13599 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13601 commit ff4b8a03a40f7eb1d71b96aebb5daa1d8df83d2a
13602 Author: Chris Wilson <chris@chris-wilson.co.uk>
13603 Date: Tue Jan 24 17:51:35 2012 +0000
13605 sna: Use RepeatPad to avoid sampling the border color when downsampling
13607 References: https://bugs.freedesktop.org/show_bug.cgi?id=45086
13608 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13610 commit 09bddddaf94b848f417e1c9449cfe4b21d2554dd
13611 Author: Chris Wilson <chris@chris-wilson.co.uk>
13612 Date: Tue Jan 24 15:50:02 2012 +0000
13614 sna: handle negative modulus for tiled blits
13616 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13618 commit f3b11ebc28114062581c39089781114323946654
13619 Author: Chris Wilson <chris@chris-wilson.co.uk>
13620 Date: Tue Jan 24 16:05:09 2012 +0000
13622 sna: Add more assertions for copy errors
13624 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13626 commit b56971986d2d495aa38f015beb659fb12f0d433a
13627 Author: Chris Wilson <chris@chris-wilson.co.uk>
13628 Date: Tue Jan 24 17:23:59 2012 +0000
13630 sna: Trim the clipped Segment to the right length
13632 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13634 commit 358a0861e642c26d925a69656039b4c95adba237
13635 Author: Chris Wilson <chris@chris-wilson.co.uk>
13636 Date: Tue Jan 24 13:30:19 2012 +0000
13638 sna: Do not discard CPU damage for a partial copy
13640 If we are copying a region that does not fill its extents (i.e. is not
13641 singular) then we must be care not to discard the CPU damage that is not
13642 overwritten by the copy.
13644 Fixes regression from 77ee92248582d65a03619d1bb1d93a74468eea00
13645 (sna: Use full usage flags for moving the dst pixmap for a copy).
13647 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13649 commit 672c6ad8170572c0f2562d5a8381ab2d17fd1e1b
13650 Author: Chris Wilson <chris@chris-wilson.co.uk>
13651 Date: Tue Jan 24 12:56:59 2012 +0000
13653 sna: Silence the compiler with a few DBG format warnings
13655 The size of the integers being printed changed causing warnings on 32/64
13658 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13660 commit 238f3d9bc563950475ee823f844197cf15a909e2
13661 Author: Chris Wilson <chris@chris-wilson.co.uk>
13662 Date: Tue Jan 24 12:19:38 2012 +0000
13664 sna: A deferred attachment will never have a CPU bo
13666 So we can remove that check in sna_copy_boxes.
13668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13670 commit c353a8cfde838b070f3e71d3946f3b40d1d5113a
13671 Author: Chris Wilson <chris@chris-wilson.co.uk>
13672 Date: Tue Jan 24 10:52:27 2012 +0000
13674 sna: Revise use of the inplace hint to favour small cacheable operations
13676 The goal of the heuristic is to reduce readbacks and damage tracking on
13677 active GPU bo whilst simultaneously offering the best performance for
13678 small operations which would prefer to be performed on the shadow rather
13681 This restores ShmPutImage performance.
13683 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13685 commit 29547421f392a9b589ba069cbfbba01dbe535d93
13686 Author: Chris Wilson <chris@chris-wilson.co.uk>
13687 Date: Tue Jan 24 10:23:46 2012 +0000
13689 sna: Use the reduced ROP from the fbGC to accurately determine readback
13691 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13693 commit ec794f7594ae5e2c223b7f7255ff8979d9fb06c1
13694 Author: Chris Wilson <chris@chris-wilson.co.uk>
13695 Date: Mon Jan 23 22:14:15 2012 +0000
13697 sna: Track a proxy's dirty status upon itself rather than the target
13699 As proxy's are short-lived and are not used outside of the operation for
13700 which they are created, dirtied or flushed, we can safely copy the dirty
13701 status onto the proxy object itself.
13703 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13705 commit 2574a04c940e032fdcf1cd39a0950515e8652471
13706 Author: Chris Wilson <chris@chris-wilson.co.uk>
13707 Date: Mon Jan 23 21:45:29 2012 +0000
13709 sna: Split the active bo cache into per-tiling caches
13711 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13713 commit 566cbf1ef53b1f970289fcd5b3b389a74beb7e78
13714 Author: Chris Wilson <chris@chris-wilson.co.uk>
13715 Date: Mon Jan 23 21:15:10 2012 +0000
13717 sna: Abort search for a linear bo if we encounter a purged buffer
13719 Given the rarity of encountering a purged buffer versus the frequency of
13720 scanning the list and the then likely result of allocation a new buffer,
13721 simply abort the search on the first purged bo.
13723 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13725 commit b8d3353624be51f2b618467c899a9ce3cbf3cbfe
13726 Author: Chris Wilson <chris@chris-wilson.co.uk>
13727 Date: Mon Jan 23 19:13:22 2012 +0000
13729 sna: Refactor tests for a solid gc to use the new function
13731 This way all paths can test to see if they might be able to reduce the
13732 tiled fill or the opaque fill into a solid fill.
13734 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13736 commit 78238b3a37c1c8a152bd3920e73aa6034c0f663f
13737 Author: Chris Wilson <chris@chris-wilson.co.uk>
13738 Date: Mon Jan 23 18:00:05 2012 +0000
13740 sna: Prefer mi span filling functions to call through to the GPU
13742 Having removed the double analysis for the fast paths, at least, the
13743 span filling code on the GPU is now faster than doing the same
13744 operations in cache memory for the majority of cases. So allow the
13745 driver to prefer to use those functions when it has a GPU bo.
13747 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13749 commit c5c77d04d79666962fd7c2d86bc6e7aef716084b
13750 Author: Chris Wilson <chris@chris-wilson.co.uk>
13751 Date: Mon Jan 23 18:24:21 2012 +0000
13753 sna: Reduce GC to solid if possible along general spans path
13755 Again, we should no longer be hitting this code if at all possible, but
13756 for completeness...
13758 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13760 commit 03611f662db854dc3b14f32c0315bf45bc415bbf
13761 Author: Chris Wilson <chris@chris-wilson.co.uk>
13762 Date: Mon Jan 23 19:59:06 2012 +0000
13764 sna: Guard against being passed n==0 to FillSpans
13766 Apparently this does happen, notably by miFillPolyArc.
13768 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13770 commit 7ac13a4d5aca1627b3a5fc9e7261d5dfafba970b
13771 Author: Chris Wilson <chris@chris-wilson.co.uk>
13772 Date: Tue Jan 24 00:21:12 2012 +0000
13774 sna: Provide a fast path for miZeroLine for PolySegment
13776 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13778 commit e7817a2206bd0b1cc4e4458686c328f7b41ea32c
13779 Author: Chris Wilson <chris@chris-wilson.co.uk>
13780 Date: Mon Jan 23 17:43:55 2012 +0000
13782 sna: Override sna_fill_spans for PolySegment
13784 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13786 commit 5b8db54d0fa9eab2610feed6947e917f2e400e1a
13787 Author: Chris Wilson <chris@chris-wilson.co.uk>
13788 Date: Tue Jan 24 00:21:12 2012 +0000
13790 sna: Provide a fast path for miZeroLine for PolyLine
13792 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13794 commit 280a9165db5382b93ab449158106bdd5aa7b2329
13795 Author: Chris Wilson <chris@chris-wilson.co.uk>
13796 Date: Mon Jan 23 17:43:55 2012 +0000
13798 sna: Override sna_fill_spans for PolyLine
13800 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13802 commit 72fc3122b85fe1a1a16c7f7401c300089273fd88
13803 Author: Chris Wilson <chris@chris-wilson.co.uk>
13804 Date: Tue Jan 24 01:27:08 2012 +0000
13806 sna: Prefer to always use span generation for PolyArc
13808 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13810 commit 98879e3786b7f6241378a32a10f2d46eb01baf85
13811 Author: Chris Wilson <chris@chris-wilson.co.uk>
13812 Date: Mon Jan 23 23:22:00 2012 +0000
13814 sna: Specialise PolyPoint for callback from miZeroPolyArc
13816 miZeroPolyArc may use either FillSpans for PolyPoint to generate its
13817 curves, so also provide custom point filling routines.
13819 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13821 commit 5b10cb4a5ce84bd677b0a67a17d8674defb19f31
13822 Author: Chris Wilson <chris@chris-wilson.co.uk>
13823 Date: Mon Jan 23 17:02:46 2012 +0000
13825 sna: Override sna_fill_spans for PolyArc
13827 This is the missing like that allows the GPU to outperform fbPolyArc.
13829 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13831 commit 6390a44a69d7f13a3ba16e8b48c640dc39dd227f
13832 Author: Chris Wilson <chris@chris-wilson.co.uk>
13833 Date: Mon Jan 23 15:57:53 2012 +0000
13835 sna: Pass the reduce clip region to sna_fill_spans__cpu
13837 Since we compute it for the pixmap migration, we may as well use to
13838 perform the clipping within FillSpans as well.
13840 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13842 commit 40e3745debe9738362661c7d7f498f3489c28876
13843 Author: Chris Wilson <chris@chris-wilson.co.uk>
13844 Date: Mon Jan 23 15:42:18 2012 +0000
13846 sna: Wrap gc->ops->PolyFillArc
13848 The goal is to avoid the overhead of performing multiple region analysis
13849 when calling sna_fill_spans by doing it once at the top level and then
13850 choose the most appropriate drawing method.
13852 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13854 commit 55c2211143b3be2827fd4e299c3c211f3bd7935b
13855 Author: Chris Wilson <chris@chris-wilson.co.uk>
13856 Date: Mon Jan 23 18:15:36 2012 +0000
13858 sna: Use a streamlined sna_fill_spans for when we know we want to use the gpu
13860 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13862 commit b20ae331566e92f1a995f49fd0760f020db25035
13863 Author: Chris Wilson <chris@chris-wilson.co.uk>
13864 Date: Mon Jan 23 19:49:18 2012 +0000
13866 sna: Switch to using spans based on use_wide_spans()
13868 This just gives us a hook with which to conveniently enable or disable
13871 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13873 commit 41739ff9517ca158d9f7b7ac82e5bacef62a3d68
13874 Author: Chris Wilson <chris@chris-wilson.co.uk>
13875 Date: Mon Jan 23 15:42:18 2012 +0000
13877 sna: Wrap gc->ops->FillPolygon
13879 The goal is to avoid the overhead of performing multiple region analysis
13880 when calling sna_fill_spans by doing it once at the top level and then
13881 choose the most appropriate drawing method.
13883 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13885 commit aef56b194d4de9bd71ab71e3d93d81e1750262c5
13886 Author: Chris Wilson <chris@chris-wilson.co.uk>
13887 Date: Mon Jan 23 15:38:30 2012 +0000
13889 sna: Mark GPU bo as busy when drawing to an all-damaged pixmap for core
13891 The fast path to decide whether to use the GPU bo for the core drawaing
13892 operations forgot to update the active status of the pixmap. This
13893 included forgetting to clear the is-cleared flag.
13895 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13897 commit c67e6a05ff729c0ddb26775e1c8533be1144d85a
13898 Author: Chris Wilson <chris@chris-wilson.co.uk>
13899 Date: Mon Jan 23 11:40:13 2012 +0000
13901 sna: Align the origin of downsampling subregion to an even pixel
13903 In order for us to produce stable downsampled images across multiple
13904 frames, we need to sample the same pairs of pixels every time. This
13905 requires us to align the origin of the sample region to an even pixel.
13907 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
13908 References: https://bugs.freedesktop.org/show_bug.cgi?id=45086
13909 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13911 commit 9267da93969d1727cb95956c3f63643f0039c60d
13912 Author: Chris Wilson <chris@chris-wilson.co.uk>
13913 Date: Mon Jan 23 11:20:12 2012 +0000
13915 sna: Apply composite offset to copy-clear GPU path
13917 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13919 commit 803f6558c665c16ba2d20415e778b5bd1d308c2f
13920 Author: Chris Wilson <chris@chris-wilson.co.uk>
13921 Date: Mon Jan 23 11:05:50 2012 +0000
13923 sna: Only subtract the fillrect from the CPU damage if it covers the drawable
13925 The idea behind that optimisation is for the inactive pixmap to be
13926 refreshed and allowed to be transfered back to the GPU when it is
13927 entirely redrawn. As such performing the subtraction when it does not
13928 completely remove it only incurs additional overhead.
13930 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13932 commit 267493f39137704bd52df65b7605be30d2a062b2
13933 Author: Chris Wilson <chris@chris-wilson.co.uk>
13934 Date: Mon Jan 23 10:54:13 2012 +0000
13936 sna: Assert that the subtract operation does reduce an all-damaged
13938 Somewhere somewhen it appears that I am discarding the all-damaged flag
13939 on the pointer. The only possibility I can see is for a no-op
13940 subtraction, so put an assert there just in case the impossible is
13943 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13945 commit 2908fe2df7b8f870a55fc5856f94ea541a7c7b75
13946 Author: Chris Wilson <chris@chris-wilson.co.uk>
13947 Date: Sun Jan 22 18:46:57 2012 +0000
13949 sna: Remove assertion that all flushing list members are in the GPU domain
13951 This may not be true for external buffers that are put on the flushing
13952 list because they have foreign requests pending.
13954 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13956 commit a0bc17c0fa34c9f9243817d1674d1ff235dc99d6
13957 Author: Chris Wilson <chris@chris-wilson.co.uk>
13958 Date: Sun Jan 22 18:27:34 2012 +0000
13960 sna/gen6: stall before blending dirty pixels from the render cache
13962 Another restriction on the render pipeline, it turns out, is that before
13963 the blend unit can read back the dst pixels in a subsequent primitive,
13964 we must stall the pipeline for the completion of that earlier primitive.
13966 This is demonstrated by cacomposite.
13968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13970 commit 5aa494305e90eed5df211427741a75380add8e96
13971 Author: Chris Wilson <chris@chris-wilson.co.uk>
13972 Date: Sun Jan 22 16:10:39 2012 +0000
13974 sna: Correct the offsets for performing partial downsampling on the GPU
13976 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
13977 References: https://bugs.freedesktop.org/show_bug.cgi?id=45086
13978 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13980 commit a7b24bb4342eefb2b23fa40d7c7216c38f2edc09
13981 Author: Chris Wilson <chris@chris-wilson.co.uk>
13982 Date: Sun Jan 22 15:50:47 2012 +0000
13984 sna: Correct image buffer size when uploading a downsampled image
13986 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
13987 References: https://bugs.freedesktop.org/show_bug.cgi?id=45086
13988 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13990 commit 238fc67aa945d5669eccd2a8fe8e98402b9a6223
13991 Author: Chris Wilson <chris@chris-wilson.co.uk>
13992 Date: Sun Jan 22 16:04:13 2012 +0000
13994 sna: Incude prefer-gpu hint in decision to upload CPU textures
13996 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
13998 commit aeaffb102190a8846c6341e72a4338b41639d679
13999 Author: Chris Wilson <chris@chris-wilson.co.uk>
14000 Date: Sun Jan 22 15:52:41 2012 +0000
14002 sna: Add some more DBG along the downsampling source extraction path
14004 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14006 commit 92badc9c9c2636284af8bc56b098d24ff222f167
14007 Author: Chris Wilson <chris@chris-wilson.co.uk>
14008 Date: Sun Jan 22 12:41:31 2012 +0000
14010 sna: Avoid a common readback after clearing
14012 It is surprisingly common for a pixmap to be created, cleared and then
14013 used as an upload target or, even worse, as a source for a ShmGetImage.
14014 In order to prevent this folly, we can trivially track when we clear an
14015 entire pixmap and its GPU bo and avoid the readback in such cases.
14017 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14019 commit 67a166d1247a1197fc851308c33c4970868f6a21
14020 Author: Chris Wilson <chris@chris-wilson.co.uk>
14021 Date: Sun Jan 22 13:46:05 2012 +0000
14023 sna: Pass inplace hint for sna_copy_boxes fallback
14025 In the unlikely event that this makes a difference, provide the hint as
14026 to when we do not read back from the destination and so a streaming copy
14027 would be preferable.
14029 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14031 commit 77ee92248582d65a03619d1bb1d93a74468eea00
14032 Author: Chris Wilson <chris@chris-wilson.co.uk>
14033 Date: Sun Jan 22 13:42:57 2012 +0000
14035 sna: Use full usage flags for moving the dst pixmap for a copy
14037 Now that the migration code can decide for itself when to not move
14038 damage, we can pass the hints to the code rather than perform the
14039 optimisation in sna_copy_boxes.
14041 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14043 commit dd079dd1f64e7832c6343c24958aa9eeb2f58789
14044 Author: Chris Wilson <chris@chris-wilson.co.uk>
14045 Date: Sat Jan 21 19:38:40 2012 +0000
14047 sna: Check the blt composite op is within the source bounds
14049 If we attempt to read from a drawable that is partially off its backing
14050 pixmap (such as a partially offscreen composite window) we need to fixup
14051 the read from the out-of-bounds regions to return clear. Since we don't,
14052 the easier answer is to just to switch to the render pipeline for such
14055 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14057 commit 8a91c7d85740a5adc25d2a9b1972c367780ce714
14058 Author: Chris Wilson <chris@chris-wilson.co.uk>
14059 Date: Sat Jan 21 01:18:17 2012 +0000
14061 sna: Give kgem_create_buffer_2d a unique id
14063 The gen3, among others, backend uses the unique id of a buffer to track
14064 the currently attached buffer and uses 0 as the invalid value. Linear
14065 buffers as created by kgem_create_buffer_2d were not being assigned a
14066 unique id causing mayhem when they were then being passed to the
14067 backends as render targets and sources. In particular, gen3 did not
14068 notice the switch in render target and did not emit commands to change
14069 the GPU target nor attach the buffer to the batch, causing the
14070 sna_read_boxes to fail and for us to trigger an assertion for an
14071 uncomsumed read buffer.
14073 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
14074 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42718
14075 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14077 commit d21e80433f2574fe96036ad1d563515406a1946f
14078 Author: Chris Wilson <chris@chris-wilson.co.uk>
14079 Date: Sat Jan 21 01:17:39 2012 +0000
14081 sna/gen3: Add DBG markup for tracking target state
14083 In particular for when we switch to a new destination and mark it dirty.
14085 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14087 commit 20006a9ae7f993f02c45ccbc1e414d74bfee3acd
14088 Author: Chris Wilson <chris@chris-wilson.co.uk>
14089 Date: Fri Jan 20 22:17:44 2012 +0000
14091 sna: Cleanup PolyText and ImageText fallback debugging
14093 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14095 commit 3fbd7cafb4394e0641565dc794b82bef615b4aac
14096 Author: Chris Wilson <chris@chris-wilson.co.uk>
14097 Date: Fri Jan 20 22:16:53 2012 +0000
14099 sna: Copy the glyphs metrics before checking for an empty glyph
14101 The empty glyph still needs the correct advance, and copying it too late
14102 left it as zero and so we were collapsing spaces in PolyText8 and
14105 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14107 commit f0d8f5d377460bb7da82546e8304252312e97147
14108 Author: Chris Wilson <chris@chris-wilson.co.uk>
14109 Date: Fri Jan 20 14:34:28 2012 +0000
14111 sna/gen3: Set bo to NULL for render_fill_one
14113 It had to happen, I missed one! From fc9531fc2df.
14115 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14117 commit 0286ddbd67d3b2585ffa3deccb7eeb674cd7a3e0
14118 Author: Chris Wilson <chris@chris-wilson.co.uk>
14119 Date: Fri Jan 20 11:33:05 2012 +0000
14121 sna/gen6: Cleanup non-pipelined w/a
14123 Move the workaround CS stall into the emit drawrect which is the only
14124 non-pipelined op we emit. This removes the split between deciding
14125 whether we will emit a drawrect and actual emission.
14127 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14129 commit a0c4024e32ccab70e7f7893896d30c80ad9f4240
14130 Author: Chris Wilson <chris@chris-wilson.co.uk>
14131 Date: Fri Jan 20 09:42:18 2012 +0000
14133 sna/gen6: Do not enable blending on video sources
14135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14137 commit bbbda4c408c2576927582097f7cd4f606b1761fa
14138 Author: Chris Wilson <chris@chris-wilson.co.uk>
14139 Date: Fri Jan 20 09:40:04 2012 +0000
14141 sna/gen6: Reduce the sampler stall if already stalling for drawrect
14143 If we are force to emit a stall for the non-pipelined workaround, we do
14144 not then need to emit a stall for switching samplers.
14146 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14148 commit e91601278a057f70df76b74539ff4f80d203bf00
14149 Author: Chris Wilson <chris@chris-wilson.co.uk>
14150 Date: Fri Jan 20 01:37:29 2012 +0000
14152 sna/gen6: Force a CS stall for changing samplers
14154 On closer inspection, we still need the workaround of forcing a pipeline
14155 stall if we update the samplers.
14157 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14159 commit fc9531fc2df12b896e6d9e8b3f0d7248e55bfcdd
14160 Author: Chris Wilson <chris@chris-wilson.co.uk>
14161 Date: Thu Jan 19 21:15:35 2012 +0000
14163 sna: Move the flush to the backends
14165 This allows us to implement backend specific workarounds and use the
14166 more appropriate device specific flushing.
14168 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14170 commit 2e0a534a888c59b921fd4cad69b980c6eda72212
14171 Author: Chris Wilson <chris@chris-wilson.co.uk>
14172 Date: Thu Jan 19 18:52:35 2012 +0000
14174 sna/gen7: Forward port recent changes from gen6
14176 Fixes for resubmitting batches after running out of space for vertex
14177 buffers and also a couple of trivial spans functions.
14179 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14181 commit 5caf806d42a1909ff3a7716d8bc592fa68a02377
14182 Author: Chris Wilson <chris@chris-wilson.co.uk>
14183 Date: Thu Jan 19 17:33:25 2012 +0000
14185 sna: BLT use dword pitch only for tiled surfaces
14187 The gen4+ spec is a little misleading as states that all BLT pitches for
14188 the XY commands are in dwords. Apparently not, as the upload/download
14189 functions were already demonstrating. This only became apparent when
14190 accelerating core text routines to offscreen pixmaps, such as composited
14193 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14195 commit dbc75532d5f6a10e9fc7cfe3634d5a0db7304a99
14196 Author: Chris Wilson <chris@chris-wilson.co.uk>
14197 Date: Thu Jan 19 14:36:32 2012 +0000
14199 sna: Tweak move-to-cpu to ignore inplace hint if its already on the CPU
14201 If we test the area to be drawn against the existing CPU damage and find
14202 it is already on the CPU, we may as well continue to utilize that
14205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14207 commit 7ad4a0c9423ee7f4041173c428b07ac4af312fe1
14208 Author: Chris Wilson <chris@chris-wilson.co.uk>
14209 Date: Thu Jan 19 11:15:38 2012 +0000
14211 sna: Only use the blitter to emit wide spans if we cannot stream the updates
14213 If either the region is busy on the gpu or if we need to read the
14214 destination then we would incur penalties for trying to perform the
14215 operation through the GTT. However, if we are simply streaming pixels to
14216 an unbusy bo then we can do so inplace faster than computing the
14217 corresponding GPU commands and uploading them.
14219 Note: currently it is universally slower to use the GPU here (the
14220 computation of the spans is too slow). However that is only according to
14221 micro-benchmarks, avoiding the readback is likely to be more efficient
14224 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14226 commit 9db6b9fad808175184849f3030862ea115ef0708
14227 Author: Chris Wilson <chris@chris-wilson.co.uk>
14228 Date: Thu Jan 19 11:54:14 2012 +0000
14230 sna: Also check for the inplace hint when migrating the whole pixmap
14232 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14234 commit d3f7d5d614748306cc376d1929da2bbc14494a6e
14235 Author: Chris Wilson <chris@chris-wilson.co.uk>
14236 Date: Thu Jan 19 11:15:38 2012 +0000
14238 sna: Only use the blitter to emit spans if we cannot stream the updates
14240 If either the region is busy on the gpu or if we need to read the
14241 destination then we would incur penalties for trying to perform the
14242 operation through the GTT. However, if we are simply streaming pixels to
14243 an unbusy bo then we can do so inplace faster than computing the
14244 corresponding GPU commands and uploading them.
14246 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14248 commit ff2eb116ef85182eea9ed06daaa1e9a4f7bdbad3
14249 Author: Chris Wilson <chris@chris-wilson.co.uk>
14250 Date: Wed Jan 18 00:56:07 2012 +0000
14252 sna: Micro-optimise line extents for zero line width
14254 Handling zero line widths is the common case, so avoid the extra work.
14256 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14258 commit 3c010745076204ed3b66e947c9eab6de84f670fe
14259 Author: Chris Wilson <chris@chris-wilson.co.uk>
14260 Date: Thu Jan 19 09:46:52 2012 +0000
14262 sna: filter out degenerate segments whilst drawing unclipped PolySegment
14264 The damage layer was detecting that we were asking it to accumulate a
14265 degenerate box emanating from PolySegment, as the unclipped paths made
14266 the fatal assumption that it would not need to filter out degenerate
14267 boxes. However, a degenerate line becomes a point, does the same apply
14268 to a degenerate segment?
14270 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14272 commit 35f81005f91d294e61bb4ced7cbddd1a76ccb324
14273 Author: Chris Wilson <chris@chris-wilson.co.uk>
14274 Date: Thu Jan 19 00:41:08 2012 +0000
14276 sna/damage: Always mark the damage as dirty when recording new boxes
14278 A few of the create_elts() routines missed marking the damage as dirty
14279 so that if only part of the emebbed box was used (i.e. the damage
14280 contained less than 8 rectangles that needed to included in the damage
14281 region) then those were being ignored during migration and testing.
14283 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
14284 References: https://bugs.freedesktop.org/show_bug.cgi?id=44682
14285 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14287 commit 36e691ea900d7979950c19714e15c5887d3ea039
14288 Author: Chris Wilson <chris@chris-wilson.co.uk>
14289 Date: Thu Jan 19 00:39:28 2012 +0000
14291 sna: Demote MOVE_READ if the GC operation covers the clip
14293 If the write operation fills the entire clip, then we can demote and
14294 possible avoid having to read back the clip from the GPU provided that
14295 we do not need the destination data due to arithmetic operation or mask.
14297 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14299 commit 17efdbc48c964d9d7b0af630c13a6ceb62fb1787
14300 Author: Chris Wilson <chris@chris-wilson.co.uk>
14301 Date: Thu Jan 19 00:36:52 2012 +0000
14303 sna: Clip damage area with source extents for fallback
14305 The damage tracking code asserts that it only handles clip regions.
14306 However, sna_copy_area() was failing to ensure that its damage region
14307 was being clipped by the source drawable, leading to out of bounds reads
14308 during forced fallback.
14310 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14312 commit fb07243c9a08b24e2dc03b86dc16d3cd72cbfcba
14313 Author: Chris Wilson <chris@chris-wilson.co.uk>
14314 Date: Wed Jan 18 21:30:34 2012 +0000
14316 sna: Fine grained fallback debugging for core drawing routines
14318 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14320 commit 05f9764a8860cbdd7ea2ce7f04c221ec8481e0bc
14321 Author: Chris Wilson <chris@chris-wilson.co.uk>
14322 Date: Wed Jan 18 21:29:45 2012 +0000
14324 sna/damage: Fast path singular regions
14326 Mainly for consistency, so that we treat it like the other damage
14327 addition functions.
14329 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14331 commit 96529e345d0718d253e4abfa0a02cd1f9f7c0406
14332 Author: Chris Wilson <chris@chris-wilson.co.uk>
14333 Date: Wed Jan 18 21:05:40 2012 +0000
14335 sna: Make sure we create a mappable GPU bo when streaming writes
14337 If we decide to do the CPU fallback inplace on the GPU bo through a WC
14338 mapping (because it is a large write-only operation), make sure that
14339 the new GPU bo we create is not active and so will not^W^W is less likely
14340 to cause a stall when mapped.
14342 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14344 commit efce896e1dbab6177a773a7cf1c8eaf2dab0cdee
14345 Author: Chris Wilson <chris@chris-wilson.co.uk>
14346 Date: Wed Jan 18 20:53:55 2012 +0000
14348 sna: Check number of boxes to migrate during move-to-cpu
14350 When reducing the damage we may find that it is actually empty and so
14351 sna_damage_get_boxes() returns 0, be prepared.
14353 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14355 commit 334f3f70a8b4372cb089773ecded8339d2c7d984
14356 Author: Chris Wilson <chris@chris-wilson.co.uk>
14357 Date: Wed Jan 18 20:09:26 2012 +0000
14359 sna/gen3: Set the batch mode for emitting video state
14361 The lack of kgem_set_mode() here is causing some recently added
14362 assertions to fail.
14364 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14366 commit 76203b70708b8750f52348ea8b9135191e650ff7
14367 Author: Chris Wilson <chris@chris-wilson.co.uk>
14368 Date: Wed Jan 18 18:36:43 2012 +0000
14370 sna: Almagamate writes based on the total number of bytes written
14372 Cachelines will only be dirtied for the bytes accessed so a better
14373 metric would based on the total number of pages brought into the TLB
14374 and the total number of cachelines used. Base the decision on whether
14375 to try and amalgamate the upload with others on the number of bytes
14376 copied rather than the overall extents.
14378 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14380 commit 470741e84c813a257897ee69f14500207a2674dc
14381 Author: Chris Wilson <chris@chris-wilson.co.uk>
14382 Date: Wed Jan 18 18:35:59 2012 +0000
14386 All of the asserts and debug options that lead me to believe that the
14387 tiling was completely screwy for some writes.
14389 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14391 commit ab387a89cf64ace9f7d7fa52b7743e78b625f47d
14392 Author: Chris Wilson <chris@chris-wilson.co.uk>
14393 Date: Wed Jan 18 18:32:51 2012 +0000
14395 sna: Update bo->tiling during search_linear_cache
14397 search_linear_cache() was updated to track the first good match whilst it
14398 continued to search for a better match. This resulted in the first good
14399 bo being modified and a record of those modifications lost, in
14400 particular the change in tiling.
14402 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14404 commit 4b893ab0812452238305b29b1beea224219f641d
14405 Author: Chris Wilson <chris@chris-wilson.co.uk>
14406 Date: Wed Jan 18 16:34:17 2012 +0000
14408 sna: Remove defunct debugging option
14410 FORCE_GPU_ONLY now has no effect except for marking the initial pixmap
14411 as all-damaged on the GPU, and so not testing the paths for which it was
14412 originally introduction.
14414 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14416 commit 965586544a8559b6b6b98c20e14c1266e46bf20d
14417 Author: Chris Wilson <chris@chris-wilson.co.uk>
14418 Date: Wed Jan 18 16:11:50 2012 +0000
14420 sna/gen6: Don't assume that a batch mode implies a non-empty batch
14422 Just in case we set a mode then fail to emit any dwords. Sounds
14423 inefficient and woe betide the culprit when I find it...
14425 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14427 commit d2e05750360fb19575f3becd87816184615855bf
14428 Author: Chris Wilson <chris@chris-wilson.co.uk>
14429 Date: Wed Jan 18 16:08:51 2012 +0000
14431 sna: Fix some tracking of to-be-flushed dri pixmaps
14433 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14435 commit 1ad5320fd40eaaf0bffcd209bf1eabd9a20e4bf8
14436 Author: Chris Wilson <chris@chris-wilson.co.uk>
14437 Date: Wed Jan 18 15:39:35 2012 +0000
14439 sna: Add valgrind markup for tracking CPU mmaps
14441 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14443 commit f3da610eadd1f560812603ab1a72c2e173679689
14444 Author: Chris Wilson <chris@chris-wilson.co.uk>
14445 Date: Wed Jan 18 12:59:48 2012 +0000
14447 sna: Prevent switching rings with render disabled
14449 We fudge forced used of the BLT ring unless we install a render backend
14450 and so we must also prevent the ring from being reset when the GPU is
14451 idle. Therefore we make handing the ring status a backend function.
14453 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14455 commit 6d31cb2d943be456f4ebaa663275c6a939013d9e
14456 Author: Chris Wilson <chris@chris-wilson.co.uk>
14457 Date: Wed Jan 18 11:25:30 2012 +0000
14459 sna: Restore use of shadow pixmaps by default without RENDER support
14461 If we do not have access to an accelerated render backend, only create
14462 GPU buffers for the scanout and use an accelerated blitter for
14463 upload/download and operating inplace on the scanout.
14465 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14467 commit 15a150579c3907054e186c68382873efdbadca75
14468 Author: Chris Wilson <chris@chris-wilson.co.uk>
14469 Date: Wed Jan 18 10:26:33 2012 +0000
14471 intel: Trivially remove a piece of XAA dependency for shadow
14473 The wolves are gathering at the door baying for the removal of XAA from
14476 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14478 commit 850495f956c811b1eb617d2e704e6bb7b5a86369
14479 Author: Chris Wilson <chris@chris-wilson.co.uk>
14480 Date: Tue Jan 17 22:51:29 2012 +0000
14482 sna: Fix increment of damage boxes after updating for rectangles
14485 ==13639== Conditional jump or move depends on uninitialised value(s)
14486 ==13639== at 0x5520B1E: pixman_region_init_rects (in
14487 /usr/lib/x86_64-linux-gnu/libpixman-1.so.0.24.0)
14488 ==13639== by 0x89E6ED7: __sna_damage_reduce (sna_damage.c:489)
14489 ==13639== by 0x89E7FEC: _sna_damage_contains_box (sna_damage.c:1161)
14490 ==13639== by 0x89CFCD9: sna_drawable_use_gpu_bo (sna_damage.h:175)
14491 ==13639== by 0x89D52DA: sna_poly_segment (sna_accel.c:6130)
14492 ==13639== by 0x21F87E: damagePolySegment (damage.c:1096)
14493 ==13639== by 0x1565A2: ProcPolySegment (dispatch.c:1771)
14494 ==13639== by 0x159FB0: Dispatch (dispatch.c:437)
14495 ==13639== by 0x1491D9: main (main.c:287)
14496 ==13639== Uninitialised value was created by a heap allocation
14497 ==13639== at 0x4028693: malloc (in
14498 /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
14499 ==13639== by 0x89E6BFB: _sna_damage_create_boxes (sna_damage.c:205)
14500 ==13639== by 0x89E78F0: _sna_damage_add_rectangles (sna_damage.c:327)
14501 ==13639== by 0x89CD32D: sna_poly_fill_rect_blt.isra.65
14503 ==13639== by 0x89DE23F: sna_poly_fill_rect (sna_accel.c:8366)
14504 ==13639== by 0x21E9C8: damagePolyFillRect (damage.c:1309)
14505 ==13639== by 0x26DD3F: miPaintWindow (miexpose.c:674)
14506 ==13639== by 0x18370A: ChangeWindowAttributes (window.c:1553)
14507 ==13639== by 0x154500: ProcChangeWindowAttributes (dispatch.c:696)
14508 ==13639== by 0x159FB0: Dispatch (dispatch.c:437)
14509 ==13639== by 0x1491D9: main (main.c:287)
14512 Use 'count' everywhere for consistency.
14514 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14516 commit 4b5c9affd4968fd5b3d85551cf5a0b4d600a7892
14517 Author: Chris Wilson <chris@chris-wilson.co.uk>
14518 Date: Tue Jan 17 20:23:26 2012 +0000
14520 sna: Restore orginal shadow pointer before uploading CPU damage
14522 Detected by valgrind:
14523 ==22012== Source and destination overlap in memcpy(0xd101000, 0xd101000,
14525 ==22012== at 0x402A180: memcpy (in
14526 /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
14527 ==22012== by 0x89BD4ED: memcpy_blt (blt.c:209)
14528 ==22012== by 0x89F2921: sna_write_boxes (sna_io.c:364)
14529 ==22012== by 0x89CFABF: sna_pixmap_move_to_gpu (sna_accel.c:1900)
14530 ==22012== by 0x89F49B0: sna_render_pixmap_bo (sna_render.c:571)
14531 ==22012== by 0x8A268CE: gen5_composite_picture (gen5_render.c:1908)
14532 ==22012== by 0x8A29B8A: gen5_render_composite (gen5_render.c:2252)
14533 ==22012== by 0x89E6762: sna_composite (sna_composite.c:485)
14534 ==22012== by 0x21D3C3: damageComposite (damage.c:569)
14535 ==22012== by 0x215963: ProcRenderComposite (render.c:728)
14536 ==22012== by 0x159FB0: Dispatch (dispatch.c:437)
14537 ==22012== by 0x1491D9: main (main.c:287)
14540 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14542 commit bbd6c8123635899e89911104bf84e1b7d11d66a1
14543 Author: Eugeni Dodonov <eugeni.dodonov@intel.com>
14544 Date: Tue Jan 17 16:16:37 2012 -0200
14546 sna: check for LLC support
14548 Instead of checking for CPU generation, use the libdrm-provided
14549 I915_PARAM_HAS_LLC instead.
14551 v2: use a define check to verify if we have I915_PARAM_HAS_LLC.
14553 Signed-off-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
14554 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14556 commit e4efde920bcab980451e79df7d645d1814e5f78d
14557 Author: Chris Wilson <chris@chris-wilson.co.uk>
14558 Date: Tue Jan 17 13:44:12 2012 +0000
14560 sna: Track whether damage is a complete representation of the dirt
14562 The previous commit undoes a premature optimisation that assumed that
14563 the current damage captured all pixels written. However, it happens to
14564 be a useful optimisation along that path (tracking upload of partial
14565 images), so add the necessary booking that watches for when the union
14566 of cpu and gpu damage is no longer the complete set of all pixels
14567 written, that is if we either migrate from one pixmap to the other, the
14568 undamaged region goes untracked. We also take advantage of whenever we
14569 damage the whole pixel to restore knowledge that our tracking of all
14570 pixels written is complete.
14572 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14574 commit a9b705f9a7bbd37a0543ce5fbd4c60c912d1017d
14575 Author: Chris Wilson <chris@chris-wilson.co.uk>
14576 Date: Tue Jan 17 12:58:43 2012 +0000
14578 sna: Mark GPU as all-damaged discarding the CPU bo to prevent stalls
14580 If we discard the CPU bo, we lose knowledge of whatever regions had been
14581 initialised but no longer dirty on the GPU, but instead must assume that
14582 the entirety of the GPU bo is dirty.
14584 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14586 commit 9d631e26d758eb4c2032a502604f07c5403b25ff
14587 Author: Chris Wilson <chris@chris-wilson.co.uk>
14588 Date: Tue Jan 17 12:58:03 2012 +0000
14590 sna: Mark the freshly allocated CPU bo as in the CPU domain
14592 As we immediately use it after creation, we need to inform GEM of the
14595 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14597 commit dfbf02b8778719b86e0f5d78ee833a19436d43a0
14598 Author: Chris Wilson <chris@chris-wilson.co.uk>
14599 Date: Tue Jan 17 10:55:29 2012 +0000
14601 sna: Add some DBG breadcrumbs to put_image upload paths
14603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14605 commit da90afc32f19e9168cbaa4e0cdcd941d12f2c142
14606 Author: Chris Wilson <chris@chris-wilson.co.uk>
14607 Date: Tue Jan 17 10:24:06 2012 +0000
14609 sna: Add DBG breadcrumbs to gradient initialisation
14611 Put some markers into the debug log as those functions create many
14612 proxies causing a lot of debug noise.
14614 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14616 commit d14341cb22e37f52070cd92b707fec5e08038e96
14617 Author: Chris Wilson <chris@chris-wilson.co.uk>
14618 Date: Tue Jan 17 00:31:43 2012 +0000
14620 sna: Add a render ring detiling read path
14622 For SNB, in case you really, really want to use GPU detiling and not
14623 incur the ring switch. Tweaking when to just mmap the target seems to
14624 gain most anyway...
14626 The ulterior motive is that this provides fallback paths for avoiding
14627 the use of TILING_Y with GTT mmaps which is broken on 855gm.
14629 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14631 commit 3620f9ca459f7bbbb01786c95b5583023b6d870c
14632 Author: Chris Wilson <chris@chris-wilson.co.uk>
14633 Date: Tue Jan 17 00:23:04 2012 +0000
14635 sna: Cap pwrite buffer alignment to 64k
14637 We only want to create huge pwrite buffers when populating the inactive
14638 cache for mmapped uploads. In the absence of using mmap for upload, be
14639 more conservative with the alignment value so as not to simply waste
14640 valuable aperture and memory.
14642 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14644 commit b9f59b1099eeba3fb8b25693d4f37dd13267d8a3
14645 Author: Chris Wilson <chris@chris-wilson.co.uk>
14646 Date: Tue Jan 17 00:20:51 2012 +0000
14648 sna: correct adjust of a stolen 2d read buffer
14650 If we steal a write buffer for creating a pixmap for read back, then we
14651 need to be careful as we will have set the used amount to 0 and then try
14652 to incorrectly decrease by the last row. Fortunately, we do not yet have
14653 any code that attempts to create a 2d buffer for reading.
14655 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14657 commit 6fc4cdafeba4517b94202ff292f05fdd90e52b2e
14658 Author: Chris Wilson <chris@chris-wilson.co.uk>
14659 Date: Mon Jan 16 21:34:43 2012 +0000
14661 sna: Correct assertion for a partial read buffer
14663 The batch may legitimately be submitted prior to the attachment of the
14664 read buffer, if, for example, we need to switch rings. Therefore update
14665 the assertion to only check that the bo remains in existence via either
14666 a reference from the exec or from the user
14668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14670 commit 377f5e16cd14bdf42e67b0c8f2e13b75526f3a85
14671 Author: Chris Wilson <chris@chris-wilson.co.uk>
14672 Date: Mon Jan 16 15:51:33 2012 +0000
14674 sna/gen[45]: clear the state tracker before setting the formats
14676 When backporting the patches from gen6, I didn't notice the memset that
14677 came later, and this wasn't along the paths checked by rendercheck.
14679 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14681 commit 6387f2fb8a40f32cc4a3e4228983dcf0604ac4dc
14682 Author: Chris Wilson <chris@chris-wilson.co.uk>
14683 Date: Mon Jan 16 14:12:36 2012 +0000
14685 sna/gen[4567]: x1r5g5b5 is only a render target, not sampler
14687 Whilst we can render to and blend with an depth 15 target, we cannot use
14688 it as a texture with the sampling engine.
14690 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14692 commit 8b2bb666662305ab88aad8198ad69b1c98407d75
14693 Author: Chris Wilson <chris@chris-wilson.co.uk>
14694 Date: Mon Jan 16 13:35:05 2012 +0000
14696 sna/gen6: Restore the non-pipelined op after every WM binding table update
14698 The hw wants it as demonstrated by the '>' in KDE's menus. Why is it
14699 always KDE that demonstrates coherency problems...
14701 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14703 commit a11b22d172898ebfc657339a9ca56102ae92592f
14704 Author: Chris Wilson <chris@chris-wilson.co.uk>
14705 Date: Mon Jan 16 13:06:58 2012 +0000
14707 sna/gen[23]: Remark the destination bo as dirty after flushing
14709 One of the side-effects of emitting the composite state is that it
14710 tags the destination surface as dirty as a result of the *forthcoming*
14711 operation. So emitting the flush after emitting the composite state
14712 clears that tag, so we need to restore it for future coherency.
14714 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14716 commit 2f09363a6e2c77fb23bf0e4c12c4507521fa631b
14717 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
14718 Date: Thu Jan 12 15:30:15 2012 +0800
14720 uxa/glamor: Create glamor pixmap by default.
14722 When creating native glamor pixmaps we will get much better performance
14723 than using the textured-drm pixmap, this commit is to make that the
14724 default behaviour when configured to use glamor. Another advantage
14725 of this commit is that we reduce the risk of encountering the
14726 "incompatible region exists for this name" and the associated
14727 render corruption. And since we now never intentionally allocate
14728 a reusable pixmap we could just make all (intel_glamor) allocations
14729 non-reusable without incurring too great an overhead.
14731 A side effect is that those glamor pixmaps do not have a
14732 valid BO attached to them and thus it fails to get a DRI drawable. This
14733 commit also fixes that problem by adjusting the fixup_shadow mechanism
14734 to recreate a textured-drm pixmap from the native glamor pixmap. I tested
14735 this with mutter, and it works fine.
14737 The performance gain to apply this patch is about 10% to 20% with
14738 different workload.
14740 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
14741 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
14743 commit fd4c139a3959df90c7f078817fc6e2f3db715bf8
14744 Author: Chris Wilson <chris@chris-wilson.co.uk>
14745 Date: Mon Jan 16 00:32:12 2012 +0000
14747 sna: On LLC systems quietly replace all linear mmappings using the CPU
14749 If the GPU and CPU caches are shared and coherent, we can use a cached
14750 mapping for linear bo in the CPU domain with no penalty and so avoid the
14751 penalty of using WC/UC mappings through the GTT (and any aperture
14752 pressure). We presume that the bo for such mappings are indeed LLC
14755 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14757 commit c20a729d0a57fd5a782d9114535d9474f39a8950
14758 Author: Chris Wilson <chris@chris-wilson.co.uk>
14759 Date: Sun Jan 15 22:19:22 2012 +0000
14761 sna/gen6: Force a batch submission after allocation failure during composite
14763 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14765 commit 380a2fca3cce4c99c5026ab800f7885a1959b16d
14766 Author: Chris Wilson <chris@chris-wilson.co.uk>
14767 Date: Sun Jan 15 22:13:20 2012 +0000
14769 sna: Optimise call to composite with single box
14771 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14773 commit 9f89250de1dc134f54cb68012fa5eb996c61d57e
14774 Author: Chris Wilson <chris@chris-wilson.co.uk>
14775 Date: Sun Jan 15 21:54:13 2012 +0000
14777 sna: Use the prefer-GPU hint for forcing allocation for core drawing
14779 Similar to the render paths and simpler than the current look up tiling
14782 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14784 commit 8652bf7a196f53842db70f5c70aded31b470b0ab
14785 Author: Chris Wilson <chris@chris-wilson.co.uk>
14786 Date: Sun Jan 15 19:04:01 2012 +0000
14788 sna: Don't track an unmatching tiled bo when searching the linear cache
14790 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14792 commit cc4b616990fa71e96d38512d8cdb4c3e2abb21c0
14793 Author: Chris Wilson <chris@chris-wilson.co.uk>
14794 Date: Sun Jan 15 19:55:50 2012 +0000
14796 sna/video: Increase the level of paranoia
14798 In how many different ways can we check that the scanout is allocated
14799 before we start decoding video?
14801 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14803 commit 7f480ba02c66fcc02bad483731c4c0cf6c746c0d
14804 Author: Chris Wilson <chris@chris-wilson.co.uk>
14805 Date: Sun Jan 15 18:45:06 2012 +0000
14807 sna: Tidy search through active bo cache
14809 Perform the assertions upon cache consistency upfront, and tidy the
14812 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14814 commit 6f7bc35d7f956f7c2507eabc874ead1a83c85ddb
14815 Author: Chris Wilson <chris@chris-wilson.co.uk>
14816 Date: Sun Jan 15 18:14:24 2012 +0000
14818 sna: Use indirect uploads rather than teardown existing CPU maps
14820 Allow the snoopable CPU mapping to be used inplace of the GTT map for
14823 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14825 commit 475fa67ed320f94df37ec86fe6c5dda886923751
14826 Author: Chris Wilson <chris@chris-wilson.co.uk>
14827 Date: Sun Jan 15 17:30:00 2012 +0000
14829 sna: Fast path move-area-to-cpu when the pixmap is already on the cpu
14831 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14833 commit 37ced44a53008debaf869ec9ef4ba2e5d6982e76
14834 Author: Chris Wilson <chris@chris-wilson.co.uk>
14835 Date: Sun Jan 15 15:35:57 2012 +0000
14837 sna: Be a little more lenient wrt damage migration if we have CPU bo
14839 The idea being that they facilitate copying to and from the CPU, but
14840 also we want to avoid stalling on any pixels help by the CPU bo.
14842 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14844 commit e3732a6f7f61a959521be9a668bba045591e633c
14845 Author: Chris Wilson <chris@chris-wilson.co.uk>
14846 Date: Sun Jan 15 11:06:59 2012 +0000
14848 sna: Defer ring switching until after a period of idleness
14850 Similar to the desire to flush the next batch after an overflow, we do
14851 not want to incur any lag in the midst of drawing, even if that lag is
14852 mitigated by GPU semaphores.
14854 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14856 commit 5df7147b0934dadc63459c2c0f33e0663692425a
14857 Author: Chris Wilson <chris@chris-wilson.co.uk>
14858 Date: Sun Jan 15 10:28:00 2012 +0000
14860 sna: Restore the kgem_create_map() symbol
14862 As the stub is exported to the driver even in the absence of vmapping.
14864 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14866 commit be53740c6f0bb182e3d88f0597f127732c839316
14867 Author: Chris Wilson <chris@chris-wilson.co.uk>
14868 Date: Sun Jan 15 10:16:13 2012 +0000
14870 sna: Various DBG typos
14872 Fix some mispellings inside the DBG messages
14874 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14876 commit 349e9a7b94199e759acaaccac3abf5e28f3c246f
14877 Author: Chris Wilson <chris@chris-wilson.co.uk>
14878 Date: Sun Jan 15 10:04:11 2012 +0000
14880 sna: Prefer read-boxes inplace again
14882 Using the gpu to do the detiling just incurs extra latency and an extra
14883 copy, so go back to using a fence and GTT mapping for the common path.
14885 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14887 commit 09dc8b1b358aa33836d511b75f92e8d096bc7e59
14888 Author: Chris Wilson <chris@chris-wilson.co.uk>
14889 Date: Sat Jan 14 23:21:29 2012 +0000
14891 sna/gen7: Check reused source for validity
14893 Be sure the mask picture has a valid format even though it points to the
14894 same pixels as the valid source. And also be wary if the source was
14895 converted to a solid, but the mask is not.
14897 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14899 commit d9871f01d8c4584c1f39eb09b51d5461863934e9
14900 Author: Chris Wilson <chris@chris-wilson.co.uk>
14901 Date: Sat Jan 14 23:21:29 2012 +0000
14903 sna/gen6: Check reused source for validity
14905 Be sure the mask picture has a valid format even though it points to the
14906 same pixels as the valid source. And also be wary if the source was
14907 converted to a solid, but the mask is not.
14909 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14911 commit 1d6030322e2c3bae87a0173a32fb8d341dea560c
14912 Author: Chris Wilson <chris@chris-wilson.co.uk>
14913 Date: Sat Jan 14 23:21:29 2012 +0000
14915 sna/gen5: Check reused source for validity
14917 Be sure the mask picture has a valid format even though it points to the
14918 same pixels as the valid source. And also be wary if the source was
14919 converted to a solid, but the mask is not.
14921 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14923 commit 0e4a24ef6c186909c99a501cb606994b5c10a813
14924 Author: Chris Wilson <chris@chris-wilson.co.uk>
14925 Date: Sat Jan 14 23:21:29 2012 +0000
14927 sna/gen4: Check reused source for validity
14929 Be sure the mask picture has a valid format even though it points to the
14930 same pixels as the valid source. And also be wary if the source was
14931 converted to a solid, but the mask is not.
14933 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14935 commit ea299f2523d5adeca84274bb7003cde8ec743a44
14936 Author: Chris Wilson <chris@chris-wilson.co.uk>
14937 Date: Sat Jan 14 23:21:29 2012 +0000
14939 sna/gen3: Check reused source for validity
14941 Be sure the mask picture has a valid format even though it points to the
14942 same pixels as the valid source. And also be wary if the source was
14943 converted to a solid, but the mask is not.
14945 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14947 commit 007da2f97894814eaded4d24e0481f950ca7bd00
14948 Author: Chris Wilson <chris@chris-wilson.co.uk>
14949 Date: Sat Jan 14 23:21:29 2012 +0000
14951 sna/gen2: Check reused source for validity
14953 Be sure the mask picture has a valid format even though it points to the
14954 same pixels as the valid source. And also be wary if the source was
14955 converted to a solid, but the mask is not.
14957 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14959 commit 1d55b42fbd77e85ab8930b78bb5a9e3bccbca49e
14960 Author: Chris Wilson <chris@chris-wilson.co.uk>
14961 Date: Sat Jan 14 23:44:23 2012 +0000
14963 sna: Fix read back of partial mmapped buffers
14965 Do not move a GTT mapped buffer into the CPU domain, it causes untold
14966 pain for no benefit!
14968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14970 commit 046e945173397e5fee36a231303b68a6bc9d809d
14971 Author: Chris Wilson <chris@chris-wilson.co.uk>
14972 Date: Sat Jan 14 23:25:52 2012 +0000
14974 sna: Discard read buffers after use
14976 Rather than pollute the cache with bo that are known not to be in the
14977 GTT and are no longer useful, drop the bo after we read from it.
14979 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14981 commit 421ee0bb53321e9fb2f17c850c51b80cc385bf57
14982 Author: Chris Wilson <chris@chris-wilson.co.uk>
14983 Date: Sun Jan 15 00:46:47 2012 +0000
14985 sna: Do not assume that the mappable aperture size is a power of two
14987 And instead derive a power-of-two alignment value for partial buffer
14988 sizes from the mappable aperture size and use that during
14989 kgem_create_buffer()
14991 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
14992 References: https://bugs.freedesktop.org/show_bug.cgi?id=44682
14993 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14995 commit 5d5da35c9fab4e7154921861ecbb83befb8840e5
14996 Author: Chris Wilson <chris@chris-wilson.co.uk>
14997 Date: Sat Jan 14 18:28:24 2012 +0000
14999 sna/gen[23]: Check for room in the batch before emitting pipeline flushes
15001 Use a single idiom and reuse the check built into the state emission,
15002 for both spans/boxes paths.
15004 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15006 commit f7e4799687d5010ea0b56aeb6fbbb2e854b1b363
15007 Author: Chris Wilson <chris@chris-wilson.co.uk>
15008 Date: Sat Jan 14 16:53:38 2012 +0000
15010 sna/gen6: Allow greater use of BLT
15012 Typically we will be bound to the RENDER ring as once engaged we try not
15013 to switch. However, with semaphores enabled we may switch more freely
15014 and there it is advantageous to use as much of the faster BLT as is
15017 The most contentious point here is the choice of whether to use BLT for
15018 copies by default. microbenchmarks (compwinwin) benefit from the
15019 coallescing performed in the render batch, but the more complex traces
15020 seem to prefer utilizing the blitter. The debate will continue...
15022 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15024 commit c1ce34d4509a3f3e963d82ac0569a21706892f8e
15025 Author: Chris Wilson <chris@chris-wilson.co.uk>
15026 Date: Sat Jan 14 15:08:18 2012 +0000
15028 sna/gen6: Tidy markup for when using the BLT is truly preferrable
15030 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15032 commit b64751dbdb1b88b91ad97aaf995b4261876cf860
15033 Author: Chris Wilson <chris@chris-wilson.co.uk>
15034 Date: Sat Jan 14 13:50:50 2012 +0000
15036 sna: Be more lenient wrt switching rings if the kernel supports semaphores
15038 If the kernel uses GPU semaphores for its coherency mechanism between
15039 rings rather than CPU waits, allow the ring to be chosen on the basis
15040 of the subsequent operation following a submission of batch. (However,
15041 since batches are likely to be submitted in the middle of a draw, then
15042 the likelihood is for ddx to remain on one ring until forced to switch
15043 for an operation or idle, which is the same situation as before and so
15044 the difference is miniscule.)
15046 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15048 commit 295a22d2709b2442b5254968437f897dac22a0ec
15049 Author: Chris Wilson <chris@chris-wilson.co.uk>
15050 Date: Sat Jan 14 14:12:25 2012 +0000
15052 sna: Ensure that the batch mode is always declared before emitting dwords
15054 Initially, the batch->mode was only set upon an actual mode switch,
15055 batch submission would not reset the mode. However, to facilitate fast
15056 ring switching with semaphores, reseting the mode upon batch submission
15057 is desired which means that if we submit the batch in the middle of an
15058 operation we must redeclare its mode before continuing.
15060 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15062 commit 0d2a50772200d868d094f90dc1d30c1b4d7930ba
15063 Author: Chris Wilson <chris@chris-wilson.co.uk>
15064 Date: Sat Jan 14 11:03:55 2012 +0000
15066 sna/glyphs: Cache the glyph image on the fallback path as well
15068 The glyph cache grew to accommodate the fallback pixman image for mask
15069 generation, and is equally applicable along the full fallback path.
15071 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15073 commit f3e0ba4f65db2b85f89aa3868d153434bc41c811
15074 Author: Chris Wilson <chris@chris-wilson.co.uk>
15075 Date: Sat Jan 14 10:17:27 2012 +0000
15077 sna/gen5: Disable render glyphs_to_dst
15079 Processing more than a single rectangle using the CA path on ILK is
15080 extremely hit-or-miss, often resulting in the absence of the second
15081 primitive (ie. the glyphs are cleared but not added.) This is
15082 reminiscent of the complete breakage of the BRW shaders, none of which
15083 can handle more than a single rectangle.
15085 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15087 commit fb92818ba4dc81ce62d58a87b5af6cb1e3a96708
15088 Author: Chris Wilson <chris@chris-wilson.co.uk>
15089 Date: Sat Jan 14 01:48:11 2012 +0000
15091 sna: Pass render operation to flush and avoid the implicit flush-on-batch-end
15093 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15095 commit a62429a1f79b8fa4a5ddaf61b2bc80fc8dbe576c
15096 Author: Chris Wilson <chris@chris-wilson.co.uk>
15097 Date: Fri Jan 13 19:00:01 2012 +0000
15099 sna: Upload continuation vertices into mmapped buffers
15101 In the common case, we expect a very small number of vertices which will
15102 fit into the batch along with the commands. However, in full flow we
15103 overflow the on-stack buffer and likely several continuation buffers.
15104 Streaming those straight into the GTT seems like a good idea, with the
15105 usual caveats over aperture pressure. (Since these are linear we could
15106 use snoopable bo for the architectures that support such for vertex
15107 buffers and if we had kernel support.)
15109 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15111 commit 24df8ab9742f771cfeb6d30bd8a61a17a9e22ca7
15112 Author: Chris Wilson <chris@chris-wilson.co.uk>
15113 Date: Fri Jan 13 16:15:21 2012 +0000
15115 sna: Reverse the chronological sort order of inactive vma entries
15117 The goal is to reuse the most recently bound GTT mapping in the hope
15118 that is still mappable at the time of reuse.
15120 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15122 commit 2f26bbe3dd55bfe26b6f93fc1f9e9813fc11f7c4
15123 Author: Chris Wilson <chris@chris-wilson.co.uk>
15124 Date: Fri Jan 13 15:59:51 2012 +0000
15126 sna: Remove the short-circuiting of all-damage in move_to_cpu
15128 To allow a replacement of the complete pixmap to be performed in place.
15130 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15132 commit c81dba18e6a308d716a11b90da6e53ca7f5f8a02
15133 Author: Chris Wilson <chris@chris-wilson.co.uk>
15134 Date: Fri Jan 13 14:21:13 2012 +0000
15136 sna: Hint whether we prefer to use the GPU for a pixmap
15138 This includes the condition where the pixmap is too large, as well as
15139 being too small, to be allocatable on the GPU. It is only a hint set
15140 during creation, and may be overridden if required.
15142 This fixes the regression in ocitysmap which decided to render glyphs
15143 into a GPU mask for a destination that does not fit into the aperture.
15145 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15147 commit 2bd942d55314426d7f429d7ccc0b62a622a79009
15148 Author: Chris Wilson <chris@chris-wilson.co.uk>
15149 Date: Fri Jan 13 14:08:34 2012 +0000
15151 sna/trapezoids: Quieten the debugging of the gory details of the rasteriser
15153 Hide the noise under another level of debugging so that hopefully the
15154 reason why it chose a particular path become clear.
15156 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15158 commit 5dbcfc2ee3af64846298dbcb20db27c93b3d57f2
15159 Author: Chris Wilson <chris@chris-wilson.co.uk>
15160 Date: Fri Jan 13 13:52:10 2012 +0000
15162 sna: Be more lenient in not forcing to the GPU if the sources have CPU damage
15164 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15166 commit 20ff4a1d73cc10e1f53050b19b8799ccbb5c7d04
15167 Author: Chris Wilson <chris@chris-wilson.co.uk>
15168 Date: Fri Jan 13 13:30:13 2012 +0000
15170 sna: Use top_srcdir to detect .git rather than top_builddir
15172 For srcdir != builddir builds, we need to be searching the source tree
15175 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15177 commit a4d5d725996b94e314ae7697c7a597ed2f60e8cd
15178 Author: Chris Wilson <chris@chris-wilson.co.uk>
15179 Date: Fri Jan 13 10:41:44 2012 +0000
15181 sna: Experiment with GTT mmapped upload buffers
15183 In a few places, we can stream the source into the GTT and so upload in
15184 place through the WC mapping. Notably, in many other places we want to
15185 rasterise on a partial in cacheable memory. So we need to notify the
15186 backend of the intended usage for the buffer and when we think it is
15187 appropriate we can allocate a GTT mapped pointer for zero-copy upload.
15189 The biggest improvement tends to be in the PutComposite style of
15190 microbenchmark, yet throughput for trapezoid masks seems to suffer (e.g.
15191 swfdec-giant-steps on i3 and gen2 in general). As expected, the culprit
15192 of the regression is the aperture pressure causing eviction stalls, which
15193 the pwrite paths sidesteps by doing a cached copy when there is no GTT
15194 space. This could be alleviated with an is-mappable ioctl predicting when
15195 use of the buffer would block and so falling back in those cases to
15196 pwrite. However, I suspect that this will improve dispatch latency in
15197 the common idle case for which I have no good metric.
15199 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15201 commit 252f3818250ecf4776f20afa8111d7f1f6f29c18
15202 Author: Chris Wilson <chris@chris-wilson.co.uk>
15203 Date: Fri Jan 13 11:30:32 2012 +0000
15205 sna: Relinquish the GTT mmap on inactive buffers if moved out of the aperture
15207 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15209 commit 9c73dd91e90810116c215060bac434c5e8174f5f
15210 Author: Chris Wilson <chris@chris-wilson.co.uk>
15211 Date: Sat Jan 14 17:00:01 2012 +0000
15213 Include <xorgVersion.h> to repair build
15215 intel_module.c:41:48: error: missing binary operator before token "("
15217 commit b213f6e876f29cbcaccffeeb733726be4c91529e
15218 Author: Stefan Dirsch <sndirsch@suse.de>
15219 Date: Sat Jan 14 05:43:33 2012 +0100
15221 Make driver backwards compatible for server 1.6.x.
15223 Signed-off-by: Stefan Dirsch <sndirsch@suse.de>
15225 commit 94217a4dd908f1368dfdef90797ce74a081663fb
15226 Author: Chris Wilson <chris@chris-wilson.co.uk>
15227 Date: Fri Jan 13 10:03:28 2012 +0000
15229 sna: Decouple dirty pixmaps from list if we fail to upload them
15231 Rather than iterate endlessly trying to upload the same pixmap when
15232 failing to flush dirty CPU damage, try again on the next flush.
15234 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15236 commit 65ef369c733ab45945a7d5fe4b76fe5c7167b51b
15237 Author: Chris Wilson <chris@chris-wilson.co.uk>
15238 Date: Fri Jan 13 09:58:40 2012 +0000
15240 sna: Decouple from CPU dirty list after removing all CPU damage
15242 In the paths where we discard CPU damage, we also need to remove it
15243 from the dirty list so that we do not iterate over it during flush.
15245 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15247 commit 0845fcef9e664eb100b55ad1cc20743844deee58
15248 Author: Chris Wilson <chris@chris-wilson.co.uk>
15249 Date: Thu Jan 12 23:45:03 2012 +0000
15251 sna: Correct iteration counter for stippled blits
15253 ==7215== Invalid read of size 2
15254 ==7215== at 0x51A72F3: sna_poly_fill_rect_stippled_8x8_blt
15256 ==7215== by 0x51A9CDF: sna_poly_fill_rect_stippled_blt
15258 ==7215== by 0x51A3878: sna_poly_segment (sna_accel.c:6090)
15259 ==7215== by 0x216C02: damagePolySegment (damage.c:1096)
15260 ==7215== by 0x13F6E8: ProcPolySegment (dispatch.c:1771)
15261 ==7215== by 0x1436B4: Dispatch (dispatch.c:437)
15262 ==7215== by 0x131279: main (main.c:287)
15263 ==7215== Address 0x6f851e8 is 0 bytes after a block of size 32 alloc'd
15264 ==7215== at 0x4825DEC: malloc (vg_replace_malloc.c:261)
15265 ==7215== by 0x51A3558: sna_poly_segment (sna_accel.c:6049)
15266 ==7215== by 0x216C02: damagePolySegment (damage.c:1096)
15267 ==7215== by 0x13F6E8: ProcPolySegment (dispatch.c:1771)
15268 ==7215== by 0x1436B4: Dispatch (dispatch.c:437)
15269 ==7215== by 0x131279: main (main.c:287)
15271 An example being the stippled outline in gimp, the yellow marching ants,
15272 would randomly walk over the entire image.
15274 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15276 commit 5c2c6474efd4dbc5a0fc9c68ef4e5b5e5cfee415
15277 Author: Chris Wilson <chris@chris-wilson.co.uk>
15278 Date: Thu Jan 12 23:11:05 2012 +0000
15280 sna/dri: Hook up a compile option to switch colour buffers to Y-tiling
15282 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15284 commit 59b79e5952fdc40f8c50f74a72a19363cb426140
15285 Author: Chris Wilson <chris@chris-wilson.co.uk>
15286 Date: Thu Jan 12 21:05:39 2012 +0000
15288 sna: Reorder composite-done to destroy mask bo before source bo
15290 Just in the unlikely event that we hit the delete-partial-upload path
15291 which prefers destroying the last bo first.
15293 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15295 commit 983b755313df8a0d256c59c32ec4106e35f237aa
15296 Author: Chris Wilson <chris@chris-wilson.co.uk>
15297 Date: Thu Jan 12 21:05:16 2012 +0000
15299 sna/damage: Fix union of extents with dirty damage but no region
15301 By failing to account for certain paths which would create a damage elt
15302 without fully initialisating the damage region (only the damage extents),
15303 we would later overwrite the damage extents with only the extents for
15304 this operation (rather than the union of this operation with the current
15305 damage). This fixes a regression from 098592ca5d,
15306 (sna: Remove the independent tracking of elts from boxes).
15308 Include the associated damage migration debugging code of the callers.
15310 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15312 commit 8d2f1eefe142b65db7d8821ba0f80fdb0902b2d5
15313 Author: Chris Wilson <chris@chris-wilson.co.uk>
15314 Date: Thu Jan 12 21:08:41 2012 +0000
15316 sna: Pass a hint that we may like to perform the fallback in place
15318 If we do not read back from the destination, we may prefer to utilize a
15319 GTT mapping and perform the fallback inplace. For the rare event that we
15320 wish to fallback and do not already have a shadow...
15322 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15324 commit 48ab72754d0069a3247c5fee8c353a6b593eaed9
15325 Author: Chris Wilson <chris@chris-wilson.co.uk>
15326 Date: Thu Jan 12 21:07:15 2012 +0000
15328 sna: Use the GPU bo if it is all damaged
15330 By marking the scratch upload pixmap as damaged in both domains, we
15331 confused the texture upload path and made it upload the pixmap a second
15332 time. If either bo is all-damaged, use it!
15334 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15336 commit 20a4d7181983c7fed289844be49ccaf3f56965c6
15337 Author: Chris Wilson <chris@chris-wilson.co.uk>
15338 Date: Thu Jan 12 21:05:51 2012 +0000
15340 sna: Dump batch contents for debugging before modification
15342 We need to dump the batch contents before the maps are made by the
15343 construction of the batch itself.
15345 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15347 commit 7932a2a259069bb7d19af8566f7b7704f6c2eade
15348 Author: Chris Wilson <chris@chris-wilson.co.uk>
15349 Date: Thu Jan 12 14:06:22 2012 +0000
15351 sna: Update for removal of backwards compatible miWideDash
15353 miWideDash() no longer calls miZeroLineDash() when called with
15354 gc->lineWidth==0, we need to do so ourselves.
15356 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15358 commit b7cefddd467bfd0c100219969712a0133475004c
15359 Author: Chris Wilson <chris@chris-wilson.co.uk>
15360 Date: Thu Jan 12 13:17:43 2012 +0000
15362 sna: Re-enable min-alignment workaround on pre-SNB hw
15364 Confirmed as still being required for both gen3 and gen4. One day I will
15365 get single-stream mode working, just not today apparently.
15367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15369 commit 978e1aeceab3c1a524f7d7a070fe04f37530c8d9
15370 Author: Chris Wilson <chris@chris-wilson.co.uk>
15371 Date: Thu Jan 12 11:38:56 2012 +0000
15373 sna: Only shrink a partial buffer if it is no longer used.
15375 The condition on being able to shrink a buffer is more severe than just
15376 whether we are reading from the buffer, but also we cannot swap the
15377 handles if the existing handle remains exposed via a proxy.
15379 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15381 commit d3169154d18600e0d41db5f833fad52970e17b55
15382 Author: Chris Wilson <chris@chris-wilson.co.uk>
15383 Date: Thu Jan 12 11:36:05 2012 +0000
15385 sna: Improve a DBG message
15387 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15389 commit 2a22990968aebd53a7bfeaf19e58ee609b980e5f
15390 Author: Chris Wilson <chris@chris-wilson.co.uk>
15391 Date: Thu Jan 12 11:22:06 2012 +0000
15393 sna: Prevent 60Hz wakeups if the client stops in mid-render
15395 Only continue to wake up if the scanout remains active.
15397 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15399 commit 1c0e9916caef6594511009c8cb79071ac0d82090
15400 Author: Chris Wilson <chris@chris-wilson.co.uk>
15401 Date: Thu Jan 12 11:12:56 2012 +0000
15403 sna: Align the partial buffer contents to cachelines
15405 To enable Daniel's faster pwrite paths. Only one step removed from using
15406 whole page alignment...
15408 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15410 commit 1e4080318f58fd51f37792eefd367e2e4c813ea3
15411 Author: Chris Wilson <chris@chris-wilson.co.uk>
15412 Date: Thu Jan 12 10:47:19 2012 +0000
15414 sna: Replace the open-coded bubble sort of the partial list
15416 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15418 commit 7290ced5791f9860b00901fa9a4545ab5a067fae
15419 Author: Chris Wilson <chris@chris-wilson.co.uk>
15420 Date: Thu Jan 12 02:59:43 2012 +0000
15422 sna/video: Fix for changes in damage api
15424 We can avoid both calls to modify the damage with one simple check.
15426 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15428 commit 87e6dcb3b06c274ecba5de56d4e86fec73e46dbb
15429 Author: Chris Wilson <chris@chris-wilson.co.uk>
15430 Date: Thu Jan 12 02:09:22 2012 +0000
15432 sna: Don't call RegionIntersect for the trivial PutImage
15434 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15436 commit 1bd6665093574372248a0743e5630e68aac653ed
15437 Author: Chris Wilson <chris@chris-wilson.co.uk>
15438 Date: Thu Jan 12 01:37:08 2012 +0000
15440 sna: Disable the min alignment workaround
15442 Allow all generations to use the minimum alignment of 4 bytes again as
15443 it appears to be working for me... Or at least what remains broken seems
15444 to be irrespective of this alignment.
15446 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15448 commit 112b8959261712aaa82f92af0aca4b97fa7c7f03
15449 Author: Chris Wilson <chris@chris-wilson.co.uk>
15450 Date: Thu Jan 12 01:38:18 2012 +0000
15452 sna: Prevent shrinking a partial buffer stolen for a read
15454 If we reuse a partial buffer for a read, we cannot shrink it during
15455 upload to the device as we do not track how many bytes we actually need
15456 for the read operation.
15458 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15460 commit b09ae4c20313fea7af11f77cd673039635d4dcc7
15461 Author: Chris Wilson <chris@chris-wilson.co.uk>
15462 Date: Thu Jan 12 00:58:08 2012 +0000
15464 sna: Don't drop expired partial bo immediately, wait until dispatch
15466 As the partial bo may be coupled into the execlist, we may as well hang
15467 onto the memory to service the next partial buffer request until it
15468 expires in the next dispatch.
15470 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15472 commit a3c42565a8f557b2e7f7ff7bfa45b13b606f2968
15473 Author: Chris Wilson <chris@chris-wilson.co.uk>
15474 Date: Wed Jan 11 16:39:45 2012 +0000
15476 sna: Store damage-all in the low bit of the damage pointer
15478 Avoid the function call overhead by inspecting the low bit to see if it
15479 is all-damaged already.
15481 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15483 commit c64a9d0683e047a7eb041df78db746f6dd387b5e
15484 Author: Chris Wilson <chris@chris-wilson.co.uk>
15485 Date: Wed Jan 11 19:51:42 2012 +0000
15487 sna: Choose a stride for the indirect replacement
15489 Don't blithely assume that the incoming bytes are appropriately aligned
15490 for the destination buffer. Indeed we may be replacing the destination
15491 bo with the shadow bytes out of another,larger, pixmap, in which case we
15492 do need to create a stride that is appropriate for the upload an
15493 perform the 2D copy.
15495 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15497 commit b82851e74d5010ee08938ee42fa44c29fed633b1
15498 Author: Chris Wilson <chris@chris-wilson.co.uk>
15499 Date: Wed Jan 11 15:42:00 2012 +0000
15501 sna: Mark upload pixmaps as being wholly GPU damaged
15503 So that subsequent code resists performing CPU operations with them
15504 (after they have been populated.)
15506 Marking both sides as wholly damaged breaks the rules, but should work
15507 out so long as we check whether we can perform the operation within the
15508 target damage first.
15510 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15512 commit 2a5ab05f1690484c230e8f876a3f7aefb371af71
15513 Author: Chris Wilson <chris@chris-wilson.co.uk>
15514 Date: Wed Jan 11 15:24:28 2012 +0000
15516 sna: Use a minimum alignment of 64
15518 We should be able to reduce this by disabling dual-stream mode of the
15519 GPU (which we want to achieve any way for 2D performance). Artefacts
15520 in small uploads demonstrate that we fail to do.
15522 References: https://bugs.freedesktop.org/show_bug.cgi?id=44150
15523 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15525 commit e94807759eb6cfb10bd2d372fa71cc64a730bc7c
15526 Author: Chris Wilson <chris@chris-wilson.co.uk>
15527 Date: Wed Jan 11 12:13:18 2012 +0000
15529 sna/gen6: Special case spans with no transform
15531 As the no transform is a special case of affine, we were attempting to
15532 deference the NULL transform in order to determine if it was a simple
15533 no-rotation matrix. As the operation is extremely simple, add a special
15534 case vertex program to speed it up.
15536 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15538 commit 0a5313900ec9a7c499eb5051f3a5f078a9b0bbde
15539 Author: Chris Wilson <chris@chris-wilson.co.uk>
15540 Date: Wed Jan 11 12:09:13 2012 +0000
15542 sna: Explicitly retire the bo following a serialisation point
15544 This is to keep the sanity checks in order, but conceptually should be
15547 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15549 commit 2add5991a7d186d8271cb79ae4f62dad7b4ca243
15550 Author: Chris Wilson <chris@chris-wilson.co.uk>
15551 Date: Wed Jan 11 11:33:19 2012 +0000
15553 sna: Mark the bo as no longer in the GPU domain after clearing needs_flush
15555 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15557 commit fec70985719eec75b7fa799bddc3115fb329ab86
15558 Author: Chris Wilson <chris@chris-wilson.co.uk>
15559 Date: Wed Jan 11 10:57:57 2012 +0000
15561 sna: Add assertions to track requests
15563 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15565 commit a93c93be76f6d5d2b481971349aabd15f282c3e8
15566 Author: Chris Wilson <chris@chris-wilson.co.uk>
15567 Date: Wed Jan 11 00:05:20 2012 +0000
15569 sna/gen6: Add a vertex program for a simple (affine, no rotation) spans
15571 I long for the day when this code is obsolete... Until then, this gives
15572 a nice boost in the fishtank.
15574 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15576 commit 3cf5da1090ac777044912ec24619d349d1f6b521
15577 Author: Chris Wilson <chris@chris-wilson.co.uk>
15578 Date: Tue Jan 10 23:32:29 2012 +0000
15580 sna: Amalgamate small replacements into upload buffers
15582 Similar for the standard io paths, try to reuse an upload buffer for a
15583 small replacement pixmap.
15585 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15587 commit f0e3f6b5bebf7471d3e3e84bd9b2d8469eb64093
15588 Author: Chris Wilson <chris@chris-wilson.co.uk>
15589 Date: Tue Jan 10 22:12:41 2012 +0000
15591 sna: Check needs-flush status immediately upon destroy
15593 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15595 commit b4ae6dbaedd774f250b610b1e7323897c4abe23a
15596 Author: Chris Wilson <chris@chris-wilson.co.uk>
15597 Date: Tue Jan 10 21:25:28 2012 +0000
15599 sna: Align the small upload buffers to 2 texels, and the pitch to dwords
15601 References: https://bugs.freedesktop.org/show_bug.cgi?id=44150
15602 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15604 commit 46f6c6917effbc43305367581cc1bbaa962c0251
15605 Author: Chris Wilson <chris@chris-wilson.co.uk>
15606 Date: Tue Jan 10 20:34:09 2012 +0000
15608 sna: A partial read buffer is allowed to be in the GPU domain
15610 As we can create the read buffer from an active cached bo, it may
15611 already be in the GPU domain by the time we first finish it, so fix the
15614 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15616 commit 3c26055639f789b211f7db982e65ec0b86b23d4a
15617 Author: Chris Wilson <chris@chris-wilson.co.uk>
15618 Date: Tue Jan 10 20:06:51 2012 +0000
15620 sna: Shrink the partial upload buffers before compacting the batch
15622 So that the relocation entries point into the contiguous surface/batch
15623 and can be trivially fixed up.
15625 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15627 commit 7b077a4d3d549cde79d5b876289203788294493c
15628 Author: Chris Wilson <chris@chris-wilson.co.uk>
15629 Date: Tue Jan 10 18:40:53 2012 +0000
15631 sna: Make the check for a single fill-rectangle clearer before modifying damage
15633 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15635 commit ca2a07adc45273dc1abeb4b3ba7f88461aaf9c00
15636 Author: Chris Wilson <chris@chris-wilson.co.uk>
15637 Date: Tue Jan 10 17:13:38 2012 +0000
15639 sna: Release the stale GTT mapping after recreating the bo with new tiling
15641 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15643 commit 8dd913fd3a093b54fc024fea2b31a3db2518db8d
15644 Author: Chris Wilson <chris@chris-wilson.co.uk>
15645 Date: Tue Jan 10 10:38:43 2012 +0000
15647 sna: Add reminder about possible future tiling optimisations
15649 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15651 commit 21948578d0d87f38447ef1ae44fada406949bca8
15652 Author: Chris Wilson <chris@chris-wilson.co.uk>
15653 Date: Tue Jan 10 03:19:16 2012 +0000
15655 sna: Disable the inline xRGB to ARGB upload conversion
15657 As we have to upload the dirty data anyway, setting the
15658 alpha-channel to 0xff should be free. Not so for firefox-asteroids on
15661 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15663 commit 87f73b043426c47efa7670fb65bdcc7dfcf71fc3
15664 Author: Chris Wilson <chris@chris-wilson.co.uk>
15665 Date: Tue Jan 10 01:42:37 2012 +0000
15667 sna/gen[23]: Tile render fill to oversized bo
15669 If we are forced to perform a render operation to a bo too large to fit
15670 in the pipeline, copy to an intermediate and split the operation into
15671 tiles rather than fallback.
15673 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15675 commit 2ccb31c5a4120efb940168c5e52154add01b0586
15676 Author: Chris Wilson <chris@chris-wilson.co.uk>
15677 Date: Tue Jan 10 01:41:09 2012 +0000
15679 sna: Shrink upoads buffers
15681 If we do not fill the whole upload buffer, we may be able to reuse a
15682 smaller buffer that is currently bound in the GTT. Ideally, this will
15685 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15687 commit 572cc76be5c78102166f2b3640640e21acdfa129
15688 Author: Chris Wilson <chris@chris-wilson.co.uk>
15689 Date: Tue Jan 10 09:13:34 2012 +0000
15691 sna: Destroy the counter damage after a reduction-to-all
15693 If, for instance, we reduce the GPU damage to all we know that there can
15694 be no CPU damage even though it may still have a region with a list of
15695 subtractions. Take advantage of this knowledge and cheaply discard that
15696 damage without having to evaluate it.
15698 This should prevent a paranoid assertion that there is no cpu damage
15699 when discarding the CPU bo for an active pixmap.
15701 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15703 commit 4a255e13164440b797e5ac11dcbf72f0827a3094
15704 Author: Chris Wilson <chris@chris-wilson.co.uk>
15705 Date: Tue Jan 10 16:34:23 2012 +0000
15707 sna: Replace the free-inactive-gpu-bo with the generic code
15709 The function was semantically equivalent to moving the pixmap to the CPU
15710 for writing, so replace it with a call to the generic function.
15712 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15714 commit c1d403266a611a68081690d19f6debb8e343095f
15715 Author: Chris Wilson <chris@chris-wilson.co.uk>
15716 Date: Tue Jan 10 00:41:05 2012 +0000
15718 sna: Allow for xRGB uploads to replace their target bo
15720 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15722 commit 406776cd955c3c384d4a537300e21eebe4413666
15723 Author: Chris Wilson <chris@chris-wilson.co.uk>
15724 Date: Tue Jan 10 00:25:14 2012 +0000
15726 sna: Rearrange buffer allocation to prefer llc bo over vmaps
15728 If we can create snoopable bo, we prefer to use those as creating a vmap
15729 forces a new bo creation increasing GTT pressure.
15731 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15733 commit b76865fa3deff2f44a1158914a124b9c81d67eca
15734 Author: Chris Wilson <chris@chris-wilson.co.uk>
15735 Date: Mon Jan 9 22:11:21 2012 +0000
15737 sna/gen2: Try to avoid creating a bo for solid colours
15739 As we try to use the diffuse/specular and only resort to using a texture
15740 operation for convenience in the rare case of a solid mask.
15742 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15744 commit 981aae104a96b41db88cc381cc7592818f3e1298
15745 Author: Chris Wilson <chris@chris-wilson.co.uk>
15746 Date: Mon Jan 9 22:02:40 2012 +0000
15748 sna/gen2: Eliminate some switching between logic op and blend
15750 If the new mode can be done either using a logic op or with the blend
15751 unit, prefer the currently enabled unit.
15753 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15755 commit d65b7f9cf46a48e3bfb37c0b75df55aa1e7bff41
15756 Author: Chris Wilson <chris@chris-wilson.co.uk>
15757 Date: Mon Jan 9 21:58:03 2012 +0000
15759 sna/blt: Rearrange to reduce a out-of-bounds copy to a clear
15761 If we asked to use the BLT, try to avoid trigging a context switch for
15762 a trivial case where we sample outside of a NONE source and so can
15763 reduce the operation to a clear.
15765 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15767 commit 09e54c553680cbc2f3b4319cdab0f3d1af1574a8
15768 Author: Chris Wilson <chris@chris-wilson.co.uk>
15769 Date: Mon Jan 9 14:04:41 2012 +0000
15771 sna/gen2: Add poor-man's linear gradient support
15773 Convert the linear gradient to a texture ramp and compute the texture
15774 coordinates in the standard manner.
15776 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15778 commit 6c70558ae7298db94724c931d88a730ef0151608
15779 Author: Chris Wilson <chris@chris-wilson.co.uk>
15780 Date: Sun Jan 8 20:48:48 2012 +0000
15782 sna: mark the cpu bo used for the upload buffer as in CPU domain
15784 For correctness we need to inform GEM of the change of domain for the
15785 buffer so that it knows to invalidate any caches when it is next used by
15788 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15790 commit 9ec31af02922bb016d0dfba07bc60cdca35b36f8
15791 Author: Chris Wilson <chris@chris-wilson.co.uk>
15792 Date: Sun Jan 8 15:52:17 2012 +0000
15794 sna/io: Combine small uploads into single writes
15796 For a small update, try and amalgamate the upload buffer.
15798 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15800 commit 4db1bb3fd81b51e74b7f3e90078627d9d96fbefe
15801 Author: Chris Wilson <chris@chris-wilson.co.uk>
15802 Date: Sun Jan 8 14:20:40 2012 +0000
15804 Removed deprecated xf86PciInfo.h includes
15806 The driver should and does provide its own PCI-IDs.
15808 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15810 commit 54232d1a5da51d0f68fe099436bb0a1a2e249954
15811 Author: Chris Wilson <chris@chris-wilson.co.uk>
15812 Date: Sun Jan 8 15:22:58 2012 +0000
15814 sna: Add ricer stripes to memcpy_xor
15816 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15818 commit c037b4f542a7b21cbaecedec259da3589db10039
15819 Author: Chris Wilson <chris@chris-wilson.co.uk>
15820 Date: Sun Jan 8 08:58:41 2012 +0000
15822 sna: Tune cache size for cpu bo cache
15824 This helps SNB on cairo-traces that utilize lots of temporary uploads
15825 (rasterised sources and masks for instance), but comes at a cost of
15826 regressing others...
15828 In order to counter the regression from increasing the GTT cache size,
15829 the CPU/GTT vma cache are split and accounted separately.
15831 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15833 commit 26042b2660d87044e1920a1267d9984c00c9566a
15834 Author: Chris Wilson <chris@chris-wilson.co.uk>
15835 Date: Sun Jan 8 11:45:11 2012 +0000
15837 sna: Bubble sort the partial buffer list back into order after trimming padding
15839 After reducing the used size in the partial buffer, we need to resort
15840 the list to maintain the list in decreasing amount of available space.
15842 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15844 commit 3f7ea44bf19a03ee81b683885c9c2416092254a3
15845 Author: Chris Wilson <chris@chris-wilson.co.uk>
15846 Date: Sun Jan 8 02:27:28 2012 +0000
15848 sna/gen[67]: Hook into the clear operation for glyph masks
15850 Allow SandyBridge to specialise its clear routine to reduce the number
15851 of ring switches. It may be interesting to specialise the clear routines
15852 even further and use the special render clear commands...
15854 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15856 commit 803ac5c6b992cb5448c67b11345b87a5d2b9c60d
15857 Author: Chris Wilson <chris@chris-wilson.co.uk>
15858 Date: Sun Jan 8 01:38:09 2012 +0000
15860 sna/trapezoids: Don't risk stalling for inplace SRC trapezoids
15862 Optimistically we would replace the GPU damage with the new set of
15863 trapezoids. However, if any partial damage remains then the next
15864 operation which is often to composite another layer of trapezoids (for
15865 complex clipmasks) using IN will then stall.
15867 This fixes a regression in firefox-fishbowl (and lesser regressions
15868 throughout the cairo-traces).
15870 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15872 commit 0229841c0d705138d6317a22ba80f4fa10d26acc
15873 Author: Chris Wilson <chris@chris-wilson.co.uk>
15874 Date: Sun Jan 8 01:46:24 2012 +0000
15876 sna: Do not upload an untiled GPU bo if we already have a CPU bo
15878 Continuing the tuning for sna_copy_boxes.
15880 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15882 commit 42eb9b7c4b55b620d9652ad3a54c8ab0a76fd7c2
15883 Author: Chris Wilson <chris@chris-wilson.co.uk>
15884 Date: Sun Jan 8 00:21:26 2012 +0000
15886 sna: Trim usage of vmapping
15888 The first, and likely only, goal is to support SHMPixmap efficiently
15889 (and without compromising SHMImage!) which we want to preserve as vmaps
15890 and never create a GPU bo. For all other use cases, we will want to
15891 create snoopable CPU bo ala the LLC buffers on SandyBridge.
15893 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15895 commit e52f0204939924181480feec6d1511259169c8f1
15896 Author: Chris Wilson <chris@chris-wilson.co.uk>
15897 Date: Sun Jan 8 00:20:11 2012 +0000
15899 sna: Do not move-to-gpu for sna_copy_boxes if we have a cpu bo
15901 We trade-off the extra copy in the hope that as we haven't used the GPU
15902 bo before then, we won't need it again.
15904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15906 commit c05e90aa99c093431a74a94ceba27d7dce49d05b
15907 Author: Chris Wilson <chris@chris-wilson.co.uk>
15908 Date: Sat Jan 7 18:23:43 2012 +0000
15910 sna: Missing chunks from last commit
15912 And update the check for reusing the blit!
15914 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15916 commit 292aebfcdc042b08d3a5a21f839e6aec1175359e
15917 Author: Chris Wilson <chris@chris-wilson.co.uk>
15918 Date: Sat Jan 7 18:09:47 2012 +0000
15920 sna: Prevent reuse of scanlines after the buffer is destroyed
15922 Once the buffer is destroyed, it may be reallocated with a new pitch. We
15923 could track handle and pitch, but it is easier to simply restart the
15924 blit after the buffer is freed.
15926 References: https://bugs.freedesktop.org/show_bug.cgi?id=44277
15927 References: https://bugs.freedesktop.org/show_bug.cgi?id=44555
15928 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15930 commit d7d07d1df3ffc889ccc32c2ed44ce374e37f851d
15931 Author: Chris Wilson <chris@chris-wilson.co.uk>
15932 Date: Sat Jan 7 10:37:01 2012 +0000
15934 sna: Pad upload buffers to ensure there are an even number of rows
15936 One restriction common to all generations is that samplers access pairs
15937 of rows and so we need to pad the buffer to accommodate access to that
15938 second row. Do so unconditionally along paths that may be used by the
15941 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15943 commit e2ad0f6272c776939499c6202fb60bdcc2e955fa
15944 Author: Chris Wilson <chris@chris-wilson.co.uk>
15945 Date: Fri Jan 6 17:11:00 2012 +0000
15947 sna/blt: Amalgamate many PolyFillRect of single boxes
15949 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15951 commit c085de905c30cb5bae6b339841badf1e105c6ee6
15952 Author: Chris Wilson <chris@chris-wilson.co.uk>
15953 Date: Fri Jan 6 15:47:47 2012 +0000
15955 sna: Also mark a bo created by force-to-gpu as being all-damaged
15957 Similar to the action taken into move-to-gpu so that we forgo the
15958 overhead of damage tracking when the initial act of creation is on the
15961 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15963 commit 9f1935bb4e894264053d94e53c99d5ad607700fb
15964 Author: Chris Wilson <chris@chris-wilson.co.uk>
15965 Date: Fri Jan 6 15:26:11 2012 +0000
15967 sna: Support performing alpha-fixup on the source
15969 By inlining the swizzling of the alpha-channel we can support BLT copies
15970 from an alpha-less pixmap to an alpha-destination.
15972 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15974 commit 141001df6c9c3485c500ed531a214c09b46c1d3b
15975 Author: Chris Wilson <chris@chris-wilson.co.uk>
15976 Date: Fri Jan 6 13:35:27 2012 +0000
15978 sna: always skip active search when requested for find an inactive bo
15980 References: https://bugs.freedesktop.org/show_bug.cgi?id=44504
15981 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15983 commit eac0d9652b2399f8c36ba0288db6fe347ed78dc9
15984 Author: Chris Wilson <chris@chris-wilson.co.uk>
15985 Date: Fri Jan 6 12:39:49 2012 +0000
15987 sna: Optimise sna_poly_segment() for the frequent no-op case
15989 Strange as it may seem... But the principle of doing less work with
15990 greater locality should help everywhere, just not as noticeable when
15991 real work is performed.
15993 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15995 commit a3699fff5ada85e4dea739aade25ebbb728e18f4
15996 Author: Chris Wilson <chris@chris-wilson.co.uk>
15997 Date: Fri Jan 6 12:15:46 2012 +0000
15999 sna: Only force a pipeline flush for a change of destination, not sources
16001 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16003 commit 2fefee6015ed2df52c8513ae180ca83c01ff83c0
16004 Author: Chris Wilson <chris@chris-wilson.co.uk>
16005 Date: Fri Jan 6 12:10:27 2012 +0000
16007 sna/gen6: Reuse current no-blending setup for PictOpClear
16009 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16011 commit 800ca0b4d1194544fe2461f91cbdc632c4d3dd7a
16012 Author: Chris Wilson <chris@chris-wilson.co.uk>
16013 Date: Fri Jan 6 12:00:43 2012 +0000
16015 sna/gen6: Tidy emission of CC state (blending)
16017 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16019 commit b9c9e9970cdb542173e3ed0da2bef614abedd3f7
16020 Author: Chris Wilson <chris@chris-wilson.co.uk>
16021 Date: Fri Jan 6 11:38:31 2012 +0000
16023 sna/trapezoids: Add debug markers for move-to-cpu
16025 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16027 commit 2841c5fee79c42bca3e098ec620755d341b6888f
16028 Author: Chris Wilson <chris@chris-wilson.co.uk>
16029 Date: Fri Jan 6 11:32:54 2012 +0000
16031 sna: fast path move-to-cpu of an all-damaged CPU bo
16033 When the bo is already completely damaged on the CPU, all we need to do
16034 is to sync with the CPU bo.
16036 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16038 commit 07f1724ba5759772043fc4c5d1163ddd640004c2
16039 Author: Chris Wilson <chris@chris-wilson.co.uk>
16040 Date: Fri Jan 6 11:00:38 2012 +0000
16042 sna: Pass the real destination pixmap to sna_write_boxes()
16044 Faking it for the render upload simply isn't good enough, since we need
16045 the correct drawrect.
16047 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16049 commit 5c2e9a98715eeddd8dfdff9fe3871bde659484cc
16050 Author: Chris Wilson <chris@chris-wilson.co.uk>
16051 Date: Fri Jan 6 02:26:39 2012 +0000
16053 sna/trapezoids: Use a slightly faster miTrapezoidBounds()
16055 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16057 commit 75f61678249018c84872c61b4609ba9d3d0d73dc
16058 Author: Chris Wilson <chris@chris-wilson.co.uk>
16059 Date: Fri Jan 6 01:38:42 2012 +0000
16061 sna: Only need to remove the pixmap from the dirty CPU list when clearing damage
16063 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16065 commit 86cc2cd3343c16b12e35e2048c113481f49196cf
16066 Author: Chris Wilson <chris@chris-wilson.co.uk>
16067 Date: Fri Jan 6 01:31:15 2012 +0000
16069 sna: The condition for automatically using the GPU source is no CPU damage
16071 With no CPU damage to upload, we know that there is no reason not to use
16074 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16076 commit 69c556c310da544ddb66c2441f5d06dca29fb9f3
16077 Author: Chris Wilson <chris@chris-wilson.co.uk>
16078 Date: Fri Jan 6 00:21:29 2012 +0000
16080 sna: Rearrange use-cpu-bo to avoid reducing GPU damage
16082 If we spot that the region is wholly contained within the CPU damage
16083 initially, we can conclude that is not in the GPU damage without
16086 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16088 commit 34c4dbe825a45315b05f0bd49539cbcf5de276f1
16089 Author: Chris Wilson <chris@chris-wilson.co.uk>
16090 Date: Thu Jan 5 23:18:37 2012 +0000
16092 sna/gen6: Only force BLT if the src and dst overlaps for self-copy
16094 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16096 commit fc79af7a4abf1f0ceccacf23e8467ee2872eec09
16097 Author: Chris Wilson <chris@chris-wilson.co.uk>
16098 Date: Thu Jan 5 23:06:17 2012 +0000
16100 sna: Only force a batch continuation if the scanout is written to
16102 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16104 commit c16c16d2553aa9a8bf9b0b0b367715b59020a357
16105 Author: Chris Wilson <chris@chris-wilson.co.uk>
16106 Date: Thu Jan 5 22:51:55 2012 +0000
16108 sna/gen6: Enable reuse of source Picture
16110 Check if the source and mask are identical pictures and just copy the
16111 source channel to the mask in that case.
16113 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16115 commit f642b08870c6d7686368e053962075c153f76f86
16116 Author: Chris Wilson <chris@chris-wilson.co.uk>
16117 Date: Thu Jan 5 22:26:09 2012 +0000
16119 sna: Quickly handle the common case of using a CPU source pixmap
16121 For the common case (at least with llc bo) where we are immediately
16122 using an uploaded image from its linear buffer, check upfront before
16123 computing the sampled region for transfer to the GPU.
16125 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16127 commit 8a9baa59e9d0c0761b82f23734514f6847a9bcc1
16128 Author: Chris Wilson <chris@chris-wilson.co.uk>
16129 Date: Thu Jan 5 22:16:31 2012 +0000
16131 sna: Quickly handle the common case of using a GPU source pixmap
16133 For the common case of glyphs, the pixmap is entirely on the GPU which
16134 can be quickly tested before performing the more complex transformations
16135 to determine how much pixel data we need to upload.
16137 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16139 commit d4dad6e3723d70d78247f29f86079862af175266
16140 Author: Chris Wilson <chris@chris-wilson.co.uk>
16141 Date: Thu Jan 5 21:39:24 2012 +0000
16143 sna: Check that the GTT is no longer mapped before ensuring the shadow pointer
16145 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16147 commit 352f0a3ffb8ae8503bb2fab779ea84f4e6d76317
16148 Author: Chris Wilson <chris@chris-wilson.co.uk>
16149 Date: Thu Jan 5 21:33:26 2012 +0000
16151 sna: Only mark the pixmap as GTT mapped if the private owns the pixels
16153 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16155 commit 3061382f98c0b284c250c1a36f7a4fee22046664
16156 Author: Chris Wilson <chris@chris-wilson.co.uk>
16157 Date: Thu Jan 5 21:05:52 2012 +0000
16159 sna: Make sure that we do not attach to SHM pixmaps in ignorance
16161 Without xserver support for notification of when scratch pixmaps are
16162 reused, we simply cannot attach our privates to them lest we cause
16163 corruption with SHM pixmaps.
16165 This is a recent regression back unto an old, old xserver issue.
16167 Reported-by: Paul Neumann <paul104x@yahoo.de>
16168 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=44503
16169 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16171 commit f7431726338e72d7a989e838f48aecdab15ab366
16172 Author: Chris Wilson <chris@chris-wilson.co.uk>
16173 Date: Thu Jan 5 20:26:30 2012 +0000
16175 sna: Add more debugging to use-gpu-bo
16177 To aide debugging in conjunction with compositors and their crazy
16180 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16182 commit 53c943b551b86402067ea96cd9fcca2848e84387
16183 Author: Chris Wilson <chris@chris-wilson.co.uk>
16184 Date: Thu Jan 5 18:41:26 2012 +0000
16186 sna: Check for usable pixmap before deference its private
16188 Beware the NULL pointer and early deference.
16190 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16192 commit 307276d7da99ceb84c175e412452e5d86d615cf4
16193 Author: Chris Wilson <chris@chris-wilson.co.uk>
16194 Date: Thu Jan 5 14:26:05 2012 +0000
16196 sna: Enable using the render pipeline for uploads to Y-tiled pixmaps
16198 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16200 commit f9451aaa39e1d77b08d2201f5c4a44ab834f2237
16201 Author: Chris Wilson <chris@chris-wilson.co.uk>
16202 Date: Thu Jan 5 14:22:21 2012 +0000
16204 sna: Mark the upload as having acquired the GTT mapping
16206 In order for the entire PutImage to be performed inplace, we need to
16207 maintain the tendency to keep doing inplace operations. This hint is
16208 provided by tracking whether or not the last operation used the GTT
16209 mapping. However, that hint was not being provided by zpixmap_blt.
16211 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16213 commit ce93c00a33a395986ffc0d33aae85d2c23aabc5d
16214 Author: Chris Wilson <chris@chris-wilson.co.uk>
16215 Date: Thu Jan 5 14:20:55 2012 +0000
16217 sna: Avoid operating inplace if we have pending CPU damage
16219 If the operation does not replace existing CPU damage, we are likely to
16220 want to reuse the pixmap again on the CPU, so avoid mixing CPU/GPU
16223 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16225 commit f5d02d90f3349f33341954fc55824d12a7829e6c
16226 Author: Chris Wilson <chris@chris-wilson.co.uk>
16227 Date: Thu Jan 5 12:56:21 2012 +0000
16229 sna: Avoid switching rings for uploading data
16231 On systems that incur painful overhead for ring switches, it is usually
16232 better to create a large buffer and perform a sparse copy on the same
16233 ring than create a compact buffer and use the BLT.
16235 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16237 commit a42111ff94911d5fad34e6dbac8aa4a11b2d8032
16238 Author: Chris Wilson <chris@chris-wilson.co.uk>
16239 Date: Thu Jan 5 12:12:22 2012 +0000
16241 sna/trapezoids: Try creating the trapezoids mask inplace if not using spans
16243 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16245 commit 4c6ccbd46a1681c73382b79a0246f7d8cb097606
16246 Author: Chris Wilson <chris@chris-wilson.co.uk>
16247 Date: Thu Jan 5 11:09:01 2012 +0000
16249 sna: Throttle, throttle, throttle.
16251 This reverts 281425551bdab7eb38ae167a3205b14ae3599c49 as it was causing
16252 insufferable lag in firefox.
16254 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16256 commit dbb495c3aeef34eab2aabfdfaafb81facdea4100
16257 Author: Chris Wilson <chris@chris-wilson.co.uk>
16258 Date: Thu Jan 5 10:34:08 2012 +0000
16260 sna: Attach to the pixmap for source counting when deferring uploads
16262 If we decide to defer the upload for this instance of the source pixmap,
16263 mark it so. Then if we do use it again we will upload it to a GPU bo and
16264 hopefully reuse those pixels.
16266 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16268 commit 292097590ba0f81b7505aedb2a4fbd18021641a5
16269 Author: Chris Wilson <chris@chris-wilson.co.uk>
16270 Date: Thu Jan 5 10:31:15 2012 +0000
16272 sna: Try doing a cheap is-contained check before reducing damage
16274 If the pixmap is entirely within the current CPU damage, we can forgo
16275 reducing either the GPU or CPU damage when checking whether we need to
16276 upload dirty pixels for a source texture.
16278 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16280 commit 5ddae8bf21067e7d2a031a6b9020c02ab977cfed
16281 Author: Chris Wilson <chris@chris-wilson.co.uk>
16282 Date: Thu Jan 5 10:27:37 2012 +0000
16284 sna: Check the composite-rectangles will be on the GPU before forcibly attaching
16286 If we think that the operation is better performed on the CPU, avoid the
16287 overhead of manipulating our privates.
16289 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16291 commit 6ffd3605640bacaf0d744075622463aacf78d797
16292 Author: Chris Wilson <chris@chris-wilson.co.uk>
16293 Date: Thu Jan 5 10:25:12 2012 +0000
16295 sna: Deferred attachment to a pixmap needs to mark the pixmap as entirely dirty
16297 As demonstrated with oversized glyphs and a chain of catastrophy, when
16298 attaching our private to a pixmap after creation we need to mark the
16299 entire CPU pixmap as dirty as we never tracked exactly which bits were
16302 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16304 commit a09ebe0b6cc66d08e52b57851d9b8a6a1f71df2b
16305 Author: Chris Wilson <chris@chris-wilson.co.uk>
16306 Date: Thu Jan 5 10:23:25 2012 +0000
16308 sna: Immediately upload oversized glyphs
16310 Glyphs, even large ones, we suspect will be reused and so the deferred
16311 upload is counterproductive. Upload them immediately and mark them as
16312 special creatures for later debugging.
16314 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16316 commit 797b27365dc09a3f84349a33cffbfdc67cba0baa
16317 Author: Chris Wilson <chris@chris-wilson.co.uk>
16318 Date: Thu Jan 5 10:30:47 2012 +0000
16320 sna: Don't perform a deferred attachment for vmapping if not supported
16322 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16324 commit 23e3959e1f603c3ba0bd4f01128290241cf93edf
16325 Author: Chris Wilson <chris@chris-wilson.co.uk>
16326 Date: Thu Jan 5 04:00:32 2012 +0000
16328 sna: Use the created cpu-bo for upload and download
16330 As we explicitly create CPU bo when wanted, we no longer desire to
16331 spontaneously create vmaps for simply uploading to the GPU bo.
16333 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16335 commit 4530b87e445aea30ccbe751699963118ad5672d4
16336 Author: Chris Wilson <chris@chris-wilson.co.uk>
16337 Date: Thu Jan 5 00:01:27 2012 +0000
16339 sna: Fix assertion as we may want to create active bo with cpu mappings
16341 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16343 commit 4119e68fb157fc612bce5e9c5669112ce35b4ca1
16344 Author: Chris Wilson <chris@chris-wilson.co.uk>
16345 Date: Wed Jan 4 19:35:30 2012 +0000
16347 sna/damage: Fix reduction to copy the boxes correctly
16349 We need to be carefully to copy the boxes in a strict lifo order so as
16350 to avoid overwritting the last boxes when reusing the array allocations.
16352 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16354 commit 71b0924b586d9a60397e92e941e3d0cfa636ee61
16355 Author: Chris Wilson <chris@chris-wilson.co.uk>
16356 Date: Wed Jan 4 19:34:41 2012 +0000
16358 sna: Mark tiny CPU pixmaps as completely dirty
16360 Avoid the overhead of tracking damage on small pixmaps when using CPU
16361 rasterisation; the extra cost of sending the whole pixmap compared to
16362 the damage is negligble should it ever be required on the GPU.
16364 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16366 commit fc14e63256a9b756c7c77ffe73f2f13784396c5a
16367 Author: Chris Wilson <chris@chris-wilson.co.uk>
16368 Date: Wed Jan 4 19:33:30 2012 +0000
16370 sna: Only request page size allocations for the replacement buffer
16372 A precondition on bo creation is that the size must be page aligned.
16374 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16376 commit 89739b711f42c3dbed7d3f4e6da0cdd61a5205ae
16377 Author: Chris Wilson <chris@chris-wilson.co.uk>
16378 Date: Wed Jan 4 17:41:16 2012 +0000
16380 sna: Expand small stipples into 8x8 patterns
16382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16384 commit 878fbfe509da0a25dff8515d6d3b38ca3826466e
16385 Author: Chris Wilson <chris@chris-wilson.co.uk>
16386 Date: Wed Jan 4 16:56:55 2012 +0000
16388 sna: Align tiled stipple uploads with the stipple pixmap correctly
16390 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16392 commit 28a222a66bb450c7f6d61aef16fe73332854c8ce
16393 Author: Chris Wilson <chris@chris-wilson.co.uk>
16394 Date: Wed Jan 4 12:48:25 2012 +0000
16396 sna: Always prefer to use cacheable operations after the GPU is wedged
16398 As rasterisation will be performed upon the CPU we need to avoid the
16399 readbacks form uncached memory and so we should restrict ourselves to
16400 only create further damage within the CPU pixmap.
16402 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16404 commit 99ff83a3bebf1aa25c44a87f2c344307d20bc062
16405 Author: Chris Wilson <chris@chris-wilson.co.uk>
16406 Date: Wed Jan 4 12:35:12 2012 +0000
16408 sna: Prevent creation of bo purely for GTT mapping when wedged
16410 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16412 commit a7c35fa4c3155d10ad75ac33aefb03dafa010aba
16413 Author: Chris Wilson <chris@chris-wilson.co.uk>
16414 Date: Wed Jan 4 12:33:09 2012 +0000
16416 sna: Only allow mappable pixmaps to be mapped
16418 If we did not allocate the pixel data, such as for wedged pixmaps or
16419 scratch buffers, then we cannot perform the pointer dance nor do we want
16420 to create the GPU buffer.
16422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16424 commit 2cbc97d28aae138566e49df05e16d274b533ea86
16425 Author: Chris Wilson <chris@chris-wilson.co.uk>
16426 Date: Wed Jan 4 12:32:06 2012 +0000
16428 sna: Force creation of ordinary pixmaps when wedged.
16430 If the pixmap was intended for scanout, then the GPU bo will be created
16431 upon attachment to the fb.
16433 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16435 commit f6cc73ba66f34d190b7c17421a62aa1256698431
16436 Author: Chris Wilson <chris@chris-wilson.co.uk>
16437 Date: Wed Jan 4 12:56:31 2012 +0000
16439 sna: The batch is counted in dwords, not bytes!
16441 Yikes, I choose the wrong units for the max_batch_size.
16443 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16445 commit 227fbb90c4884bbc58c6c0cfff9663ec9ca54171
16446 Author: Chris Wilson <chris@chris-wilson.co.uk>
16447 Date: Wed Jan 4 11:47:32 2012 +0000
16449 sna: Carefully free the freed_pixmap upon exit
16451 As the contents of the pixmap are now rubbish, we need to manually
16452 destroy it rather than pass it to the normal sna_pixmap_destroy()
16455 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16457 commit 0ed758cd2176ee4f34e03d05d05130d52d75e577
16458 Author: Chris Wilson <chris@chris-wilson.co.uk>
16459 Date: Wed Jan 4 12:18:20 2012 +0000
16461 sna: Limit batch to a single page on 865g
16463 Verified on real hw, this undocumented (at least in the bspec before me)
16466 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16468 commit 3449f1cbe184a618e662d2a662167f05362b82e5
16469 Author: Chris Wilson <chris@chris-wilson.co.uk>
16470 Date: Wed Jan 4 02:38:16 2012 +0000
16472 sna: Guard against NULL damage deref
16474 The fast version of damage checking assumes that the damage has already
16475 been determined to be non-NULL, so make sure it is.
16477 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16479 commit 04a6260016fb5ebefc1056ed6acf5001ec535d5c
16480 Author: Chris Wilson <chris@chris-wilson.co.uk>
16481 Date: Wed Jan 4 01:14:19 2012 +0000
16483 sna: Fix typo during partial list deletion
16485 And keep the asserts that lead to its discovery.
16487 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16489 commit f1dc1eadd84097fc691e85c636535ceeeb601a18
16490 Author: Chris Wilson <chris@chris-wilson.co.uk>
16491 Date: Wed Jan 4 00:09:30 2012 +0000
16493 sna/gen3: Remove incorrect premultiplication of solid component-alpha mask
16495 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16497 commit 8cb9b8d7d7a1eb62eb3b20e6a50d3f1c9bde40c1
16498 Author: Chris Wilson <chris@chris-wilson.co.uk>
16499 Date: Tue Jan 3 23:13:24 2012 +0000
16501 sna: Discard mask and source for PictOpClear
16503 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16505 commit feb75c55d6105a9ecc5f5ec7830276d1654918b9
16506 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
16507 Date: Sat Dec 31 21:18:15 2011 +0800
16509 uxa/glamor: Route some missing drawing function to glamor
16511 We have to route all the drawing function to glamor first, when
16512 glamor is enabled. This adds a few more functions that were previously
16513 just falling back to swrast and passes them to glamor instead.
16515 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
16516 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16518 commit f4ea732d7181d4ef7f97fb2585e87cdcaef99d46
16519 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
16520 Date: Tue Dec 27 17:09:17 2011 +0800
16522 uxa/glamor: Let glamor do the GC validation
16524 If we are using GLAMOR, then a tile pixmap or stipple pixmap
16525 may be pure glamor pixmap and thus UXA will not know how to
16526 render to them, and we need to let glamor do the validation.
16528 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
16529 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16531 commit e8aa9cedbfca09e22d46ec8882b9bebb3eb83022
16532 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
16533 Date: Tue Dec 27 17:09:16 2011 +0800
16535 uxa/glamor: Remove extraneous flush
16537 When glamor is enabled, a pixmap will not be accessed by UXA's
16538 accelerated functions. Only unaccelerated functions may access those
16539 pixmaps, and before each unaccelerated rendering, it calls
16540 uxa_prepare_access which will do a glFlush. Combined with a flush before
16541 sending to DRI clients, we no longer need to flush after every
16544 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
16545 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16547 commit 4ca94ed91fab7a9de5eed1053b891e8420cdf43d
16548 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
16549 Date: Tue Dec 27 17:09:15 2011 +0800
16551 uxa/glamor: Remove dead code.
16553 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
16554 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16556 commit 5576afc7d12ec0f028e416f4b4a076fc6d3c29d9
16557 Author: Chris Wilson <chris@chris-wilson.co.uk>
16558 Date: Tue Jan 3 16:48:52 2012 +0000
16560 sna: Inline checks for over-sized buffers during pixmap creation
16562 Make the tests for acceptable GPU pixmaps explicit and upfront.
16564 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16566 commit b2dc17678fa3a548be61e1055d4d321dbd6494c3
16567 Author: Chris Wilson <chris@chris-wilson.co.uk>
16568 Date: Tue Jan 3 15:50:21 2012 +0000
16570 sna: Always search for an exact active match first from the bo-cache
16572 And accept second-best only if permitted by flags.
16574 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16576 commit caf41ae23d263c7baed6b23a80693690c4dfe563
16577 Author: Chris Wilson <chris@chris-wilson.co.uk>
16578 Date: Tue Jan 3 15:11:13 2012 +0000
16580 sna: Force tiling on medium pitches if not reduced for other reasons
16582 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16584 commit 1b688ca0cfc1e623f896e60a6a6abf724d57efb4
16585 Author: Chris Wilson <chris@chris-wilson.co.uk>
16586 Date: Tue Jan 3 14:18:05 2012 +0000
16588 sna: Delete the last buffer if destroyed without being used on the GPU
16590 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16592 commit 4a68e7277104d7dccbdc4739eb5651902b9ccc20
16593 Author: Chris Wilson <chris@chris-wilson.co.uk>
16594 Date: Tue Jan 3 12:19:24 2012 +0000
16596 sna/gen3: Prefer to use the blitter for solid fills
16598 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16600 commit 199e357dc693cfe232464e4faa82d9223a8325d6
16601 Author: Chris Wilson <chris@chris-wilson.co.uk>
16602 Date: Tue Jan 3 12:16:48 2012 +0000
16604 sna/trapezoids: Pass the sna pointer to the callees
16606 Avoid the repeated multiple indirect dereferences.
16608 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16610 commit bd7aa1754dcf9d58d7deb92d99da9b78c91a0eb2
16611 Author: Chris Wilson <chris@chris-wilson.co.uk>
16612 Date: Tue Jan 3 12:46:40 2012 +0000
16614 sna: Avoid unnecessary call overhead for read-only move-to-cpu
16616 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16618 commit 294c8791c435acf5933109fbcab85623a55b2c6c
16619 Author: Chris Wilson <chris@chris-wilson.co.uk>
16620 Date: Tue Jan 3 12:09:51 2012 +0000
16622 sna: Use a cheaper no-reduction damage check for simply discarding further damage
16624 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16626 commit 7ff42e9c83cdb17660a64901497704b38032af8c
16627 Author: Chris Wilson <chris@chris-wilson.co.uk>
16628 Date: Tue Jan 3 19:15:07 2012 +0000
16630 sna/damage: Mark the box as packed so that the embedded_box is aligned correctly
16632 valgrind was complaining about an overlapping memcpy on a 64-bit
16633 platform as gcc padded the sna_damage_box to 28 bytes...
16635 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16637 commit 4652a20de5fee0d5d1a8baadc4a93b6a93525564
16638 Author: Chris Wilson <chris@chris-wilson.co.uk>
16639 Date: Tue Jan 3 21:20:09 2012 +0000
16641 sna/gen3: Check for a non-NULL bo before incrementing its ref in reuse_source()
16643 For gen3, we may reduce a source into a constant operator and so
16644 dispense with keeping a bo. When duplicated into the mask channel, we
16645 then need to be careful not to dereference the NULL pointer.
16647 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16649 commit 777fcc9b11e8cbfe942aa6bf898749f552acb3cf
16650 Author: Chris Wilson <chris@chris-wilson.co.uk>
16651 Date: Sun Jan 1 16:03:42 2012 +0000
16653 sna: Prefer not to force the creation of gpu bo for pixmap uploads
16655 As this causes a significant regression when benchmarking firefox on SNB
16656 with firefox-planet-gnome if we already have CPU buffers.
16658 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16660 commit 1a6c16a24cb74f82df9757c034c50f2cf141523f
16661 Author: Chris Wilson <chris@chris-wilson.co.uk>
16662 Date: Sun Jan 1 14:32:48 2012 +0000
16664 sna: Keep a freelist of buffers
16666 As reallocation of bo is the most frequent cause of malloc/free.
16668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16670 commit f008a2a90c632d843f07846426f6d14af728c13b
16671 Author: Chris Wilson <chris@chris-wilson.co.uk>
16672 Date: Sun Jan 1 10:24:49 2012 +0000
16674 sna: Tune blt routine
16676 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16678 commit b57c7248acc2ad6a51b7eac74449d155defc998f
16679 Author: Chris Wilson <chris@chris-wilson.co.uk>
16680 Date: Sun Jan 1 14:10:20 2012 +0000
16682 sna: Delete some redundant code
16684 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16686 commit 83d16455b70e2f0348e51f95128d4fcae3736e7f
16687 Author: Chris Wilson <chris@chris-wilson.co.uk>
16688 Date: Sun Jan 1 13:58:44 2012 +0000
16690 sna: Eliminate one union pass when reducing damage-add
16692 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16694 commit 2f5b769e7144c7b85010429e6d8aebf9c9fc1628
16695 Author: Chris Wilson <chris@chris-wilson.co.uk>
16696 Date: Sun Jan 1 10:00:37 2012 +0000
16698 sna: Move some checks against programming errors to asserts
16700 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16702 commit 24bf59c43fbedd9d85c913ac3c7f0e752c232eb4
16703 Author: Chris Wilson <chris@chris-wilson.co.uk>
16704 Date: Sat Dec 31 18:57:57 2011 +0000
16706 Revert "sna: Increase the minimum alignment constraint for g33"
16708 This reverts commit 2934e778f01cdf1307732b248b11a31c0e79e866. The actual
16709 cause of the bug I was seeing on my PNV box turned out to be
16710 a1f585a3d0a, so time to reinvestigate the alignment issues.
16712 commit 770a953ff03bb8328c3f29e274d225528840f30c
16713 Author: Chris Wilson <chris@chris-wilson.co.uk>
16714 Date: Sat Dec 31 17:19:08 2011 +0000
16716 uxa/video: Clear all state structures before uploading
16718 A regression from eb859f644633e left some of the state uninitialised
16719 before uploading to the GPU leading to undefined behaviour.
16721 Reported-by: Alexey Shumitsky <alexey.shumitsky@gmail.com>
16722 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=44338
16723 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=44252
16724 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16726 commit f1995c684d2e074ae04c9e42727e59cc77e16104
16727 Author: Chris Wilson <chris@chris-wilson.co.uk>
16728 Date: Sat Dec 31 15:57:32 2011 +0000
16730 uxa: the video destination should be in the render write domain
16732 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16734 commit 511146d2e40e267f74e18c0dd06b65f5d6a69f5d
16735 Author: Chris Wilson <chris@chris-wilson.co.uk>
16736 Date: Sat Dec 31 13:27:13 2011 +0000
16738 sna: Implement inplace trapezoidal addition (for clipmasks)
16740 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16742 commit a1f585a3d0ae72e896684f62960a9ae96dc259f5
16743 Author: Chris Wilson <chris@chris-wilson.co.uk>
16744 Date: Sat Dec 31 12:43:04 2011 +0000
16746 sna: Update pixmap->devKind for GTT mappings
16748 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16750 commit 343de19ef98eea016601dbbe2d1f601dbff9de20
16751 Author: Chris Wilson <chris@chris-wilson.co.uk>
16752 Date: Sat Dec 31 11:19:44 2011 +0000
16754 sna: Implement tiled uploads for small stipples larger than patterns
16756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16758 commit fb15cf2294da20bfc375ae62443b822770850316
16759 Author: Chris Wilson <chris@chris-wilson.co.uk>
16760 Date: Sat Dec 31 10:24:58 2011 +0000
16762 sna: Make sure the shadow pixmap is suitable for scanout
16764 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16766 commit 281425551bdab7eb38ae167a3205b14ae3599c49
16767 Author: Chris Wilson <chris@chris-wilson.co.uk>
16768 Date: Sat Dec 31 10:10:00 2011 +0000
16770 sna: Tweak throttling to be run before going idle
16772 The goal is to avoid introducing extra latency in the middle of a
16773 command sequence and to delay it until we have to wait for the clients.
16775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16777 commit 2f5448f1329f344702e4e29b2509d7190677bbf0
16778 Author: Chris Wilson <chris@chris-wilson.co.uk>
16779 Date: Fri Dec 30 22:13:05 2011 +0000
16781 sna: compare tiling against minimum width for relevancy
16783 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16785 commit e82b3d8069629ae5776e039d6ea1f3c901a3a094
16786 Author: Chris Wilson <chris@chris-wilson.co.uk>
16787 Date: Fri Dec 30 09:16:00 2011 +0000
16789 sna: Defer the initial source upload if the GPU bo will be untiled
16791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16793 commit 72624b0b5ba6549cb8d96f992d2656a28ec6078c
16794 Author: Chris Wilson <chris@chris-wilson.co.uk>
16795 Date: Thu Dec 29 22:43:35 2011 +0000
16797 sna: Remove redundant use-gpu-bo clause
16799 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16801 commit 6dc9bbc0161f071bbaf5c7e379974158a5ad666f
16802 Author: Chris Wilson <chris@chris-wilson.co.uk>
16803 Date: Thu Dec 29 22:43:44 2011 +0000
16805 sna: Fix DBG for no intersection with damage
16807 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16809 commit 0dc5c0651cb691fb8811cdf3075b3d322f9d37f8
16810 Author: Chris Wilson <chris@chris-wilson.co.uk>
16811 Date: Thu Dec 29 13:49:23 2011 +0000
16813 sna: Tune disarming of the scanout flush timer
16815 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16817 commit ccd895c61d9f409a36c70eaf02d7ab43cc6c03c2
16818 Author: Chris Wilson <chris@chris-wilson.co.uk>
16819 Date: Thu Dec 29 12:09:04 2011 +0000
16821 sna: Allow uploading inplace to a freshly created GPU bo
16823 If the operation is favoured to be performed using a WC upload, presume
16824 that we will use the uploaded pixmap on the GPU and so prefer to create
16825 a GPU buffer to hold the fresh data.
16827 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16829 commit f85a853455249fd782e907ce1d78d11f75362def
16830 Author: Chris Wilson <chris@chris-wilson.co.uk>
16831 Date: Thu Dec 29 11:59:57 2011 +0000
16833 sna: Only update the damage region as necesary for contains-box
16835 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16837 commit f53936407bd68f3d36ab525b4e0531fad58098e3
16838 Author: Chris Wilson <chris@chris-wilson.co.uk>
16839 Date: Thu Dec 29 11:59:37 2011 +0000
16841 sna: Reset the source counter upon dirtying the shadow pixmap
16843 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16845 commit c9f7f10bc528ac783f4e89bdb36fc56ccdcd0088
16846 Author: Chris Wilson <chris@chris-wilson.co.uk>
16847 Date: Thu Dec 29 09:03:21 2011 +0000
16849 sna: Be paranoid and check for errors from mmapping the bo
16851 Now that the error propagation is actually in place, we may as well use
16854 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16856 commit 98e053503aed292beb2006043e8508d59262c7c2
16857 Author: Chris Wilson <chris@chris-wilson.co.uk>
16858 Date: Thu Dec 29 08:50:53 2011 +0000
16860 uxa: Depend upon DRI2 not DRI
16862 The symbols required for building intel_dri.c are checked during
16863 configure under the DRI2 defines.
16865 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16867 commit 58cee9d0def82d19f03c840325b6e84c20e77ace
16868 Author: Chris Wilson <chris@chris-wilson.co.uk>
16869 Date: Wed Dec 28 22:52:58 2011 +0000
16871 sna/dri: Update set_bo() for change of bo ownership transfer during page-flip
16873 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16875 commit f9ee02be269d1658d3a8a5166e9bd2a49ab6dea0
16876 Author: Chris Wilson <chris@chris-wilson.co.uk>
16877 Date: Wed Dec 28 22:52:44 2011 +0000
16879 sna: Fix debug message for read-boxes
16881 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16883 commit 698bd43e0f7a59bcce6601b0c96df85935e54bb3
16884 Author: Chris Wilson <chris@chris-wilson.co.uk>
16885 Date: Wed Dec 28 21:45:12 2011 +0000
16887 sna: Do not create glyph caches if the gpu is wedged
16889 As these will only be created in normal memory and never used.
16891 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16893 commit 78425a15597079c0549de0fa1bbbdbebcf9eaf28
16894 Author: Chris Wilson <chris@chris-wilson.co.uk>
16895 Date: Wed Dec 28 15:26:14 2011 +0000
16897 sna: Refactor common code for testing gpu busyness of a pixmap
16899 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16901 commit 9105b7e03ab7e3602c7200ebdc44e89f873afe1f
16902 Author: Chris Wilson <chris@chris-wilson.co.uk>
16903 Date: Wed Dec 28 13:46:18 2011 +0000
16905 sna: Perform clip mask compositing inplace
16907 Avoid the extra composite-in pass for simple clipmask construction.
16909 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16911 commit a5583165da0d2ae5eb1e5a2e11ee6e245d4b5aa4
16912 Author: Chris Wilson <chris@chris-wilson.co.uk>
16913 Date: Wed Dec 28 13:57:09 2011 +0000
16915 uxa: Remove concept of global FB ownership made obsolete by GEM and KMS
16917 With the introduction of GEM, we can continue to submit batch buffers
16918 irrespective of ownership of the console, so do so.
16920 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16922 commit 337635ab97bbfc9b4455eadb63214783bb90bb2b
16923 Author: Chris Wilson <chris@chris-wilson.co.uk>
16924 Date: Wed Dec 28 10:56:57 2011 +0000
16926 sna: Add some more debugging to find the source of a read-boxes error
16928 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16930 commit 3bf81241cbc97763e5f3cbc82772437ccbff5a6c
16931 Author: Chris Wilson <chris@chris-wilson.co.uk>
16932 Date: Tue Dec 27 17:09:59 2011 +0000
16934 sna: Skip zero-sized glyphs
16936 Similar to the empty glyphs make sure that we just advance by the
16937 character width without drawing anything.
16939 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16941 commit 8fc21328a0bdf87fde35d68d2b27834011acde7b
16942 Author: Chris Wilson <chris@chris-wilson.co.uk>
16943 Date: Tue Dec 27 15:26:46 2011 +0000
16945 sna: Don't discard empty glyphs, just skip them
16947 A space is encoded as a 1x1 blank glyph, but we still need to advance by
16948 its character width and so we cannot simply discard the glyph.
16950 References: https://bugs.freedesktop.org/show_bug.cgi?id=44091id=44091
16951 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16953 commit 726290ea2a54716d3d8fb998fe7d2ab252421de7
16954 Author: Chris Wilson <chris@chris-wilson.co.uk>
16955 Date: Tue Dec 27 13:16:34 2011 +0000
16957 sna: Rearrange damage for tail-call optimisations
16959 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16961 commit 04396b6e5039c0cd1ac655a8a9afe88bb8a1ebde
16962 Author: Chris Wilson <chris@chris-wilson.co.uk>
16963 Date: Tue Dec 27 09:58:40 2011 +0000
16965 sna: Don't force the creation of a GPU bo if it will be untiled
16967 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16969 commit 655a96cd5f12ff7fe593e7da5f34c7c92ef9b87d
16970 Author: Chris Wilson <chris@chris-wilson.co.uk>
16971 Date: Mon Dec 26 14:37:37 2011 +0000
16973 sna: Mark the glyph cache as being unpageable
16975 In order to prevent it falling foul of the inactive pixmap reaper, we
16976 need to mark the pixmap as pinned.
16978 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16980 commit 2934e778f01cdf1307732b248b11a31c0e79e866
16981 Author: Chris Wilson <chris@chris-wilson.co.uk>
16982 Date: Mon Dec 26 11:30:39 2011 +0000
16984 sna: Increase the minimum alignment constraint for g33
16986 Still not getting the single-stream mode that I am aiming for!
16988 Reported-by: Matti Hamalainen <ccr@tnsp.org>
16989 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=44150
16990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
16992 commit f7593a995a829978ce81397dde5c9ea6f9bb7681
16993 Author: Chris Wilson <chris@chris-wilson.co.uk>
16994 Date: Sun Dec 25 09:40:46 2011 +0000
16996 sna: Move the is-mappable check into the callers of region_inplace()
16998 As they slightly differ in their requirements.
17000 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17002 commit aff32e3e08fce3c7b8dab3a25c96d69c409471fc
17003 Author: Chris Wilson <chris@chris-wilson.co.uk>
17004 Date: Sun Dec 25 00:36:33 2011 +0000
17006 sna: Tweak damage not to reduce if it will not affect the outcome of reducing to all
17008 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17010 commit be21a09fef0a3499153f2c123b65f1d5c872d437
17011 Author: Chris Wilson <chris@chris-wilson.co.uk>
17012 Date: Sun Dec 25 00:28:55 2011 +0000
17014 sna: Only call move-to-gpu on scanout flush if we have cpu damage to move
17016 Reduce the number of redundant calls.
17018 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17020 commit 7b69dc88094ee2bb017364cd4356f3ae69cbc1f1
17021 Author: Chris Wilson <chris@chris-wilson.co.uk>
17022 Date: Sun Dec 25 00:21:21 2011 +0000
17024 sna: Inline the common portion of sna_get_pixel_from_rgba()
17026 The function overhead completely dominates for the common case.
17028 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17030 commit b117f65520919f4ba36010cfe913a8c53166bf23
17031 Author: Chris Wilson <chris@chris-wilson.co.uk>
17032 Date: Sat Dec 24 21:03:01 2011 +0000
17034 sna: Jump straight to the fallback copy routines if the dst is not attached
17036 Marginally simplify the convoluted logic for choosing the most
17037 appropriate path and help prevent further errors.
17039 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17041 commit 72217790ee2c080d618274456360b481d015e898
17042 Author: Chris Wilson <chris@chris-wilson.co.uk>
17043 Date: Sat Dec 24 19:14:09 2011 +0000
17045 sna: Use shadow if the GPU is busy or not immediately mappable
17047 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17049 commit 0be136c21f0373d1eb2259b83c598655f4eb841e
17050 Author: Chris Wilson <chris@chris-wilson.co.uk>
17051 Date: Sat Dec 24 18:44:15 2011 +0000
17053 sna: use indirect uploads if the bo was last known to be unmappable
17055 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17057 commit e764a52ee8f0c552e218b3110318be9ba06634ae
17058 Author: Chris Wilson <chris@chris-wilson.co.uk>
17059 Date: Sat Dec 24 15:17:06 2011 +0000
17061 sna: Encourage large operations to be migrated to the GPU
17063 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17065 commit f9f8535db6a9e3affba9ba2c2a9314dfe12ab270
17066 Author: Chris Wilson <chris@chris-wilson.co.uk>
17067 Date: Sat Dec 24 21:01:39 2011 +0000
17069 sna: Fix damage reduction by adding new boxes to the tail of the box list
17071 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17073 commit f8c91f229a9ca18fc58b0cabbea595d26627db8b
17074 Author: Chris Wilson <chris@chris-wilson.co.uk>
17075 Date: Sat Dec 24 12:17:53 2011 +0000
17077 sna: We can always compare the box to be subtracted against the dirty extents
17079 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17081 commit cc21d3fa045209861564cf240a9a082e0bdb63f8
17082 Author: Chris Wilson <chris@chris-wilson.co.uk>
17083 Date: Sat Dec 24 12:17:43 2011 +0000
17085 sna/damage: Fix the is-contained test
17087 It was a non-overlapping test which is almost the reverse of what was
17090 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17092 commit b86e4f59299f935d5a0ea8375da97e6fc57571f9
17093 Author: Chris Wilson <chris@chris-wilson.co.uk>
17094 Date: Sat Dec 24 11:45:27 2011 +0000
17096 sna: Check that the copy dst is attached before replacing damage
17098 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17100 commit 3bcc9f08e1309725c18dfa5ca0e905bb88ebf4ea
17101 Author: Chris Wilson <chris@chris-wilson.co.uk>
17102 Date: Sat Dec 24 09:59:24 2011 +0000
17104 sna: Create a GPU for a streaming write
17106 If we are about to do a write-only drawing operation that will exceed
17107 our cache, allocate a GPU bo and perform the operation inplace.
17109 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17111 commit efc8d04fc114e9966e5ca00600f9663ecf03a5ca
17112 Author: Chris Wilson <chris@chris-wilson.co.uk>
17113 Date: Sat Dec 24 09:32:18 2011 +0000
17115 sna: Promote an unattached pixmap to the GPU for inplace copies
17117 If we would prefer to perform the copy on the GPU and if the pixmap is
17118 virgin, create a GPU bo for the operation.
17120 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17122 commit 8c0e7ff5ac996a5a5c0790966084d1c9e1b1f859
17123 Author: Chris Wilson <chris@chris-wilson.co.uk>
17124 Date: Sat Dec 24 09:19:18 2011 +0000
17126 sna: Clear the CPU damage for other overwriting rectangular blits as well
17128 And make the accessing of the sna_pixmap safe -- it may not even exist!
17130 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17132 commit 7f9ed3ca56466f319d9ec06aebb7055552fb1fe1
17133 Author: Chris Wilson <chris@chris-wilson.co.uk>
17134 Date: Sat Dec 24 08:57:49 2011 +0000
17136 sna: reset the pixmap on the initial clear after being swapped out
17138 Pixmaps are swapped out into the CPU after a period of inactivity. This
17139 then prevents the core rendering routines from migrating the pixmap back
17140 to the GPU until it gets used again on a Render path. However, we can
17141 clear that CPU damage and enable migration before a number of key steps
17142 in the expose process.
17144 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17146 commit 8c987b6ccfa83b99a316a7da49a6d7e30fcdf295
17147 Author: Chris Wilson <chris@chris-wilson.co.uk>
17148 Date: Sat Dec 24 09:09:58 2011 +0000
17150 sna: Cheaply reduce damage when subtracting and removing the entire damage
17152 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17154 commit 098592ca5d79af2e4bdcd82ee598c4b2ba08df6b
17155 Author: Chris Wilson <chris@chris-wilson.co.uk>
17156 Date: Sat Dec 24 02:49:24 2011 +0000
17158 sna: Remove the independent tracking of elts from boxes
17160 Following the switch to a global mode for damage, the elts array became
17161 redundant and all that is required is the list of boxes.
17163 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17165 commit 73df0c7ab7c3a9edf0be2439c7e7ab07c0d75ecf
17166 Author: Chris Wilson <chris@chris-wilson.co.uk>
17167 Date: Sat Dec 24 02:48:18 2011 +0000
17169 sna: Tune region upload inplace threshold
17171 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17173 commit ef66c5b5ad221211948ec795ade031591c3f0ac7
17174 Author: Chris Wilson <chris@chris-wilson.co.uk>
17175 Date: Sat Dec 24 00:15:33 2011 +0000
17177 sna: Search the inactive VMA cache first for a linear mapping
17179 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17181 commit 1cc07fa2d24b10ac95c7a84908290ec06539d447
17182 Author: Chris Wilson <chris@chris-wilson.co.uk>
17183 Date: Fri Dec 23 20:31:21 2011 +0000
17185 sna: Avoid forced creation of GPU bo for tiny operations and dirty pixmaps
17187 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17189 commit e5bfea5826117bdd29664d83b3d478ff9db29ca3
17190 Author: Chris Wilson <chris@chris-wilson.co.uk>
17191 Date: Fri Dec 23 20:30:58 2011 +0000
17193 sna: Discard any GPU damage when overwriting with trapezoids
17195 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17197 commit 819dc93e52533afc9b9005db1472672c3d42229f
17198 Author: Chris Wilson <chris@chris-wilson.co.uk>
17199 Date: Fri Dec 23 20:29:59 2011 +0000
17201 sna/glyphs: Create GPU bo for large enough destination surfaces with glyphs
17203 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17205 commit 9580ae8490a4119dac2fefe0085326db350209d7
17206 Author: Chris Wilson <chris@chris-wilson.co.uk>
17207 Date: Fri Dec 23 19:11:43 2011 +0000
17209 sna: Pass usage-hint to move-to-gpu
17211 When simply creating a source GPU bo it is preferrable not to mark it as
17214 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17216 commit 98f15fc61361b7f1e01969f8d4237c13e93e3fb0
17217 Author: Chris Wilson <chris@chris-wilson.co.uk>
17218 Date: Fri Dec 23 17:25:42 2011 +0000
17220 sna: Don't align pwrite to cachelines for doing discontiguous copies
17222 The batch compaction breaks the 1:1 mapping between the cpu buffer and
17223 the bo, so we can no longer safely align the transfer to whole
17226 References: https://bugs.freedesktop.org/show_bug.cgi?id=44091
17227 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17229 commit 3850f4ad48986691e1fb98038ae921deb6c25423
17230 Author: Chris Wilson <chris@chris-wilson.co.uk>
17231 Date: Fri Dec 23 12:42:02 2011 +0000
17233 sna: Silence unsigned comparison against 0
17235 The good news is if this ever failed, the kernel is far too broken...
17237 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17239 commit d9ca113a83177562cde3537bf955164c6950cbd0
17240 Author: Chris Wilson <chris@chris-wilson.co.uk>
17241 Date: Fri Dec 23 10:06:58 2011 +0000
17243 sna: Trim the unused pages from the batch between the last command and surface
17245 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17247 commit 013eda17e91ff421cd1ef70470e3575f1183fcbd
17248 Author: Chris Wilson <chris@chris-wilson.co.uk>
17249 Date: Fri Dec 23 10:40:58 2011 +0000
17251 sna: Free the additional bindings on proxies
17253 commit 84d97bdba02b909369b54de21425ffc9f6ad581a
17254 Author: Chris Wilson <chris@chris-wilson.co.uk>
17255 Date: Fri Dec 23 00:01:50 2011 +0000
17257 sna/gen2+: Reuse source channel for mask where possible
17259 GTK+ has a clever trick for premultiplying its images by loading the
17260 same pixel data into both the source and mask, and then performing the
17261 composite. This causes us to upload the same pixel data twice!
17263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17265 commit 281f620573917faef52d9226b12737ce1e2dffdc
17266 Author: Chris Wilson <chris@chris-wilson.co.uk>
17267 Date: Thu Dec 22 22:23:41 2011 +0000
17269 sna: Age active wholly damaged GPU buffers more slowly
17271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17273 commit f8575b8bd150ae11147fbf6b59171cbbb072dfec
17274 Author: Chris Wilson <chris@chris-wilson.co.uk>
17275 Date: Thu Dec 22 22:15:42 2011 +0000
17277 sna: Cancel the immediate batch submit once handled in the block handler
17279 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17281 commit 609e1d9da282323ab7abc1baad4d3f24ac55588f
17282 Author: Chris Wilson <chris@chris-wilson.co.uk>
17283 Date: Thu Dec 22 15:54:39 2011 +0000
17285 sna: Bump the inactivity timeout
17287 With the goal of removing inactive pixmap froms the GPU after 5 minutes
17290 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17292 commit 62602209e54c3568795122ee1c902c3b81985303
17293 Author: Chris Wilson <chris@chris-wilson.co.uk>
17294 Date: Thu Dec 22 14:20:20 2011 +0000
17296 sna: discard damage-all even for width|height==0 operations
17298 Even if we don't know the extents of the render operation, if the entire
17299 pixmap is damaged we can still reduce the damage tracking.
17301 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17303 commit f68a99a55e1e63d7e76a13a7b77b4bb3f7802b9e
17304 Author: Chris Wilson <chris@chris-wilson.co.uk>
17305 Date: Thu Dec 22 15:04:22 2011 +0000
17307 sna: Mark fresh GPU only bo as being all damaged
17309 Presume that we will not fallback and so treat a GPU only bo (one that
17310 is initially created on the GPU) as being all-damaged. This makes future
17311 operations cheaper as the damage tracking overhead is much reduced, and
17312 the cost of the first readback will mainly be in the synchronisation
17315 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17317 commit 1f2cd536bc712843233de3f5c50802c388d63c72
17318 Author: Chris Wilson <chris@chris-wilson.co.uk>
17319 Date: Thu Dec 22 14:30:04 2011 +0000
17321 sna: Reduce damage after complete solid fills
17323 RenderFillRectangles is often used to initially clear a pixmap after
17324 creation by flooding it with a solid colour, as is PolyRect. We can
17325 reduce further damage operations by checking for this condition.
17327 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17329 commit 7a6a2c5d4cc98da76c73152cb7eafcbdb0c088ec
17330 Author: Chris Wilson <chris@chris-wilson.co.uk>
17331 Date: Thu Dec 22 13:24:07 2011 +0000
17333 sna: Setup the pixmap correctly in order to attach our private
17335 During creation of sna_pixmap we validate that we can use a GPU bo with
17336 the target pixmap. This fails if we pass in a raw pixmap header, so
17337 make sure the scratch pixmap is fully initialised first.
17339 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17341 commit f621b3de841f6037d387ca1439a0abe12ef29811
17342 Author: Chris Wilson <chris@chris-wilson.co.uk>
17343 Date: Thu Dec 22 11:19:45 2011 +0000
17345 sna: flatten source alphamaps
17347 Replace the source picture+alpha with a bo that contains the RGB
17348 channels from source and A from the alpha map.
17350 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17352 commit 6c08eb4d6f8789e692ef018e007d1ae97a57c25f
17353 Author: Chris Wilson <chris@chris-wilson.co.uk>
17354 Date: Thu Dec 22 11:42:39 2011 +0000
17356 sna/gen2+: Prefer to use the CPU if we have a source alphamap and CPU pictures
17358 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17360 commit 354dc3c65b9cb86885b6927ee2b80cd2ee6a82ff
17361 Author: Chris Wilson <chris@chris-wilson.co.uk>
17362 Date: Wed Dec 21 21:35:06 2011 +0000
17364 sna: Avoid fallbacks for convolutions by rendering the convolved texture
17366 If we have no shader support for generic convolutions, we currently
17367 create the convolved texture using pixman. A multipass accumulation
17368 algorithm can be implemented on top of CompositePicture, so try it!
17370 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17372 commit 34cfa15e522ba486d591d37f37d494e399f70354
17373 Author: Chris Wilson <chris@chris-wilson.co.uk>
17374 Date: Wed Dec 21 00:47:47 2011 +0000
17376 sna: Drop suggested tiling parameter when creating bo
17378 The only place where we did anything other than use the default was when
17379 creating a new bo for CopyArea. In that case, basing the choice on the
17380 src GPU bo was not only wrong but a potential segfault.
17382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17384 commit 4d348425592251f49fd31b9259651e8b05e92d1e
17385 Author: Chris Wilson <chris@chris-wilson.co.uk>
17386 Date: Tue Dec 20 21:35:58 2011 +0000
17388 sna: Intiialise pixmap size before use
17390 After 0c12f7cb0 we were setting the width/height of the pixmap *after*
17391 trying to use them to determine if the pixmap could be created on the
17392 GPU. Normally this would be corrected when we attempt to render, except
17393 for the core drawing protocol.
17395 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17397 commit e71058c9c81067cd72b1193242acbd85272ab73a
17398 Author: Chris Wilson <chris@chris-wilson.co.uk>
17399 Date: Tue Dec 20 21:20:48 2011 +0000
17401 sna: Always readback untiled bo in place
17403 We are going to gain no advantage from the kernel detiling the busy
17406 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17408 commit 6690ff51ec939367051985172a00bf8ed65ac4e3
17409 Author: Chris Wilson <chris@chris-wilson.co.uk>
17410 Date: Tue Dec 20 21:09:01 2011 +0000
17412 sna: Treat all exported bo as potentially active
17414 Don't assume that a read/write will clear the active flag if the bo has
17415 been exported to another DRI client.
17417 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17419 commit 743084bbc4c411694190f60cf5161086ada6454f
17420 Author: Chris Wilson <chris@chris-wilson.co.uk>
17421 Date: Tue Dec 20 20:42:08 2011 +0000
17423 sna: Only upload tiny pixmaps on the second use
17425 Most small pixmaps appear to be single shot, so amalgamate them into one
17426 buffer and trim our memory usage.
17428 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17430 commit ca24ffe8b9acef034a5610376c73ebbdce87d41c
17431 Author: Chris Wilson <chris@chris-wilson.co.uk>
17432 Date: Tue Dec 20 20:11:12 2011 +0000
17434 sna: Improve a debug message
17436 Give the units when saying the pixmap is too small to bother tiling.
17438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17440 commit f837b9bcc7d384aeb37c2e9ebdac95571deaedc6
17441 Author: Chris Wilson <chris@chris-wilson.co.uk>
17442 Date: Tue Dec 20 20:06:25 2011 +0000
17444 sna/render: If the pixmap already has a GPU bo, use it for the source
17446 The usage hint not to create a GPU bo is obviously superceded if we
17447 already have a GPU bo.
17449 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17451 commit dfafe76a57655784c313bc2078d0b28a7ac2675b
17452 Author: Chris Wilson <chris@chris-wilson.co.uk>
17453 Date: Tue Dec 20 18:39:20 2011 +0000
17455 sna: Do not use Y-tiling for composite back pixmaps
17457 These are treated by the core drawing routines as replacements for the
17458 front-buffer attached to Windows, and so expect the usual BLT
17459 accelerations are available, for example overlapping blits to handle
17460 scrolling. If we create these pixmaps with Y-tiling and then they are
17461 pinned by the external compositor we are forced to perform a double copy
17462 through the 3D pipeline as it does not handle overlapping blits and the
17463 BLT does not handle Y-tiling.
17465 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17467 commit 0cda7b4fa82870d57fcd0036a3b734cc73e2bf28
17468 Author: Chris Wilson <chris@chris-wilson.co.uk>
17469 Date: Tue Dec 20 17:58:33 2011 +0000
17471 sna: Implement extended fallback handling for src == dst copies
17473 Only marginally better than falling all the way back to using the CPU,
17474 is to perform a double copy to workaround the overlapping copy.
17476 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17478 commit d257a967396c517146cfb12bbec5cd28418752a3
17479 Author: Chris Wilson <chris@chris-wilson.co.uk>
17480 Date: Tue Dec 20 12:53:42 2011 +0000
17482 sna: Explicitly handle depth==1 scratch pixmaps
17484 Short-cut the determination of whether it can be tiled and accelerated
17485 -- we know it can't! This is mainly to cleanup the debug logs.
17487 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17489 commit 0c12f7cb01e75de3bf9c2af8ac6d5b4152566457
17490 Author: Chris Wilson <chris@chris-wilson.co.uk>
17491 Date: Tue Dec 20 10:19:01 2011 +0000
17493 sna: Tidy up some recent valgrind complaints with reuse of scratch pixmaps
17495 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17497 commit d3a4f5db149888239b78df1b9fcc5379b1603005
17498 Author: Chris Wilson <chris@chris-wilson.co.uk>
17499 Date: Tue Dec 20 12:12:39 2011 +0000
17501 sna: Fixup the refcnt to avoid an assert
17503 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17505 commit ac52a1fcd11a3ab10e1843ed2d19f9cf818de4c1
17506 Author: Chris Wilson <chris@chris-wilson.co.uk>
17507 Date: Tue Dec 20 02:20:48 2011 +0000
17509 sna: Don't immediately check for region intersection after subtract
17511 In the READ==0 case we know that the region does not intersect damage
17512 because we have just subtracted, and checking the intersection causes us
17513 to immediately apply the subtraction operation defeating the
17514 optimisation and forcing the expensive operation each time.
17516 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17518 commit 4071dca0ef8b2b1605f16bed5c42991885a35efd
17519 Author: Chris Wilson <chris@chris-wilson.co.uk>
17520 Date: Tue Dec 20 00:25:29 2011 +0000
17522 sna: Don't mark mapping as synchronous by default
17524 Only those that point into scratch memory need to synchronized before
17525 control is handed back to the client.
17527 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17529 commit 4c2a97e9d2b1073f0e4b3f7b6670939e1b1c5121
17530 Author: Chris Wilson <chris@chris-wilson.co.uk>
17531 Date: Mon Dec 19 23:10:45 2011 +0000
17533 sna: Always pass the damage to sna_drawable_use_gpu_bo()
17535 As it now assumes that the damage is always writable.
17537 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17539 commit 06a1792f91d3ae2ebb17e4ad55caec10c5407aa6
17540 Author: Chris Wilson <chris@chris-wilson.co.uk>
17541 Date: Mon Dec 19 21:04:02 2011 +0000
17543 sna: Avoid the GPU readback with READ==0 move_to_cpu
17545 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17547 commit 0b5fec3f80c8d2e4d3b504af5bb02fb76be6b3a6
17548 Author: Chris Wilson <chris@chris-wilson.co.uk>
17549 Date: Mon Dec 19 20:13:05 2011 +0000
17551 sna: Drop the is-mapped flag after operating via the GPU
17553 Mark the end of a sequence of CPU operations and force the decision to
17554 map again to be based on the current upload operation.
17556 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17558 commit 351c8f16336e757b8a4d218b3afda0698781d143
17559 Author: Chris Wilson <chris@chris-wilson.co.uk>
17560 Date: Mon Dec 19 19:51:12 2011 +0000
17562 sna: Discard all damage when replacing pixmap contents
17564 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17566 commit addf66dda7cecaee8e58f53bb4e8d725dfdd2717
17567 Author: Chris Wilson <chris@chris-wilson.co.uk>
17568 Date: Mon Dec 19 18:37:08 2011 +0000
17570 sna: Tweak the rendering priorities
17572 If the last operation was on the GPU, continue on the GPU if this
17573 operation overlaps any GPU damage or does not overlap CPU damage.
17574 Otherwise, if the last operation was on the CPU only switch to the GPU
17575 if we do not overlap any CPU damage and overlap existing GPU damage.
17577 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17579 commit 9b6ade12346efa5b9cc095ad44c7d71880a19ab0
17580 Author: Chris Wilson <chris@chris-wilson.co.uk>
17581 Date: Mon Dec 19 17:21:26 2011 +0000
17583 sna: Create a GPU bo for accelerated core drawing
17585 As we now can accelerate most of the common core drawing operations, we
17586 can create GPU bo for accelerated drawing on first use without undue
17587 fear of readbacks. This benefits Qt especially which heavily uses core
17588 the drawing operations.
17590 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17592 commit eeb81dd6b4651c6b3e8676007d37480888726efa
17593 Author: Chris Wilson <chris@chris-wilson.co.uk>
17594 Date: Mon Dec 19 16:48:02 2011 +0000
17596 sna: Remove the forced inplace upload
17598 Make the decision per-operation, with a tendency to remain mapped.
17600 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17602 commit c3a8d77a2b6806a69de87fbaadea8612ef5f91dd
17603 Author: Chris Wilson <chris@chris-wilson.co.uk>
17604 Date: Mon Dec 19 16:51:29 2011 +0000
17606 sna: Tune the inplace cross-over point to be half-cache size
17608 The theory being that we will also require cache space to copy from when
17609 uploading into the shadow.
17611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17613 commit d53b1b28951fd7558e9913a2973c17983e0bcecf
17614 Author: Chris Wilson <chris@chris-wilson.co.uk>
17615 Date: Mon Dec 19 15:33:39 2011 +0000
17617 configure: Bump the required pixman version
17619 UXA now also uses pixman_triangle_t in order for its fallback, so we
17620 need to bump the required pixman version for UXA as well as SNA.
17622 Reported-by: Fabio Pedretti <fabio.ped@libero.it>
17623 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43946
17624 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17626 commit 1fa5721f064a8d1f34e4032b52f24597f4015313
17627 Author: Chris Wilson <chris@chris-wilson.co.uk>
17628 Date: Mon Dec 19 00:37:03 2011 +0000
17630 sna: Reset the GTT mapping flag when freeing the shadow pointers
17632 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17634 commit 7326d3098662688b8040c6e1261064caed1f5d06
17635 Author: Chris Wilson <chris@chris-wilson.co.uk>
17636 Date: Mon Dec 19 00:35:14 2011 +0000
17638 sna: Restore CPU shadow after a GTT memory
17640 When mixing operations and switching between GTT and CPU mappings we
17641 need to restore the original CPU shadow rather than accidentally
17644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17646 commit ae32aaf4b20dafef138dc9c28dbddbfe49f24b83
17647 Author: Chris Wilson <chris@chris-wilson.co.uk>
17648 Date: Mon Dec 19 00:34:12 2011 +0000
17650 sna/gen[23]: We need to check the batch before doing an inline flush
17652 A missing check before emitting a dword into the batch opened up the
17653 possibility of overflowing the batch and corrupting our state.
17655 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17657 commit e32ad646762ccc7f22f938454e222d43abfb38ed
17658 Author: Chris Wilson <chris@chris-wilson.co.uk>
17659 Date: Sun Dec 18 23:42:07 2011 +0000
17661 sna: Continue searching the linear lists for CPU mappings
17663 Prefer to reuse an available CPU mapping which are considered precious
17664 and reaped if we keep too many unused entries availabled.
17666 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17668 commit 15a769a66fa1afbcffc642ef980387cffefc6bef
17669 Author: Chris Wilson <chris@chris-wilson.co.uk>
17670 Date: Sun Dec 18 22:18:02 2011 +0000
17672 sna: Distinguish between GTT and CPU maps when searching for VMA
17674 Similarly try to avoid borrowing any vma when all we intend to do is
17677 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17679 commit d0ee695ef091671e2cc69b773f517030ebe961b2
17680 Author: Chris Wilson <chris@chris-wilson.co.uk>
17681 Date: Sun Dec 18 20:28:18 2011 +0000
17683 sna: the active cache is not marked as purgeable, so skip checking it
17685 Otherwise we do a lot of list walking for no-ops.
17687 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17689 commit 8df9653135c6390ed699ba9f8cbf374b2cc84396
17690 Author: Chris Wilson <chris@chris-wilson.co.uk>
17691 Date: Sun Dec 18 19:53:14 2011 +0000
17693 sna: clear the request list when reusing a flushing bo
17695 That the rq is NULL when on the flushing list is no longer true, but
17696 now it points to the static request instead.
17698 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17700 commit b51e3de66292e4b29483324e1dbfeb2495817351
17701 Author: Chris Wilson <chris@chris-wilson.co.uk>
17702 Date: Sun Dec 18 19:51:17 2011 +0000
17704 sna: When freeing vma, first see if the kernel evicted any
17706 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17708 commit fed8d145c148bfa8a8a29f4088902377f9a10440
17709 Author: Chris Wilson <chris@chris-wilson.co.uk>
17710 Date: Sun Dec 18 19:26:38 2011 +0000
17712 sna: Use a safe iterator whilst searching for inactive linear bo
17714 As we may free a purged bo whilst iterating, we need to keep the next bo
17717 Include the debugging that led to this find.
17719 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17721 commit 2a98dabcabf25067abcda60f233656e19e83493a
17722 Author: Chris Wilson <chris@chris-wilson.co.uk>
17723 Date: Sun Dec 18 18:48:30 2011 +0000
17725 sna: Purge all evicted bo
17727 After we find a single bo that has been evicted from our cache, the
17728 kernel is likely to have evicted many more so check our caches for any
17731 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17733 commit 8ae105b2c7a06fc81757f06ec1aaa2447b53498a
17734 Author: Chris Wilson <chris@chris-wilson.co.uk>
17735 Date: Sun Dec 18 18:20:11 2011 +0000
17737 sna: Only retire for the VMA search if there are cached VMA
17739 If there are no VMA that might become inactive, there is no point
17740 scanning the inactive lists if we are searching for VMA.
17742 This prevents the regression in firefox-fishbowl whilst maintaining most
17743 of the improvement with PutComposite.
17745 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17747 commit a0c0a3765ca348c74096fb157885da5b1258ee08
17748 Author: Chris Wilson <chris@chris-wilson.co.uk>
17749 Date: Sun Dec 18 16:59:15 2011 +0000
17751 sna: Retire if the inactive vma list is empty
17753 Try to recycle vma by first trying to populate the inactive list before
17754 scanning for a vma bo to harvest.
17756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17758 commit 34efb7314612cedcda4f866bc33f3ad5b6929ae2
17759 Author: Chris Wilson <chris@chris-wilson.co.uk>
17760 Date: Sun Dec 18 16:58:04 2011 +0000
17762 sna: Hint likely usage of CPU bo
17764 If we are going to transfer GPU damage back to the CPU bo, then we can
17765 reuse an active buffer and so improve the recycling.
17767 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17769 commit 301896743867de0cbd22063c4bf2e1a8cf491656
17770 Author: Chris Wilson <chris@chris-wilson.co.uk>
17771 Date: Sun Dec 18 16:17:04 2011 +0000
17773 sna: Only upload to the source GPU bo if we need tiling to avoid TLB misses
17775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17777 commit b7f5d75aa54db7d0542b09f4b3081f7692ede602
17778 Author: Chris Wilson <chris@chris-wilson.co.uk>
17779 Date: Sun Dec 18 16:11:34 2011 +0000
17781 Silence uxa-only compilation
17783 Kill the stray warning for the undeclared extern used by the module
17786 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17788 commit a73cc4bf1e554806f403c6704d1cf98491f4d444
17789 Author: Chris Wilson <chris@chris-wilson.co.uk>
17790 Date: Sun Dec 18 14:54:12 2011 +0000
17792 sna/gen5: Tidy checking against hardcoded maximum 3D size
17794 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17796 commit b43548af39f8773283b744e979ee575032183cdc
17797 Author: Chris Wilson <chris@chris-wilson.co.uk>
17798 Date: Sun Dec 18 12:11:01 2011 +0000
17800 sna: Explicitly handle errors from madv
17802 In order to avoid conflating whether a bo was marked purgeable with its
17803 retained state, we need to carefully handle the errors from madv.
17805 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17807 commit 954cf5129d462f7f4ff5d4ff44b256e0f091667c
17808 Author: Chris Wilson <chris@chris-wilson.co.uk>
17809 Date: Sun Dec 18 10:40:35 2011 +0000
17811 sna/gen[67]: check for context switch after preparing source
17813 If we used the BLT to prepare the source, see if we can continue the
17814 operation on the BLT.
17816 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17818 commit 90a432431c8fc293db611289c3369669c2975706
17819 Author: Chris Wilson <chris@chris-wilson.co.uk>
17820 Date: Sun Dec 18 10:23:11 2011 +0000
17822 sna/gen[23]: Try BLT if the source/target do no fit in the 3D pipeline
17824 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17826 commit eeb9741981be528d1fba3021a30ab24084a5ce8c
17827 Author: Chris Wilson <chris@chris-wilson.co.uk>
17828 Date: Sun Dec 18 10:04:08 2011 +0000
17830 sna/gen3: Tidy checks against hardcoded maximum 3D pipeline size
17832 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17834 commit dcfcf438a511ee43e5969d01e41b3040a08cac3a
17835 Author: Chris Wilson <chris@chris-wilson.co.uk>
17836 Date: Sun Dec 18 09:58:08 2011 +0000
17838 sna/gen2+: If we use the BLT to prepare the target, try using BLT for op
17840 If we incurred a context switch to the BLT in order to prepare the
17841 target (uploading damage for instance), we should recheck whether we can
17842 continue the operation on the BLT rather than force a switch back to
17845 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17847 commit 507debe8015ee10734a0b8a6ddf4aa251b7dddc5
17848 Author: Chris Wilson <chris@chris-wilson.co.uk>
17849 Date: Sun Dec 18 01:45:47 2011 +0000
17851 sna/gen5: If we need to flush the composite op, check to see if we can blit
17853 If we need to halt the 3D engine in order to flush the pipeline for a
17854 dirty source, we may as well re-evaluate whether we can use the BLT
17857 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17859 commit de530f89a36a80b652ed01001e6f0d4c8b1dc11a
17860 Author: Chris Wilson <chris@chris-wilson.co.uk>
17861 Date: Sun Dec 18 01:41:03 2011 +0000
17863 sna/gen5+: First try a blt composite if the source/dest are too large
17865 If we will need to extract either the source or the destination, we
17866 should see if we can do the entire operation on the BLT.
17868 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17870 commit 7b88f87945f23b3f80ae44b2aa19f00dff1904c8
17871 Author: Chris Wilson <chris@chris-wilson.co.uk>
17872 Date: Sun Dec 18 00:44:00 2011 +0000
17874 sna: Upload images in place from CopyArea
17876 As for PutImage, if the damage will be immediately flushed out to the
17877 GPU bo, we may as well do the write directly to the GPU bo and not
17878 staged via the shadow.
17880 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17882 commit 1418e4f3156a17ff6c5cd0f653918160e30b7c0e
17883 Author: Chris Wilson <chris@chris-wilson.co.uk>
17884 Date: Sat Dec 17 23:11:03 2011 +0000
17886 sna: Tune the default pixmap upload paths
17888 One issue with the heuristic is that it is based on total pixmap size
17889 whereas the goal is to pick the placement for the next series of
17890 operations. The next step in refinement is to combine an overall
17891 placement to avoid frequent migrations along with a per-operation
17894 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17896 commit 25c353503a25d20e7db5acdc63d83564804efdf4
17897 Author: Chris Wilson <chris@chris-wilson.co.uk>
17898 Date: Sat Dec 17 17:16:07 2011 +0000
17900 sna: Simplify write domain tracking
17902 Replace the growing bitfield with an enum marking where it was last
17905 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17907 commit d20d167a753d8e4fe581950e1bc49f29e0ec9f1f
17908 Author: Chris Wilson <chris@chris-wilson.co.uk>
17909 Date: Sat Dec 17 16:28:04 2011 +0000
17911 sna: Upload to large pixmaps inplace
17913 When the pixmap is large, larger than L2 cache size, we are unlikely to
17914 benefit from first copying the data to a shadow buffer -- as that shadow
17915 buffer itself will mostly reside in main memory. In such circumstances
17916 we may as perform the write to the GTT mapping of the GPU bo. As such,
17917 it is a fragile heuristic that may require further tuning.
17919 Avoiding that extra copy gives a 30% boost to putimage500/shmput500 at
17920 ~10% cost to putimage10/shmput10 on Atom (945gm/PineView), without any
17921 noticeable impact upon cairo.
17923 Reported-by: Michael Larabel <Michael@phoronix.com>
17924 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17926 commit dd8fd6c90612ada39eb32b98adc5acc97e7902aa
17927 Author: Chris Wilson <chris@chris-wilson.co.uk>
17928 Date: Sat Dec 17 12:38:09 2011 +0000
17930 sna: Search through the inactive VMA cache for potential upload bo
17932 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17934 commit 8ef5d8c1955e2e2ee19c64730f600639ac42de55
17935 Author: Chris Wilson <chris@chris-wilson.co.uk>
17936 Date: Thu Dec 15 17:52:20 2011 +0000
17938 sna: Map the upload buffer using an LLC bo
17940 In order to avoid having to perform a copy of the cacheable buffer into
17941 GPU space, we can map a bo as cacheable and write directly to its
17942 contents. This is only a win on systems that can avoid the clflush, and
17943 also we have to go to greater measures to avoid unnecessary
17944 serialisation upon that CPU bo. Sadly, we do not yet go to enough length
17945 to avoid negatively impacting ShmPutImage, but that does not appear to
17946 be a artefact of stalling upon a CPU buffer.
17948 Note, LLC is a SandyBridge feature enabled by default in kernel 3.1 and
17949 later. In time, we should be able to expose similar support for
17950 snoopable buffers for other generations.
17952 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17954 commit 6e47f283711d122d96384a1a82854c11644e6d68
17955 Author: Chris Wilson <chris@chris-wilson.co.uk>
17956 Date: Fri Dec 16 23:49:18 2011 +0000
17958 sna/gen3: Enforce a minimum width of 2 elements for the render target
17960 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17962 commit 2ff0826f94ca16e95cd662385f7091be750dec30
17963 Author: Chris Wilson <chris@chris-wilson.co.uk>
17964 Date: Fri Dec 16 23:33:25 2011 +0000
17966 sna: Discard GPU damage first before choosing where to fill_boxes()
17968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17970 commit 55520bab578865f878965aa362ec4933f4b26050
17971 Author: Chris Wilson <chris@chris-wilson.co.uk>
17972 Date: Fri Dec 16 23:22:38 2011 +0000
17974 sna/gen3: Initialise missing value of need ca pass for fill_boxeS()
17976 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17978 commit e56d5081ea2da930bc82036ac7c4c78b433117ed
17979 Author: Chris Wilson <chris@chris-wilson.co.uk>
17980 Date: Fri Dec 16 23:18:02 2011 +0000
17982 sna: Wrap I915_GEM_GET_PARAM with valgrind markup
17984 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17986 commit e0399ec1619fe4f87f0578791e697a7e2a8c86dc
17987 Author: Chris Wilson <chris@chris-wilson.co.uk>
17988 Date: Fri Dec 16 23:07:37 2011 +0000
17990 sna: Suppress an overwritten XY_SRC_COPY
17992 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
17994 commit 1684ed6a5e8c26ecb48cc1a5025107466526fe94
17995 Author: Chris Wilson <chris@chris-wilson.co.uk>
17996 Date: Fri Dec 16 19:35:35 2011 +0000
17998 sna: Clean up compiler warnings for shadowed variables
18000 No outright bug, just plenty of unwanted noise.
18002 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18004 commit 300586b229ae941ac31850af00d120d8c441c583
18005 Author: Chris Wilson <chris@chris-wilson.co.uk>
18006 Date: Fri Dec 16 17:56:53 2011 +0000
18008 sna/gen4+: disable the blend unit for PictOpSrc
18010 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18012 commit 0de7604d8ecfc73ef9e92059340d00ce5cfcdd75
18013 Author: Chris Wilson <chris@chris-wilson.co.uk>
18014 Date: Fri Dec 16 17:40:01 2011 +0000
18016 src/gen4+: Add support for depth 15 render copies/fills
18018 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18020 commit a8fe50ab6503d9cb63931771318dc3e84d002092
18021 Author: Chris Wilson <chris@chris-wilson.co.uk>
18022 Date: Sat Dec 17 11:15:36 2011 +0000
18024 uxa: Explicitly check for libdrm_intel in configure
18026 And remove the excess dependencies from the common files.
18028 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18030 commit 3320b459d6fe90d1145e3874c840b95279fb16c8
18031 Author: Chris Wilson <chris@chris-wilson.co.uk>
18032 Date: Sat Dec 17 12:08:24 2011 +0000
18034 sna: Clean up caches stored upon the batch bo
18036 Until the advent of the VMA cache, we were safe to reap the batch bo by
18037 hand. However, as we continue to add additional data to the bo, it is
18038 wise to use the common free function.
18040 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
18041 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43899
18042 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18044 commit e2c8bac972cf46982bf64bd786ecc001009081f0
18045 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
18046 Date: Fri Dec 16 15:11:16 2011 +0800
18048 uxa/glamor: Fallback to new glamor pixmap if failed to create textured pixmap.
18050 If we failed to create textured pixmap from BO's handle, we
18051 turn to create a new glamor pixmap by call glamor_create_pixmap
18052 rather than fallback to in-memory pixmap. Have to introduce
18053 a new wrapper function intel_glamor_create_pixmap.
18055 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
18057 commit 19c184b7e4f8de747ed6fb1f6f910238193cf2a1
18058 Author: Chris Wilson <chris@chris-wilson.co.uk>
18059 Date: Thu Dec 15 18:18:19 2011 +0000
18061 sna/gen3: Check for upload failure of video bo
18063 And propagate that failure back to the client.
18065 Reported-by: Paul Neumann <paul104x@yahoo.de>
18066 References: https://bugs.freedesktop.org/show_bug.cgi?id=43716
18067 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18069 commit e68a8748f3418179eb41e8d81ba1050b883b7e47
18070 Author: Chris Wilson <chris@chris-wilson.co.uk>
18071 Date: Thu Dec 15 09:44:20 2011 +0000
18073 uxa: Move the region creation beyond the call into glamor
18075 So that we avoid leaking the region if hooking into glamor.
18077 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18079 commit e3153a779ed220ad71958ce56be179d458794964
18080 Author: Chris Wilson <chris@chris-wilson.co.uk>
18081 Date: Thu Dec 15 09:30:28 2011 +0000
18083 sna: cap the number of times we attempt to resubmit the batch upon EBUSY
18085 Just in case the kernel gets stuck in a loop and we fail to make any
18088 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18090 commit 1cc43dc97bd6438d7fc808cb5ee449889a966bdf
18091 Author: Chris Wilson <chris@chris-wilson.co.uk>
18092 Date: Thu Dec 15 00:00:56 2011 +0000
18094 sna: More missing move-to-cpu allocation checks
18096 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18098 commit 12448b56060fb9c14da6b4ab02c95c7834fd3874
18099 Author: Chris Wilson <chris@chris-wilson.co.uk>
18100 Date: Thu Dec 15 00:01:37 2011 +0000
18102 sna: silence warning for unused 'priv'
18104 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18106 commit eb8e979b4de5d854f15ae4c86d3db29371b386c1
18107 Author: Chris Wilson <chris@chris-wilson.co.uk>
18108 Date: Wed Dec 14 23:02:28 2011 +0000
18110 uxa/glamor: Allocate a fbPixmap with storage for fallbacks
18112 When we try to create a glamor pixmap and fail we need to create a real
18113 pixmap along with its pixel allocation, instead of detaching ourselves
18114 and returning the fake pixmap header.
18116 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18118 commit 232fa93a4bf1760e16e913e8fb4d5159c4be9295
18119 Author: Chris Wilson <chris@chris-wilson.co.uk>
18120 Date: Wed Dec 14 20:11:39 2011 +0000
18122 uxa/glamor: Always notify glamor that the glyph has been unrealized
18124 The danger of the early return when UXA is not using glyphs is evident
18125 in the eventual crash when glamor begins evicting and reusing its glyph
18128 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18130 commit 95cceb5ae5503af0ac50a923fa47e134f0da8743
18131 Author: Chris Wilson <chris@chris-wilson.co.uk>
18132 Date: Wed Dec 14 19:27:53 2011 +0000
18134 sna: Fix DBG crash whilst pruning inactive GPU buffers
18136 Don't attempt to dereference the NULL gpu_bo after having just freed it.
18137 Here in lies the folly of trying to blindly silence the compiler.
18139 Instead we should heed the error return as it means that we didn't
18140 decouple the pixmap from the inactive list and so we choose to place it
18141 back on the active list to purge again in the near future.
18143 Reported-by: Paul Neumann <paul104x@yahoo.de
18144 References: https://bugs.freedesktop.org/show_bug.cgi?id=43716
18145 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18147 commit e7f4b7fd91a41cac77e5eb1cb4f185141b09a09e
18148 Author: Chris Wilson <chris@chris-wilson.co.uk>
18149 Date: Wed Dec 14 18:21:22 2011 +0000
18151 sna: Add some DBG() around Y-to-X fallbacks
18153 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18155 commit 32bb2c89b8cab92f16af32e7a084bfefa406661d
18156 Author: Chris Wilson <chris@chris-wilson.co.uk>
18157 Date: Wed Dec 14 16:28:54 2011 +0000
18159 sna: Check allocation of pixman_image_t
18161 And just fail to perform the copy, clearing the dst instead.
18163 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18165 commit 37c525a11c0fab4e279e80ad24e5cf791239e005
18166 Author: Chris Wilson <chris@chris-wilson.co.uk>
18167 Date: Wed Dec 14 16:26:46 2011 +0000
18169 sna: Skip glyphs if we fail to allocate pixel data for them
18171 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18173 commit 6a8188bb4dd0eba58e29803566023ad5c3d4d58a
18174 Author: Chris Wilson <chris@chris-wilson.co.uk>
18175 Date: Wed Dec 14 16:24:54 2011 +0000
18177 sna: Close any handles after bo allocation failures
18179 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18181 commit 43a22743124a83310379122d509c35840b583b2e
18182 Author: Chris Wilson <chris@chris-wilson.co.uk>
18183 Date: Wed Dec 14 13:26:05 2011 +0000
18185 sna: Use the provided bo for blitting rather than assume priv->gpu_bo
18187 Reported-by: nkalkhof@web.de
18188 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43802
18189 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18191 commit 30f5ee11f8ec3688807bbaded92561e96f9a439b
18192 Author: Chris Wilson <chris@chris-wilson.co.uk>
18193 Date: Wed Dec 14 12:30:40 2011 +0000
18195 sna: Use a static request and synchronous rendering in case of malloc failure
18197 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18199 commit 23fb2cebbe1d4d7df7403a64635339efa08a4dfe
18200 Author: Chris Wilson <chris@chris-wilson.co.uk>
18201 Date: Wed Dec 14 12:34:03 2011 +0000
18203 sna/blt: Add a missing allocation check upon the source CPU pixmap
18205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18207 commit 5b0f3ff9a83b7cc932e96400999dc5e4a49369cb
18208 Author: Chris Wilson <chris@chris-wilson.co.uk>
18209 Date: Wed Dec 14 12:23:04 2011 +0000
18211 sna/damage: Guard against malloc failures
18213 In the event of failure, we choose to loose track of the damage and
18214 choose rendering corruption over crashing.
18216 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18218 commit f8b9a2a5473efd0fefb3d2878e70c24c320245ca
18219 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
18220 Date: Tue Dec 13 21:46:14 2011 +0800
18222 uxa/glamor: Enable the rest of the glamor rendering routines
18224 This commit hooks up all the remaining rendering routines to call into
18225 glamor; the takeover is nearly complete! When tested with the latest
18226 glamor master branch, it passes rendercheck.
18228 One thing need to be pointed out is the picture's handling.
18229 Pictures support many different color formats, but glamor's
18230 texture only support a few color formats. And the most common
18231 scenario is that we create a pixmap with a color depth and
18232 then attach it to a picture which has a specific color format
18233 with the same color depth. But there is no way to change a
18234 texture's internal format after the texture was allocated.
18235 If you do that, the OpenGL will allocate a new texture. And
18236 then the glamor side and UXA side will be inconsitent. So
18237 for all the picture related operations, we can't fallback to
18238 UXA path directly, even it is rather a straight forward
18239 operation. So for the get_image, Addtraps.., we have to add
18240 wrappers function for them to jump into glamor firstly.
18242 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
18243 [ickle: prefer access; ok = glamor(); finish; if (!ok) goto fallback; return; ]
18244 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18246 commit d5456e40d9bebc9d6a121abdb124a90a2a8a6958
18247 Author: Chris Wilson <chris@chris-wilson.co.uk>
18248 Date: Wed Dec 14 10:41:33 2011 +0000
18250 uxa/glamor: Silence a compiler warning for some unused code
18252 intel_glamor.c: In function 'intel_glamor_create_screen_image':
18253 intel_glamor.c:192:12: warning: variable 'pixmap' set but not used
18254 [-Wunused-but-set-variable]
18256 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18258 commit 4f1a99a70e76ea5637c5ee8226b2e52a464f5948
18259 Author: Chris Wilson <chris@chris-wilson.co.uk>
18260 Date: Tue Dec 13 19:44:15 2011 +0000
18262 sna: Protect against deferred malloc failures for pixel data
18264 As we now defer the allocation of pixel data until first use, it can
18265 fail in the middle of a rendering routine. In order to prevent chasing
18266 us passing a NULL pointer into the fallback routines, we need to propagate
18267 the failure from the malloc and suppress the failure, discarding the
18268 operation, which is less than ideal.
18270 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18272 commit d2c6d950ed2c5882e7d501b6974e72be4d6da8a8
18273 Author: Chris Wilson <chris@chris-wilson.co.uk>
18274 Date: Tue Dec 13 15:04:10 2011 +0000
18276 sna: Mark upload buffers as unaccessible upon submission
18278 Use valgrind to catch use-after-finish bugs.
18280 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18282 commit e39ea29bcc5d1d76dd5cb3db555516ef848647a5
18283 Author: Chris Wilson <chris@chris-wilson.co.uk>
18284 Date: Tue Dec 13 13:18:56 2011 +0000
18286 sna: Allow the debugger to map bo from the batch during kgem_submit()
18288 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18290 commit 2fabb5068d36ce97242844066b5121fcf6d9c24e
18291 Author: Chris Wilson <chris@chris-wilson.co.uk>
18292 Date: Tue Dec 13 13:15:51 2011 +0000
18294 sna: Debug fixup for non-LLC systems
18296 The cpu bo is only allocated on LLC systems, so do avoid the NULL deref on
18297 debugging for others.
18299 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18301 commit e037379c8e57f8e42956863c10a4e2b18057a6d5
18302 Author: Chris Wilson <chris@chris-wilson.co.uk>
18303 Date: Tue Dec 13 13:14:15 2011 +0000
18305 sna: Fix a debugging assert
18307 The bo is allowed to be NULL, so defer the assert until after it is
18308 known to be non-NULL.
18310 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18312 commit 5d5b2b8ee203ae2274fc7d13ed38d2945facca9e
18313 Author: Chris Wilson <chris@chris-wilson.co.uk>
18314 Date: Tue Dec 13 11:30:05 2011 +0000
18316 uxa: Cap the maximum number of VMA cached
18318 Since we can not keep an unlimited number of vma cached due to the hard
18319 per-process limits on the number of mappings and recreating mappings is
18320 slow due to excruciatingly slow GTT pagefaults, we need to compromise
18321 and keep a small MRU cache of inactive mmaps.
18323 This uses the new API in libdrm-2.4.29 to specify the limit upon the VMA
18324 cache maintained by libdrm.
18326 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18328 commit 1128825efb3179a5a5d951fa24db6b769ee41219
18329 Author: Chris Wilson <chris@chris-wilson.co.uk>
18330 Date: Tue Dec 13 11:20:25 2011 +0000
18332 uxa: Wakeup 3s after the last rendering to reap the bo-cache
18334 libdrm expires its bo 2s after entry into the cache, but we need to free
18335 a buffer to trigger the reaper. So schedule a timer event to trigger 3s
18336 after the last rendering is submitted to free any resident bo during
18337 long periods of idleness.
18339 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18341 commit db7c9e8561afcc1ca7ab16b3bf2d5b49938e26d9
18342 Author: Chris Wilson <chris@chris-wilson.co.uk>
18343 Date: Tue Dec 13 10:05:25 2011 +0000
18345 configure: Link the extra valgrind debugging to --enable-debug
18347 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18349 commit d02dc0fd84194b6acc607d17e4396dd1fdc681f8
18350 Author: Chris Wilson <chris@chris-wilson.co.uk>
18351 Date: Tue Dec 13 01:18:59 2011 +0000
18353 sna: Set the refcnt on the replacement bo
18355 The paranoia wasn't in vain.
18357 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18359 commit 7472db8c8c9994798ea36de1ce2f51725d8b387a
18360 Author: Chris Wilson <chris@chris-wilson.co.uk>
18361 Date: Mon Dec 12 21:55:04 2011 +0000
18363 sna: Double-check that the submitted buffers were not purged
18365 More paranoia is good for the soul.
18367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18369 commit 0bbd6a08fe485c80bd5c9a1b7027618a03a26f84
18370 Author: Chris Wilson <chris@chris-wilson.co.uk>
18371 Date: Mon Dec 12 21:52:56 2011 +0000
18373 sna/gen2: Tidy checking against too large pixmaps for the 3D pipeline
18375 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18377 commit b392474f3a052fefc410383725b64f0fdbe7788e
18378 Author: Chris Wilson <chris@chris-wilson.co.uk>
18379 Date: Mon Dec 12 21:28:09 2011 +0000
18381 sna: Force a suitable minimum stride for 3D temporaries
18383 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18385 commit 3c22baaba97a89d24e507e3d832d53d764902040
18386 Author: Chris Wilson <chris@chris-wilson.co.uk>
18387 Date: Mon Dec 12 21:16:56 2011 +0000
18389 sna/gen2: Check for unhandled pitches in the render pipeline
18391 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18393 commit f6a30df8dc75852d1f6daeabe4511b275594e52e
18394 Author: Chris Wilson <chris@chris-wilson.co.uk>
18395 Date: Mon Dec 12 20:40:29 2011 +0000
18397 sna: Enable memcpy uploads to SHM pixmaps
18399 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18401 commit 3c163d105e964a1084d665500ef917254d8f2179
18402 Author: Chris Wilson <chris@chris-wilson.co.uk>
18403 Date: Mon Dec 12 16:26:13 2011 +0000
18405 sna: Use the CPU bo as a render source if compatible and no GPU bo
18407 This is principally to catch the cases of compositing after a fresh
18410 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18412 commit c481bec356b2e40e66a000dbaaf261bf7aae930d
18413 Author: Chris Wilson <chris@chris-wilson.co.uk>
18414 Date: Mon Dec 12 14:10:57 2011 +0000
18416 sna: Experiment with creating the CPU pixmap using an LLC BO
18418 A poor cousin to vmap is to instead allocate snooped bo and use a CPU
18419 mapping for zero-copy uploads into GPU resident memory. For maximum
18420 performance, we still need tiled GPU buffers so CPU bo are only useful
18421 in situations where we are frequently migrating data.
18423 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18425 commit 6c9aa6f9cf8e59ca6aa1866b83690a1de8cfb757
18426 Author: Chris Wilson <chris@chris-wilson.co.uk>
18427 Date: Mon Dec 12 11:47:25 2011 +0000
18429 sna: Defer allocation of memory for larger pixmap until first use
18431 In the happy scenario where the pixmap only resides upon the GPU we can
18432 forgo the CPU allocation entirely. The goal is to reduce the number of
18433 needless mmaps performed by the system memory allocator and reduce
18434 overall memory consumption.
18436 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18438 commit 4b48d28f6e85f345730cb7242f2152ac115b267a
18439 Author: Chris Wilson <chris@chris-wilson.co.uk>
18440 Date: Mon Dec 12 10:52:34 2011 +0000
18442 sna: Fix a typo, end statements with semi-colons
18444 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18446 commit 4d20798c7871ffe8581e2cf509b6aa2e40b3ae5f
18447 Author: Chris Wilson <chris@chris-wilson.co.uk>
18448 Date: Mon Dec 12 10:10:16 2011 +0000
18450 sna: We need to remap the gpu_only mmap prior to every use
18452 Since the VMA may be reaped at any time whilst the mapping is idle.
18454 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18456 commit 2682308c10c9622247409dee9c6bb6a862b73c3f
18457 Author: Chris Wilson <chris@chris-wilson.co.uk>
18458 Date: Mon Dec 12 09:24:11 2011 +0000
18460 sna: Remove bo transference for whole XCopyArea
18462 In benchmarking firefox this performs whose - it would appear the
18463 sources are indeed used more often than not.
18465 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18467 commit 7703424222f8bd08450c1b109ea64c6b0ee901d8
18468 Author: Chris Wilson <chris@chris-wilson.co.uk>
18469 Date: Mon Dec 12 00:15:10 2011 +0000
18471 sna/gen6: Only use CPU bo for a render target if untiled
18473 For large render targets, we prefer to use tiled bo in order to avoid
18474 severe performance degradation. However, if we don't have a GPU bo but
18475 do have a CPU bo and the operation would be untiled, then simply use the
18478 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18480 commit a92a41ba3291a4304948ed1961d1242ca7d3981e
18481 Author: Chris Wilson <chris@chris-wilson.co.uk>
18482 Date: Sun Dec 11 19:03:54 2011 +0000
18484 sna/gen6: Tidy the usage of the max pipeline size
18486 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18488 commit e9e6d6f7c8af300716c64b041f96c09d58b3eaa2
18489 Author: Chris Wilson <chris@chris-wilson.co.uk>
18490 Date: Sun Dec 11 17:25:46 2011 +0000
18492 sna/gen3: Move the video dst_bo to make the conditional clearer
18494 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18496 commit 118ef0781c5064aa0696af59a5dac67091b10046
18497 Author: Chris Wilson <chris@chris-wilson.co.uk>
18498 Date: Sun Dec 11 16:56:53 2011 +0000
18500 sna/composite: Make the check for a no-op earlier and clearer
18502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18504 commit 2674ef864c93034d728ee11ee0b7ab96cd4cba57
18505 Author: Chris Wilson <chris@chris-wilson.co.uk>
18506 Date: Sun Dec 11 16:23:13 2011 +0000
18508 sna: Enable hooking up of valgrind during debugging
18510 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18512 commit c83fd4e24de187700bc125037cd9ae311514213e
18513 Author: Chris Wilson <chris@chris-wilson.co.uk>
18514 Date: Sun Dec 11 16:14:38 2011 +0000
18516 sna: Add some more debug messages for VMA caching
18518 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18520 commit 3ae7fb918a2ca47d3944521a1ae30143dcb14eeb
18521 Author: Chris Wilson <chris@chris-wilson.co.uk>
18522 Date: Sun Dec 11 13:52:42 2011 +0000
18524 sna: Restrict pitch alignment on 945gm to 64 bytes
18526 In theory we should be able to disable dual-stream mode and so be
18527 subject to much looser restrictions (such as the pitch need only be
18528 dword aligned). However, achieving single-stream mode seems quite
18531 Reported-by: Paul Neumann <paul104x@yahoo.de>
18532 References: https://bugs.freedesktop.org/show_bug.cgi?id=43706
18533 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18535 commit 2f35d77cd07a4974051c922f4809ce44939b0556
18536 Author: Chris Wilson <chris@chris-wilson.co.uk>
18537 Date: Sun Dec 11 13:37:18 2011 +0000
18539 sna: Update computation of untiled pitch to cater for CREATE_SCANOUT
18541 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18543 commit 5a0139487f5a5caada56121395aa166fc9f08663
18544 Author: Chris Wilson <chris@chris-wilson.co.uk>
18545 Date: Sun Dec 11 12:34:40 2011 +0000
18547 sna/gen3: Ensure that depth read/writes are disabled before first use
18549 Our goal is to achieve "single-stream" rendering where the entire
18550 RenderCache is allocated to the colour buffer (rather than split between
18551 colour and depth). In theory all that is required is for the pipeline
18552 not to reference the depth buffer at all, however it is not made clear
18553 when that evaluation is made.
18555 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18557 commit a02bbd8700690c98c6cf3ae98dd7ee3da9887b0a
18558 Author: Chris Wilson <chris@chris-wilson.co.uk>
18559 Date: Sun Dec 11 10:34:37 2011 +0000
18561 sna: Only transfer the bo if the src/dst are of matching size
18563 If the src replaces the dst, it could just be a much larger pixmap!
18565 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18567 commit 43a99648635fa76e0d6d71acaeb6ba919d24eb26
18568 Author: Chris Wilson <chris@chris-wilson.co.uk>
18569 Date: Sun Dec 11 10:30:48 2011 +0000
18571 sna: Only transfer unpinned buffers
18573 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18575 commit eb859f644633ee716083d253a5b7ff95163380e5
18576 Author: Chris Wilson <chris@chris-wilson.co.uk>
18577 Date: Sun Dec 11 01:34:05 2011 +0000
18579 uxa/video: Correct the offset of the binding table in the surface buffer
18581 The binding table is intended to be after all the surface descriptions,
18582 so make sure we write it with the appropriate offset into the buffer.
18584 Fixes regression from 699888a64 (uxa/video: Use the common bo
18585 allocations and upload)
18587 Reported-by: Cyril Brulebois <kibi@debian.org>
18588 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43704
18589 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18591 commit 051a18063df075536cb1ac0dc4dfc3c1306ab74e
18592 Author: Chris Wilson <chris@chris-wilson.co.uk>
18593 Date: Sat Dec 10 22:45:25 2011 +0000
18595 sna: Implement a VMA cache
18597 A VMA cache appears unavoidable thanks to compiz and an excrutiatingly
18598 slow GTT pagefault, though it does look like it will be ineffectual
18599 during everyday usage. Compiz (and presumably other compositing
18600 managers) appears to be undoing all the pagefault minimisation as
18601 demonstrated on gen5 with large XPutImage. It also appears the CPU to
18602 memory bandwidth ratio plays a crucial role in determining whether
18603 going straight to GTT or through the CPU cache is a win - so no trivial
18606 x11perf -putimage10 -putimage500 on i5-2467m:
18608 bare: 1150,000 2,410
18609 compiz: 438,000 2,670
18611 bare: 1190,000 2,730
18612 compiz: 437,000 2,690
18614 bare: 658,000 2,670
18615 compiz: 389,000 2,520
18619 bare: 537,000 1,080
18620 compiz: 263,000 398
18622 bare: 606,000 1,360
18623 compiz: 203,000 985
18625 bare: 294,000 1,070
18626 compiz: 197,000 821
18631 compiz: 106,000 123
18634 compiz: 103,000 197
18639 Reported-by: Michael Larabel <Michael@phoronix.com>
18640 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18642 commit 735a15208dd600eefa3090f344186df9cac0462d
18643 Author: Chris Wilson <chris@chris-wilson.co.uk>
18644 Date: Sat Dec 10 23:45:56 2011 +0000
18646 sna/gen5: Remove a redundant format check
18648 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18650 commit c5584252c34b792313578cc31d56248d8990b571
18651 Author: Chris Wilson <chris@chris-wilson.co.uk>
18652 Date: Sat Dec 10 23:34:16 2011 +0000
18654 sna: Remember to assign a new unique id for the replaced bo
18656 Missed from the previous patch.
18658 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18660 commit 9c764dc13be40fc14238a4c130f7d8b44dbdf7db
18661 Author: Chris Wilson <chris@chris-wilson.co.uk>
18662 Date: Sat Dec 10 22:38:57 2011 +0000
18664 sna: Be more pessimistic with CPU sources
18666 Try to avoid a few more unnecessary context switches.
18668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18670 commit 358aaef6dbff0737f026046b0c9a2e21d8cfdf2a
18671 Author: Chris Wilson <chris@chris-wilson.co.uk>
18672 Date: Sat Dec 10 22:38:33 2011 +0000
18674 sna/dri: Prefer using the BLT for DRICopyRegion on pre-SNB
18676 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18678 commit c295ad8da91e39c8fffa540901097651df5d24b2
18679 Author: Chris Wilson <chris@chris-wilson.co.uk>
18680 Date: Sat Dec 10 22:37:31 2011 +0000
18682 sna: Transfer the whole bo for a replacement XCopyArea
18684 If we are copying over the entire source onto the destination,just copy
18685 across the GPU bo. This is often used for caching images as pixmaps.
18687 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18689 commit ece7fc8afeb8eefcf0ad1a054f02e7fac8db6327
18690 Author: Chris Wilson <chris@chris-wilson.co.uk>
18691 Date: Sat Dec 10 21:04:08 2011 +0000
18693 sna: Only use the 64-byte pitch alignment for scanout
18695 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18697 commit b3816cf3a99d23d0c3ab4cd716b24ea544a07283
18698 Author: Chris Wilson <chris@chris-wilson.co.uk>
18699 Date: Sat Dec 10 20:32:56 2011 +0000
18701 sna: Remove assertions that external bo are not busy
18703 We have to be careful to assume bo via exposed are under our full
18704 control, in particular not to assert their state. :(
18706 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18708 commit b5a6bc9e33a2797c926969e1f98b4a9a796248c5
18709 Author: Chris Wilson <chris@chris-wilson.co.uk>
18710 Date: Sat Dec 10 13:14:45 2011 +0000
18712 sna/gen[23]: Fixup render targets with pitches below hw minimum
18714 gen2/3 have a restriction that the 3D pipeline cannot render to a pixmap
18715 with a pitch less than 8/16 respectively. Rather than mandating all
18716 pixmaps to be created with a stride greater than 16, fixup the bo for
18717 the rare occasions when it is necessary.
18719 Reported-by: Paul Neumann <paul104x@yahoo.de>
18720 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43688
18721 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18723 commit c0dab7b1cf17fe751c86ad2b3fabce682eb50366
18724 Author: Chris Wilson <chris@chris-wilson.co.uk>
18725 Date: Sat Dec 10 12:46:46 2011 +0000
18727 sna/trapezoids: Try to render traps onto a8 destinations in place
18729 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18731 commit c73b14cabb059075af0f1727847074a89941c9df
18732 Author: Chris Wilson <chris@chris-wilson.co.uk>
18733 Date: Sat Dec 10 11:41:18 2011 +0000
18735 sna/trapezoids: First try the scan converter for fallbacks
18737 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18739 commit 22d9bc0bc1ccd869ebcc93f9a57efc8a85840cb8
18740 Author: Chris Wilson <chris@chris-wilson.co.uk>
18741 Date: Fri Dec 9 23:48:11 2011 +0000
18743 sna: Use a single definition for the inactive cache timeout
18745 And share it between the timer and the expiration function, just to
18748 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18750 commit eb3e04d960cd749adf696ba15c501a17cd1a184e
18751 Author: Chris Wilson <chris@chris-wilson.co.uk>
18752 Date: Fri Dec 9 23:36:56 2011 +0000
18754 sna: Fallback to ordinary monotonic clock if coarse is not supported
18756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18758 commit 1c202cc0746541b541d0af22eab5c9798d229a69
18759 Author: Chris Wilson <chris@chris-wilson.co.uk>
18760 Date: Fri Dec 9 17:25:19 2011 +0000
18762 sna: s/MONOTONICE/MONOTONIC/
18764 A late addition to be flexible for compiling on different systems
18767 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18769 commit c51626ccb660c777da8d84bb425de0d590864448
18770 Author: Chris Wilson <chris@chris-wilson.co.uk>
18771 Date: Fri Dec 9 15:45:29 2011 +0000
18773 sna: Use the coarse monotonic clock to coalesce wakeup events
18775 For the long interval events (such as expiring the caches), we do not
18776 need precise timing and so can use a coarse timer to allow the system
18777 to coalesce and reduce wakeup events.
18779 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18781 commit c22197f25bc0419d9f2abfcc978df5ef439feb47
18782 Author: Chris Wilson <chris@chris-wilson.co.uk>
18783 Date: Fri Dec 9 15:12:03 2011 +0000
18785 sna: Discard bo for idle private pixmaps
18787 If a pixmap lies around for a couple of minutes not being used, it is
18788 unlikely to be used again in the near future. Reap the GPU buffers of
18789 any of those idle pixmaps (copying to a more compact buffer in system
18790 memory) in order to free up resources for use elsewhere. Any object
18791 that is exposed via DRI is obviously exempt from this reaping.
18793 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18795 commit 429a36f7481b9bfd5ed137642d2916d69a713557
18796 Author: Chris Wilson <chris@chris-wilson.co.uk>
18797 Date: Fri Dec 9 09:54:12 2011 +0000
18799 uxa: Fix clip processing for uxa_fill_spans()
18801 Fixes regression from e0066e77e026b0dd0daa0c3765473c7d63aa6753
18802 (uxa: Simplify Composite solid acceleration for spans by only clipping
18805 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43649
18806 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18808 commit 699888a6410b0699c69a7f8a8d82dc4fde6fcc7f
18809 Author: Chris Wilson <chris@chris-wilson.co.uk>
18810 Date: Thu Dec 8 16:55:20 2011 +0000
18812 uxa/video: Use the common bo allocations and upload
18814 In order to avoid inconsistent usage of coherency domains and to avoid
18815 completely unnecessary clflushing during video playback, use the same
18816 buffer allocation and upload functions as the rest of the driver.
18818 Reported-by: Christophe Roland <roll68@gmail.com>
18819 Bugzilla: http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=60;bug=651316
18820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18822 commit 706d3a97bd3049e60c08d121a3d4d81c03029a87
18823 Author: Chris Wilson <chris@chris-wilson.co.uk>
18824 Date: Thu Dec 8 18:05:14 2011 +0000
18826 sna/trapezoids: Fix detection of rectilinearity after projection
18828 A typo confused left and right, rejecting true vertical edges, and worse
18829 might have incurred false positives.
18831 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18833 commit bc081420a5d6d28c4e6fd5527a70caaa3a8361c4
18834 Author: Simon Que <sque@chromium.org>
18835 Date: Wed Dec 7 16:20:41 2011 -0800
18837 xf86-video-intel: change order of DPMS operations
18839 The operations when setting dpms on should be in the order opposite
18840 of what's done when setting dpms off.
18842 This is because of potentially conflicting effects:
18843 ~ drmModeConnectoSetProperty() enables/disables the backlight driver.
18844 Some backlight drivers such as intel_backlight set the backlight to 0
18845 when disabled and to max when enabled.
18846 ~ intel_output_dpms_backlight() saves the backlight value when turning
18847 DPMS off and restores it when turning DPMS on.
18849 Here's the current order of operations:
18851 xset dpms force off (backlight is nonzero)
18852 drmModeConnectoSetProperty(DPMSModeOff)
18853 kernel: disable backlight, backlight=0
18854 intel_output_dpms_backlight(DPMSModeOff)
18855 save backlight value (0) <-- it has been set to 0 by kernel
18859 drmModeConnectoSetProperty(DPMSModeOn)
18860 kernel: enable backlight, backlight=max
18861 intel_output_dpms_backlight(DPMSModeOn)
18862 set backlight to saved value (0)
18864 The correct way to do this would be to reverse the operations during
18865 xset dpms force off:
18866 intel_output_dpms_backlight(DPMSModeOff)
18867 save backlight value (nonzero)
18869 drmModeConnectoSetProperty(DPMSModeOff)
18870 kernel: enable backlight, backlight=0
18872 This restores the saved nonzero backlight value during the force on.
18874 Signed-off-by: Simon Que <sque@chromium.org>
18875 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18877 commit 84aaf1537cbd29e163346d03debc39f4623c69eb
18878 Author: Chris Wilson <chris@chris-wilson.co.uk>
18879 Date: Thu Dec 8 12:36:08 2011 +0000
18881 sna/gen7: Reduce dst readbacks for unsupported sources
18883 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18885 commit 440ac68ec074e82818713773f3e2cb5d363862aa
18886 Author: Chris Wilson <chris@chris-wilson.co.uk>
18887 Date: Thu Dec 8 12:35:54 2011 +0000
18889 sna/gen6: Reduce dst readbacks for unsupported sources
18891 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18893 commit bc68211d18356c4a0e011ac360665d600de9fa30
18894 Author: Chris Wilson <chris@chris-wilson.co.uk>
18895 Date: Thu Dec 8 12:35:39 2011 +0000
18897 sna/gen5: Reduce dst readbacks for unsupported sources
18899 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18901 commit a5df7c28e458c2f2b173fe5d745f447087eb8c6f
18902 Author: Chris Wilson <chris@chris-wilson.co.uk>
18903 Date: Thu Dec 8 12:35:22 2011 +0000
18905 sna/gen4: Reduce dst readbacks for unsupported sources
18907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18909 commit cc8cab649c0a4788803d2c62f186142ec7152b89
18910 Author: Chris Wilson <chris@chris-wilson.co.uk>
18911 Date: Thu Dec 8 12:34:58 2011 +0000
18913 sna/gen3: Reduce readbacks on dst for unsupported sources
18915 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18917 commit e5bc0c823b8fd46b3f851c5e9a4b4de39cda9a91
18918 Author: Chris Wilson <chris@chris-wilson.co.uk>
18919 Date: Thu Dec 8 12:34:34 2011 +0000
18921 sna/gen2: Avoid readbacks for unsupported sources
18923 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18925 commit 874c722c86b06ac0aa10e6943d3faf76fcb751be
18926 Author: Chris Wilson <chris@chris-wilson.co.uk>
18927 Date: Thu Dec 8 12:07:28 2011 +0000
18929 sna: Beware flushing partial buffers before they are written
18931 A partial buffer is considered finished upon the next batch submission,
18932 so one needs to be careful that it is completely written to before such
18933 an event is triggered. move-to-cpu is such a trigger as demonstrated by
18934 the picture fixup routine for handling convolution filters.
18936 Reported-by: Victor Machado <machado.prx@gmail.com>
18937 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43607
18938 Signed-off-by: Chris Wilson <ickle@crestline.(none)>
18940 commit 6ccb114a7e685c69fb388ebd119393455b315c36
18941 Author: Chris Wilson <chris@chris-wilson.co.uk>
18942 Date: Wed Dec 7 21:09:31 2011 +0000
18944 sna: Prefer to use our pixmap upload paths
18946 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18948 commit 101942d41df7efaa6103e31e738775fafdb63159
18949 Author: Chris Wilson <chris@chris-wilson.co.uk>
18950 Date: Mon Dec 5 10:11:04 2011 +0000
18952 uxa: Unmap the buffer after swrast
18954 Otherwise we may exhaust the per-process vma limit and cause
18955 applications to stop rendering and eventually crash the X server.
18957 Will only work in conjunction with a new libdrm (2.4.28) and commit
18958 c549a77c (intel: Unmap buffers during drm_intel_gem_bo_unmap)
18961 Reported-by: nobled@dreamwidth.org
18962 References: https://bugs.freedesktop.org/show_bug.cgi?id=43075
18963 References: https://bugs.freedesktop.org/show_bug.cgi?id=40066
18964 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18966 commit b424b10e771b1d3d041efdd2b77f576251364744
18967 Author: Chris Wilson <chris@chris-wilson.co.uk>
18968 Date: Sun Dec 4 13:01:10 2011 +0000
18970 sna: use tight pitches for a8
18972 As we never use these with a depth nor attach them to scanout, we can
18973 safely relax the multiple-of-64 byte pitch restriction. In the unlikely
18974 event that we do need A8 surfaces with depthbuffers, this is broken...
18976 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18978 commit 46c7df803881fa0e733c0d0fdd37567ebdccd6ac
18979 Author: Chris Wilson <chris@chris-wilson.co.uk>
18980 Date: Sat Dec 3 18:02:00 2011 +0000
18982 sna: Remove one redundant retire
18984 There is no need to retire immediately after a batch and no indication
18985 that it will be useful.
18987 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18989 commit b99c6b13ebba9521333b8dd5982ac37b6e244b54
18990 Author: Chris Wilson <chris@chris-wilson.co.uk>
18991 Date: Sat Dec 3 18:01:11 2011 +0000
18993 sna: Pass the current value of the batch offset to the kernel relocator
18995 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
18997 commit 735219cd59e6184a6622d3d429a704ca3f58b9cd
18998 Author: Chris Wilson <chris@chris-wilson.co.uk>
18999 Date: Fri Dec 2 10:42:00 2011 +0000
19001 uxa: Ensure that we can fallback with all of (src, mask, dst) as GTT mappings
19003 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19005 commit f6c82c73b673ec3c9cce432fe38d5e0076234efd
19006 Author: Chris Wilson <chris@chris-wilson.co.uk>
19007 Date: Fri Dec 2 10:34:10 2011 +0000
19009 uxa: Fix runtime linking of previous commit
19011 So much for relying on compiler warnings.
19013 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19015 commit 85d3dc5910a2eea3a10b822e01443e11eaae9291
19016 Author: Chris Wilson <chris@chris-wilson.co.uk>
19017 Date: Fri Dec 2 10:22:51 2011 +0000
19019 uxa: Reset size limits based on AGP size
19021 The basis for the constraints are what we can map into the aperture for
19022 direct writing with the CPU, so use the size of the mappable region as
19023 opposed to the size of the total GTT.
19025 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19027 commit e55198746102afb7427f577bd5bfc76667438da9
19028 Author: Chris Wilson <chris@chris-wilson.co.uk>
19029 Date: Thu Dec 1 13:49:03 2011 +0000
19031 sna: Reuse the full size of an old handle for io
19033 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19035 commit c5632369cbd6473304c06e4230347abbe46513ec
19036 Author: Chris Wilson <chris@chris-wilson.co.uk>
19037 Date: Thu Dec 1 13:23:56 2011 +0000
19039 sna: Move the preservation of the io handle into the common destroy path
19041 In order to capture and reuse all io buffers.
19043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19045 commit 95f4da647a4055545b09cae0834df0fa2127a458
19046 Author: Chris Wilson <chris@chris-wilson.co.uk>
19047 Date: Wed Nov 30 11:59:31 2011 +0000
19049 sna: Align pwrite to transfer whole cachelines
19051 Daniel claims that this is will be faster, or will be once he has
19052 completed rewriting pwrite!
19054 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19056 commit ecd6cca617ac29cf2b1b2a4d33fca19b84fea2a9
19057 Author: Chris Wilson <chris@chris-wilson.co.uk>
19058 Date: Tue Nov 29 19:27:46 2011 +0000
19060 sna/gen5: Handle cpu-bo for render targets
19062 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19064 commit d8f2e87473eae81e07922271b5f608fdceb4ae5e
19065 Author: Chris Wilson <chris@chris-wilson.co.uk>
19066 Date: Tue Nov 29 10:59:48 2011 +0000
19068 sna/render: Fix check for "migrate whole pixmap"
19070 The whole pixmap means the sample covers the full width and height, not
19073 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19075 commit 20e57914089d6959529ad222e04853f5ffcb364f
19076 Author: Chris Wilson <chris@chris-wilson.co.uk>
19077 Date: Tue Nov 29 10:59:20 2011 +0000
19079 sna: Fix assertion around flushing of mmap(PROT_READ)
19081 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19083 commit 56155c91afb5a0dcafe4dbc22ade3b7c0c977674
19084 Author: Chris Wilson <chris@chris-wilson.co.uk>
19085 Date: Tue Nov 29 10:27:18 2011 +0000
19087 sna/gen6: Set the batch mode prior to checking limits and flushing
19089 If we change contexts, then we will submit the batch obsoleting the
19090 earlier resource checks.
19092 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19094 commit 5b1e9e15738b9001346ab6e0166f861ce308008e
19095 Author: Chris Wilson <chris@chris-wilson.co.uk>
19096 Date: Mon Nov 28 22:01:00 2011 +0000
19098 sna: Always reduce tiling for thin pixmaps
19100 Benchmarking on the current code base, says this is now a win. A
19101 reversal of older benchmarks, so expect further tuning.
19103 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19105 commit dacb301c05dd1686e788d6ed8101887c27968389
19106 Author: Chris Wilson <chris@chris-wilson.co.uk>
19107 Date: Mon Nov 28 18:59:10 2011 +0000
19109 sna: Use the correct pitch when creating an untiled active buffer
19111 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19113 commit 8657128fa7e758a2dde93340d6e58928d5f11255
19114 Author: Chris Wilson <chris@chris-wilson.co.uk>
19115 Date: Mon Nov 28 18:22:01 2011 +0000
19117 sna: Pass the pixmap to sna_replace()
19119 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19121 commit 4e38d22105da2bd97db005dc505e75dcd22291d3
19122 Author: Chris Wilson <chris@chris-wilson.co.uk>
19123 Date: Mon Nov 28 18:21:19 2011 +0000
19125 sna: Tidy kgem_choose_tiling()
19127 Reduce the calls to compute the surface size down to one.
19129 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19131 commit 04f47e52da06cc20c2a266da92ba3f8e961433cb
19132 Author: Chris Wilson <chris@chris-wilson.co.uk>
19133 Date: Sun Nov 27 17:45:05 2011 +0000
19135 sna: Don't try to guess when the kernel believes a buffer to be flushed
19137 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19139 commit d87c332653001db4413c7dd81116ee03830db491
19140 Author: Chris Wilson <chris@chris-wilson.co.uk>
19141 Date: Sun Nov 27 10:18:11 2011 +0000
19143 sna: Eliminate empty glyphs during format conversion
19145 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19147 commit 413776c79ee7c55b3a48b1ed29af63f8cea39337
19148 Author: Chris Wilson <chris@chris-wilson.co.uk>
19149 Date: Sun Nov 27 10:17:40 2011 +0000
19151 sna: Use 64-bit mov (when available) for bitmap glyphs
19153 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19155 commit 507f99eba2c72bc109c9267c698f3d5f9fb6f539
19156 Author: Chris Wilson <chris@chris-wilson.co.uk>
19157 Date: Sat Nov 26 17:57:57 2011 +0000
19159 sna/gen6+: Only use BLT if the untiled bo will cause per-pixel TLB misses
19161 i.e. only force the BLT if using the sampler is going to be incredibly
19164 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19166 commit f11e9f189111f53ace20381a0bd4f559ccd3605f
19167 Author: Chris Wilson <chris@chris-wilson.co.uk>
19168 Date: Sat Nov 26 17:12:03 2011 +0000
19170 sna: Only check for all-damaged CPU bo if we may discard the GPU bo
19172 The goal of the optimisation is to discard the GPU bo early, so we
19173 can skip the extra damage reduction if there is no gpu bo.
19175 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19177 commit ccee730e45d8c2490e578d0b85d9e76f3a44f76f
19178 Author: Chris Wilson <chris@chris-wilson.co.uk>
19179 Date: Sat Nov 26 16:47:04 2011 +0000
19181 sna: Discard the gpu-bo is the entire pixmap is written to by the CPU.
19183 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19185 commit df68b20a17e28c0898226f970315bc8e515e9f56
19186 Author: Chris Wilson <chris@chris-wilson.co.uk>
19187 Date: Sat Nov 26 16:45:22 2011 +0000
19189 sna: Disabling tiling if the bo would be smaller than a page
19191 Once again experiment with untiled smalled buffers.
19193 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19195 commit d88ad2e60ac3d9f5608db1653a7af789f7f1b446
19196 Author: Chris Wilson <chris@chris-wilson.co.uk>
19197 Date: Sun Nov 27 09:01:09 2011 +0000
19199 sna: Round up stride to alignment for TLB miss checking
19201 We want to avoid the condition of reducing the tiling mode (when reusing
19202 an active untiled buffer in preference to creating a new) for a wide buffer
19203 when doing will force a TLB miss on each sample.
19205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19207 commit 16f5e224dcfd97012b38ee1af6c72dbe3c0f3304
19208 Author: Chris Wilson <chris@chris-wilson.co.uk>
19209 Date: Fri Nov 25 16:21:41 2011 +0000
19211 sna: Avoid the double application of drawable offsets for tiled spans
19213 As we handle tiled spans indirectly, we need to avoid applying the
19214 drawable offsets twice (once in the mi layer generating the spans, and
19215 then once more in the tiled rect renderer).
19217 Reported-by: Ulrich Müller <ulm@gentoo.org>
19218 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43245
19219 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19221 commit e06789871b5cdf0c5535d46622b8ec3f85eb4386
19222 Author: Chris Wilson <chris@chris-wilson.co.uk>
19223 Date: Fri Nov 25 15:24:29 2011 +0000
19225 sna/gen5+: Prefer using the BLT when either src or dst is untiled
19227 The cost of the TLB miss on every sample far outweighs the impact of the
19228 context (and ring) switch.
19230 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19232 commit d0cce0da71fae275a0f1be33b2bebad9d41212c9
19233 Author: Chris Wilson <chris@chris-wilson.co.uk>
19234 Date: Fri Nov 25 13:16:37 2011 +0000
19236 sna/gen5: Use the same prefer_blt_fill() switch for fill-one
19238 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19240 commit 7ad2d3b695115439f685a12d43ab240cf5e18ef7
19241 Author: Chris Wilson <chris@chris-wilson.co.uk>
19242 Date: Fri Nov 25 09:44:21 2011 +0000
19244 sna/gen2: Wire up NO_COMPOSITE for debugging
19246 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19248 commit 394ad51c421f41e4bc6545fcee10b1757cb5f518
19249 Author: Chris Wilson <chris@chris-wilson.co.uk>
19250 Date: Thu Nov 24 17:35:36 2011 +0000
19252 sna/video: Check for Xv presence in the server before initialising
19254 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19256 commit a2fab9560ec5dd6efc352917fe7626e59bed69e3
19257 Author: Chris Wilson <chris@chris-wilson.co.uk>
19258 Date: Thu Nov 24 17:14:08 2011 +0000
19260 sna: Downgrade Y-tiling if we must use the BLT
19262 There are many operations, usually the core drawing acceleration, where
19263 the BLT is much more preferable than using the CPU. However, the BLT is
19264 limited to only using X-tiling, so if we encounter a Y-tiled pixmap
19265 target we need to recreate it as X-tiling before proceeding. Hopefully,
19266 the pixmap is then kept around and rendered multiple times to amoritize
19267 the cost of the copy.
19269 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19271 commit e0a4492c8b51d6ae09cde0b895da6177284e95e0
19272 Author: Chris Wilson <chris@chris-wilson.co.uk>
19273 Date: Thu Nov 24 17:00:18 2011 +0000
19275 sna: Use Y-tiling for source pixmaps
19277 Y-tiling is slightly faster with RENDER operations, so attempt to
19278 allocate source-only pixmaps using this tiling mode. Actually using
19279 Y-tiling is a delicate balance because it then prevents the use of the
19280 BLT. For instance, enabling Y-tiling by default gives a 30% performance
19281 improvement on the fish-demo (compositing benchmark) at 2560x1440 on
19282 Ironlake but regresses tiger-demo by 2x (spans benchmark).
19284 So experiment with this compromise and allow for changing the default
19287 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19289 commit 7e4a1b7ed25a881113e197acde62b07f26f2d6ef
19290 Author: Chris Wilson <chris@chris-wilson.co.uk>
19291 Date: Thu Nov 24 11:34:51 2011 +0000
19293 src/sna/gen5: Replace the precompiled shaders
19295 Take advantage of a couple of new instructions introduced with Cantiga
19296 to reduce the instruction count inside the shaders and improve
19297 performance by around 10% in the fish-demo.
19299 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19301 commit 67ea8808d7f0ce47733a66d8708dcfa0dbec14c1
19302 Author: Chris Wilson <chris@chris-wilson.co.uk>
19303 Date: Thu Nov 24 11:45:38 2011 +0000
19305 sna/gen5: Remove the unused SIP
19307 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19309 commit 499ed74e979d99bbe1670ad54875fba6055fce5d
19310 Author: Chris Wilson <chris@chris-wilson.co.uk>
19311 Date: Thu Nov 24 11:02:01 2011 +0000
19313 sna: Tweak cancellation of deferred flush
19315 The goal is to keep running until the tick after every stops,
19316 irrespective of forced flushes.
19318 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19320 commit 35239d207c38e1ae0cb843ed7a72621711cb001e
19321 Author: Chris Wilson <chris@chris-wilson.co.uk>
19322 Date: Thu Nov 24 02:51:51 2011 +0000
19324 sna: Always call retire following a mmap
19326 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19328 commit 11e0b456de93dae9f7ba4298d8a32b6b60edc503
19329 Author: Chris Wilson <chris@chris-wilson.co.uk>
19330 Date: Thu Nov 24 03:04:09 2011 +0000
19332 sna/gen5: Use the BLT rather than flush when copying
19334 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19336 commit a44663af9335136cba59b12078c6717992283b62
19337 Author: Chris Wilson <chris@chris-wilson.co.uk>
19338 Date: Thu Nov 24 02:54:43 2011 +0000
19340 sna/gen4,gen5: Fix typo and only emit a flush if the video source is dirty
19342 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19344 commit 4a27dd287cdb85d80ceaf6b54e7de0c17e01c00d
19345 Author: Chris Wilson <chris@chris-wilson.co.uk>
19346 Date: Mon Nov 21 12:56:32 2011 +0000
19348 uxa: Make the glamor/uxa transition more verbose
19350 And so hopefully make it clearer. In the process we restore the flushing
19351 behaviour for UXA back to before the glamor intervention.
19353 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19355 commit b0d64a9567d9cb749568448e4ce090b187308028
19356 Author: Chris Wilson <chris@chris-wilson.co.uk>
19357 Date: Mon Nov 21 11:49:49 2011 +0000
19359 sna: Correct end-point adjustment for stippling PolySegment
19361 Reported-by: Roman Jarosz <kedgedev@gmail.com>
19362 Reported-by: da_fox@mad.scientist.com
19363 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43134
19364 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19366 commit a5353a7c1d1f02b887f61f543c8e4bd4e8bd104f
19367 Author: Chris Wilson <chris@chris-wilson.co.uk>
19368 Date: Mon Nov 21 10:58:17 2011 +0000
19370 sna/gen4: Use the special composite rect emission for fill_one()
19372 In order to workaround a bug in the shaders on gen4, we need to flush
19373 the pipeline after every rectangle. The recently introduced fill-one
19374 mechanism for gen4, missed this vital step triggering a random hang with
19375 an otherwise sane batchbuffer (the missing flush is hard to spot!).
19377 Fixes regression from 86f99379ee5 (sna/gen4: Add fill-one).
19379 Reported-by: Albert Damen <albrt@gmx.net>
19380 Reported-by: Fryderyk Dziarmagowski <fdziarmagowski@gmail.com>
19381 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43083
19382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19384 commit 3b9479dc39d32fd97f80c1e5e0fac67d36ee5e40
19385 Author: Chris Wilson <chris@chris-wilson.co.uk>
19386 Date: Sat Nov 19 19:12:06 2011 +0000
19388 sna: Micro-optimise sna_push_pixels_solid_blt()
19390 Unroll the byte reversal as we know the row length is word aligned.
19392 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19394 commit 297f7680cfd73ab6eb099adca7aefe00e61c9e85
19395 Author: Chris Wilson <chris@chris-wilson.co.uk>
19396 Date: Sat Nov 19 16:11:42 2011 +0000
19398 sna: Avoid the NULL damage deref
19400 Fixes x11perf -wdcircle100 -time 1 -repeat 1 -rop GXxor
19402 Reported-by: Fryderyk Dziarmagowski <fdziarmagowski@gmail.com>
19403 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43084
19404 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19406 commit 4a86ec5ddbaf119fd9a32ecd963f7238152d7d28
19407 Author: Chris Wilson <chris@chris-wilson.co.uk>
19408 Date: Sat Nov 19 00:33:28 2011 +0000
19410 sna/gen7: Apply recent SNB tuning as applicable
19412 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19414 commit fe944d61d53c275a168da2e9515d1e42be81630b
19415 Author: Chris Wilson <chris@chris-wilson.co.uk>
19416 Date: Sat Nov 19 00:02:17 2011 +0000
19418 sna/gen7: Correct shifts for surface state
19420 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19422 commit 70042400fd95f613b9547d0a72d6288abcc2357d
19423 Author: Chris Wilson <chris@chris-wilson.co.uk>
19424 Date: Fri Nov 18 23:41:33 2011 +0000
19426 sna/gen7: minor tidy of redundant defines
19428 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19430 commit f938cd15990d1a851b9d92414545dda37f00c5db
19431 Author: Chris Wilson <chris@chris-wilson.co.uk>
19432 Date: Thu Nov 17 13:43:37 2011 +0000
19434 glamor: Include glamor cflags and libs for atypical builds
19436 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19438 commit eb6148a36a318eb0b4aad865c34b50fedf4b709e
19439 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
19440 Date: Wed Nov 16 15:04:37 2011 +0800
19442 uxa/glamor: Turn on glamor for fill spans and rects
19444 This commit only enables two glamor functions for
19445 uxa_fill_spans and uxa_poly_fill_rects.
19447 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
19448 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19450 commit fbabe60f48006ee664c983082498a863e8beec71
19451 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
19452 Date: Wed Nov 16 15:04:36 2011 +0800
19454 glamor: Initial commit to introduce glamor acceleration.
19456 Added one configuration option --enable-glamor to control
19457 whether use glamor. Added one new file intel_glamor.c to
19458 wrap glamor egl API for intel driver's usage.
19459 This commit doesn't really change the driver's control path.
19460 It just adds necessary files for glamor and change some
19463 Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
19464 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
19465 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19467 commit c4c2eb1fae2518578a56a278afeaec66d544439d
19468 Author: Zhigang Gong <zhigang.gong@linux.intel.com>
19469 Date: Wed Nov 16 15:04:37 2011 +0800
19471 uxa/glamor: Introduce additional access modes to wrap glamor acceleration
19473 Integrate glamor acceleration into UXA framework. Add
19474 necessary flushing at the following points:
19476 1. Flush UXA batch buffer before call into glamor.
19477 2. Flush GL operations after return from a glamor function.
19478 3. The point we need to flush UXA batch buffer, we also
19479 need to flush GL operations, for example, in
19480 intel_flush_callback and couple of places in intel_display.c.
19482 Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
19483 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19485 commit ac7df2919b92c9fe47c6745975e1848932f989e8
19486 Author: Chris Wilson <chris@chris-wilson.co.uk>
19487 Date: Wed Nov 16 23:25:16 2011 +0000
19489 Revert "Disable adding normal RTF modes for an eDP"
19491 This reverts commit 212fa9868767637e8f430485eeb522c99e63fd16.
19493 The underlying register programming for eDP is now believed to be fixed
19496 References: https://bugs.freedesktop.org/show_bug.cgi?id=38012
19497 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41070
19498 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19500 commit 24ece4a87e34c27115ffb55f91294bce2bdf1502
19501 Author: Chris Wilson <chris@chris-wilson.co.uk>
19502 Date: Wed Nov 16 23:14:25 2011 +0000
19504 configure: Version bump for 2.17.0 release
19506 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19508 commit 14b3176962bd535933623d8ed77ba286b077d6f0
19509 Author: Chris Wilson <chris@chris-wilson.co.uk>
19510 Date: Wed Nov 16 23:13:44 2011 +0000
19512 NEWS: Release notes for 2.17.0
19514 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19516 commit c259144e3fc52d078b0a78107c38f0f3d3a2bbc1
19517 Author: Chris Wilson <chris@chris-wilson.co.uk>
19518 Date: Wed Nov 16 10:28:23 2011 +0000
19520 sna: The block handler is passed an indirect pointer to the timeval
19522 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19524 commit bfd2bb40274d1242001d295a4010211fd51b0fc3
19525 Author: Chris Wilson <chris@chris-wilson.co.uk>
19526 Date: Tue Nov 15 10:38:09 2011 +0000
19528 sna: Correct dependencies for DRI2
19530 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19532 commit 3771387ad11b5842a83e58a4b373c2acdd827bd2
19533 Author: Chris Wilson <chris@chris-wilson.co.uk>
19534 Date: Tue Nov 15 10:32:34 2011 +0000
19536 Compile out UXA if so desired
19538 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19540 commit edbeab8c4edf9e0e89d85add485fe659795b6350
19541 Author: Chris Wilson <chris@chris-wilson.co.uk>
19542 Date: Tue Nov 15 10:17:06 2011 +0000
19544 sna: Reduce and clarify dependencies
19546 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19548 commit 78d4e99fc916e6477edb01c6f24b69ad73adc552
19549 Author: Chris Wilson <chris@chris-wilson.co.uk>
19550 Date: Tue Nov 15 11:06:04 2011 +0000
19552 sna: And keep unity happy
19554 Rewrite the DRI layer to avoid the various bugs and shortcomings of the
19555 Xserver and interfacing with mesa.
19557 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38732
19558 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=39044
19559 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19561 commit aac022cbb3342e7027301a3f8c9add8929edbad6
19562 Author: Chris Wilson <chris@chris-wilson.co.uk>
19563 Date: Mon Nov 14 19:49:29 2011 +0000
19565 sna: Optimise single pixel transfers
19567 Surprisingly frequent.
19569 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19571 commit 5647e2981d81fd2b5cdb22d544f83d2d0f9445c9
19572 Author: Chris Wilson <chris@chris-wilson.co.uk>
19573 Date: Mon Nov 14 14:11:56 2011 +0000
19575 sna/damage: Always reduce damage for testing PIXMAN_REGION_OUT
19577 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
19578 References: https://bugs.freedesktop.org/show_bug.cgi?id=42414
19579 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19581 commit 405b015fe255c75f91b0f3fe93f06bd14faee5a1
19582 Author: Chris Wilson <chris@chris-wilson.co.uk>
19583 Date: Mon Nov 14 14:05:10 2011 +0000
19585 sna: Be explicit in all sna_damage_contains_box() tests
19587 Clarify the exact in/out/maybe expected result when testing.
19589 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19591 commit 5d23149647b1645057dbd89087691e873494f763
19592 Author: Chris Wilson <chris@chris-wilson.co.uk>
19593 Date: Mon Nov 14 13:53:37 2011 +0000
19595 sna: Remove redundant 'can_fill_spans()'
19597 Spans are almost always accelerated now...
19599 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19601 commit ae9de984adb275e028e7fbcb8b74a3ce11920eed
19602 Author: Chris Wilson <chris@chris-wilson.co.uk>
19603 Date: Mon Nov 14 13:39:34 2011 +0000
19605 sna: Convert non-FillSolid rectilinear lines to boxes
19607 And render using the tiled/stippled rect routines instead.
19609 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19611 commit e861e816f53989ce3ab3ec89dd1d9a35155f999a
19612 Author: Chris Wilson <chris@chris-wilson.co.uk>
19613 Date: Mon Nov 14 13:39:34 2011 +0000
19615 sna: Convert non-FillSolid rectilinear segments to boxes
19617 And render using the tiled/stippled rect routines instead.
19619 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19621 commit 6fc2928f578ae1e278c968d5c778d99d56bd667a
19622 Author: Chris Wilson <chris@chris-wilson.co.uk>
19623 Date: Mon Nov 14 13:05:00 2011 +0000
19625 sna: tidy assignment of composite damage
19627 Make sure that the damage is always set, even if only to NULL, so that
19628 we are safe if in future the operation state is not initially cleared.
19630 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19632 commit 99338ebe61917a07611cbc81cfaf70c0ec905014
19633 Author: Chris Wilson <chris@chris-wilson.co.uk>
19634 Date: Mon Nov 14 12:04:49 2011 +0000
19636 sna: Check whether damage can be reduced to all-damage on moving to GPU
19638 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19640 commit 00d1c539e4b69636bfc35c0c47a6f1db3a33e731
19641 Author: Chris Wilson <chris@chris-wilson.co.uk>
19642 Date: Mon Nov 14 11:19:42 2011 +0000
19644 sna/damage: Add a little more verbosity to debugging
19646 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19648 commit 60c0fc101adbc457a7c1927602175f8079175e10
19649 Author: Chris Wilson <chris@chris-wilson.co.uk>
19650 Date: Mon Nov 14 11:12:47 2011 +0000
19652 sna: Refactor some common code into the common destroy-gpu-bo function
19654 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19656 commit 571457e9a4b0941ec956ef98391e96fe90fb4abc
19657 Author: Chris Wilson <chris@chris-wilson.co.uk>
19658 Date: Mon Nov 14 10:31:21 2011 +0000
19660 sna: Be more thorough discarding fences for large objects
19662 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19664 commit e33dfbe3c00e634328adc7319261e5501f4c7a7a
19665 Author: Chris Wilson <chris@chris-wilson.co.uk>
19666 Date: Mon Nov 14 09:41:53 2011 +0000
19668 sna/damage: Verify that all-damage covers the entire pixmap after reduction
19670 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19672 commit f033172d2f383c7e17362a10f56c4958fb9a3a6b
19673 Author: Chris Wilson <chris@chris-wilson.co.uk>
19674 Date: Mon Nov 14 09:32:53 2011 +0000
19676 sna: Fix debug compilation
19678 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
19679 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42887
19680 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19682 commit 0c405595fb8739b22569f81f400100b5dfa472cf
19683 Author: Chris Wilson <chris@chris-wilson.co.uk>
19684 Date: Mon Nov 14 09:27:49 2011 +0000
19686 sna: Downgrade tiling in order to fit within fence constraints
19688 Fixes a later assert that the fenced size is valid for kgem_create_2d()
19690 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
19691 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42888
19692 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19694 commit 1baff9f8cf79fdb4a5649eabcc66cf73487c873e
19695 Author: Chris Wilson <chris@chris-wilson.co.uk>
19696 Date: Mon Nov 14 08:44:35 2011 +0000
19698 uxa/gen4+: Re-emit composite invariant after video
19700 Reported-by: Ben Hutchings <ben@decadent.org.uk>
19701 Bugzilla: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=635953
19702 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19704 commit 7ad3b377a394c9342c3ab2f003d48b7cb3c0891d
19705 Author: Chris Wilson <chris@chris-wilson.co.uk>
19706 Date: Sun Nov 13 17:53:11 2011 +0000
19708 sna/video: Use the right pointer for unmapping
19710 Don't just deference any old random pointer, use the one we actually
19711 mapped in the first place!
19713 Reported-by: Matti Hamalainen <ccr@tnsp.org>
19714 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42880
19715 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19717 commit 8da8809d03734d92c3bcf478fe051077ce3caa29
19718 Author: Chris Wilson <chris@chris-wilson.co.uk>
19719 Date: Sun Nov 13 17:49:23 2011 +0000
19721 sna/video: Simplify check for 915G[M] which is simply gen==30
19723 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19725 commit cac335b979a95e31e201f9766d19d737361e09b2
19726 Author: Chris Wilson <chris@chris-wilson.co.uk>
19727 Date: Sun Nov 13 13:13:03 2011 +0000
19729 sna/video: Constify a couple of attribute arrays
19731 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19733 commit e592eb93dbd7df177993986f01ba6f65c4e20d4a
19734 Author: Chris Wilson <chris@chris-wilson.co.uk>
19735 Date: Sun Nov 13 11:16:45 2011 +0000
19737 configure: Version bump for 2.16.902 snapshot
19739 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19741 commit 56707b97dc746f7b46dd6cf94ba6c1db85ed619c
19742 Author: Chris Wilson <chris@chris-wilson.co.uk>
19743 Date: Sun Nov 13 11:16:12 2011 +0000
19745 NEWS: Details for 2.16.902
19747 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19749 commit f0acc6c0f15875fd51f6df6c8f13d589f9149771
19750 Author: Chris Wilson <chris@chris-wilson.co.uk>
19751 Date: Sun Nov 13 09:39:47 2011 +0000
19753 sna/composite: Attempt to reduce the damage is the operation is contained
19755 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19757 commit 87147b63e905c3566ac1c12e0712f068e90d5e2e
19758 Author: Chris Wilson <chris@chris-wilson.co.uk>
19759 Date: Sat Nov 12 19:35:43 2011 +0000
19761 sna/damage: Reduce the damage for evaluating sna_damage_is_all
19763 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19765 commit e8799cdea461df5102d421fda26fecceae79b929
19766 Author: Chris Wilson <chris@chris-wilson.co.uk>
19767 Date: Sat Nov 12 12:19:31 2011 +0000
19769 sna: Be stricter and disallow allocation of large fenced objects
19771 When allocating objects, we need to check the size of the full fenced
19772 regions against the mappable limits in order to be able to mmap the
19775 References: https://bugs.freedesktop.org/show_bug.cgi?id=42813
19776 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19778 commit 0269ec5533ecf7bec0f01c682e085861a3d2ab00
19779 Author: Chris Wilson <chris@chris-wilson.co.uk>
19780 Date: Sat Nov 12 11:49:32 2011 +0000
19782 sna: Handle incremental uploads from PutImage
19784 PutImage rarely uploads the entire image inside a single request,
19785 instead breaking up into scanline segments to fit within the protocol
19786 limits. A few optimisations are based on detecting when we can discard
19787 the GPU bo for an all-dirty CPU pixmap, which are useful in this case so
19788 check for an entirely dirty pixmap following an incremental PutImage.
19790 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19792 commit ed68f1b00ae7c0fb3be1f0e758b3683f3c26675b
19793 Author: Chris Wilson <chris@chris-wilson.co.uk>
19794 Date: Sat Nov 12 11:13:59 2011 +0000
19796 sna/gen7: Fix PRIMITIVE command
19798 The topology is now an extra dword rather than an embedded field in the
19801 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19803 commit 991ffcb60a15f08ebb396d4d4ec86d5aff0ded31
19804 Author: Chris Wilson <chris@chris-wilson.co.uk>
19805 Date: Fri Nov 11 23:20:36 2011 +0000
19807 sna: Reset the source counter if we completely dirty the CPU pixmap
19809 For a long lived pixmap which we are repeatedly using for upload and
19810 copying to other pixmaps, we don't want to keep to maintain a GPU
19811 buffer. So instead, reset the source counter if we discard the GPU bo
19812 and treat the pixmap as purely CPU.
19814 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19816 commit 403a4dd353eef67decfba965e40623e8137e1944
19817 Author: Chris Wilson <chris@chris-wilson.co.uk>
19818 Date: Fri Nov 11 14:59:03 2011 +0000
19820 sna: Defer source migration for a solitary upload via CopyArea
19822 Try to avoid allocating a GPU pixmap for PutImage followed by CopyArea.
19824 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19826 commit 11acfaa62362660ff303f3afe90d1afa36b9ce73
19827 Author: Chris Wilson <chris@chris-wilson.co.uk>
19828 Date: Fri Nov 11 14:46:33 2011 +0000
19830 sna/glyph: Tweak the small mask threshold
19832 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19834 commit c6e6ae1829e06ee8fe8eb063f2433cce603c9f96
19835 Author: Chris Wilson <chris@chris-wilson.co.uk>
19836 Date: Fri Nov 11 14:22:38 2011 +0000
19838 sna/glyphs: Cache the glyph pixman_image_t wrapper
19840 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19842 commit 8f50950f467eb2440009a807081f3ba2c9db209b
19843 Author: Chris Wilson <chris@chris-wilson.co.uk>
19844 Date: Thu Nov 10 15:58:36 2011 +0000
19846 sna/gen7: Remove stray no-op from GEN7_3DSTATE_SBE
19848 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19850 commit 6a8338fc08a1b7e5e2c6722baa87cff3915b6ef9
19851 Author: Chris Wilson <chris@chris-wilson.co.uk>
19852 Date: Thu Nov 10 15:11:13 2011 +0000
19854 sna: Begin debugging gen7
19856 This is the stub of the decoder, sufficient to give details of the ops
19857 within the batch and to keep the debugger happy.
19859 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19861 commit 2309f19638f8b2c35eb60fb44fa988aa5aaab57f
19862 Author: Chris Wilson <chris@chris-wilson.co.uk>
19863 Date: Thu Nov 10 11:57:21 2011 +0000
19865 sna/dri: Apply the damage for fullscreen async blits
19867 Otherwise gnome-shell forgets to update. Eventually, I'll get the async
19868 pageflipping bits merged into the Xserver and this path and its extra
19869 bw wastage will be history! But still I'll be undermined by the
19872 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19874 commit e65c2c54f5efcec16e97085dbc41042007c39151
19875 Author: Chris Wilson <chris@chris-wilson.co.uk>
19876 Date: Thu Nov 10 11:37:46 2011 +0000
19878 sna: Store the sna pointer on the pixmap for cheap lookups
19880 A large part of the function preamble overhead is the multi-indirection
19881 lookup for retrieving the sna pointer. We can eliminate most of these by
19882 storing a pointer on the Pixmap as well, which we often need to lookup
19885 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19887 commit bddac63de2d26c6ad4ade2f0a038d756f7a41424
19888 Author: Chris Wilson <chris@chris-wilson.co.uk>
19889 Date: Thu Nov 10 10:55:13 2011 +0000
19891 sna: Defer purging a target buffer
19893 This is to keep the sanity checks upon the caches happy by keeping a
19894 potential flush out of the inactive cache.
19896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19898 commit 9083f7edcb05b1deb7a77c9c5a8b54fe73176648
19899 Author: Chris Wilson <chris@chris-wilson.co.uk>
19900 Date: Thu Nov 10 10:51:32 2011 +0000
19902 sna/gen3: Fixup some false asserts
19904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19906 commit ea01f7681901e3181204d3117fc0a23655de631f
19907 Author: Chris Wilson <chris@chris-wilson.co.uk>
19908 Date: Thu Nov 10 10:51:01 2011 +0000
19910 configure: Force full-debug builds to disable compiler optimisations
19912 commit a72563464f53f24246379afcf1365206fae9aad3
19913 Author: Chris Wilson <chris@chris-wilson.co.uk>
19914 Date: Thu Nov 10 00:55:55 2011 +0000
19916 sna/gen3: Clear the video shader flag upon batch reset
19918 Fixes regression from 3252b46e1.
19920 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19922 commit 3252b46e115e6cc46abbcc4f502feeb9edc31ad6
19923 Author: Chris Wilson <chris@chris-wilson.co.uk>
19924 Date: Wed Nov 9 15:16:04 2011 +0000
19926 sna/gen3: Pack video frames
19928 If you're going to benchmark, you may as well do so favourably.
19930 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19932 commit 8e926759528a12076bd7565672acd82a37aa3eb1
19933 Author: Chris Wilson <chris@chris-wilson.co.uk>
19934 Date: Wed Nov 9 14:44:54 2011 +0000
19936 sna/video: Pass texture video limits to the client
19938 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19940 commit bca58ff6f02b20ccf88937df8e2688f18cabac80
19941 Author: Chris Wilson <chris@chris-wilson.co.uk>
19942 Date: Wed Nov 9 14:18:57 2011 +0000
19944 sna: Clear kgem->busy when all requests have been processed
19946 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19948 commit 8fdbd01c94c0f6f56247feed8f225901ba39e18e
19949 Author: Chris Wilson <chris@chris-wilson.co.uk>
19950 Date: Wed Nov 9 13:14:04 2011 +0000
19952 sna/video: Use the normal bo cache for texture video streams
19954 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19956 commit 436a78684507129ca1245e065ff68870eb27624b
19957 Author: Chris Wilson <chris@chris-wilson.co.uk>
19958 Date: Wed Nov 9 12:48:27 2011 +0000
19960 sna: Track purgeability separately from reusability
19962 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19964 commit 2699c01a396b644a35a9acb969ab556454be3517
19965 Author: Chris Wilson <chris@chris-wilson.co.uk>
19966 Date: Wed Nov 9 12:27:52 2011 +0000
19968 sna: Handle non-reusable bo more correctly during retire
19970 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19972 commit fcc89d26600a57317007035fa720915e8a2ade81
19973 Author: Chris Wilson <chris@chris-wilson.co.uk>
19974 Date: Wed Nov 9 12:25:43 2011 +0000
19976 sna: Safety first; make all flinked bo unreusable
19978 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19980 commit f62b9f94f64f402e731463205635d386386e274e
19981 Author: Chris Wilson <chris@chris-wilson.co.uk>
19982 Date: Wed Nov 9 10:15:11 2011 +0000
19984 sna/video: Pass cropped source dimensions along with frame data
19986 So pack all the relevant details into the same structure.
19988 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19990 commit 5809b1ad934d6e19dbcf00f1b59e156761ec16b6
19991 Author: Chris Wilson <chris@chris-wilson.co.uk>
19992 Date: Wed Nov 9 09:41:53 2011 +0000
19994 sna/video: Tidy up copy routines
19996 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
19998 commit 1cd06100185a37ee33209fb18362da89f9646e6b
19999 Author: Chris Wilson <chris@chris-wilson.co.uk>
20000 Date: Tue Nov 8 17:37:44 2011 +0000
20002 sna: Protect against NULL deference of damage after reduction
20004 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20006 commit 4b100b099a4d5877c79989e8b3b4ffcd3aa2306f
20007 Author: Chris Wilson <chris@chris-wilson.co.uk>
20008 Date: Tue Nov 8 16:30:22 2011 +0000
20010 sna: Convert stippled spans to rects
20012 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20014 commit 7b95f87b26675af3a1923fef824c45e087098d61
20015 Author: Chris Wilson <chris@chris-wilson.co.uk>
20016 Date: Tue Nov 8 16:18:58 2011 +0000
20018 sna: Use the blitter for overlapping lines if the alu doesn't read dst
20020 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20022 commit 9a8d3a9dfff8be65b4ab28312ec49dbca36019e0
20023 Author: Chris Wilson <chris@chris-wilson.co.uk>
20024 Date: Tue Nov 8 13:51:00 2011 +0000
20026 sna: Avoid the penalty of only writing partial channels with glyphs
20028 Yes, writing 3 channels is slower than writing 4. But it's okay we
20029 simply ignore the alpha value anyway.
20031 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20033 commit 576c5169efd068c364fb7d266fbb6b1dd2ad3f4d
20034 Author: Chris Wilson <chris@chris-wilson.co.uk>
20035 Date: Tue Nov 8 13:27:05 2011 +0000
20037 sna: Tidy sna_copy_bitmap_blt
20039 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20041 commit 32bbeefb8563ae2ab3a105d0b1c9b55d9587b788
20042 Author: Chris Wilson <chris@chris-wilson.co.uk>
20043 Date: Tue Nov 8 13:17:20 2011 +0000
20045 sna: Pack clipped stippled uploads
20047 This are even more likely to fit inside the immediate payload.
20049 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20051 commit 3620cd2d157f3d81ccb76bce2ab813bd1e058878
20052 Author: Chris Wilson <chris@chris-wilson.co.uk>
20053 Date: Tue Nov 8 11:25:14 2011 +0000
20055 sna: Begin hooking up valgrind/memcheck
20057 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20059 commit aeaadb16093f83c9b830958c5946d4d048ffd26b
20060 Author: Chris Wilson <chris@chris-wilson.co.uk>
20061 Date: Tue Nov 8 11:03:19 2011 +0000
20063 sna: Handle creation of large Screen pixmap when the GPU is wedged
20065 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20066 References: https://bugs.freedesktop.org/show_bug.cgi?id=42606
20067 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20069 commit 7577d6ea45794b83d2f84b175d453d8a460b0e9b
20070 Author: Chris Wilson <chris@chris-wilson.co.uk>
20071 Date: Tue Nov 8 10:24:34 2011 +0000
20073 sna: Add some error messages to explain why we failed to create the screen
20075 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20077 commit 33256af40b3ce2cf8a899ced1fcbf40e316772e4
20078 Author: Chris Wilson <chris@chris-wilson.co.uk>
20079 Date: Tue Nov 8 09:55:07 2011 +0000
20081 sna: Fixes for DBG_NO_HW (i.e. simulated GPU hanges);
20083 A couple of the recent GPU paths were failing to check for !wedged.
20085 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20087 commit 5f0886dae29429f498fb10a12d5dc8de6bd798fc
20088 Author: Chris Wilson <chris@chris-wilson.co.uk>
20089 Date: Tue Nov 8 00:02:18 2011 +0000
20091 sna/trapezoids: Use ints for the offsets to accommodate multiplication
20093 Although the original precison need only 16-bits to store the offsets,
20094 after projecting on to the sample grid we need a few more bits of
20095 precision and so need a larger integer type to avoid overflow and render
20098 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20099 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42680
20100 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20102 commit 33351d5c3dd912534c54e64ccfc7adc4c6f6ecae
20103 Author: Chris Wilson <chris@chris-wilson.co.uk>
20104 Date: Mon Nov 7 22:57:44 2011 +0000
20106 sna/glyphs: Bypass masks for single glyphs
20108 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20110 commit e4872225ace7e3b129292f5822aa75b84c52beff
20111 Author: Chris Wilson <chris@chris-wilson.co.uk>
20112 Date: Mon Nov 7 22:57:05 2011 +0000
20114 sna/io: Minor tidy of setting command flags
20116 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20118 commit 5ba8ba7421e36bcdf83e84423595326939dbe7cd
20119 Author: Chris Wilson <chris@chris-wilson.co.uk>
20120 Date: Mon Nov 7 22:55:10 2011 +0000
20122 sna: Avoid overwriting an upload buffer during readback
20124 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20125 Buzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42677
20126 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20128 commit 8f7a8a80237db77452f02273bd8ade68dfba575f
20129 Author: Chris Wilson <chris@chris-wilson.co.uk>
20130 Date: Mon Nov 7 20:08:25 2011 +0000
20132 sna/composite: Minor fixes in operator and colour reduction
20134 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20136 commit b1234f3d3a27f326b8048e3d6b476021a26e9101
20137 Author: Chris Wilson <chris@chris-wilson.co.uk>
20138 Date: Mon Nov 7 20:07:37 2011 +0000
20140 sna: Expand multiplies of two 16-bit values to a full 32-bit range
20142 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20144 commit afdf931e61821985b31b339d1f346ddd7c4e9e3c
20145 Author: Chris Wilson <chris@chris-wilson.co.uk>
20146 Date: Mon Nov 7 20:06:35 2011 +0000
20148 sna: Add some more debugging in the hunt for overflows
20150 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20152 commit 8f3c845782fdb2fa8bdf751bdd7dd83ca02c42ac
20153 Author: Chris Wilson <chris@chris-wilson.co.uk>
20154 Date: Mon Nov 7 20:05:41 2011 +0000
20156 sna/blt: Small cleanups
20158 Whilst perusing for overflows, remove some redundant conditionals.
20160 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20162 commit 8e775cecccebe543d344721d45b2d43ee9f122b3
20163 Author: Chris Wilson <chris@chris-wilson.co.uk>
20164 Date: Mon Nov 7 20:01:33 2011 +0000
20166 sna/gen3: Fix false reduction of ComponentAlpha with white source
20168 The principle behind the opertator reduction of WHITE * maskca is valid,
20169 except that we failed to account for the src/mask transposition when
20170 emitting the vertices - garbage ensued.
20172 Given that we agressively reduce the shader required for WHITE * maskca,
20173 it does not seem worthwhile to special case the primitive emitter as
20176 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20177 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42676
20178 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20180 commit 65a440543b13e3e605a4a2d6209a460fbbe55736
20181 Author: Chris Wilson <chris@chris-wilson.co.uk>
20182 Date: Mon Nov 7 20:00:20 2011 +0000
20184 sna: Fix 16-bit overflow of rowlength for memcpy
20186 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20187 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42619
20188 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20190 commit d4edbd480445bc6aadd2c9f17262bd4b3eefbca6
20191 Author: Chris Wilson <chris@chris-wilson.co.uk>
20192 Date: Mon Nov 7 11:16:06 2011 +0000
20194 sna/glyph: Fallback if we fail to clear the scratch pixmap for the glyph mask
20196 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42663
20197 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20198 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20200 commit 05f3e96a69af7750d2d62cf5ba991b89f39bc88a
20201 Author: Chris Wilson <chris@chris-wilson.co.uk>
20202 Date: Mon Nov 7 09:28:19 2011 +0000
20204 test: Compile fix for make check
20206 Reported-by: Jeremy Huddleston <jeremyhu@apple.com>
20207 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20209 commit cd3d86037746f3a03c4004c1e3ba6e8281344745
20210 Author: Chris Wilson <chris@chris-wilson.co.uk>
20211 Date: Sun Nov 6 11:13:27 2011 +0000
20213 sna: Beware unsigned promotion of int16_t to uint32_t
20215 Mmakes for an unhappy mempy!
20217 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20219 commit 0b9408d972050cb02c1024926c406cd45508a158
20220 Author: Chris Wilson <chris@chris-wilson.co.uk>
20221 Date: Sun Nov 6 09:55:09 2011 +0000
20223 sna: Self-intersection of wide PolyLine are only drawn once
20225 We need to process the union of the PolyLine command if lineWidth!=0 so
20226 we cannot generally feed lineWidth==1 into our special case handler.
20227 Proving the lines do not intersect is as difficult as finding the
20228 intersections and thereby finding the union of the path - so there is no
20229 advantage in adding a check whether a wide line could be special cased.
20231 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20233 commit c76714c29d29687f941a9112e80223c817185d53
20234 Author: Chris Wilson <chris@chris-wilson.co.uk>
20235 Date: Sun Nov 6 09:43:47 2011 +0000
20237 test: Add a basic line tester
20239 Starting with exercising drawing of a single segment.
20241 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20243 commit c1e1e20fe7875262530a4d522e970985642c3f13
20244 Author: Chris Wilson <chris@chris-wilson.co.uk>
20245 Date: Sat Nov 5 20:59:17 2011 +0000
20247 sna: Add the pixman version to the debug output
20249 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20251 commit e2165f0e6b0620e3d788546924a2174506fbbde5
20252 Author: Chris Wilson <chris@chris-wilson.co.uk>
20253 Date: Sat Nov 5 19:46:13 2011 +0000
20255 sna: For a 32k max window size, we need to handle up to 128k strides
20257 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20258 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42619
20259 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20261 commit c5d94b21d5c65dcae3e294412a8f20bad347a6d8
20262 Author: Chris Wilson <chris@chris-wilson.co.uk>
20263 Date: Sat Nov 5 14:32:16 2011 +0000
20265 sna: Extend the zero-line segments by one
20267 Fixes misrendering of the wine tabs
20269 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20270 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42606
20271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20273 commit 0cf29afda9b061455bc438a22bbbfb1d6db6528e
20274 Author: Chris Wilson <chris@chris-wilson.co.uk>
20275 Date: Sat Nov 5 12:36:53 2011 +0000
20277 sna/glyphs: Fix clip detection for small masks
20281 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20283 commit d21c30d0b8b08ec85c3b5d9a02a6969f7c8b9315
20284 Author: Chris Wilson <chris@chris-wilson.co.uk>
20285 Date: Sat Nov 5 11:56:17 2011 +0000
20287 sna: Constrain GPU pixmaps to always fit within the blitter
20289 Otherwise we end up always doing expensive readbacks where we would
20290 obviously prefer it if we simply used cached memory for the CPU
20291 operation and then upload.
20293 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20295 commit e309cea3ae5c22c58a918008a52e3d8d694c9ff2
20296 Author: Chris Wilson <chris@chris-wilson.co.uk>
20297 Date: Sat Nov 5 11:28:06 2011 +0000
20299 sna: Do the supported PictOp check first
20301 There is no point even attempting a BLT operation if we know that it is
20302 an unusual render operation.
20304 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20306 commit ab3d6d878fa27515eb760107d963d522cc87cd77
20307 Author: Chris Wilson <chris@chris-wilson.co.uk>
20308 Date: Sat Nov 5 10:08:02 2011 +0000
20310 sna: Render small glyph masks on the CPU
20312 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20314 commit c25cf434d81ac1b4da29c9c75ca1d7c28be71d04
20315 Author: Chris Wilson <chris@chris-wilson.co.uk>
20316 Date: Sat Nov 5 01:04:54 2011 +0000
20318 sna/gen3: Initiailse the op for fill_boxes()
20320 After removing the memset(0), a couple of fields where left
20321 uninitialised, causing potential rendering glitches.
20323 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20325 commit 96e760ea7589cbe953bd0a6f2f41980a7f1c4795
20326 Author: Chris Wilson <chris@chris-wilson.co.uk>
20327 Date: Sat Nov 5 01:04:24 2011 +0000
20329 sna: s/flush/vblank/ fixes for DBG()
20331 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20333 commit 676cb4e38dc381b2ef4fb092b66db80687aa5013
20334 Author: Chris Wilson <chris@chris-wilson.co.uk>
20335 Date: Fri Nov 4 23:30:09 2011 +0000
20337 sna: Run the deferred flush at vrefresh
20339 This helps to reduce the perceived jerkiness of the redraw.
20341 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20342 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42413
20343 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20345 commit 8052c3904a1ac0308846ac13534c4c76bb96d5dd
20346 Author: Chris Wilson <chris@chris-wilson.co.uk>
20347 Date: Fri Nov 4 21:14:23 2011 +0000
20349 sna: Drain the delayed timer after forcing a flush
20351 The goal is to remove a spurious wakeup when we have no work outstanding
20352 afterwards and so would end up prematurely disabling the timer.
20354 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20356 commit 112c61f6557c2337d5f1b25eb71d0dc9ada6ec89
20357 Author: Chris Wilson <chris@chris-wilson.co.uk>
20358 Date: Fri Nov 4 21:13:55 2011 +0000
20360 sna: Convert GXcopy with -1 to GXset
20362 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20364 commit 573b23c6b618af808d3236395bbf4f4993b406e5
20365 Author: Chris Wilson <chris@chris-wilson.co.uk>
20366 Date: Fri Nov 4 15:57:24 2011 +0000
20368 sna: Add some asserts to detect buffer overflow.
20370 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20372 commit 4ba55c3d110ba51400a72da7d4f4c62efcea1335
20373 Author: Chris Wilson <chris@chris-wilson.co.uk>
20374 Date: Fri Nov 4 15:55:15 2011 +0000
20376 sna/gen5: Prefer using the BLT for many solid fills
20378 Even if it means incurring a context switch, the BLT unit is
20379 significantly faster so long as we do enough fills. And there is the
20382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20384 commit 703cf8abdbcda783e2bb0ceb423909ee8ca98176
20385 Author: Chris Wilson <chris@chris-wilson.co.uk>
20386 Date: Fri Nov 4 15:24:57 2011 +0000
20388 sna: Add earlier guards against to wedged to sna_glyph_blt
20390 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20392 commit fc415ec3d3e982d89d7a25db8d5fd18e0412a99b
20393 Author: Chris Wilson <chris@chris-wilson.co.uk>
20394 Date: Fri Nov 4 13:53:17 2011 +0000
20396 sna/gen5: Also experiment with lazy spans here
20398 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20400 commit 567cfa508f788b4dd5de9aafbe2f148bb6a8dbc6
20401 Author: Chris Wilson <chris@chris-wilson.co.uk>
20402 Date: Fri Nov 4 13:43:03 2011 +0000
20404 sna/gen6: Enable spans interface for boxes
20406 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20408 commit 465515a1446fd05fdc4691522803c94515187b4c
20409 Author: Chris Wilson <chris@chris-wilson.co.uk>
20410 Date: Fri Nov 4 13:40:13 2011 +0000
20412 sna: Silly compile fix, escaped when testing uxa
20414 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20416 commit a26c5d44ccaa7e9ff931948032b073f0a550343a
20417 Author: Chris Wilson <chris@chris-wilson.co.uk>
20418 Date: Fri Nov 4 12:36:40 2011 +0000
20420 uxa: Ensure that alphaMaps are mapped into the CPU for fallbacks
20422 Reported-by: Hans-Peter Budek <peter.budek@gmx.de>
20423 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20425 commit 34758895cdd93bd7671a78464e79b3891bca113d
20426 Author: Chris Wilson <chris@chris-wilson.co.uk>
20427 Date: Fri Nov 4 11:57:09 2011 +0000
20429 sna: Ensure operations on a ShmPixmap are synchronous with clients
20431 If we are rendering to or from a ShmPixmap, we need to be sure that the
20432 operation is complete prior to sending an XSync response to client in
20433 order to preserve mixed rendering coherency.
20435 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20437 commit 5525691eb024f2a04b486652c24e0e34176fdd5e
20438 Author: Chris Wilson <chris@chris-wilson.co.uk>
20439 Date: Fri Nov 4 10:49:48 2011 +0000
20441 sna/gen6: Poor man's spans layered on top of the exisiting composite
20443 Performance of this lazy interface looks inconclusive:
20447 xlib swfdec-giant-steps 1063.56 -> 710.68: 1.50x speedup
20448 xlib firefox-asteroids 3612.55 -> 3012.58: 1.20x speedup
20449 xlib firefox-canvas-alpha 15837.62 -> 13442.98: 1.18x speedup
20450 xlib ocitysmap 1106.35 -> 970.66: 1.14x speedup
20451 xlib firefox-canvas 33140.27) -> 30616.08: 1.08x speedup
20452 xlib poppler 629.97 -> 585.95: 1.08x speedup
20453 xlib firefox-talos-gfx 2754.37 -> 2562.00: 1.08x speedup
20456 xlib gvim 1363.16 -> 1439.64: 1.06x slowdown
20457 xlib midori-zoomed 758.48 -> 904.37: 1.19x slowdown
20458 xlib firefox-fishbowl 22068.29 -> 26547.84: 1.20x slowdown
20459 xlib firefox-planet-gnome 2995.96 -> 4231.44: 1.41x slowdown
20461 It remains off and a curiosity for the time being.
20463 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20465 commit 5dba7028cd5024282b601c41ddfc9081eaa1d189
20466 Author: Chris Wilson <chris@chris-wilson.co.uk>
20467 Date: Fri Nov 4 11:24:03 2011 +0000
20469 sna: Make sure the alpha map is accessible by the CPU on fallbacks
20471 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20473 commit 70c9e70f35af21de2b5b56d72f2a9e7fe9ed49d6
20474 Author: Chris Wilson <chris@chris-wilson.co.uk>
20475 Date: Fri Nov 4 10:19:21 2011 +0000
20477 sna: Don't reuse partial vmapped bo
20479 A fun use after free.
20481 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20483 commit 7e8c9a5b8b1625fdfe885740c36da3f4c1373ee6
20484 Author: Chris Wilson <chris@chris-wilson.co.uk>
20485 Date: Wed Nov 2 14:02:56 2011 +0000
20487 sna: Submit the batch on the next blockhander if operation overflows
20489 If an operation overflows from one batch into another, we submit the
20490 complete batch and begin a new. That new batch will not be submitted
20491 unless it is filled or on the next delayed flush update. This can cause
20492 a flicker as a large operation is broken up, such as performing a
20493 CopyArea through a Clipmask. So if we submit a full batch during a flush
20494 interval, immediately flush any partial batch at the next blockhandler.
20496 This stops rude Santa flashing Rudolf in xsnow!
20498 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20500 commit 5b21838471651a59d3e425c05a389c3f9196e21d
20501 Author: Chris Wilson <chris@chris-wilson.co.uk>
20502 Date: Thu Nov 3 22:51:24 2011 +0000
20504 sna: Compute the correct extents for the PolyRectangle
20506 Otherwise we may leave one behind...
20508 A regression from the introduction of sna_poly_rectangles:
20509 40af32a0e9ed971a1f2c2a45266f32016bda7ed3 (sna: Execute blits directly
20512 Reported-by: Matti Hamalainen <ccr@tnsp.org>
20513 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42568
20514 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20516 commit 2174f840158aa9cfa370ade38be28f8dc8e4b526
20517 Author: Chris Wilson <chris@chris-wilson.co.uk>
20518 Date: Thu Nov 3 20:41:31 2011 +0000
20520 uxa: Remove caching of surface binding location
20522 If the pixmap were to be used multiple times within a batch with
20523 mulitple formats, the cache would only return the initial location with
20524 the incorrect format and so cause rendering glitches. For instance, GTK+
20525 uses the same pixmap as an xrgb source and as an argb mask in order to
20526 premultiply and composite in a single pass. Rather than introduce an
20527 overly complication caching (handle, format) mechanism, kiss and remove
20528 the invalid implementation.
20530 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40926
20531 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20533 commit a1b40a20bbba4b70990a8983a2916d3d5850d828
20534 Author: Chris Wilson <chris@chris-wilson.co.uk>
20535 Date: Thu Nov 3 19:27:23 2011 +0000
20537 sna: Support binding of a bo for multiple formats
20539 Applications may use the same pixmap with multiple formats within the
20540 same operation. For instance, you can premultiply and composite a normal
20541 pixmap in this manner. However, as we reused the sampler binding
20542 locations of the source (without an alpha channel) for the mask, we
20543 failed to read and multiply by the alpha channel causing it to remain
20544 black instead of transparent.
20546 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40926
20547 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20549 commit 31c5eb8e906bf8e59743372edb2d703b50cd311e
20550 Author: Chris Wilson <chris@chris-wilson.co.uk>
20551 Date: Thu Nov 3 17:07:48 2011 +0000
20553 sna: Clean up the fallback code for glyphs
20555 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20557 commit fa0fefd638d148d48760c95fe05520cddd96e5a4
20558 Author: Chris Wilson <chris@chris-wilson.co.uk>
20559 Date: Thu Nov 3 15:33:25 2011 +0000
20561 sna: Restore the lowlevel glyph routines for the sake of Damage
20563 Damage bypasses the Text interface, preventing the backend from hooking
20564 into the font and storing private glyph representations, and calls
20565 directly into the Glyph routines. So to prevent a segfault we have to
20568 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20570 commit 1677b273af67e2b690bd21a1b43a7d9f9d5a70c1
20571 Author: Chris Wilson <chris@chris-wilson.co.uk>
20572 Date: Thu Nov 3 14:31:53 2011 +0000
20574 sna: Skip encoding zero sized glyphs
20576 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20578 commit 239cfb99f9697392d7c25328093e6662ad04bdc9
20579 Author: Chris Wilson <chris@chris-wilson.co.uk>
20580 Date: Thu Nov 3 14:25:54 2011 +0000
20582 sna: Unroll the quadword upload of the glyph data
20584 We know that the length is nicely aligned and so can avoid a relatively
20585 expensive call into memcpy.
20587 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20589 commit e2542bad88573e590ef4cbeed6b5884166b56adb
20590 Author: Chris Wilson <chris@chris-wilson.co.uk>
20591 Date: Thu Nov 3 14:17:05 2011 +0000
20593 sna: Add the missing returns to prevent fbImageGlyphBlt fallbacks
20595 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20597 commit bc032c9be1bab7477fbf4b2a64fb7aca6d75ac18
20598 Author: Chris Wilson <chris@chris-wilson.co.uk>
20599 Date: Thu Nov 3 11:47:18 2011 +0000
20601 sna: Coalesce reduction of cpu damage
20603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20605 commit c92671b33ef30a8eb723eb366692ae3a5878b8f9
20606 Author: Chris Wilson <chris@chris-wilson.co.uk>
20607 Date: Thu Nov 3 11:47:18 2011 +0000
20609 sna: Coalesce reduction of gpu damage
20611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20613 commit 3b4a508eb022e9ade5929d3ebc97e0f9216a80b1
20614 Author: Chris Wilson <chris@chris-wilson.co.uk>
20615 Date: Thu Nov 3 00:32:29 2011 +0000
20617 sna: Translate glyphs into MSBFirst upon initial load
20619 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20621 commit 8a259e34d3295af3be864e4fae96e6c3a9a632c7
20622 Author: Chris Wilson <chris@chris-wilson.co.uk>
20623 Date: Thu Nov 3 01:05:52 2011 +0000
20625 sna: gc->miTranslate is always 1
20627 So we can perform some constant folding and eliminate dead code.
20629 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20631 commit 8f68f9e5f8341e20c4e1e46044f79806a44ecd03
20632 Author: Chris Wilson <chris@chris-wilson.co.uk>
20633 Date: Thu Nov 3 00:59:46 2011 +0000
20635 sna: Trimming is redundant given that we always have a CompositeClip
20637 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20639 commit f4bdd84b846e6e778b7c496e7ed0d51e77310f36
20640 Author: Chris Wilson <chris@chris-wilson.co.uk>
20641 Date: Thu Nov 3 00:48:10 2011 +0000
20643 sna: Simplify the uncommon check for gpu-only damage by using damage-all
20645 The use of a gpu-only scratch bo is uncommon with the core acceleration
20646 routines, and we can eliminate the check for not incrementing the damage
20647 by allocating a damage-all and using the common optimisation of
20650 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20652 commit 353fa4218cd904857fe95ef140e7f38aefe07578
20653 Author: Chris Wilson <chris@chris-wilson.co.uk>
20654 Date: Wed Nov 2 22:53:51 2011 +0000
20656 sna: Don't call into retire unless there are outstanding requests
20658 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20660 commit 6fdd0f254d2e58d425d6d23023e9db3d2924c6be
20661 Author: Chris Wilson <chris@chris-wilson.co.uk>
20662 Date: Wed Nov 2 22:23:16 2011 +0000
20664 sna: Use an integer value for the sentinel
20666 clang complains otherwise.
20668 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20670 commit 1073c78f6cebfd6380b53dd891b7a72e50f398d6
20671 Author: Chris Wilson <chris@chris-wilson.co.uk>
20672 Date: Wed Nov 2 22:14:40 2011 +0000
20674 sna: Pack small 1-bpp uploads into immediate buffers
20676 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20678 commit 73b2ef5a7de8d733fa1821e5010480ede305e25e
20679 Author: Chris Wilson <chris@chris-wilson.co.uk>
20680 Date: Wed Nov 2 22:02:51 2011 +0000
20682 sna: gc->pCompositeClip always exists after validate
20684 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20686 commit 6fd0754e9c81164ca0fdcf96ba2bfb419568187f
20687 Author: Chris Wilson <chris@chris-wilson.co.uk>
20688 Date: Wed Nov 2 18:03:07 2011 +0000
20690 sna: defer calling fbValidateGC until we fallback
20692 A secondary effect is that this prevents needless migration of the
20693 tiling pixmap which we want to optimistically keep on the GPU.
20695 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20697 commit 6f8875e216194474e818ba6e743e7c22b817f96e
20698 Author: Chris Wilson <chris@chris-wilson.co.uk>
20699 Date: Wed Nov 2 15:41:40 2011 +0000
20701 sna: Wait an extra interval before switching off periodic updates
20703 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20705 commit def6b5e733e623333abf596bbdbabf8dd0ac8b00
20706 Author: Chris Wilson <chris@chris-wilson.co.uk>
20707 Date: Wed Nov 2 14:35:45 2011 +0000
20709 sna: Only execute the accel block handler if we are about to block
20711 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20713 commit 088b875affb8655010bc678ca4a3e87703c92daf
20714 Author: Chris Wilson <chris@chris-wilson.co.uk>
20715 Date: Wed Nov 2 13:11:04 2011 +0000
20717 sna: Use the packed 8x8 stipple for non-repeating small rectangles
20719 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20721 commit 0c373db73d1d2b7cf844653307c60a3891a3cc10
20722 Author: Chris Wilson <chris@chris-wilson.co.uk>
20723 Date: Wed Nov 2 12:43:55 2011 +0000
20725 sna: Fix the 8x8 stippled pattern origin
20727 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20729 commit 9cdf8be55029f30b47cfc39ddd2754d40602cf2c
20730 Author: Chris Wilson <chris@chris-wilson.co.uk>
20731 Date: Wed Nov 2 12:17:31 2011 +0000
20733 sna: Add clipped stippled upload support
20737 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20739 commit 6553c9e1cbfcecf781ba106de99f9e70c6d373ba
20740 Author: Chris Wilson <chris@chris-wilson.co.uk>
20741 Date: Wed Nov 2 10:03:45 2011 +0000
20743 sna: Quieten a fewer compiler sign compare warnings
20745 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20747 commit 42e2036cda6b76f2b64e28a81bdaa661e4cbffa5
20748 Author: Chris Wilson <chris@chris-wilson.co.uk>
20749 Date: Tue Nov 1 20:56:48 2011 +0000
20751 sna: Use bo for scratch pixmaps allocated by the dix/mi routines
20753 Now that we have the rudiments of accelerated deep-plane copies, we can
20754 begin to benefit from using BO for the core dix/mi routines like
20757 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20759 commit 45cc952994e169590ac4300f4ccf54810377f7f2
20760 Author: Chris Wilson <chris@chris-wilson.co.uk>
20761 Date: Tue Nov 1 20:20:35 2011 +0000
20763 sna: Fix debugging assertions for Composite
20765 When comparing drawable clip extents against pixmap boundaries we need
20766 to include the pixmap screen offset on a Composited desktop.
20768 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20770 commit ebc9f3966d28343aebeba6886500abf511631785
20771 Author: Chris Wilson <chris@chris-wilson.co.uk>
20772 Date: Tue Nov 1 19:47:51 2011 +0000
20774 sna: Accelerate deep-plane copy
20776 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20778 commit e0fd07bc251296784bf70f02877765171a053cc6
20779 Author: Chris Wilson <chris@chris-wilson.co.uk>
20780 Date: Tue Nov 1 17:02:17 2011 +0000
20782 sna: Accelerate XYPixmap upload when using GXcopy
20784 Mostly for the lols.
20786 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20788 commit 22c43efe6b9b5f669593aa9f3af6ee437426c5d2
20789 Author: Chris Wilson <chris@chris-wilson.co.uk>
20790 Date: Tue Nov 1 13:43:23 2011 +0000
20792 sna: Implement 8x8 stippled rect fills
20794 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20796 commit c58b7643e9b86599a41ede516a778bd9869adc8e
20797 Author: Chris Wilson <chris@chris-wilson.co.uk>
20798 Date: Tue Nov 1 10:13:20 2011 +0000
20800 sna: Accelerate 1bpp uploads
20802 These still get used (see Wine and Swing) by applications which like to
20803 do "crisp" 1-bit rendering on the client side and then put onto the
20804 scanout. So avoid the readbacks, and push them through the BLT instead. It
20805 turns out to be faster than using fb too, bonus!
20807 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20809 commit 8939ddbc27efedbb3e446dd26425a6bed39577f2
20810 Author: Chris Wilson <chris@chris-wilson.co.uk>
20811 Date: Mon Oct 31 17:45:02 2011 +0000
20813 sna: Improve rendering of thin segments
20815 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20817 commit 98b830a975e04685455b244a8414a293857cbd00
20818 Author: Chris Wilson <chris@chris-wilson.co.uk>
20819 Date: Mon Oct 31 17:41:27 2011 +0000
20821 sna: Improve rendering of thin lines
20823 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20824 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42443
20825 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20827 commit 6a5834ae560ef3b05c42b58ee0654316a66eaed0
20828 Author: Chris Wilson <chris@chris-wilson.co.uk>
20829 Date: Mon Oct 31 16:28:59 2011 +0000
20831 sna/gen3: Fix invalid assert
20833 SHADER_CONSTANT is expected here, the other IMMEDIATES however should
20834 have already been handled.
20836 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20838 commit e12375d05bfb32933897b62829d82c6f1aa81d70
20839 Author: Chris Wilson <chris@chris-wilson.co.uk>
20840 Date: Mon Oct 31 14:24:19 2011 +0000
20842 sna: Fix a few minor cut'n'paste errors
20844 Use the constant value [last_box-box] where applicable.
20846 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20848 commit 20a4f533958991557a3b05af00f2856657579d68
20849 Author: Chris Wilson <chris@chris-wilson.co.uk>
20850 Date: Mon Oct 31 14:14:40 2011 +0000
20852 sna: Apply the GPU damage for clipped PolyFillRectangles
20854 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20855 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42425
20856 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20858 commit 7172f28c2b36aaa8c3a74ae4a7f699d3d4af10eb
20859 Author: Chris Wilson <chris@chris-wilson.co.uk>
20860 Date: Mon Oct 31 12:53:35 2011 +0000
20862 sna: Fix placement of clipped wide PolyRectangle
20864 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20866 commit 59535d0e1c587cfe9249ab1baf9acf8470b49b8b
20867 Author: Chris Wilson <chris@chris-wilson.co.uk>
20868 Date: Mon Oct 31 12:10:52 2011 +0000
20870 sna: Set the flush interval based on output vrefresh
20872 Rather than a blank 25Hz, use twice the vblank interval to hopefully
20875 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20877 commit 9a2e59bfba6748325918f895b3e58351c7eb4baa
20878 Author: Chris Wilson <chris@chris-wilson.co.uk>
20879 Date: Mon Oct 31 11:56:53 2011 +0000
20881 sna: Fix reversal of zero/wide for PolyRectangle
20883 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20884 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42414
20885 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20887 commit b1287e49347e0fff67f43d157fa87758d5cbe588
20888 Author: Chris Wilson <chris@chris-wilson.co.uk>
20889 Date: Mon Oct 31 11:27:36 2011 +0000
20891 sna/video: Fix copy region for cropped video
20893 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20894 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42412
20895 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20897 commit b4ab412cb3eb8d37a33f5eeae5c613090ad2df0f
20898 Author: Chris Wilson <chris@chris-wilson.co.uk>
20899 Date: Mon Oct 31 10:37:43 2011 +0000
20901 sna/blt: Optimise fill with GXcopy and pixel==0 to a GXclear
20903 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20905 commit 088728ad26c7efc2d65b97ff22d526c65a0fa442
20906 Author: Chris Wilson <chris@chris-wilson.co.uk>
20907 Date: Mon Oct 31 10:04:35 2011 +0000
20909 sna/blt: Don't idiotically override alu for pixel==0 and track alu
20911 Another classic and utterly stupid mistake. At least I was consistent in
20914 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
20915 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42414
20916 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20918 commit 9c399c74adc5e5bc7d425382a74d84315d2cef02
20919 Author: Chris Wilson <chris@chris-wilson.co.uk>
20920 Date: Mon Oct 31 09:20:54 2011 +0000
20922 sna: Add an option to disable the delayed flush
20925 Option "DelayedFlush" "False"
20928 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20930 commit c8860ce26eaa8f9e876af2354d61c35338076ac8
20931 Author: Chris Wilson <chris@chris-wilson.co.uk>
20932 Date: Sun Oct 30 23:25:48 2011 +0000
20934 sna: Report whether anything was retired to the inactive list
20936 So that we can skip a fresh search of the inactive list if nothing
20939 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20941 commit a7adcc8e47ba550e7c36a6543a6e0d99040c7354
20942 Author: Chris Wilson <chris@chris-wilson.co.uk>
20943 Date: Sun Oct 30 21:19:35 2011 +0000
20945 sna: Stash the last scratch pixmap
20947 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20949 commit c8d4f5d32af94f712cc81ff495b19907261759ea
20950 Author: Chris Wilson <chris@chris-wilson.co.uk>
20951 Date: Sun Oct 30 21:21:25 2011 +0000
20953 sna: Correct PolyRectangle verticals to be the end-point not the length
20955 Fixes an off-by-one in the drawing of the rectangles caused by a
20956 too-literal translation of the rectangles into boxes.
20958 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20960 commit 5c45622b889ba66b6476241601e1458d219c45ba
20961 Author: Chris Wilson <chris@chris-wilson.co.uk>
20962 Date: Sun Oct 30 17:49:19 2011 +0000
20964 uxa/glyphs: Fallback rather than fail to render with a non-solid Source
20966 Reported-by: Uli Schlachter <psychon@znc.in>
20967 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31819
20968 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20970 commit e6eb803cb44b01e300fb4b08304227a430912beb
20971 Author: Chris Wilson <chris@chris-wilson.co.uk>
20972 Date: Sun Oct 30 17:44:26 2011 +0000
20974 sna: Loop over all clip rects for glyph blt rather than fallback
20976 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20978 commit 1eefb75c91fb769551b485f1b2c6b59023c69906
20979 Author: Chris Wilson <chris@chris-wilson.co.uk>
20980 Date: Sun Oct 30 17:17:36 2011 +0000
20982 sna: Don't mess with NDEBUG
20984 This is set in configure and redefining it later inside the C files just
20985 leads to trouble and broken compilation.
20987 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
20989 commit 7237374a03d1e0c2e6c0b2d0eb42afc8df25df0d
20990 Author: Chris Wilson <chris@chris-wilson.co.uk>
20991 Date: Sun Oct 30 16:46:10 2011 +0000
20993 configure: Bump version to 2.16.901
20995 commit 0a78a75e75225d89e465804c482a667840a2c8de
20996 Author: Chris Wilson <chris@chris-wilson.co.uk>
20997 Date: Sun Oct 30 16:45:46 2011 +0000
20999 NEWS: Add release notes for 2.16.901
21001 commit c7552cb819ae8c90975193c159236e0070628dd9
21002 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
21003 Date: Sun Oct 30 16:35:14 2011 +0100
21005 i810: kill mmio vga access helpers
21007 Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
21009 commit e8b8adeab3b2e69a8684217271604a5f658f617d
21010 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
21011 Date: Sun Oct 30 16:34:05 2011 +0100
21013 i810: use vgahw functions instead of roll-your-own
21015 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
21017 commit 4aeea489ed3189d7ebcae4199153c1ae736ebf68
21018 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
21019 Date: Sun Oct 30 16:07:45 2011 +0100
21021 i810: rip out PIO access functions
21023 All the io done at init time is already converted to libpciaccess
21025 Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
21027 commit e735636194671e221eca3e86b9bd06fd0d4b6c5a
21028 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
21029 Date: Sun Oct 30 15:51:49 2011 +0100
21031 i810: kill dead i830 defines
21033 Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
21035 commit bb840ee09fa620648065884306fbe3e0da37ad77
21036 Author: Chris Wilson <chris@chris-wilson.co.uk>
21037 Date: Sun Oct 30 16:13:04 2011 +0000
21039 sna: Convert PolyRectangles to use Box32
21041 For the same overflow protection.
21043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21045 commit af52e234a78bbde964a93ef8cce21f41a5c959f3
21046 Author: Chris Wilson <chris@chris-wilson.co.uk>
21047 Date: Sun Oct 30 15:34:34 2011 +0000
21049 sna: Fix compilation
21051 Remember to drink coffee and compile "obvious" fixes before pushing.
21053 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21055 commit 23b314729c43101e515e2625db2a2401c72c2596
21056 Author: Chris Wilson <chris@chris-wilson.co.uk>
21057 Date: Sun Oct 30 15:24:21 2011 +0000
21059 sna: Rewrite box32_add_rect() to enable int32_t promotion
21061 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21063 commit 427622aca31b7463eb2dcd52e352dc8997e84baa
21064 Author: Chris Wilson <chris@chris-wilson.co.uk>
21065 Date: Sun Oct 30 12:02:42 2011 +0000
21067 sna: Push overflow detection to the end of extents
21069 Removes a couple of conditionals from the middle of the hotpath and on
21070 Intel we are not realising the benefit of only utilising 16-bit values.
21072 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21074 commit 2d3aba8518ae790325c6123dacc3ce73b6e91cbd
21075 Author: Chris Wilson <chris@chris-wilson.co.uk>
21076 Date: Sun Oct 30 11:22:26 2011 +0000
21078 Revert "sna: Don't flush the render caches if in the process of writing again"
21080 This reverts commit 15266e1b9500f6b348661c60d1982bde911f2d0e.
21082 KDE relies upon the ability to render into a sampler and then render
21083 upon itself. Not the first sign of madness...
21085 Will have to find another way of winning back the compwinwin
21088 commit d4062705178856cffc83030fa40f758e9d884566
21089 Author: Chris Wilson <chris@chris-wilson.co.uk>
21090 Date: Sun Oct 30 10:40:32 2011 +0000
21092 sna/composite: Fix incorrect operator reduction for RenderFillRectangles
21094 As exemplified by KDE (using Kate) on gen3, it would attempt to render a
21095 large set of boxes using OVER and a transparent colour. As gen3 copied
21096 across some of the BLT assumptions, it was incorrectly reducing that to
21097 a CLEAR and thus rendering incorrectly.
21099 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21101 commit da3f8656744f7c801608c7c89a859c0f7dc9e6e0
21102 Author: Jeremy Huddleston <jeremyhu@apple.com>
21103 Date: Sat Oct 29 20:15:11 2011 -0700
21105 Use malloc/calloc/realloc/free directly
21107 Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
21109 commit 17ba07a7befc7f442ae1c917a087ac026516d886
21110 Author: Chris Wilson <chris@chris-wilson.co.uk>
21111 Date: Sat Oct 29 20:03:37 2011 +0100
21113 sna/gen3: Set the src flags for fill
21115 A regression from c2040fb8e62076d87f7a5b407505ee3f3482945e (sna: Remove
21116 the memset(0) of the fill op) left the flags for the source channel
21117 uninitalised leaving the shader construction undefined and causing
21120 Reported-by: Paul Neumann <paul104x@yahoo.de>
21121 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42367
21122 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21124 commit 4835fe5b2e19ac456773565399f334845f7442c0
21125 Author: Chris Wilson <chris@chris-wilson.co.uk>
21126 Date: Sat Oct 29 18:56:29 2011 +0100
21128 sna/genX: Improve reduction of Render operator to BLT alu
21130 This appeared to introduce a visual gitch into the xfce4 selection box
21133 References: https://bugs.freedesktop.org/show_bug.cgi?id=42367
21134 Reported-by: Paul Neumann <paul104x@yahoo.de>
21135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21137 commit 33383fde55609fb4bd372d2d41a6254799499225
21138 Author: Chris Wilson <chris@chris-wilson.co.uk>
21139 Date: Sat Oct 29 12:53:15 2011 +0100
21141 sna: Tweak PolyRectangle blitting
21143 A box of height/width 1 and lineWidth 0 can also be drawn with a single
21144 box as the far edge is contiguous with far side of the near edge.
21146 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21148 commit 64a57c8dbec81f6420ff032985237eb16cebfb87
21149 Author: Chris Wilson <chris@chris-wilson.co.uk>
21150 Date: Sat Oct 29 11:01:51 2011 +0100
21152 sna: Further clipping bugs
21154 Make sure the line is clipped to the current clip box and discard if the
21155 clip is reduced to nothing.
21157 References: https://bugs.freedesktop.org/show_bug.cgi?id=42361
21158 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21160 commit 719626ef7918d5cd69d13f983b8b615aefe215d6
21161 Author: Chris Wilson <chris@chris-wilson.co.uk>
21162 Date: Sat Oct 29 10:26:08 2011 +0100
21164 sna: Increment clip box when short-circuiting the iteration
21166 Reported-by: nkalkhof@web.de
21167 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi
21168 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21170 commit 32a9e0352a1e28509580d84bfc5377d8fae913c4
21171 Author: Chris Wilson <chris@chris-wilson.co.uk>
21172 Date: Fri Oct 28 23:14:24 2011 +0100
21174 sna: Retire and search inactive again before creating new buffer
21176 Attempt to retire an active buffer into the inactive cache before giving
21177 up and creating a new buffer.
21179 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21181 commit 15266e1b9500f6b348661c60d1982bde911f2d0e
21182 Author: Chris Wilson <chris@chris-wilson.co.uk>
21183 Date: Fri Oct 28 18:00:56 2011 +0100
21185 sna: Don't flush the render caches if in the process of writing again
21187 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21189 commit d3c7ee921172d4898ebdc6e944812a02426008e9
21190 Author: Chris Wilson <chris@chris-wilson.co.uk>
21191 Date: Fri Oct 28 16:44:16 2011 +0100
21193 sna: Faster unclipped rectilinear segments
21195 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21197 commit 3cd909cfe9e7d7e97a1d51513a049e312fd9ddae
21198 Author: Chris Wilson <chris@chris-wilson.co.uk>
21199 Date: Fri Oct 28 15:48:49 2011 +0100
21201 sna: Call miZeroArcLine where appropriate for GPU paths
21203 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21205 commit 9921c98df070032e082570f0663f6d3ad0cccd56
21206 Author: Chris Wilson <chris@chris-wilson.co.uk>
21207 Date: Fri Oct 28 14:09:11 2011 +0100
21209 sna: Faster unclipped PolyFillRect
21211 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21213 commit c1718f96f520874c1fb2af0eb975c7f548762d34
21214 Author: Chris Wilson <chris@chris-wilson.co.uk>
21215 Date: Fri Oct 28 14:01:24 2011 +0100
21217 sna: Faster unclipped PolyPoint
21219 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21221 commit a4762ed891a50238abddf6c762a053db5c34b87f
21222 Author: Chris Wilson <chris@chris-wilson.co.uk>
21223 Date: Fri Oct 28 13:39:17 2011 +0100
21225 sna: Call directly into tiled rects for tiled spans.
21227 As we already know the extents and that this is a candidate for
21228 GPU-acceleration, we can skip over those steps and emit the tiled rects.
21230 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21232 commit a02069df2ddc259d65ad468834f7968d283b713f
21233 Author: Chris Wilson <chris@chris-wilson.co.uk>
21234 Date: Fri Oct 28 11:30:37 2011 +0100
21236 sna: Faster unclipped spans
21238 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21240 commit 48eff1676c4066c8b8e069fd579a1b6fc2329960
21241 Author: Chris Wilson <chris@chris-wilson.co.uk>
21242 Date: Thu Oct 27 23:30:06 2011 +0100
21244 sna/gen2: Reset BLT registers across 2D/3D context switches
21246 It appears the hardware trashes the BLT registers after a 3D context
21247 switch, so we need to reload.
21249 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21251 commit f006fac72a733e047b9468704e9c9206f62f96f5
21252 Author: Chris Wilson <chris@chris-wilson.co.uk>
21253 Date: Thu Oct 27 21:50:25 2011 +0100
21255 sna: Fast path typically unclipped tiled rects
21257 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21259 commit d6f543011bf3d03b9298b5ceeddfc1fc20fc534b
21260 Author: Chris Wilson <chris@chris-wilson.co.uk>
21261 Date: Thu Oct 27 20:16:35 2011 +0100
21263 sna: Use private identifier for internal scratch pixmaps
21265 This was meant to be a part of the previous commit.
21267 These are the scratch pixmaps for which we truly do want to allocate
21270 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21272 commit 14a065a267523f9a31be07daaab19b67e9f02290
21273 Author: Chris Wilson <chris@chris-wilson.co.uk>
21274 Date: Thu Oct 27 18:07:53 2011 +0100
21276 sna: Create scratch pixmap on the CPU
21278 Contrary to earlier belief, the only usage of
21279 CREATE_PIXMAP_USAGE_SCRATCH in the dix/mi are either along paths unused
21280 by us or more importantly for operations which are detrimental to create
21281 GPU-only buffers. So create CPU pixmaps instead.
21283 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21285 commit 5d52d5009375618bb703909fa0fad04e6d54a2ba
21286 Author: Chris Wilson <chris@chris-wilson.co.uk>
21287 Date: Thu Oct 27 17:06:51 2011 +0100
21289 sna: Tweak zero line spans to include missing last segment
21291 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21293 commit 72e00096351137bd7eb0ae91474657a597836252
21294 Author: Chris Wilson <chris@chris-wilson.co.uk>
21295 Date: Thu Oct 27 16:36:39 2011 +0100
21297 sna: Fix advance through clip boxes for fill->boxes
21299 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21301 commit c7bd89fb54686c14b687c007e42c34c23423b0fd
21302 Author: Chris Wilson <chris@chris-wilson.co.uk>
21303 Date: Thu Oct 27 16:23:35 2011 +0100
21305 sna: Fix PolyPoint not to always send 512 boxes
21307 Just emit the remaining boxes!
21309 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21311 commit cfd5b7f6ce80af911bfc8b71bba776ea814e307a
21312 Author: Chris Wilson <chris@chris-wilson.co.uk>
21313 Date: Thu Oct 27 15:28:11 2011 +0100
21315 sna: Use fill->boxes to pack PolyFillRectangle
21317 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21319 commit 8d304f47e35be344550f389158eac3d4c21d9440
21320 Author: Chris Wilson <chris@chris-wilson.co.uk>
21321 Date: Thu Oct 27 14:43:46 2011 +0100
21323 sna: Use fill->boxes for rectilinear blits
21325 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21327 commit 5d51ff0cf901b808c3cd5aea761d51e6d15b73bf
21328 Author: Chris Wilson <chris@chris-wilson.co.uk>
21329 Date: Thu Oct 27 13:57:59 2011 +0100
21331 sna: Clip PolyLines to the current clip box
21333 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21335 commit 990043f6b8cf8e580aad2e52f8d236a6aec838b5
21336 Author: Chris Wilson <chris@chris-wilson.co.uk>
21337 Date: Thu Oct 27 12:33:55 2011 +0100
21339 sna: Handle singular clip regions for PolyRectangles
21341 Otherwise X will crash!
21343 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21345 commit 991582f6678581e3d92c0497e10ecee67536efc2
21346 Author: Chris Wilson <chris@chris-wilson.co.uk>
21347 Date: Thu Oct 27 09:27:41 2011 +0100
21349 sna/io: Update batch mode after submitting partial requests
21351 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21353 commit 91745de113340c8a281e8cb724b4a7e0b7a0711a
21354 Author: Chris Wilson <chris@chris-wilson.co.uk>
21355 Date: Thu Oct 27 09:17:50 2011 +0100
21357 sna: Further improve use of fill->boxes for PolyRectangle
21359 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21361 commit ffe1bdea82c07a2738743855dfa83494531b41aa
21362 Author: Chris Wilson <chris@chris-wilson.co.uk>
21363 Date: Wed Oct 26 18:37:29 2011 +0100
21365 sna/blt: After submitting the batch, it will be empty
21367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21369 commit b34e7454574f36796560955fb9d3d33ca00924d6
21370 Author: Chris Wilson <chris@chris-wilson.co.uk>
21371 Date: Wed Oct 26 18:11:25 2011 +0100
21373 sna/blt: Convert Copy with color 0 into a Clear
21375 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21377 commit 389241aa6469b2e64a58cc0527e3310fdc69ffb7
21378 Author: Chris Wilson <chris@chris-wilson.co.uk>
21379 Date: Wed Oct 26 17:50:38 2011 +0100
21381 sna: Suppress an overwritten fill
21383 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21385 commit baf78fcddb87eba987163485ad8fd227fb244f29
21386 Author: Chris Wilson <chris@chris-wilson.co.uk>
21387 Date: Wed Oct 26 12:54:13 2011 +0100
21389 sna/gen2: Cache the last solid colour for spans
21391 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21393 commit ff4715a36999a43cee20cf3474b4ae0af646342c
21394 Author: Chris Wilson <chris@chris-wilson.co.uk>
21395 Date: Wed Oct 26 12:40:46 2011 +0100
21397 sna/composite: Discard opaque masks
21399 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21401 commit 0079579ebd548fc1b490a42928373507a93229bd
21402 Author: Chris Wilson <chris@chris-wilson.co.uk>
21403 Date: Wed Oct 26 11:49:53 2011 +0100
21405 sna: Fix debug compilation, again.
21407 commit 5988422455ebb2a3b1cafadc1ff302c35a7a5039
21408 Author: Chris Wilson <chris@chris-wilson.co.uk>
21409 Date: Wed Oct 26 11:47:45 2011 +0100
21411 sna: Handle degenerate copy area whilst wedged
21413 We want to return early to avoid handing the empty box as the lower code
21414 makes the assumption that it has work to do.
21416 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21418 commit 3fc7cd2dc8e2b9c0d4143a9f6657f84050ae249a
21419 Author: Chris Wilson <chris@chris-wilson.co.uk>
21420 Date: Sun Oct 23 20:59:50 2011 +0100
21422 sna/blt: Rename the composite paths
21424 Whilst re-reading, I found the old naming scheme slightly muddled and
21425 not nearly as descriptive as it should be. Hopefully this slight tweak
21426 reads better in 6 months time.
21428 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21430 commit 280325900bf3c17c3177f6e3f52236d6b452e0f3
21431 Author: Chris Wilson <chris@chris-wilson.co.uk>
21432 Date: Sun Oct 23 20:58:48 2011 +0100
21434 sna/blt: Share the common fill->boxes between blt and render paths
21436 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21438 commit 56f65e65204f1ecbad598419e525e69f6179f9c7
21439 Author: Chris Wilson <chris@chris-wilson.co.uk>
21440 Date: Sun Oct 23 20:40:21 2011 +0100
21442 sna: Inline box_intersect()
21444 Most of this branchy function is spent in manipulating the stack for the
21445 function call, so inline it.
21447 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21449 commit f65e73f8a3a013e79b753d625f79994408e717f7
21450 Author: Chris Wilson <chris@chris-wilson.co.uk>
21451 Date: Sun Oct 23 19:55:40 2011 +0100
21453 sna: Use fill->boxes to unroll FillSpans
21455 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21457 commit 94f9cd40dcdd444d3b5a5598deb0f3d04026df3a
21458 Author: Chris Wilson <chris@chris-wilson.co.uk>
21459 Date: Sun Oct 23 19:11:35 2011 +0100
21461 sna: Use fill->boxes to unroll PolyPoint
21463 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21465 commit 5e08ca75a5eb852c09b9acf6004668bc63f1fe6c
21466 Author: Chris Wilson <chris@chris-wilson.co.uk>
21467 Date: Sun Oct 23 17:53:54 2011 +0100
21469 sna: Use the new fill->boxes for PolyRectangle
21471 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21473 commit efdc7985e2f619ca94cf6ad77ead4f6867ceedc5
21474 Author: Chris Wilson <chris@chris-wilson.co.uk>
21475 Date: Wed Oct 26 14:57:28 2011 +0100
21479 commit 0ed3426a810336b666604d34c10f996f318ebf82
21480 Author: Chris Wilson <chris@chris-wilson.co.uk>
21481 Date: Fri Oct 21 23:45:28 2011 +0100
21483 sna: Convert diagonal zero-width lines into blits
21485 This is slower than falling back to swrast for x11perf (up to 4x slower
21486 on SNB), it is still faster than doing that rasterisation through a
21487 WC-mapping and much faster in ordinary usage due to avoiding the
21490 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21492 commit c12371d9e7b3bbff7f318186a0933d6108db0bc8
21493 Author: Chris Wilson <chris@chris-wilson.co.uk>
21494 Date: Fri Oct 21 22:28:56 2011 +0100
21496 sna: Flatten the branching for fill-spans
21498 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21500 commit e7f19d85fead62df21eedc87409b5abacfec4a17
21501 Author: Chris Wilson <chris@chris-wilson.co.uk>
21502 Date: Fri Oct 21 22:15:09 2011 +0100
21504 sna/blt: Upload the box using a single 64-bit instruction
21506 So long as we have 64-bit instructions of course!
21508 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21510 commit 40af32a0e9ed971a1f2c2a45266f32016bda7ed3
21511 Author: Chris Wilson <chris@chris-wilson.co.uk>
21512 Date: Fri Oct 21 20:10:02 2011 +0100
21514 sna: Execute blits directly for PolyRectangle
21516 By constructing the batch buffer directly for PolyRectangle, rather than
21517 via miPolyRectangle->(PolyFillRectangle/PolyLine), we dramatically
21518 reduce the CPU overhead and can saturate the GPU.
21520 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21522 commit a18f559961135fa288dda3b94207abb0b6d4d302
21523 Author: Chris Wilson <chris@chris-wilson.co.uk>
21524 Date: Fri Oct 21 16:37:09 2011 +0100
21526 sna: Fix debug compilation
21528 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21530 commit a3a075201c8e47cfad6ec296b7baf549deb44cb3
21531 Author: Chris Wilson <chris@chris-wilson.co.uk>
21532 Date: Fri Oct 21 15:30:52 2011 +0100
21534 sna: Treat a bo marked as being last seen on the GPU as requiring a flush
21536 We are hitting this path upon retiring the old scanout buffer following
21537 a page-flip. We want to treat this as being hot and available for reuse,
21538 so mark it as such until the next retirement pass.
21540 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21542 commit 999d13f32162ce11fe4042825b5b849073944ca8
21543 Author: Chris Wilson <chris@chris-wilson.co.uk>
21544 Date: Fri Oct 21 12:43:02 2011 +0100
21546 sna: Fast path unclipped points
21548 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21550 commit eb80013f4ea50ad4f54b4d1d90c5e9da27e1c34c
21551 Author: Chris Wilson <chris@chris-wilson.co.uk>
21552 Date: Fri Oct 21 11:49:55 2011 +0100
21554 sna: Fast path for unclipped rectangles
21556 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21558 commit d6ae86a51cc49dbba579838edd24a94e8f68294c
21559 Author: Chris Wilson <chris@chris-wilson.co.uk>
21560 Date: Fri Oct 21 03:06:20 2011 +0100
21562 sna: Optimise reduce_damage() to handle all-damaged pixmaps
21564 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21566 commit 5b945b85ee938688412df6adfea488766f4656af
21567 Author: Chris Wilson <chris@chris-wilson.co.uk>
21568 Date: Fri Oct 21 02:45:50 2011 +0100
21570 sna: Fast-path unclipped fill spans
21572 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21574 commit e9a193e9803bb69fc4a7e712f33a36ba395b3c89
21575 Author: Chris Wilson <chris@chris-wilson.co.uk>
21576 Date: Fri Oct 21 02:20:22 2011 +0100
21578 sna: Short-circuit GetWindowPixmap() to speed-up pixmap retrieval
21580 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21582 commit 662402e41e5c7976002a71a7f961ca97c771b1f3
21583 Author: Chris Wilson <chris@chris-wilson.co.uk>
21584 Date: Fri Oct 21 02:11:19 2011 +0100
21586 sna: Micro-optimise PolyLine blt
21588 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21590 commit c2040fb8e62076d87f7a5b407505ee3f3482945e
21591 Author: Chris Wilson <chris@chris-wilson.co.uk>
21592 Date: Fri Oct 21 02:00:44 2011 +0100
21594 sna: Remove the memset(0) of the fill op
21596 The backends are all expected to initialise the state required.
21598 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21600 commit 7e7759db4c1dbf8b782fab9b9ca95cfa1db6cf59
21601 Author: Chris Wilson <chris@chris-wilson.co.uk>
21602 Date: Fri Oct 21 01:41:56 2011 +0100
21604 sna: Pass a BoxRec to the fill op
21606 For many of the core drawing routines, passing a BoxRec for the fill is
21607 more convenient since they already have one generated by the clip
21610 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21612 commit 98bca2dbb9c33c548b1d9789ddb633ae0389c8f1
21613 Author: Chris Wilson <chris@chris-wilson.co.uk>
21614 Date: Fri Oct 21 01:07:52 2011 +0100
21616 sna: Check whether we can blt whilst determining the PolyLine extents
21618 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21620 commit 91f49549750aac99d38ae8f3a132c955a2c8d8f5
21621 Author: Chris Wilson <chris@chris-wilson.co.uk>
21622 Date: Fri Oct 21 00:42:43 2011 +0100
21624 sna: Speed-up blitting of unclipped lines
21626 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21628 commit 06d7574d7da047a6314be5af27f4ac47ccd5a350
21629 Author: Chris Wilson <chris@chris-wilson.co.uk>
21630 Date: Thu Oct 20 21:38:30 2011 +0100
21632 sna/damage: Stash the last freed damage for quick reallocation
21634 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21636 commit 1471ef82b595a207dc8873c45176ec31f277a43f
21637 Author: Chris Wilson <chris@chris-wilson.co.uk>
21638 Date: Thu Oct 20 21:32:14 2011 +0100
21640 sna/damage: Only track the mode globally
21642 As damage accumulation is handled modally, we do not need to track the
21643 mode per elt and so attempt to simplify the code.
21645 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21647 commit 1229bf6a690e7f315ab22957da23481f58515ff9
21648 Author: Chris Wilson <chris@chris-wilson.co.uk>
21649 Date: Thu Oct 20 17:37:41 2011 +0100
21651 sna: Actually apply the composite offset for the self-copy
21653 I translated the region to copy by the composite pixmap offset, only
21654 failed to use the translated region for the actual copy command (using
21655 instead the original boxes). Fix that mistake by avoiding the temporary
21656 region entirely and applying the translation inplace.
21658 We also have to be careful in the case of copying between two composited
21659 windows that have different offsets into the same screen pixmap.
21661 This fixes the regression introduced with a3466c8b69af (sna/accel:
21662 Implement a simpler path for CopyArea between the same pixmaps).
21664 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21666 commit 3526d83e460ce6410f23f59d1315793ff9607253
21667 Author: Chris Wilson <chris@chris-wilson.co.uk>
21668 Date: Thu Oct 20 00:29:52 2011 +0100
21670 sna/dri: Perform a little dance to restore the gpu status on release of the bo
21672 As the bo is outside of our control whilst it is under the influence of
21673 an external renderer, we try to maintain it on the gpu so as to avoid
21674 unnecessary ping-pong. But once it is wholly back under our control, we
21675 want to stop paying the penalty for sharing it.
21677 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21679 commit 50b980f12e02401cdd4bc21b970d92e7bd1e6459
21680 Author: Chris Wilson <chris@chris-wilson.co.uk>
21681 Date: Wed Oct 19 21:09:01 2011 +0100
21683 sna: Reuse any partial write buffer for readback
21685 Take advantage of any available temporary buffer that we reuse for
21686 readback knowing that it is the last operation in the batch.
21688 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21690 commit 972989276dd3f84c1cedca0494d04f907875f8f3
21691 Author: Chris Wilson <chris@chris-wilson.co.uk>
21692 Date: Wed Oct 19 20:41:14 2011 +0100
21694 sna: Add some debug to discern the nature of the Cr fallback
21696 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21698 commit fc224e2f8c0200c114f8dbddfc4f5395a3ef0fad
21699 Author: Chris Wilson <chris@chris-wilson.co.uk>
21700 Date: Wed Oct 19 20:20:55 2011 +0100
21702 sna: Clean up some debug messages for 64bit
21704 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21706 commit f4346e5d255f419ee6148f7d69f02560732dd4de
21707 Author: Chris Wilson <chris@chris-wilson.co.uk>
21708 Date: Wed Oct 19 19:26:19 2011 +0100
21710 sna/blt: Use SCANLINE_BLT for multiple fill boxes
21712 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21714 commit d7fb98efdcc1bcc2cca6deb54b16d425b0350196
21715 Author: Chris Wilson <chris@chris-wilson.co.uk>
21716 Date: Wed Oct 19 17:52:56 2011 +0100
21718 sna: The initial aperture check for a set of bo is unlikely to fail
21720 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21722 commit 3c8e1c5c4cbe703781f2845926f76a9537a971ea
21723 Author: Chris Wilson <chris@chris-wilson.co.uk>
21724 Date: Wed Oct 19 16:35:08 2011 +0100
21726 sna: Reset the ring flag upon idling
21728 We track the last ring used when active so as to avoid stalling between
21729 batches. Once the GPU has retired all the pending requests, we can use
21730 whichever ring is preferrable for the next operation without any danger
21731 of stalling upon submission.
21733 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21735 commit 594c5f86bb2e1cd0390b360c926ead58ba49979f
21736 Author: Chris Wilson <chris@chris-wilson.co.uk>
21737 Date: Wed Oct 19 16:25:42 2011 +0100
21739 sna: Don't rewrite the solid colour cache if it hasn't changed
21741 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21743 commit 5515f75647bb148d9e720dcc4713a93b59ffbd49
21744 Author: Chris Wilson <chris@chris-wilson.co.uk>
21745 Date: Wed Oct 19 13:49:55 2011 +0100
21747 sna: Enlarge the minimum pixmap size to migrate for Render
21749 This is to workaround a ping-pong issue involving small icons. The
21750 horrible sequence of operations appears to use a tiled FillRect to copy
21751 from the scanout onto to a temporary pixmap, which causes us to
21752 readback from the scanout. We are destined to hit the fallback path there
21753 anyway until we implement stippling...
21755 References: https://bugs.freedesktop.org/show_bug.cgi?id=41718
21756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21758 commit 87ba33bc0bcaf3c5160edfc61cff3ba0d28d3fd8
21759 Author: Chris Wilson <chris@chris-wilson.co.uk>
21760 Date: Wed Oct 19 13:09:19 2011 +0100
21762 sna/damage: Avoid reducing the damage for simple tests
21764 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21766 commit 63aa84ef57ff1e55d2df6d680d6bbf88be0619a6
21767 Author: Chris Wilson <chris@chris-wilson.co.uk>
21768 Date: Wed Oct 19 13:08:50 2011 +0100
21770 sna/gen6: Apply the unknown blitter death workaround
21772 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21774 commit 25f2d181946738d1ac2deb724eb48453c8692ed1
21775 Author: Chris Wilson <chris@chris-wilson.co.uk>
21776 Date: Tue Oct 18 23:03:10 2011 +0100
21778 sna: Compute region extents in place to eliminate redundant stack space
21780 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21782 commit 1a43b2a46a67fdea215719c4446224c9d2e26a20
21783 Author: Chris Wilson <chris@chris-wilson.co.uk>
21784 Date: Tue Oct 18 17:52:56 2011 +0100
21786 sna: Use the unlikely wedged() throughout
21788 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21790 commit 7a9c76e1812d106fedf160c959e6e502998e4ce8
21791 Author: Chris Wilson <chris@chris-wilson.co.uk>
21792 Date: Tue Oct 18 15:51:33 2011 +0100
21794 sna: Micro-optimise fill-spans
21796 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21798 commit 0b83abfb2bc3f65447205048ae0af1a94fa9ef15
21799 Author: Chris Wilson <chris@chris-wilson.co.uk>
21800 Date: Tue Oct 18 14:43:37 2011 +0100
21802 sna/gen6: Don't modify composite arguments whilst checking for BLT
21804 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21806 commit ae7bd2bf70b42b8c12e5e5ae074891595827a1c0
21807 Author: Chris Wilson <chris@chris-wilson.co.uk>
21808 Date: Tue Oct 18 11:54:05 2011 +0100
21810 sna/blt: Move the conditional out of the fill-many hotpath
21812 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21814 commit 86f99379ee55b157634174c4c8aad61a30b4f57f
21815 Author: Chris Wilson <chris@chris-wilson.co.uk>
21816 Date: Tue Oct 18 11:35:12 2011 +0100
21818 sna/gen4: Add fill-one
21820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21822 commit 26c082dbd3f89b987168d321f72e78db8c5ddc2c
21823 Author: Chris Wilson <chris@chris-wilson.co.uk>
21824 Date: Tue Oct 18 10:50:53 2011 +0100
21826 sna/gen6: Precompute floats_per_rect
21828 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21830 commit c69a7989b3c9cedc961e29b1a569b174207964ae
21831 Author: Chris Wilson <chris@chris-wilson.co.uk>
21832 Date: Tue Oct 18 10:48:37 2011 +0100
21834 sna/gen5: Add fill-one implementation
21836 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21838 commit 26a1918290b2d23de72a67805fe5e766da0c5309
21839 Author: Chris Wilson <chris@chris-wilson.co.uk>
21840 Date: Tue Oct 18 10:42:48 2011 +0100
21842 sna/gen6: Try continuing with the BLT if the last batch was also BLT
21844 In the vain hope of reducing switching between rings and introducing
21845 stalls between batches.
21847 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21849 commit 4b1398f26d4de44167096c6548ff7384c24d5e06
21850 Author: Chris Wilson <chris@chris-wilson.co.uk>
21851 Date: Tue Oct 18 00:29:18 2011 +0100
21853 sna/gen6: Micro-optimise gen6_rectangle_begin
21855 We can only emit state between primitives, ergo we need only check for
21856 state updates if we've finished the vbo or are starting a new operation.
21858 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21860 commit b2d842df5e3198dab401fbd450d4af83d44917dd
21861 Author: Chris Wilson <chris@chris-wilson.co.uk>
21862 Date: Mon Oct 17 23:34:19 2011 +0100
21864 sna/gen2: Precompute floats-per-rect
21866 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21868 commit 729f1ec7869167ff91fa969d06dbdfbf0e8b4126
21869 Author: Chris Wilson <chris@chris-wilson.co.uk>
21870 Date: Mon Oct 17 23:20:20 2011 +0100
21872 sna/gen2: Prefer to use the BLT
21874 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21876 commit adef3a56cde1210554dc8327baa37c36fc5f18e3
21877 Author: Chris Wilson <chris@chris-wilson.co.uk>
21878 Date: Mon Oct 17 23:08:32 2011 +0100
21880 sna/gen3: Prefer to use the BLT where possible for composite ops
21882 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21884 commit f40e6ab50b264ef7eb7750f72ab7418c0d3dc9c1
21885 Author: Chris Wilson <chris@chris-wilson.co.uk>
21886 Date: Mon Oct 17 23:06:30 2011 +0100
21888 sna/gen3: Prefer to use the BLT to clear the scratch glyph pixmaps
21890 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21892 commit 941e3504e049be8af50ca9df2d0aafe9eb477983
21893 Author: Chris Wilson <chris@chris-wilson.co.uk>
21894 Date: Mon Oct 17 19:51:33 2011 +0100
21896 sna: Some more debug output for request/bo retirement
21898 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21900 commit 1d82fe3e528578d11d2a258b586b67c807eb7cb4
21901 Author: Chris Wilson <chris@chris-wilson.co.uk>
21902 Date: Mon Oct 17 19:47:03 2011 +0100
21904 sna: Only retire the flush handler if nothing was submitted within the period
21906 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21908 commit 7758e333409409393c4c974adb2831a7b5b18fe4
21909 Author: Chris Wilson <chris@chris-wilson.co.uk>
21910 Date: Mon Oct 17 19:32:44 2011 +0100
21912 sna/gen3: Use immediates for black/white solid sources
21914 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21916 commit 29ca1a3922cb0e6f3d7b71857a252e5de81941b5
21917 Author: Chris Wilson <chris@chris-wilson.co.uk>
21918 Date: Mon Oct 17 17:16:16 2011 +0100
21920 sna/gen3: More removal of memset(0)
21922 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21924 commit 88bc10878830dffefda6ed905c1f598d7a7d67ef
21925 Author: Chris Wilson <chris@chris-wilson.co.uk>
21926 Date: Sun Oct 16 20:05:49 2011 +0100
21928 sna/gen5: Rearrange the BLT->RENDER workaround
21930 So that we can simply use the pending DRAWRECT as the non-pipelined
21931 flush required following use of a BLT command.
21933 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21935 commit 18aed47ba37ccfbcdd72ed251fc97222741c66a7
21936 Author: Chris Wilson <chris@chris-wilson.co.uk>
21937 Date: Mon Oct 17 16:41:14 2011 +0100
21939 sna: Fast path common colour conversions
21941 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21943 commit a3466c8b69afeda95180fcdc97b56d31f7a1d1ad
21944 Author: Chris Wilson <chris@chris-wilson.co.uk>
21945 Date: Sun Oct 16 20:52:47 2011 +0100
21947 sna/accel: Implement a simpler path for CopyArea between the same pixmaps
21949 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21951 commit 8afa05191a9a337ce6e02b1064de2876f5635435
21952 Author: Chris Wilson <chris@chris-wilson.co.uk>
21953 Date: Mon Oct 17 00:55:51 2011 +0100
21955 sna: Improve debug output for mi/fb fallbacks
21957 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21959 commit 2b118658d5b9cfa56132f91238f2ea61fba8b357
21960 Author: Chris Wilson <chris@chris-wilson.co.uk>
21961 Date: Mon Oct 17 15:00:01 2011 +0100
21963 sna/gen3: Remove memset(0) from fill_boxes
21965 And initialise only the state used for the operation.
21967 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21969 commit f1bb4ebfd8991f2f9eb9c38b9259792c11e7c86a
21970 Author: Chris Wilson <chris@chris-wilson.co.uk>
21971 Date: Mon Oct 17 14:22:58 2011 +0100
21973 sna/damage: Clear damage->all when subtracting
21975 I missed clearing the all-damaged flag on the subtract fast paths,
21976 causing us to disregard further GPU damage and losing track of render
21979 Reported-by: Roman Jarosz <kedgedev@gmail.com>
21980 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41769
21981 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21983 commit 46f97127c22ea42bc8fdae59d2a133e4b8b6c997
21984 Author: Chris Wilson <chris@chris-wilson.co.uk>
21985 Date: Sun Oct 16 21:40:15 2011 +0100
21987 snb,ivb: Workaround unknown blitter death
21989 The first workaround was a performance killing MI_FLUSH_DW after every
21990 op. This workaround appears to be a stable compromise instead, only
21991 requiring a redundant command after every BLT command with little
21992 impact on throughput.
21994 Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=27892
21995 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=39524
21996 Tested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
21997 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
21999 commit 71bf291e563ec5224777b4907a5347a3fbfe64c5
22000 Author: Chris Wilson <chris@chris-wilson.co.uk>
22001 Date: Mon Oct 17 12:53:19 2011 +0100
22003 sna/gen3: Micro-optimise gen3_rectangle_begin()
22005 We only need to emit the vbo description either at the beginning of a
22006 new op (when the state may have changed) or after finishing a full vbo.
22008 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22010 commit a032feb2531439b9945d7ae1e0c2e3ab95a960e1
22011 Author: Chris Wilson <chris@chris-wilson.co.uk>
22012 Date: Mon Oct 17 12:41:21 2011 +0100
22014 sna/gen3: Store floats_per_rect alongside floats_per_vertex
22016 Moves a multiply out of the hot path.
22018 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22020 commit 2cbe79b85b46796d561105afa980e7a52f4f1889
22021 Author: Chris Wilson <chris@chris-wilson.co.uk>
22022 Date: Mon Oct 17 11:57:07 2011 +0100
22024 sna/gen3: Improve reduction of render operator to blt
22026 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22028 commit 606e18bd360d904a8aed01aef43644fb4a51d929
22029 Author: Chris Wilson <chris@chris-wilson.co.uk>
22030 Date: Mon Oct 17 10:50:33 2011 +0100
22032 configure: Add a check for pixman version
22034 As SNA requires a fairly recent release and calls directly into pixman.
22036 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22038 commit 17acaf8e232a8bad9bc2c653a9fceb06f0f65f6b
22039 Author: Chris Wilson <chris@chris-wilson.co.uk>
22040 Date: Mon Oct 17 00:55:51 2011 +0100
22042 sna: Defer CPU fallback if fb calls into mi
22044 If the fb routine does not access the pixel data directly, but instead
22045 calls into an mi routine, we can defer the readback and possibly avoid
22048 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22050 commit c348b69f66cbe1cafc88396523da9768e1006889
22051 Author: Chris Wilson <chris@chris-wilson.co.uk>
22052 Date: Sun Oct 16 18:23:33 2011 +0100
22054 sna: Prefer to accelerate non-RENDER operations if already using the GPU
22056 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22058 commit 6fdc9e3fd3f8defb7ad62de11f8cb069a10e5736
22059 Author: Chris Wilson <chris@chris-wilson.co.uk>
22060 Date: Sun Oct 16 17:57:55 2011 +0100
22062 sna: Simplify busy tracking by trusting the bo->gpu flag
22064 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22066 commit 30dff81cd62b5fc5a28d5175a08c5a3ee09667bd
22067 Author: Chris Wilson <chris@chris-wilson.co.uk>
22068 Date: Sun Oct 16 17:41:34 2011 +0100
22070 sna/accel: Upload source to GPU for !copy CopyAreas
22072 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22074 commit eefa925e3d3e009b21ecc0b428d93a6c732bfa14
22075 Author: Chris Wilson <chris@chris-wilson.co.uk>
22076 Date: Sun Oct 16 17:39:06 2011 +0100
22078 sna/accel: Fall-forward for handling a non-copy CopyArea to a dst gpu bo
22080 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22082 commit 01f41e624b31af88d433a25eaefce557e7ff3d62
22083 Author: Chris Wilson <chris@chris-wilson.co.uk>
22084 Date: Sun Oct 16 17:38:26 2011 +0100
22086 sna/gen2: Eliminate redundant diffuse and rectlist emission
22088 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22090 commit d9e3dbffcb7e6cff35e10100d81544936813095a
22091 Author: Chris Wilson <chris@chris-wilson.co.uk>
22092 Date: Sun Oct 16 12:48:23 2011 +0100
22094 sna/gen2: Fix fill-one-box
22096 Lets only have one special gen2 value for the source channel pixel
22097 colour and so remove the confusion and misrendering.
22099 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22101 commit b89592b16e373da1fe259ad067f65ba95021dc49
22102 Author: Chris Wilson <chris@chris-wilson.co.uk>
22103 Date: Sun Oct 16 19:39:14 2011 +0100
22105 sna/gen7: Add missing tail to 3DSTATE_SBE
22107 I was a few NULs short of the declared amount.
22109 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22111 commit 33f17b996dd0fcbb8e07875770e11571380d6509
22112 Author: Chris Wilson <chris@chris-wilson.co.uk>
22113 Date: Fri Oct 14 12:01:34 2011 +0100
22115 Remove a couple of trivial compile warnings for unused variables
22117 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22119 commit e5e27528ffeb74f404823c678e04e847cf2109aa
22120 Author: Chris Wilson <chris@chris-wilson.co.uk>
22121 Date: Fri Oct 14 11:51:44 2011 +0100
22123 sna/blt: Add a pair of operator reductions for opaque fills
22125 Extend the coverage of the BLT for a couple more composite operations,
22126 though they be relatively infrequent.
22128 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22130 commit 3f5289b29962c3e5d94deb30782a156292d20053
22131 Author: Chris Wilson <chris@chris-wilson.co.uk>
22132 Date: Fri Oct 14 11:31:46 2011 +0100
22134 sna/trapezoids: Fallback if the composite operation is not supported
22136 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22138 commit 3d79264ef8a8be225d9b13efd14f37a7478e6b1b
22139 Author: Chris Wilson <chris@chris-wilson.co.uk>
22140 Date: Fri Oct 14 11:31:00 2011 +0100
22142 sna/tiling: Hook up composite_box
22144 Now used for spans and so triggering death by NULL function pointer on
22147 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22149 commit f5f3135e3cc1f4740508773cfefca0954aed2fd3
22150 Author: Chris Wilson <chris@chris-wilson.co.uk>
22151 Date: Fri Oct 14 10:40:39 2011 +0100
22153 sna: Ensure that the scanout is flushed before we sleep
22155 So that we do not see any delayed rendering on i965+.
22157 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22159 commit 020d84d2d09c71a8209dd774a0735a3c11cb148d
22160 Author: Chris Wilson <chris@chris-wilson.co.uk>
22161 Date: Thu Oct 13 22:32:00 2011 +0100
22163 sna/accel: Fix sense of nothing-to-do due to clipping
22165 A missing not due to a hasty conversion from an assert.
22167 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22169 commit b7fd6906c41e328649b97e16c42848a39f6e48f8
22170 Author: Chris Wilson <chris@chris-wilson.co.uk>
22171 Date: Thu Oct 13 17:46:48 2011 +0100
22173 sna/accel: Actually apply the clip to the glyph extents
22175 References: https://bugs.freedesktop.org/show_bug.cgi?id=41718
22176 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22178 commit eb5f86b78181b174664aaa9bc689db5421b7e7bd
22179 Author: Chris Wilson <chris@chris-wilson.co.uk>
22180 Date: Thu Oct 13 17:32:54 2011 +0100
22182 sna/accel: Correctly offset the damage intersection for glyph runs
22184 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22186 commit a4766026ad0aa14283e35598d33f13d311cc029b
22187 Author: Chris Wilson <chris@chris-wilson.co.uk>
22188 Date: Thu Oct 13 17:32:30 2011 +0100
22190 sna/accel: Check for reloc overflow when adding a new glyph run
22192 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22194 commit b10850af4125b8dbc36d13fa83f1dcce13ac6684
22195 Author: Chris Wilson <chris@chris-wilson.co.uk>
22196 Date: Thu Oct 13 17:09:25 2011 +0100
22198 sna/blt: Check for reloc overflow when initialising solid fills
22200 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22202 commit d5242565ef17cb4265177338df5b564b960e3692
22203 Author: Chris Wilson <chris@chris-wilson.co.uk>
22204 Date: Wed Oct 12 21:14:43 2011 +0100
22206 sna/gen2: Trim some surplus state commands
22208 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22210 commit 7c5020532335c07d9b0f6cecb0b004de4af6d273
22211 Author: Chris Wilson <chris@chris-wilson.co.uk>
22212 Date: Wed Oct 12 20:36:53 2011 +0100
22214 Remove vestigial includes from DRI1
22216 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22218 commit b6b3bb9b5e56949cbd9189d9857ffa1ff46377fd
22219 Author: Chris Wilson <chris@chris-wilson.co.uk>
22220 Date: Tue Sep 20 19:41:03 2011 +0100
22222 sna: compare against the right array of cache buckets
22224 Fortunately harmless today as there are the same number of
22225 inactive/active buckets.
22227 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22229 commit 1f70095837a30d6a88b9e313d3583d38ea55221d
22230 Author: Chris Wilson <chris@chris-wilson.co.uk>
22231 Date: Tue Oct 11 16:06:31 2011 +0100
22233 sna: And free the clip after creation for PolySegments
22235 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22237 commit 92f4d978c8b1f5e9a6cf3affa15d90bfb3d4e7b9
22238 Author: Chris Wilson <chris@chris-wilson.co.uk>
22239 Date: Tue Oct 11 15:52:13 2011 +0100
22241 sna: More micro-optimisation of messing around with clip regions
22243 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22245 commit 57151f654752ed64716358ecb5e9217b59784da8
22246 Author: Chris Wilson <chris@chris-wilson.co.uk>
22247 Date: Tue Oct 11 15:23:17 2011 +0100
22249 sna: Micro-optimise checking for singular clip boxes
22251 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22253 commit 823a4272c50247482428a16cb08741bf87a302ea
22254 Author: Chris Wilson <chris@chris-wilson.co.uk>
22255 Date: Tue Oct 11 13:51:41 2011 +0100
22257 sna/gen3: Avoid RENDER/BLT context switch for fill boxes
22259 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22261 commit 887361de173e082841453902089427447c5a158a
22262 Author: Chris Wilson <chris@chris-wilson.co.uk>
22263 Date: Tue Oct 11 13:51:25 2011 +0100
22265 sna: Enable single fill fast path for PolySegment
22267 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22269 commit 721cf30e9eeb0102c77f337332ad8259463b7fef
22270 Author: Chris Wilson <chris@chris-wilson.co.uk>
22271 Date: Tue Oct 11 13:28:45 2011 +0100
22273 sna/accel: If the data is already on the GPU, use it for the source
22275 Fixes regression from 1ec6a0e2775 (sna: Move the source to the GPU
22278 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22280 commit 15a4410cec111c3b3bd15ec33f29bc1f4cb709c1
22281 Author: Chris Wilson <chris@chris-wilson.co.uk>
22282 Date: Tue Oct 11 13:28:30 2011 +0100
22284 sna: use correct insertion point for sorting partials
22286 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22288 commit a9b53c48774078d404149e86267c39eb82234c3a
22289 Author: Chris Wilson <chris@chris-wilson.co.uk>
22290 Date: Tue Oct 11 10:37:25 2011 +0100
22292 sna: Mark the spans render functions as fastcall
22294 This reduces the amount of dancing required to call into the span
22295 functions as we can pass the arguments in both the integer and floating
22298 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22300 commit 208fa8e6b88daf8446239a1a001a68ceb7d02e76
22301 Author: Chris Wilson <chris@chris-wilson.co.uk>
22302 Date: Tue Oct 11 10:08:42 2011 +0100
22304 sna/trapezoid: Perform the NULL check for damage in the caller
22306 Save the function call overhead in the common case.
22308 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22310 commit 5050fead0ef81785a2a9be735852921463377e04
22311 Author: Chris Wilson <chris@chris-wilson.co.uk>
22312 Date: Tue Oct 11 09:59:13 2011 +0100
22314 sna/gen3: avoid applying zero offset to common spans
22316 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22318 commit d0184b59095d5b8fab1a65ceba075d29189130d4
22319 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
22320 Date: Sun Oct 9 18:43:14 2011 +0200
22322 snb: implement PIPE_CONTROL workaround
22324 Sandybdrige requires an elaborate dance to flush caches without
22325 hanging the gpu. See public docs Vol2Part1 1.7.4.1 PIPE_CONTROL
22326 or the corrensponding code in mesa/kernel.
22328 This (together with the corresponding patch for the kernel) seems to
22329 fix the hangs in cairo-perf-traces I'm seeing on my snb machine.
22331 v2: Incorporate review from Chris Wilson. For paranoia keep all three
22332 PIPE_CONTROL cmds in the same batchbuffer to avoid upsetting the gpu.
22334 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
22336 commit 4a2e833ab17e4facf0f90166b82a5696a1deef91
22337 Author: Chris Wilson <chris@chris-wilson.co.uk>
22338 Date: Tue Oct 11 08:45:47 2011 +0100
22340 sna/gen7: Add render support for fill one
22342 To prevent the RENDER to BLT transition and potential stall.
22344 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22346 commit 41f525fab5a82c24adafc7e8c8409417d16a0e9a
22347 Author: Chris Wilson <chris@chris-wilson.co.uk>
22348 Date: Tue Oct 11 00:22:11 2011 +0100
22350 sna/gen6: Add render support for fill-one-box
22352 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22354 commit 5b6575bdded4b24ec1c9515203f44798225c10b6
22355 Author: Chris Wilson <chris@chris-wilson.co.uk>
22356 Date: Mon Oct 10 16:39:02 2011 +0100
22358 sna: Support a fast composite hook for solitary boxes
22360 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22362 commit c5414ec992d935e10156a2b513d5ec2dded2f689
22363 Author: Chris Wilson <chris@chris-wilson.co.uk>
22364 Date: Sun Oct 2 12:02:41 2011 +0100
22366 sna: Use BLT operations to avoid fallbacks in core glyph rendering
22368 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22370 commit 6b62b9d7c4301457a024f840383fb4515704ebdb
22371 Author: Chris Wilson <chris@chris-wilson.co.uk>
22372 Date: Tue Oct 4 19:30:06 2011 +0100
22374 sna: Increase reserved space in batch to accommodate gen5 workaround
22376 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22378 commit 6351d8eb823dbfffb4792af8592d1ce565211e23
22379 Author: Chris Wilson <chris@chris-wilson.co.uk>
22380 Date: Tue Oct 4 19:29:46 2011 +0100
22382 sna/gen[23]: Fix compilation with debugging enabled
22384 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22386 commit 1ec6a0e27753027a68beeaff0276ae65c658d468
22387 Author: Chris Wilson <chris@chris-wilson.co.uk>
22388 Date: Mon Oct 3 16:14:18 2011 +0100
22390 sna: Move the source to the GPU if it is reused
22392 We attempt to skip upload a source pixmap to the GPU in the event it is
22393 used only once (for example during image upload by firefox). However, if
22394 we continue to use the CPU source pixmap then it obviously was worth
22395 uploading to the GPU. So if we use the CPU pixmap a second time, do the
22396 upload and then blit.
22398 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22400 commit 413c9f7111c605a3842d4d0a85f8cf01470e5faa
22401 Author: Chris Wilson <chris@chris-wilson.co.uk>
22402 Date: Mon Oct 3 00:24:50 2011 +0100
22404 sna/blt: Suppress repeated SETUP_BLT
22406 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22408 commit 1067335305a6b8123217518085f70a77b7f0334f
22409 Author: Chris Wilson <chris@chris-wilson.co.uk>
22410 Date: Mon Oct 3 00:11:55 2011 +0100
22412 sna/blt: SETUP_BLT needs 9 dwords of batch space, not 3!
22414 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22416 commit d8c96a6a1d50cd3dd40fd4b78958c8f78337a2fb
22417 Author: Chris Wilson <chris@chris-wilson.co.uk>
22418 Date: Sun Oct 2 12:57:53 2011 +0100
22420 sna/blt: Use SETUP_MONO to reduce the number of fill relocations
22422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22424 commit 4d227d43f0dfca1fa1822207d68539635aac37b3
22425 Author: Chris Wilson <chris@chris-wilson.co.uk>
22426 Date: Sun Oct 2 12:03:21 2011 +0100
22428 sna/accel: Correct syntax for constifying BoxPtr
22430 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22432 commit 3dd80524160d6b6cf0fb668afcc766f2bc6f8383
22433 Author: Chris Wilson <chris@chris-wilson.co.uk>
22434 Date: Sun Oct 2 09:59:03 2011 +0100
22436 sna/accel: Only throttle after flushing
22438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22440 commit 04b8f0a5a16ff71c5e55e0281dbcaa32554703ec
22441 Author: Chris Wilson <chris@chris-wilson.co.uk>
22442 Date: Sun Oct 2 09:36:38 2011 +0100
22444 sna/accel: Add a compile option to force flushing on every blockhandler
22446 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22448 commit 32cef71efe231692f7b24f66913a8ae57bc00aa8
22449 Author: Chris Wilson <chris@chris-wilson.co.uk>
22450 Date: Sun Oct 2 08:55:55 2011 +0100
22452 sna/accel: Add an compile option to disable use of spans
22454 Using spans has a tremendous effect (~100x) on x11perf, some good but
22455 mostly bad. However, in reality operations are mixed and so preventing
22456 migration on alternate opertaions is a win. In the x11perf slowdowns, it
22457 appears to be CPU bound and so it seems like there should be plenty of
22458 scope for recovering the lost performance.
22460 However, for the time being, just go back to the old fallbacks.
22462 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22464 commit dc1ec0dd1a5f66114ee4249309d3a283ad84d584
22465 Author: Chris Wilson <chris@chris-wilson.co.uk>
22466 Date: Sat Oct 1 20:11:47 2011 +0100
22468 sna/accel: Only disable the flush migitration if nothing changed
22470 Previously we ignored updating the scanout in place, and so we were not
22471 amoritizing the shadow cost of common core rendering operations.
22473 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22475 commit c6acf1325833b8679ef09ab74f0cb0fd82a8cd92
22476 Author: Chris Wilson <chris@chris-wilson.co.uk>
22477 Date: Sat Oct 1 11:26:54 2011 +0100
22479 sna/accel: Micro-optimise sna_fill_spans_blt
22481 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22483 commit 8029765515399b130bee18db0a2830eb83f47a07
22484 Author: Chris Wilson <chris@chris-wilson.co.uk>
22485 Date: Sat Oct 1 11:19:49 2011 +0100
22487 sna/accel: Don't attempt converting to spans if we will only fallback
22489 As the span code does not yet handle plane masks or stippling, it is
22490 disadvantageous to convert to spans only to fallback.
22492 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22494 commit cd11bd69f4c1b961593886945946c7d7c9269900
22495 Author: Chris Wilson <chris@chris-wilson.co.uk>
22496 Date: Sat Oct 1 10:45:18 2011 +0100
22498 sna/accel: Use miPolyArc to convert arcs into spans for gpu bo
22500 This is actually tricker than it looks since miPolyArc() sometimes uses
22501 an intermediate bitmap which performs worse than the fbPolyArc() fallback.
22503 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22505 commit d07256cc3360cf94d131d9a1b24d9f1c6a70ee04
22506 Author: Chris Wilson <chris@chris-wilson.co.uk>
22507 Date: Sat Oct 1 10:42:45 2011 +0100
22509 sna/accel: Convert segments into spans similarly to PolyLine
22511 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22513 commit d09a229e32468c810f5056bfd0043f2fb68038d3
22514 Author: Chris Wilson <chris@chris-wilson.co.uk>
22515 Date: Sat Oct 1 10:01:21 2011 +0100
22517 sna/accel: Use the mi*Line routines to convert the line into spans for gpu bo
22519 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22521 commit e7a662b92e3012735f7a3d72d4f7860ca61a3810
22522 Author: Chris Wilson <chris@chris-wilson.co.uk>
22523 Date: Sat Oct 1 09:26:52 2011 +0100
22525 sna: Sort partials by remaining space
22527 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22529 commit 13b9b5d8d6b6f6db59fe2418270ac93d9b74436c
22530 Author: Chris Wilson <chris@chris-wilson.co.uk>
22531 Date: Sat Oct 1 09:23:41 2011 +0100
22533 sna/io: Only mark the buffer as LAST if we know we will flush the IO
22535 Otherwise we can continue to batch up the data upload into larger
22538 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22540 commit 7ecc6993b8c8205e5896ee87c14d6a76310c9aa7
22541 Author: Chris Wilson <chris@chris-wilson.co.uk>
22542 Date: Fri Sep 30 16:11:05 2011 +0100
22544 sna/gen6: Fix offset of Scan-Line-Compare register
22546 Reported-by: Frank Mariak <fmariak@macrosystem.de>
22547 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22549 commit d8fe941bc245e24c83c417ccff5c57e83baac3f7
22550 Author: Chris Wilson <chris@chris-wilson.co.uk>
22551 Date: Wed Sep 28 23:04:03 2011 +0100
22553 sna: Check for request retires after every batch
22555 In the beginning, I did perform a retire after ever batch. Then I
22556 decided that it was too much CPU overhead for too little gain. On
22557 reflection, i.e. further benchmarking, we do see a performance
22558 improvement for recycling active buffers faster.
22560 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22562 commit e74a39b45446e2c43df5419efc4a2e0e73275c45
22563 Author: Chris Wilson <chris@chris-wilson.co.uk>
22564 Date: Wed Sep 28 14:47:33 2011 +0100
22566 sna/gen7: Confusion reigns as trying to fix errors found by an outdated checker
22568 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22570 commit 6395894ada6b9c14deb62814ccf55848eaa80527
22571 Author: Chris Wilson <chris@chris-wilson.co.uk>
22572 Date: Tue Sep 27 23:26:33 2011 +0100
22574 sna/gen7: Fix up a couple instances of my inability to count
22576 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22578 commit a53538659d5e18cb5a7a7a7dbd460f35985183fd
22579 Author: Chris Wilson <chris@chris-wilson.co.uk>
22580 Date: Sun Sep 25 12:40:01 2011 +0100
22582 sna/accel: Fix s/x/y/ typo in computing relative drawing coordinates
22584 Reported-by:Roman Jarosz <kedgedev@gmail.com>
22585 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41165
22586 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22588 commit 960688d1684dd062c345aa8df1ff4f105015f467
22589 Author: Chris Wilson <chris@chris-wilson.co.uk>
22590 Date: Sun Sep 25 12:23:10 2011 +0100
22592 sna/accel: Debug option to force CPU fallbacks
22594 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22596 commit 4fd46b8bb7e7a104a0afa0e5dee92993e043ef57
22597 Author: Chris Wilson <chris@chris-wilson.co.uk>
22598 Date: Sun Sep 25 10:02:02 2011 +0100
22600 sna/glyphs: Add glyphs directly onto a client temporary buffer
22602 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22604 commit 5154e281ed66334a8ff6342236c1203a1e4ac5a3
22605 Author: Chris Wilson <chris@chris-wilson.co.uk>
22606 Date: Sun Sep 25 09:56:16 2011 +0100
22608 sna/glyph: Avoid useless attempt at GPU glyph rendering to a1 destinations
22610 The actual bug is a little involved as we don't damage the temporary
22611 glyph mask correctly presuming that we only hit GPU paths. However,
22612 should we fail to prepare the composite operation that paints the mask
22613 on to the destination, things fail horribly.
22615 One particular example is that wine like to create its own temporary a1
22616 buffer for the glyphs (which we render to via another temporary mask...)
22617 which triggers the delayed fallback and then sw compositing with a random
22620 Reported-by: Roman Jarosz <kedgedev@gmail.com>
22621 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41165
22622 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22624 commit 46fedf0cf1a413381109c1de34619e84ad614b8a
22625 Author: Chris Wilson <chris@chris-wilson.co.uk>
22626 Date: Sun Sep 25 09:38:43 2011 +0100
22628 sna/kgem: Check all operation bo in a single amalgamation
22630 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22632 commit ccf6547a8f1b77ddf4e79bdefe2a0878c0e9b256
22633 Author: Chris Wilson <chris@chris-wilson.co.uk>
22634 Date: Sat Sep 24 23:16:57 2011 +0100
22636 sna: Paranoid debug flush after every op (as well as before)
22638 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22640 commit 02337600345951c0ea629db11f0484403a260991
22641 Author: Chris Wilson <chris@chris-wilson.co.uk>
22642 Date: Sat Sep 24 23:03:52 2011 +0100
22644 sna/gen5: Debug option to disable state caching
22646 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22648 commit af4d3853ae8f834fbf2fc55fb2b397b3f7f17656
22649 Author: Chris Wilson <chris@chris-wilson.co.uk>
22650 Date: Sat Sep 24 19:26:46 2011 +0100
22652 sna/glyphs: Convert all sub-8bpp masks to a8
22654 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22656 commit c79e90da71ae58cb3804f8abaf0541350e47eb14
22657 Author: Chris Wilson <chris@chris-wilson.co.uk>
22658 Date: Sat Sep 24 19:26:31 2011 +0100
22660 sna: Add a debug option to disable caching
22662 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22664 commit 6bbb88af096e054877409a54d0e0a4ccf5ee317e
22665 Author: Kenneth Graunke <kenneth@whitecape.org>
22666 Date: Sat Sep 24 09:27:33 2011 +0100
22668 Fix incorrect maximum PS thread count on IvyBridge
22670 I mistakenly set GEN7_PS_MAX_THREAD_SHIFT to 23; it's actually 24 on
22671 Ivybridge. Not only did this halve our thread count, it caused us to
22672 write 1 into a bit 23, which is marked as MBZ (must be zero).
22673 Furthermore, it made us write an even number into this field, which is
22674 apparently not allowed. Apparently we were just lucky it worked.
22676 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
22677 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22679 commit 7f7f95abbf57e6e71f6a30d917f97c2f2bd6cea9
22680 Author: Chris Wilson <chris@chris-wilson.co.uk>
22681 Date: Fri Sep 23 11:38:18 2011 +0100
22683 sna/accel: Use the PolyFillRect to handle tiled spans
22685 Would be preferrable to duplicate the tiling logic. Leave the task of
22686 reimplementing XAA to another day!
22688 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22690 commit 964c96b181e6b55f4de6a4011a5beb6cc2cac33d
22691 Author: Chris Wilson <chris@chris-wilson.co.uk>
22692 Date: Fri Sep 23 11:37:48 2011 +0100
22694 sna/accel: Always subtract the enlarged region from the outstanding GPU damage
22696 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22698 commit c68856f34653ac3e7af900dfbba41a108ffe119e
22699 Author: Chris Wilson <chris@chris-wilson.co.uk>
22700 Date: Fri Sep 23 11:35:50 2011 +0100
22702 sna/accel: Only skip undamaging the GPU for reads
22704 Introduced with ac1b83240e5 (sna/accel: Simplify single pixel read-back)
22706 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22708 commit 5913c90967091124e7c7b262782f0e99cf400eab
22709 Author: Chris Wilson <chris@chris-wilson.co.uk>
22710 Date: Thu Sep 22 17:13:20 2011 +0100
22712 sna/accel: fix assert to include the offset of copy
22714 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22716 commit 31857068026138fe5edd8656381fc1ae59d3058b
22717 Author: Chris Wilson <chris@chris-wilson.co.uk>
22718 Date: Thu Sep 22 17:12:53 2011 +0100
22720 sna: move an assert
22722 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22724 commit 61764af13aa3c770d19d51c8ad198cab8a5866f1
22725 Author: Chris Wilson <chris@chris-wilson.co.uk>
22726 Date: Thu Sep 22 12:59:32 2011 +0100
22728 sna/dri: Bump DRI2INFOREC_VERSION
22730 An intervening dri2 patch landed in the sever.
22732 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22734 commit 12070b8eb1e98cb158802fff3e59f906b1547877
22735 Author: Chris Wilson <chris@chris-wilson.co.uk>
22736 Date: Thu Sep 22 12:18:26 2011 +0100
22738 sna/dri: Play with marking bo reusable
22740 Since we have no recycling of wc pages in the kernel, we try hard to
22741 recycle buffers in userspace to avoid GTT thrashing. This requires
22742 co-operation between DRI clients and X, which is sadly lacking and so we
22743 need to discard any buffer given out to a client after it is finished.
22744 We cheat slightly for page-flips and access to the scanout.
22746 A further compromise.
22748 References: https://bugs.freedesktop.org/show_bug.cgi?id=38732
22749 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22751 commit 902a0234191d77816259affe331339caaa9131d6
22752 Author: Chris Wilson <chris@chris-wilson.co.uk>
22753 Date: Mon Sep 19 20:34:15 2011 +0100
22755 sna: Use pot buckets for active cache
22757 On x11perf one of the major hotspots is the search through the active
22758 list for an object large enough to reuse as the target surface. We can
22759 eliminate that overhead by keeping those active objects in pre-sorted
22762 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22764 commit 7bb2c5c4d55560f7901fd8036cab164460f81edc
22765 Author: Chris Wilson <chris@chris-wilson.co.uk>
22766 Date: Mon Sep 19 19:12:46 2011 +0100
22768 sna: Move an assertion to the point of import
22770 All the guesswork is so that when we require an inactive bo, we do
22771 actually get a buffer that is not currently on a GPU active list. For
22772 some unresolved reason, this assertion was firing when putting the
22773 buffer onto the inactive list - so just workaround the worrisome issue
22774 by delaying the check until use.
22776 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22778 commit 686a5ec52c9ae475fac22c36fff18814372def71
22779 Author: Chris Wilson <chris@chris-wilson.co.uk>
22780 Date: Mon Sep 19 18:14:05 2011 +0100
22782 sna/trapezoids: Fix overflow during sorting of mono edge step
22784 We were tracking the 32bit value of the prev_x using only a 16bit
22785 variable, and so failing to sort the edges after advancing to the next
22788 Fixes cairo a1-clip-fill-rule.
22790 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22792 commit 0390105bc239bf2ac22189f39fccc9d98bae4992
22793 Author: Chris Wilson <chris@chris-wilson.co.uk>
22794 Date: Mon Sep 19 12:48:38 2011 +0100
22796 sna: compile fixes for debugging
22798 Update the DBG messages to reflect changes in function parameters.
22800 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22802 commit 7f0cba3a255eb9e76d56201b26d15f2d49134b58
22803 Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
22804 Date: Fri Sep 16 18:56:00 2011 -0300
22806 Remove useless assertion
22808 Removes 17 instances of:
22809 warning: comparison of unsigned expression >= 0 is always true
22811 Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
22812 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22814 commit 2d9effca142b1d82cd1a7ad3b7d2a61c4567801b
22815 Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
22816 Date: Fri Sep 16 18:53:01 2011 -0300
22818 Fix "always false" conditionals
22820 Enums are unsigned by default in gcc and we can't rely on any specific
22821 signedess for the other compilers.
22823 i965_render.c: In function ‘i965_prepare_composite’:
22824 i965_render.c:2018:2: warning: comparison of unsigned expression < 0 is always false
22825 i965_render.c:2025:2: warning: comparison of unsigned expression < 0 is always false
22826 i965_render.c:2050:3: warning: comparison of unsigned expression < 0 is always false
22827 i965_render.c:2057:3: warning: comparison of unsigned expression < 0 is always false
22829 Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
22830 [ickle: take advantage and rename the enum values]
22831 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22833 commit 7bcd1d2b75d280cecfc23877bb13426a28c07a1f
22834 Author: Chris Wilson <chris@chris-wilson.co.uk>
22835 Date: Sat Sep 17 21:38:46 2011 +0100
22837 sna: perform a warnings reduction pass
22839 Didn't spot anything that might have led to a genuine bug, but this
22840 should help improve the signal-to-noise ratio of warnings in the future.
22842 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22844 commit a5b5b4ebb14cab19258dd1cbc12721aaa9df226d
22845 Author: Chris Wilson <chris@chris-wilson.co.uk>
22846 Date: Sat Sep 17 00:28:30 2011 +0100
22848 sna: Apply the screen offset for direct upload of a composited pixmap
22850 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38733
22851 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22853 commit daeb9291e22d94cd2f81b3b56418f0f879531da5
22854 Author: Chris Wilson <chris@chris-wilson.co.uk>
22855 Date: Sat Sep 17 00:27:55 2011 +0100
22857 sna/trapezoids: Ensure we do not overrun the vertical edges for mono
22859 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22861 commit 481df2618af3c10877db0610c4cadf6fc509e72f
22862 Author: Chris Wilson <chris@chris-wilson.co.uk>
22863 Date: Sat Sep 17 00:27:32 2011 +0100
22865 sna: Add debug option to force allocaiton of gpu-only pixmaps
22867 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22869 commit 52f93c95465151d16f588d648a58b1f979e946a3
22870 Author: Chris Wilson <chris@chris-wilson.co.uk>
22871 Date: Fri Sep 16 23:24:16 2011 +0100
22873 sna/glyphs: Disable the glyph cache for debugging
22875 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22877 commit c8a144d6f84a00a26785fd518cb890ad77edd45f
22878 Author: Chris Wilson <chris@chris-wilson.co.uk>
22879 Date: Fri Sep 16 22:56:36 2011 +0100
22881 sna/glyphs: Correct offset of fallback-via-mask glyphs
22883 Under a compositing manager where we have fun values for both
22884 drawable->x/y and pixmap->screen.x/y, we were not drawing either the
22885 glyphs into the mask correctly and then failed to composite the mask in
22886 the right position on top of the pixmap.
22888 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22890 commit 1eea2c4e60af535ac6500b1b62367f9f63f2a784
22891 Author: Chris Wilson <chris@chris-wilson.co.uk>
22892 Date: Fri Sep 16 20:35:03 2011 +0100
22894 sna: Can't free bo from the active list
22896 As the active bo is still referenced in the request list, we can not
22897 simply free it but need to wait for it to be purged on expiration.
22899 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22901 commit d743c456db11bd3c0d21948dc3a2eb4b33e46342
22902 Author: Chris Wilson <chris@chris-wilson.co.uk>
22903 Date: Fri Sep 16 20:34:33 2011 +0100
22905 sna: Debug option to disable tiling
22907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22909 commit 8c35df4493c910ef62c66a411c14e5be3b26bcf0
22910 Author: Chris Wilson <chris@chris-wilson.co.uk>
22911 Date: Fri Sep 16 20:33:57 2011 +0100
22913 sna: Remove bad assert
22915 The caller intentionally violates this assert, so it is useless.
22917 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22919 commit 72ca1fdf2fea80ff4cf003bb025424d15d8bac58
22920 Author: Chris Wilson <chris@chris-wilson.co.uk>
22921 Date: Thu Sep 15 17:52:45 2011 +0100
22923 configure: Introduce --with-builderstring
22925 Julien Cristau disliked my broadcasting of the git tree used to build
22926 his distribution package as it bore little relevance to his users. As it
22927 is only useful for people installing their own drivers (as a means of
22928 sanity checking that they are running the right driver), we introduce
22929 the --with-builderstring idiom borrowed from the xserver. This allows
22930 the builder to override the use of `git describe` and either leave it
22931 blank or to fill it with something useful for their own purposes.
22933 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22935 commit 544a4167c156eac7b630f90ddef30a6f00829835
22936 Author: Chris Wilson <chris@chris-wilson.co.uk>
22937 Date: Wed Sep 14 18:39:37 2011 +0100
22939 sna/trapezoids: Amalgamate adjacent mono spans
22941 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22943 commit 3565c48c4bb77c836d817de75d098791dbb529d3
22944 Author: Chris Wilson <chris@chris-wilson.co.uk>
22945 Date: Wed Sep 14 17:45:41 2011 +0100
22947 sna: Yet another s/x/y/ typo
22949 Every time I do a transformation into pixmap space I like to include one
22950 of these copy'n'paste errors.
22952 Reported-by: Paul Neumann <paul104x@yahoo.de>
22953 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40850
22954 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22956 commit 4d4aca21fd32a367d78dfaf1c23b4faeabe36428
22957 Author: Chris Wilson <chris@chris-wilson.co.uk>
22958 Date: Wed Sep 14 13:06:26 2011 +0100
22960 sna: assert that the memcpy upload path points to valid regions
22962 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22964 commit 83961051789e418f6022096b1ed2302cb2f05e5b
22965 Author: Chris Wilson <chris@chris-wilson.co.uk>
22966 Date: Tue Sep 13 18:37:01 2011 +0100
22968 sna/gen3: Reset the shader after playing video
22970 Another piece of state we zap without marking as dirty when playing
22973 Reported-by: Paul Neumann <paul104x@yahoo.de>
22974 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40842
22975 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22977 commit 34c03f22158d8fc1ec2fc56cb5e87f74e42d9a8f
22978 Author: Chris Wilson <chris@chris-wilson.co.uk>
22979 Date: Tue Sep 13 17:24:14 2011 +0100
22981 sna: Clear the list of buffers upon server regen
22983 Or else we may try to clear the new framebuffer with an invalid batch,
22984 because it will reuse the same bo as last time and that bo may still
22985 think it is part of the old batch.
22987 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22989 commit 831cdb837189e92aee80030dec553f28e53ba0bc
22990 Author: Chris Wilson <chris@chris-wilson.co.uk>
22991 Date: Tue Sep 13 17:09:38 2011 +0100
22993 sna/debug: Assert that BLT commands on gen2/3 are properly fenced
22995 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
22997 commit a9242a77f0952d029be57d75615dd107af660de1
22998 Author: Chris Wilson <chris@chris-wilson.co.uk>
22999 Date: Tue Sep 13 16:20:40 2011 +0100
23001 sna: Add a couple of debug statements for why we may not flush the scanout
23003 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23005 commit 0056bfd11e9930c1bcad6c186cd80fbab3694a8d
23006 Author: Chris Wilson <chris@chris-wilson.co.uk>
23007 Date: Tue Sep 13 14:44:48 2011 +0100
23009 video: Stop advertising unsupported Xv attributes
23011 We have yet to implement a yuv-shader that applies
23012 contrast/brightness/saturation and so should not advertise such
23013 features, potentially allowing the client to fallback and perform the
23014 adjustments itself.
23016 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23018 commit 84a7c11a8134dfd040d2f90bb1e0670aa2c89962
23019 Author: Chris Wilson <chris@chris-wilson.co.uk>
23020 Date: Tue Sep 13 14:42:33 2011 +0100
23022 sna/video: Stop advertising unsupported Xv attributes
23024 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23026 commit a0d533ea8d098dbbd86af166de9c07ce67050c35
23027 Author: Chris Wilson <chris@chris-wilson.co.uk>
23028 Date: Tue Sep 13 12:23:10 2011 +0100
23030 sna/trapezoids: Implement a mono-rasteriser to skip opacity spans
23032 For mono-rasteriser we can simply emit the composite spans without
23033 requiring an opacity shader.
23035 For single trapezoids, it will be more interesting to emit triangles
23036 directly. However, we still need to compute the union of many
23037 trapezoids, and this builds upon existing code.
23039 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23041 commit 343b99d2a85c520c81d1408534b2531fcbedfaae
23042 Author: Chris Wilson <chris@chris-wilson.co.uk>
23043 Date: Tue Sep 13 11:39:39 2011 +0100
23045 sna: Reuse the smallest bo that fits the request
23047 When scanning the active list, pick the smallest that we can reuse.
23049 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23051 commit 90071454d24e7230d00c49f5a897296298fc1f22
23052 Author: Chris Wilson <chris@chris-wilson.co.uk>
23053 Date: Tue Sep 13 11:37:17 2011 +0100
23055 sna: Don't clear the kernel-needs-flush flag if we emit a MI_FLUSH
23057 We need to mirror the kernel active lists in order to predict when it
23058 will stall upon an access to a bo, and so we cannot clear the
23059 needs_flush for our own MI_FLUSH.
23061 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23063 commit 2deafe2d95d07b574f5a92471f7d7f13aebc790f
23064 Author: Chris Wilson <chris@chris-wilson.co.uk>
23065 Date: Tue Sep 13 11:34:43 2011 +0100
23067 sna/accel: Use correct enum for requesting a write-buffer
23069 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23071 commit 544b1a678d5bf49559d8422750866685c36cb152
23072 Author: Chris Wilson <chris@chris-wilson.co.uk>
23073 Date: Tue Sep 13 11:34:25 2011 +0100
23075 sna/gen5: Fix debug compilation
23077 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23079 commit 6b1ed58d63e9ac80d7d028fa3036633436154816
23080 Author: Chris Wilson <chris@chris-wilson.co.uk>
23081 Date: Mon Sep 12 19:07:25 2011 +0100
23083 sna/gen5: Avoid bitfields for simple assignments
23085 commit 03a7fc16f9ef00ca5591655337621ec67bc37cba
23086 Author: Chris Wilson <chris@chris-wilson.co.uk>
23087 Date: Mon Sep 12 12:20:24 2011 +0100
23089 sna: Avoid the call overhead for the trivial clip case
23091 Profile guided micro-optimisation. /o\
23093 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23095 commit 2540c877d4811b318e1c2cd707745e5b06c9fa4b
23096 Author: Chris Wilson <chris@chris-wilson.co.uk>
23097 Date: Mon Sep 12 11:54:46 2011 +0100
23099 sna/gen6: Prefer RENDER for copies as it compacts better
23101 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23103 commit f2a617dcc5fbeaf503db4777f0ec7b3d862085b9
23104 Author: Chris Wilson <chris@chris-wilson.co.uk>
23105 Date: Mon Sep 12 11:54:02 2011 +0100
23107 sna: Use single precision for computing the texcoord scale factor
23109 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23111 commit c2e19987c333de31b258a3af37a2ed1f63ef949a
23112 Author: Chris Wilson <chris@chris-wilson.co.uk>
23113 Date: Mon Sep 12 10:37:24 2011 +0100
23115 sna: Provide a direct lookup cache of alpha values
23117 To reflect recent changes in cairo.
23119 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23121 commit 04c5a3df02f6f40a904ff4edb927ae6ff0ce6408
23122 Author: Chris Wilson <chris@chris-wilson.co.uk>
23123 Date: Mon Sep 12 09:20:58 2011 +0100
23125 sna: Fix compilation without timerfd()
23127 Based on the patch by Konstantin Belousov.
23129 Reported-by: Konstantin Belousov <konstantin.belousov@zoral.com.ua>
23130 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23132 commit ae887c9d1cb2a06c515e054b36464bcf65440f9f
23133 Author: Chris Wilson <chris@chris-wilson.co.uk>
23134 Date: Sun Sep 11 13:59:08 2011 +0100
23136 sna/trapezoids: Prevent a memleak if feed lots of degenerate boxes
23138 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23140 commit 72fafdfd374eb165e703b791c70592c47e5772fc
23141 Author: Chris Wilson <chris@chris-wilson.co.uk>
23142 Date: Fri Sep 9 23:15:58 2011 +0100
23144 gitignore: add git_version.h
23146 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23148 commit ac4d4cdbc1336c86a16a2ec3da2ef6e9842c1079
23149 Author: Chris Wilson <chris@chris-wilson.co.uk>
23150 Date: Fri Sep 9 23:12:59 2011 +0100
23152 sna: Mark the bo as reusable after extracting the handle from the buffer
23154 The whole purpose for that little dance was so that we could reuse the
23155 bo. However, we left it marked as non-reusable in order for us not to
23156 tie up memory with too many buffers and so defeated the purpose of
23157 trying to place it into the inactive cache.
23159 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23161 commit b3429cf12d6e3e8dd89372f8f70c64205d835b0c
23162 Author: Chris Wilson <chris@chris-wilson.co.uk>
23163 Date: Fri Sep 9 18:09:55 2011 +0100
23165 sna/gen3: Use a clear pattern for ill-defined radial gradients
23167 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23169 commit b6837c21b459a99cb51b0d41906d82604e515dc0
23170 Author: Chris Wilson <chris@chris-wilson.co.uk>
23171 Date: Fri Sep 9 14:57:09 2011 +0100
23173 sna/gen5: Prefer BLT for solids
23175 And Ironlake also fails to live up to the promise that its GPU is fast
23176 enough to run simple programs at memory rates.
23178 x11perf -trap300 5x fold improvement. No obvious improvement elsewhere
23181 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23183 commit 27e42b4e1216099f4a6aa7fddeb1e83b87124b6e
23184 Author: Chris Wilson <chris@chris-wilson.co.uk>
23185 Date: Fri Sep 9 14:06:53 2011 +0100
23187 sna: Prefer memcpy_blt over fbBlt
23189 We know we have compatible formats since we have a gpu_bo attached to
23190 the pixmap, so we can use the simpler direct memcpy rather than calling
23193 On my i3-330m, this improves putimage500 from 730 to 1100 ops/s.
23195 Reported-by: Michael Larabel <Michael@phoronix.com>
23196 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23198 commit 2e1bf7e1b44db16d0c322f17535fc6a6fa07353b
23199 Author: Chris Wilson <chris@chris-wilson.co.uk>
23200 Date: Fri Sep 9 12:36:32 2011 +0100
23202 sna: Record git-tree used for compilation
23204 Hopefully, I have all the dependencies correct for auto-updating and
23205 should continue to work with tarballs...
23207 The next step is to perhaps include it in the usual version number,
23208 perhaps as patch level?
23210 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23212 commit f73cd955e760064f57f414dce5f39e43022e4ea6
23213 Author: Chris Wilson <chris@chris-wilson.co.uk>
23214 Date: Fri Sep 9 11:37:32 2011 +0100
23216 sna/trapezoids: Hook up Imprecise AddTraps in lieu of spans
23218 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23220 commit adde6eab5d0e532633c864830f8edfa63b36ef4e
23221 Author: Chris Wilson <chris@chris-wilson.co.uk>
23222 Date: Thu Sep 8 23:35:33 2011 +0100
23224 sna/trapezoids: Fast upload path for gpu busy bo
23226 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23228 commit e9ca05331db9849fba39fbd0583db7052fc80a31
23229 Author: Chris Wilson <chris@chris-wilson.co.uk>
23230 Date: Thu Sep 8 22:54:22 2011 +0100
23232 sna/traps: Use the trapezoid path for AddTraps
23234 Usually this will be to CPU-only pixmap, but just on the off-chance that
23235 we are stalling for a GPU pixmap just the faster path developed for
23238 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23240 commit afdb8aa89a88c62ccd5e26a3ed9adb70fec310a8
23241 Author: Chris Wilson <chris@chris-wilson.co.uk>
23242 Date: Wed Sep 7 20:10:02 2011 +0100
23244 sna/gen3: Do not assume video updates are always vsync'ed
23246 In case the video is running async, then there may be subsequent
23247 instructions within the batch and so we do need to mark the clobbered
23248 state as dirty when setting up the video frame.
23250 Reported-by: Paul Neumann <paul104x@yahoo.de>
23251 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40693
23252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23254 commit 6aee152cb8444bf63c0009a75c6537e9886b6070
23255 Author: Chris Wilson <chris@chris-wilson.co.uk>
23256 Date: Sun Sep 4 12:57:17 2011 +0100
23258 sna/gen2: Flush the batch when we run out of vertex space
23260 Unlike the later gen, we do not yet use a separate vertex buffer and so
23261 when can no longer fit a rectangle (and its CA ghost) we must flush the
23262 batch. Due to the duplication required for the CA pass, the normal
23263 checks to see whether we had sufficient space to add the new command
23264 were passing as they failed to take into account the need to submit the
23265 whole primitive again.
23267 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23269 commit 48bfe4e6ded746f32080b83079235a7576f50d7a
23270 Author: Chris Wilson <chris@chris-wilson.co.uk>
23271 Date: Sun Sep 4 12:46:32 2011 +0100
23273 sna/gen2: Improve batch decoder.
23275 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23277 commit 2cda0aaf397de1a0ca049508c6fa76f2dd4e61e8
23278 Author: Chris Wilson <chris@chris-wilson.co.uk>
23279 Date: Sat Sep 3 14:46:57 2011 +0100
23281 sna/trapezoids: Check for alignment after projection
23283 If after projection onto the Imprecise fast sample grid, the trapezoid
23284 becomes a pixel-aligned box, treat it as such and send it down the fast
23287 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23289 commit db0fb368c135d4fef4ae993df67ed4610a80fd52
23290 Author: Chris Wilson <chris@chris-wilson.co.uk>
23291 Date: Fri Sep 2 13:13:38 2011 +0100
23293 sna: Add missing implementation for Triangles
23295 Feed both into spans and as a mask fallback.
23297 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23299 commit 695e7115efab96c4467dc926be07dc2609011720
23300 Author: Chris Wilson <chris@chris-wilson.co.uk>
23301 Date: Thu Sep 1 15:50:29 2011 +0100
23303 sna/trapezoids: Edges may lie out of bounds
23305 We cannot assume that the edge lies completely within the target, so we
23306 must make sure that the initial prev_x is truly less than any possible
23307 value whilst sorting intersections.
23309 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23311 commit 9a563ea03b6ad87d41bc091c5819e6c462100450
23312 Author: Chris Wilson <chris@chris-wilson.co.uk>
23313 Date: Wed Aug 31 23:29:00 2011 +0100
23315 sna: Use the shadow buffer for PutImage
23317 This is optimising for the x11perf putimage benchmark, but nevertheless,
23318 uploading the PutImage directly into the uncached scanout is between
23319 2-20x slower than making a temporary copy in the shaodw buffer and
23320 doing a deferred update. Most of the overhead is in the kernel, and
23321 should be addressed there (rather than worked around) and a portion is
23322 due to the overdraw in the benchmark (which is not likely to be
23323 realistic, but then again neither should PutImage be!).
23325 The argument for uploading inplace when possible is that given that the
23326 buffer already exists on the GPU implies that is likely to be used again
23327 in future by the GPU and so we will be uploading it at some point.
23328 Deferring that upload incurs an extra copy. The putimage benchmark does
23329 not actually use the pixel data and so that extra cost is not being
23332 Reported-by: Michael Larabel <Michael@phoronix.com>
23333 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23335 commit 32fc0c896e0dfd06617c12beda1ccacedf69fb4a
23336 Author: Chris Wilson <chris@chris-wilson.co.uk>
23337 Date: Wed Aug 31 23:21:54 2011 +0100
23339 sna/gen6: Prefer the BLT ring, except for copies on behalf of DRI
23341 As demonstrated by the all-important trap300, using the BLT is 2x faster
23342 than the RENDER ring for the simple case of solid fills. (Though note
23343 that performing the relocations costs 3x as much CPU for 2x GPU
23344 performance.) One case that may regress from this change is copywinpix
23345 which should benefit from the batching in the RENDER commands, and might
23346 warrant revisiting in the future (with realistic and synthetic
23347 benchmarks in hand!)
23349 However, due to the forced stall when switching rings, we still want to
23350 perform RENDER copies on behalf of DRI clients and before page-flips.
23352 Checking against cairo-perf-trace indicated no major impact -- I had
23353 worried that setting the BLT flag for some clears might have had a
23354 knock-on effect causing too many operations that could be pipelined on
23355 the RENDER ring to be sent to the BLT ring instead.
23357 Reported-by: Michael Larabel <Michael@phoronix.com>
23358 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23360 commit 5586dd729b153e37f942a285d328a07f3fe7ae16
23361 Author: Chris Wilson <chris@chris-wilson.co.uk>
23362 Date: Wed Aug 31 09:25:48 2011 +0100
23364 sna/trapezoids: Refactor to project the trapezoid only once
23366 And doing so means that we can go back to using the common validity
23369 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23371 commit 3507437cdba62c28772822cce8c4b656448546db
23372 Author: Chris Wilson <chris@chris-wilson.co.uk>
23373 Date: Tue Aug 30 22:13:24 2011 +0100
23375 sna/trapezoids: Reject invalid traps after projecting onto the sample grid
23377 If either of the edges are degenerate on the sample grid, then the trap
23378 has zero height and must be skipped. (Otherwise if just one edge becomes
23379 degenerate than the polygon becomes unbalanced and the rasteriser will
23382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23384 commit 150a0612d545b7fd77bf393efdef1cc92ed91e61
23385 Author: Chris Wilson <chris@chris-wilson.co.uk>
23386 Date: Tue Aug 30 19:45:22 2011 +0100
23388 sna/trapezoids: Allocate sufficient space for a8 mask for mono traps
23390 Oops, a silly cut'n'paste from caused us to allocate an A1 pixmap for
23391 mono traps instead of the A8 pixmap that we tried to write to; mayhem
23394 Reported-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
23395 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23397 commit aafe03d3d1cfd5cbdf345cd51436de268ac27b6f
23398 Author: Chris Wilson <chris@chris-wilson.co.uk>
23399 Date: Mon Aug 29 16:47:48 2011 +0100
23401 sna: Retain the GTT space used for an upload buffer
23403 In order to retain the GTT space without keeping hold of the memory used
23404 for the upload buffer, we have to create a new bo and copy the relevant
23407 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23409 commit 28c8c5ca14c5b4eab7e2dc58c87050c1019037bb
23410 Author: Chris Wilson <chris@chris-wilson.co.uk>
23411 Date: Mon Aug 29 16:27:35 2011 +0100
23413 sna: Free the buffers immediately upon release
23415 They do not appear to have been leaked per-se, but we end up
23416 accumulating the unused buffers. A more complicated solution would be to
23417 reallocate the handle for retained buffers so that the GTT region can be
23420 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=39184
23421 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23423 commit 4f2fc00944272243af343a8b99f6eb8e0dfda006
23424 Author: Chris Wilson <chris@chris-wilson.co.uk>
23425 Date: Mon Aug 29 15:14:41 2011 +0100
23427 sna: Cleanup up the cache upon close
23429 To help with leak-chasing under valgrind.
23431 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23433 commit 0ac4b974b90a5614d1114d5e211cb4b986a89454
23434 Author: Chris Wilson <chris@chris-wilson.co.uk>
23435 Date: Mon Aug 29 10:46:46 2011 +0100
23437 sna/video: Defend against PutImage to a broken screen
23439 Similar to the previous commit, check that the Screen Pixmap is bound to
23440 a bo before proceeding.
23442 [Note that in this case, the absence of the bo would have been picked
23443 up much later after doing all of the setup...]
23445 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23447 commit 0a74cd77a3f462d5d2707bf1f35663279bcb9be5
23448 Author: Chris Wilson <chris@chris-wilson.co.uk>
23449 Date: Mon Aug 29 10:41:26 2011 +0100
23451 video: check that the pixmap exists before use
23453 Now, the pixmap being used is meant to the Screen pixmap and by rights
23454 that has to exists in a GPU buffer! Evidence contrary to the above
23455 exists and so we had better check that we have a bo before using...
23457 Reported-by: Toralf Förster <toralf.foerster@gmx.de>
23458 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40439
23459 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23461 commit 8216c92d5c49d4b8de5c16b615bca0acece49a70
23462 Author: Chris Wilson <chris@chris-wilson.co.uk>
23463 Date: Sun Aug 28 22:32:09 2011 +0100
23465 sna/trapezoids: Use the tor scan converter to compute the low precision mask
23467 Take of the advantage of the faster mask computation available using the
23468 imprecise tor scan converter for chipsets non yet supporting spans.
23469 In doing so, limit the ability to full step only for vertical only rows
23470 as the small sample grid reduces the benefits of the computationally
23471 more expensive full-step.
23473 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23475 commit aeee6db798e9ded746b399bc08716cff0bd16ef8
23476 Author: Chris Wilson <chris@chris-wilson.co.uk>
23477 Date: Sat Aug 27 11:43:48 2011 +0100
23479 sna/trapezoids: Reduce imprecise sampling to 4x4
23481 Note this also revealed a subtle bug in the handling of degenerate
23482 trapezoids after shrinking to the raster grid.
23484 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23486 commit ac1b83240e5d4dc5d5bacad3ed49446143cc5589
23487 Author: Chris Wilson <chris@chris-wilson.co.uk>
23488 Date: Sat Aug 27 11:42:19 2011 +0100
23490 sna/accel: Simplify single pixel read-back
23492 The single pixel case is usually assocated with synchronisation of perf
23493 clients and so we do not want to incur extra complication along that
23494 path. Also the cost of tracking a single pixel of non-damage outweighs
23497 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23499 commit 786a770f528a0daee2971494352672cb89f48384
23500 Author: Chris Wilson <chris@chris-wilson.co.uk>
23501 Date: Thu Aug 25 19:52:13 2011 +0100
23503 sna/video: Flush the video state at the end of the operation
23505 Or in the case where a second command is received prior to the batch
23506 being flushed, the vertex data is not flushed and leads to the a
23507 miscompution of the number of vertices emitted.
23509 Reported-by: Elias Probst <mail@eliasprobst.eu>
23510 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=40332
23511 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23513 commit bd98001a49be061f0f2cb008b515f5505c8a63a4
23514 Author: Chris Wilson <chris@chris-wilson.co.uk>
23515 Date: Thu Aug 25 14:49:30 2011 +0100
23517 sna: Clear structures across server reset
23519 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23521 commit 0865acb3ad03239ffdf8cbf3c9fc204b6c97121b
23522 Author: Chris Wilson <chris@chris-wilson.co.uk>
23523 Date: Thu Aug 25 12:47:27 2011 +0100
23525 sna/dri2: Add some debug around the use of the Resource database
23527 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23529 commit 98b67457caea583de50bff1b11a3c5a2bd694e26
23530 Author: Chris Wilson <chris@chris-wilson.co.uk>
23531 Date: Thu Aug 25 12:45:43 2011 +0100
23533 sna/display: Destroy shadow data
23535 Under certain circumstances the shadow can be destroy after being
23536 allocated but before being created. The pixmap is a NULL pointer at that
23537 time, but we know that its value should be data, so just use the data
23540 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23542 commit b9ae4e7e71527f1cf2b6dd6890420da22a22fad9
23543 Author: Chris Wilson <chris@chris-wilson.co.uk>
23544 Date: Wed Aug 24 21:38:28 2011 +0100
23546 sna/gen3: reset blend state after applying CA pass
23548 Otherwise we use the stale value when rendering CA glyphs directly to
23549 the front-buffer and subsequent rendering have a tendency to become
23550 invisible. (Rendering via a temporary glyph mask has a fortunate
23551 side-effect of reseting sufficient state to force the re-emission of the
23554 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23556 commit ef52f6c8c3421a31552d5965c40b4d9cf68f6a5c
23557 Author: Chris Wilson <chris@chris-wilson.co.uk>
23558 Date: Wed Aug 24 11:04:28 2011 +0100
23560 sna/render: allow CLAMP_TO_EDGE for outside samples of extract regions as well
23562 When clipping the sample region to the edge of the texture we can also
23563 allow the GPU to use CLAMP_TO_EDGE (as well as CLAMP_TO_BORDER) to
23564 emulate the RepeatPad mode of the parent texture. (Only the
23565 RepeatNormal, RepeatReflect need special treatment with regard to tiling
23566 that is not yet handled.)
23568 This fixes the recent performance regression due to a slight change in
23569 the fish benchmark that caused it to sample outside of the texture atlas
23570 for one of its little fish.
23572 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23574 commit d330f3751e484d28293fc593a20375a31b068bc7
23575 Author: Stefan Dirsch <sndirsch@suse.de>
23576 Date: Fri Jul 22 12:21:01 2011 +0200
23578 Fix array size calculation for intel_pci_probe().
23580 commit ccddff087df0c567c28416941b175be81190a1d3
23581 Author: Chris Wilson <chris@chris-wilson.co.uk>
23582 Date: Mon Aug 8 00:42:21 2011 +0100
23584 sna/trapezoids: Speedup tor rasteriser
23586 Faster sorts for the win.
23588 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23590 commit bfbe36cfea76337689dd8a101ec03469f6d3553d
23591 Author: Chris Wilson <chris@chris-wilson.co.uk>
23592 Date: Thu Aug 11 01:08:49 2011 +0100
23594 sna/gradient: Use a high-precision ramp for a color step rather than fallback
23596 Slightly less precise, but the difference should not be observable in
23599 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23601 commit 0e61e235bf7a926fd4e5b1f5a05b72dce4c450f3
23602 Author: Chris Wilson <chris@chris-wilson.co.uk>
23603 Date: Wed Jul 13 22:55:31 2011 +0100
23605 sna/damage: Take advantage of marking all-damaged
23607 Return early from adding new damage regions if we know that we have
23608 already marked it as all-damaged.
23610 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23612 commit 3a81bb6bafdbd37802dab96b8f05173ec6701d7f
23613 Author: Chris Wilson <chris@chris-wilson.co.uk>
23614 Date: Tue Aug 9 09:42:40 2011 +0100
23616 NEWS: 2.16.0 release
23618 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23620 commit f4bbbd1dfe59031f864c7d2bbf7bb0252a3aa6a7
23621 Author: Edward Sheldrake <ejsheldrake@gmail.com>
23622 Date: Mon Aug 1 14:46:08 2011 +0100
23624 Fix man page formatting
23626 Two option sections were not starting at the beginning of a new line.
23628 commit 63518c42234001ec96f638af5732c09079a3d682
23629 Author: Chris Wilson <chris@chris-wilson.co.uk>
23630 Date: Mon Aug 1 13:36:28 2011 +0100
23632 dri: Build fix for xserver-1.7.7
23634 Back in the olden days before the introduction of dixRegisterPrivate().
23636 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23638 commit 7976f5144d42a03ccd027908252a600db2631054
23639 Author: Chris Wilson <chris@chris-wilson.co.uk>
23640 Date: Sat Jul 30 09:21:36 2011 +0100
23642 NEWS: 2.15.901 snapshot
23644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23646 commit 2cfb703bbe7ec043d443452f4fa94f06b1ff7266
23647 Author: Chris Wilson <chris@chris-wilson.co.uk>
23648 Date: Sat Jul 30 09:23:22 2011 +0100
23650 Fix typos for distcheck
23652 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23654 commit 6f919264da69ed8bce6f2573629d89962a670783
23655 Author: Chris Wilson <chris@chris-wilson.co.uk>
23656 Date: Sat Jul 30 09:00:06 2011 +0100
23658 sna: Mark the stencil as untiled
23660 In reality, Mesa will be treating it as W-tiling, only we have no way of
23661 communicating that requirement to the kernel (as not only does the
23662 kernel not understand W-tiling, but also the GTT is incapable of fencing
23663 a W-tiled region.).
23665 Ported from Chad Versace's 3e55f3e88.
23667 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23669 commit 326a84e832c43cc200a32e091b4e60b8b6a4a7c8
23670 Author: Chris Wilson <chris@chris-wilson.co.uk>
23671 Date: Fri Jul 22 11:10:26 2011 +0100
23673 sna: Port IVB acceleration code (Xrender + Xv)
23675 Based on the superlative work by Kenneth Graunke and Xiang, Haihao.
23677 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23679 commit 10790921570fea21d6b2bcc07a0204ef315c9182
23680 Author: Chris Wilson <chris@chris-wilson.co.uk>
23681 Date: Sat Jul 30 08:46:19 2011 +0100
23683 sna: Include the pixmap size in the debug info for moving to cpu
23685 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23687 commit 5691c8cdec2a5264b8552f096b4a34b97afe4ae5
23688 Author: Kenneth Graunke <kenneth@whitecape.org>
23689 Date: Thu Jul 14 10:14:53 2011 -0700
23691 render: Enable RENDER acceleration on Ivybridge.
23693 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23694 Acked-by: Eric Anholt <eric@anholt.net>
23695 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23697 commit 0d92612b2a2782f80196a08eb9a17af906169f18
23698 Author: Kenneth Graunke <kenneth@whitecape.org>
23699 Date: Thu Jul 14 00:30:09 2011 -0700
23701 render: Update pixel shader state for Ivybridge.
23703 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23704 Acked-by: Eric Anholt <eric@anholt.net>
23705 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23707 commit 7460ee73d1fd22e6b02ce125f11ac38efff743ce
23708 Author: Kenneth Graunke <kenneth@whitecape.org>
23709 Date: Thu Jul 14 00:16:54 2011 -0700
23711 render: Use Ivybridge variants for 3D pipeline setup.
23713 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23714 Acked-by: Eric Anholt <eric@anholt.net>
23715 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23717 commit e3a09608716c1dee2554e8fc26d6f77aeec684bd
23718 Author: Kenneth Graunke <kenneth@whitecape.org>
23719 Date: Thu Jul 14 00:04:31 2011 -0700
23721 render: Refactor to use newly shared pipeline setup code in i965_3d.c.
23723 Slightly generalize the shared SF and CC code to accomodate both.
23725 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23726 Acked-by: Eric Anholt <eric@anholt.net>
23727 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23729 commit 682a690bfeeabae710b1392282163eab35b58eed
23730 Author: Kenneth Graunke <kenneth@whitecape.org>
23731 Date: Wed Jul 13 22:52:52 2011 -0700
23733 Xv: Refactor out pipeline setup functions for future reuse in render.
23735 While we're at it, make the functions simply take an intel_screen_private
23736 pointer directly instead of having to fetch it from ScrnInfoPtr.
23738 Also coalesce some gen6/gen7 functions that were 98% identical.
23740 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23741 Acked-by: Eric Anholt <eric@anholt.net>
23742 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23744 commit 54b3222658a285d26b7800bdc5f8343c918a804e
23745 Author: Kenneth Graunke <kenneth@whitecape.org>
23746 Date: Wed Jul 13 22:27:37 2011 -0700
23748 render: Update 3DPRIMITIVE for Ivybridge.
23750 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23751 Acked-by: Eric Anholt <eric@anholt.net>
23752 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23754 commit 4e491a1f6d7a15009904cb79e2c0b13cf2d9d6ec
23755 Author: Kenneth Graunke <kenneth@whitecape.org>
23756 Date: Wed Jul 13 22:19:21 2011 -0700
23758 render: Set Address Modify Enable in 3DSTATE_VERTEX_BUFFERS on Gen7.
23760 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23761 Acked-by: Eric Anholt <eric@anholt.net>
23762 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23764 commit 2787cd66f0d2907110f774392370537df63a96fa
23765 Author: Kenneth Graunke <kenneth@whitecape.org>
23766 Date: Wed Jul 13 22:10:17 2011 -0700
23768 render: Update SAMPLER_STATE for Ivybridge.
23770 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23771 Acked-by: Eric Anholt <eric@anholt.net>
23772 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23774 commit 1a7e541d0d33727953fa69f1e29fa4eda665a0d9
23775 Author: Kenneth Graunke <kenneth@whitecape.org>
23776 Date: Wed Jul 13 21:53:27 2011 -0700
23778 render: Update SURFACE_STATE for Ivybridge.
23780 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23781 Acked-by: Eric Anholt <eric@anholt.net>
23782 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23784 commit 07cc488bcf3f7653cd54928e3cedb1f4f102c5e0
23785 Author: Kenneth Graunke <kenneth@whitecape.org>
23786 Date: Sat Jul 9 01:36:52 2011 -0700
23788 render: New Ivybridge assembly programs for render acceleration.
23790 These are exactly the same as the ones for Sandybridge, but with message
23791 registers translated (hopefully) in the same way as Haihao's new
23792 programs (m1 == g65).
23794 Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
23795 Acked-by: Eric Anholt <eric@anholt.net>
23796 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
23798 commit 1b9e82b4b59dd50c1b057f58d71d5c04dcc648a7
23799 Author: Chris Wilson <chris@chris-wilson.co.uk>
23800 Date: Tue Jul 26 08:29:36 2011 +0100
23802 sna: Revert enabling scan-line wait on SNB
23804 Hanging the machine does indeed prevent video tearing. Just not quite
23805 what the user expected...
23807 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=39497
23808 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23810 commit 6dbbb74bde034f5f00aee0396ccd1e03a6625fbd
23811 Author: Chris Wilson <chris@chris-wilson.co.uk>
23812 Date: Sat Jul 16 14:05:54 2011 +0100
23814 sna: Enable gen6 scan-line waiting
23816 The code was ready and waiting, just forgot to turn it on.
23818 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23820 commit 3e55f3e88b40471706d5cd45c4df4010f8675c75
23821 Author: Chad Versace <chad@chad-versace.us>
23822 Date: Mon Jul 18 00:38:00 2011 -0700
23824 dri: Do not tile stencil buffer
23826 Until now, the stencil buffer was allocated as a Y tiled buffer, because
23827 in several locations the PRM states that it is. However, it is actually
23828 W tiled. From the PRM, 2011 Sandy Bridge, Volume 1, Part 2, Section
23829 4.5.2.1 W-Major Format:
23830 W-Major Tile Format is used for separate stencil.
23832 The GTT is incapable of W fencing, so we allocate the stencil buffer with
23833 I915_TILING_NONE and decode the tile's layout in software.
23835 This commit mutually depends on the mesa commit:
23836 intel: Fix stencil buffer to be W tiled
23837 Author: Chad Versace <chad@chad-versace.us>
23838 Date: Mon Jul 18 00:37:45 2011 -0700
23840 Signed-off-by: Chad Versace <chad@chad-versace.us>
23841 Reviewed-by: Ian Romanick <ian.romanick@intel.com>
23842 Acked-by: Kenneth Graunke <kenneth@whitecape.org>
23844 commit 212fa9868767637e8f430485eeb522c99e63fd16
23845 Author: Chris Wilson <chris@chris-wilson.co.uk>
23846 Date: Wed Jul 13 21:11:14 2011 +0100
23848 Disable adding normal RTF modes for an eDP
23850 This is causing a hard hang with 2.6.39+, we don't know why so play safe
23851 and disable for the time being.
23853 References: https://bugs.freedesktop.org/show_bug.cgi?id=38012
23854 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23856 commit 7a695c9f6b919eba8785d62094526762ba2a0449
23857 Author: Chris Wilson <chris@chris-wilson.co.uk>
23858 Date: Wed Jul 13 10:57:13 2011 +0100
23860 sna: Fast-path single span boxes
23862 These are very common when compositing unclipped trapezoids, and the
23863 majority of the overhead is in handling the arbitrary number of boxes
23864 and misses out on the constant folding the compiler can do if it is
23865 known we have just one box.
23867 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23869 commit 0190964906ad878d469d6021c10cde4e7f5799c3
23870 Author: Chris Wilson <chris@chris-wilson.co.uk>
23871 Date: Wed Jul 13 10:26:44 2011 +0100
23873 sna/damage: Avoid testing against a completey damaged region
23875 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23877 commit b929717c89352d4b47dc2b9ce58e26fbbf327201
23878 Author: Chris Wilson <chris@chris-wilson.co.uk>
23879 Date: Wed Jul 13 09:49:09 2011 +0100
23881 sna/gen3: Tune emit_spans_primitive_constant
23883 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23885 commit fbdbfaf38d4da5204750d91cf5a3f43307a6ac33
23886 Author: Chris Wilson <chris@chris-wilson.co.uk>
23887 Date: Wed Jul 13 17:38:41 2011 +0100
23889 sna/glyphs: Discard GLYPH_PICTURE hint if the glyph doesn't fit into the cache
23891 If the glyph is too big to fit into the cache, than ideally we do want
23892 to keep an associated GPU bo around for future use. As it is too large
23893 to fit into the cache, it of reasonable size and there is little wastage
23894 in allocating indiviual GPU bo for each oversized glyph.
23896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23898 commit 12f52530db1727e2f835e55dc922e5efaf4d3a16
23899 Author: Chris Wilson <chris@chris-wilson.co.uk>
23900 Date: Wed Jul 13 17:38:22 2011 +0100
23902 sna: Add some extra debugging to the texture upload fallback paths
23904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23906 commit a861094c23b5f0df7b889bcc03904e32c4c2c6be
23907 Author: Chris Wilson <chris@chris-wilson.co.uk>
23908 Date: Tue Jul 12 20:17:00 2011 +0100
23910 sna/dri: Fix a couple of typos
23912 Somehow these were lost in the rebasing.
23914 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23916 commit c221d0356db57c10f5d29ca17fdf592724b35385
23917 Author: Chris Wilson <chris@chris-wilson.co.uk>
23918 Date: Tue Jul 12 15:39:42 2011 +0100
23920 sna/dri: Remove the unused id/type members for Resource tracking
23922 ...and reduce it to a simple list.
23924 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23926 commit 644b1a903301531e1fb59f27952c15e87888c4c6
23927 Author: Chris Wilson <chris@chris-wilson.co.uk>
23928 Date: Tue Jul 12 11:50:24 2011 +0100
23930 dri: Always initialise resource members of DRI2FrameEvent
23932 As we now attempt to always decouple the lists upon freeing the frame
23933 event, we need to initialise them along all code paths.
23935 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23937 commit 32f42358140ee812984149ae52b0df3dfd1778c3
23938 Author: Chris Wilson <chris@chris-wilson.co.uk>
23939 Date: Mon Jul 11 20:29:53 2011 +0100
23941 sna/dri: Add some simple debugging
23943 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23945 commit a46598220ebf5d4e629e1e0a7baf47ce144ed2c8
23946 Author: Chris Wilson <chris@chris-wilson.co.uk>
23947 Date: Mon Jul 11 14:40:30 2011 +0100
23949 sna/dri: Refactor common code for assigning a pixmap to the DRI2 buffer
23951 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23953 commit 7538be3315b8683b05e8f6b22023baadcc0bc4da
23954 Author: Chris Wilson <chris@chris-wilson.co.uk>
23955 Date: Mon Jul 11 10:50:36 2011 +0100
23957 dri: Enable triple-bufferred pageflips
23961 Triple-buffering trade-offs output latency versus jitter. By having a
23962 pre-rendered frame ready to swap in following a pageflip, we avoid the
23963 scenario where the latency between receiving the flip complete signal
23964 from the kernel, waking up the vsynced application, it render the new
23965 frame and then for the server to process the swap request is greater
23966 than the frame interval, causing us to miss the vblank. The result is
23967 that application can become frame-locked to 30fps. Instead, we report to
23968 the application that the first frame swap is immediately completed,
23969 supply a new back buffer (or else the rendering would be blocked on
23970 waiting for the front-buffer to be swapped away from the scanout) and
23971 let them proceed to render the second frame. The second frame is added
23972 to the swap queue, and the client throttled to vrefresh. (If the client
23973 missed the vblank, the swap queue is empty and the client is immediately
23974 woken again, whilst the pageflip is pending.)
23976 Note, for practical reasons this only applies to page-flipping, for
23977 example, calls to glXSwapBuffer() on fullscreen applications.
23979 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
23981 commit 2608a367acba7247e50754c3daeed09ba2e97d05
23982 Author: Chris Wilson <chris@chris-wilson.co.uk>
23983 Date: Mon Jul 11 16:28:15 2011 +0100
23985 dri: Prevent abuse of the Resource database
23987 The Resource database is only designed to store a single value for a
23988 particular type associated with an XID. Due to the asynchronous nature
23989 of the vblank/flip requests, we would often associate multiple frame
23990 events with a particular drawable/client. Upon freeing the resource, we
23991 would not necessarily decouple the right value, leaving a stale pointer
23992 behind. Later when the client disappeared, we would write through that
23993 stale pointer upsetting valgrind and causing memory corruption. MDK.
23995 Instead, we need to implement an extra layer for tracking multiple
23996 frames within a single Resource.
23998 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=37700
23999 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24001 commit ab1000821ae881a301fb0e1f2210493ec383e681
24002 Author: Chris Wilson <chris@chris-wilson.co.uk>
24003 Date: Sat Jul 9 19:44:26 2011 +0100
24005 dri: Remove the shadow copy of attachment
24007 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24009 commit 9f22ea7ca440c788f8104647356393c96c62d155
24010 Author: Chris Wilson <chris@chris-wilson.co.uk>
24011 Date: Sat Jul 9 14:54:33 2011 +0100
24013 sna: Clamp results for computing BoxRec coords from xRectangle
24015 As the width/height in the rectangle is specified as uint16_t, the
24016 result may be larger than is storagable in the int16_t of the box. Of
24017 course it would take a really inane client to do attempt to draw
24018 something much larger than the largest possible surface... Is it strange
24019 that first example I've found to do so is a Java application?
24021 Reported-by: Nicolas Kalkhof <nkalkhof@web.de>
24022 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24024 commit f4c5dc88171435aa04a5d99372913120cdafe92b
24025 Author: Chris Wilson <chris@chris-wilson.co.uk>
24026 Date: Sat Jul 9 14:17:24 2011 +0100
24028 sna/accel: Fix fallback for depth=1 copy
24030 A little carelessness with passing down the offsets caused us to
24031 incorrectly copy depth=1 bitmaps, as exemplified by gkrellm.
24033 Reported-by: Nicolas Kalkhof <nkalkhof@web.de>
24034 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24036 commit 649ebcef0938c7cde4f3172a639d39488454999a
24037 Author: Chris Wilson <chris@chris-wilson.co.uk>
24038 Date: Fri Jul 8 17:44:58 2011 +0100
24040 sna: A buffer only needs a flush if it remains dirty at the end of the batch
24042 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24044 commit 625e37b317e4d237bdc9f80d7ab86b5d2badddc7
24045 Author: Chris Wilson <chris@chris-wilson.co.uk>
24046 Date: Fri Jul 8 18:58:32 2011 +0100
24048 sna/gen3: So we also need to ensure stippling is cleared...
24050 My theory that we used nothing that invoked polygon stippling proved
24053 Fixes regression from 3b5971bd2359383cb8326702d80e03bc15d34c69
24055 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24057 commit 1e2cae0ab339e477929e429f8ff31b11fcaf2cbf
24058 Author: Chris Wilson <chris@chris-wilson.co.uk>
24059 Date: Fri Jul 8 09:54:43 2011 +0100
24061 sna/gen3: Restore disabling the use of stencil/fog in the invariant
24063 One cleanup too far causing spurious results after rebooting. We also
24064 need to ensure that the writemask is fully enabled (ie not disabled)
24067 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24069 commit ec3dd64e739c4c2bbd8141de5275fe8a90f24a57
24070 Author: Chris Wilson <chris@chris-wilson.co.uk>
24071 Date: Fri Jul 8 00:17:27 2011 +0100
24073 sna/dri: Enable chaining of page-flips
24075 Trade off extra frames of latency for extra frames of anti-jitter
24076 buffering and loss of completion information; compiz users rejoice.
24078 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24080 commit a32694b0f07e35c22dd83bcfb828196d5d86628e
24081 Author: Chris Wilson <chris@chris-wilson.co.uk>
24082 Date: Thu Jul 7 23:11:39 2011 +0100
24084 sna/dri: Remove redundant NULL check in reference
24086 The buffer has already been dereferenced by this point...
24088 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24090 commit d180c5f5f781f4fa7f747dd7bb8073a20771cd79
24091 Author: Chris Wilson <chris@chris-wilson.co.uk>
24092 Date: Wed Jul 6 20:19:00 2011 +0100
24094 sna: Take advantage of the needs_flush tracking on the front buffer
24096 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24098 commit 0be47f459ba576ed3f93367de18732fb2367357a
24099 Author: Chris Wilson <chris@chris-wilson.co.uk>
24100 Date: Wed Jul 6 13:02:53 2011 +0100
24102 sna: Check against an execbuffer reference before discarding partials
24104 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24106 commit f6573fe7577e00a1d2706dbc39179ab658c021d2
24107 Author: Chris Wilson <chris@chris-wilson.co.uk>
24108 Date: Wed Jul 6 12:44:08 2011 +0100
24110 sna: Compute aligned tiled heights for gen2 correctly
24112 We were underestimating the height of X-tiled surfaces (and less
24113 harmfully overestimating the height of Y-tiled surfaces.)
24115 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24117 commit d6afd66461ebcdc2e8dcd94b3f46f374d8acf469
24118 Author: Chris Wilson <chris@chris-wilson.co.uk>
24119 Date: Tue Jul 5 23:12:59 2011 +0100
24121 sna: Reset unused partial buffers
24123 Whilst searching for available space on the active partial buffer list,
24124 if we discover an unreferenced one, reset its used counter to zero.
24126 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24128 commit 6e7a0c86419bf6c928837f592784333c25d8b27b
24129 Author: Chris Wilson <chris@chris-wilson.co.uk>
24130 Date: Tue Jul 5 23:05:37 2011 +0100
24132 sna: Discard unused partial buffers
24134 If we allocate a partial buffer and then fallback for the operation, the
24135 buffer would remain on the partial list waiting for another user.
24136 Discard any unused partials at the next batch submission or expiration
24139 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24141 commit 3b5971bd2359383cb8326702d80e03bc15d34c69
24142 Author: Chris Wilson <chris@chris-wilson.co.uk>
24143 Date: Tue Jul 5 22:20:28 2011 +0100
24145 sna/gen2: Restore invariant ENABLES
24147 One deletion too many, unnoticed until the next reboot. Besides the
24148 failure to disable logic op and enable colour buffer blending which
24149 causes a hang if you subsequently try to enable both, you also need
24150 to request texture caching...
24152 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24154 commit 5fa3e73f2c03bcadb87c1760f76f853c22fdba18
24155 Author: Chris Wilson <chris@chris-wilson.co.uk>
24156 Date: Tue Jul 5 21:38:08 2011 +0100
24158 sna/gen[23]: Do as the comments suggest and prefer the BLT
24160 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24162 commit f749ed618e21b56a36a2feb9b4333ec797ec2ae5
24163 Author: Chris Wilson <chris@chris-wilson.co.uk>
24164 Date: Tue Jul 5 21:37:02 2011 +0100
24166 sna: Reduce tiling if pitch is less than a tile_width/height only on pre-G33
24168 (Note this only applies to 2D pixmaps.)
24170 The rationale, borne out by experimentation with cairo-perf-trace, is
24171 that on the pre-G33 devices we always need a fence region region
24172 for tiled surfaces, i.e. at least .5/1MiB in size, and that combined
24173 with the smaller GTT on those devices, we loose the benefit of tiling to
24174 the excessive GTT thrashing.
24176 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24178 commit b9de6a98d33db660ffad3f186c2b0ab2aea71ac0
24179 Author: Chris Wilson <chris@chris-wilson.co.uk>
24180 Date: Tue Jul 5 17:48:11 2011 +0100
24182 sna: Remove unused aperture_size member
24184 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24186 commit fd3bc2af471a3f9a73064cb28818fdd3485cffc7
24187 Author: Chris Wilson <chris@chris-wilson.co.uk>
24188 Date: Mon Jul 4 16:31:20 2011 +0100
24190 sna: Clamp object size to the min of 1/4 of the whole GTT or 1/2 the mappable
24192 ... for those pesky early devices whose GTT was no larger than the AGP
24195 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24197 commit d294e41a6a49ddaa6f8d7d5cda266168095fe1f5
24198 Author: Chris Wilson <chris@chris-wilson.co.uk>
24199 Date: Mon Jul 4 13:45:06 2011 +0100
24201 sna: Update flush/retirement lists after a implicit flush for mmap
24203 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24205 commit 3e53b0f3a365c031bfef8d087bcee0ae1c2d2988
24206 Author: Chris Wilson <chris@chris-wilson.co.uk>
24207 Date: Mon Jul 4 13:44:28 2011 +0100
24209 sna: Enable relaxed-fencing for gen2 devices
24211 (Just as dependent upon non-buggy kernels as gen3...)
24213 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24215 commit 33ddaf54299979e98b196d479f1b2060d9551cc5
24216 Author: Chris Wilson <chris@chris-wilson.co.uk>
24217 Date: Mon Jul 4 13:43:48 2011 +0100
24219 sna: Fix gen2 tiled surface sizes
24221 Actually use the gen2 path for gen2 devices!
24223 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24225 commit 9eceddf69f78fc79c02ca75ed10b000beeff2033
24226 Author: Chris Wilson <chris@chris-wilson.co.uk>
24227 Date: Mon Jul 4 13:42:58 2011 +0100
24229 sna/gen2: fix batch buffer acounting
24231 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24233 commit 3f80f7edb8c511b94e66159cd59ba9362c3f69a2
24234 Author: Chris Wilson <chris@chris-wilson.co.uk>
24235 Date: Sat Jul 2 23:16:04 2011 +0100
24237 sna: Manually set to the GTT domain for mmap
24239 ...since the kernel no longer does strict coherency.
24241 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24243 commit f91ee24b2dabb48288d6e81dcdd82191f158e312
24244 Author: Chris Wilson <chris@chris-wilson.co.uk>
24245 Date: Sat Jul 2 23:05:04 2011 +0100
24247 sna: Trim number of downsample passes
24249 If we can fit the entire width or the entire height into the pipeline
24250 when downsampling, do so.
24252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24254 commit 6db93720a73f59a9857a3c5ab260fab8b957813e
24255 Author: Chris Wilson <chris@chris-wilson.co.uk>
24256 Date: Sat Jul 2 23:04:14 2011 +0100
24258 sna: Don't change tiling modes on replace
24260 This was trying to workaround a kernel bug, and instead causes a
24261 performance cliff for textures that *need* to be tiled.
24263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24265 commit 430c905ef306ece8a4cb19091711a9feae74b00b
24266 Author: Chris Wilson <chris@chris-wilson.co.uk>
24267 Date: Sat Jul 2 18:24:40 2011 +0100
24269 sna: Force tiled modes for large pitches
24271 If the surface is so big that the 2x2 texel sampling will cause a TLB
24272 miss everytime, i.e. the row pitch exceeeds 4096, then we need to
24273 encourage tiling to prevent attrocious performance.
24275 For example, try downscaling a 2560x1600 background image on a gen3
24276 device using I915_TILING_NONE...
24278 Using slideshow-demo /usr/share/backgrounds/cosmos/whirlpool.jpg, on a
24279 PineView netbook, fps goes from under 4 to over 40.
24281 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24283 commit ae567b783e5af53f17f49dbf58b9be17fcb53737
24284 Author: Chris Wilson <chris@chris-wilson.co.uk>
24285 Date: Sat Jul 2 00:11:21 2011 +0100
24287 sna: Finer-grained debugging for trapezoids
24289 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24291 commit 98f2e3855d70c02b05e2721a70ebce0c17e44283
24292 Author: Chris Wilson <chris@chris-wilson.co.uk>
24293 Date: Sat Jul 2 09:53:11 2011 +0100
24295 sna/video: Downgrade severity of "overlay not found" message
24297 We don't need to warn the user that their hardware does not support the
24298 video overlay plane (but Jesse is working on it!), but merely inform
24299 them that its presence is lacking.
24301 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24303 commit 01c258718e981f4da949a4e48595a0660d55d66e
24304 Author: Chris Wilson <chris@chris-wilson.co.uk>
24305 Date: Fri Jul 1 23:17:54 2011 +0100
24307 sna/gen2: Add missing stub debug files
24309 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24311 commit 5c8a108d2c99f542372efb1734c96ddd2af6bc76
24312 Author: Chris Wilson <chris@chris-wilson.co.uk>
24313 Date: Fri Jul 1 21:24:43 2011 +0100
24315 sna/gen2: Recompute blend pipeline for component-alpha pass
24317 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24319 commit 121511d3bd63da5eca28edf971aae3d062aed46e
24320 Author: Chris Wilson <chris@chris-wilson.co.uk>
24321 Date: Fri Jul 1 16:54:28 2011 +0100
24323 sna/gen2: Pack solid sources into the default diffuse component
24325 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24327 commit a303f85c163979276ff13e00b7c56f55b5fd1d07
24328 Author: Chris Wilson <chris@chris-wilson.co.uk>
24329 Date: Fri Jul 1 16:27:11 2011 +0100
24331 sna/gen2: Remove unused state from invariant setup
24333 ... and also some state that gets clobbered when we install the
24334 composite pipelines.
24336 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24338 commit 120c98ac10435c8e848a8337c1f544f81a05cd3a
24339 Author: Chris Wilson <chris@chris-wilson.co.uk>
24340 Date: Fri Jul 1 14:55:44 2011 +0100
24342 sna: Downsample sources 2x too large to fit in the 3D pipeline
24344 This is quite trivial to hit given the 2k limits on gen2/gen3. We
24345 compromise on image quality by pre-downscaling the source by a fixed
24346 factor to make it fit into the pipeline in preference to performing the
24347 entire operation on the CPU.
24349 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24351 commit f6c8c3bb6fd75bca6c7704b7d5869a5d44ce3832
24352 Author: Chris Wilson <chris@chris-wilson.co.uk>
24353 Date: Fri Jul 1 11:17:32 2011 +0100
24355 sna/gen2: Use specular component for solid spans
24357 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24359 commit de14e3c8595f9e315dc3ce23ad15b04e118499f4
24360 Author: Chris Wilson <chris@chris-wilson.co.uk>
24361 Date: Fri Jul 1 09:58:27 2011 +0100
24363 sna/gen2: Add missing render fallbacks for blt ops
24365 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24367 commit ecbf6bbd27b1205dcf76cfe34ae2a7a3f5ec195a
24368 Author: Chris Wilson <chris@chris-wilson.co.uk>
24369 Date: Fri Jul 1 08:50:58 2011 +0100
24371 sna/gen2: Implement composite-spans
24373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24375 commit c89b37d7b43c9e588097b7fadcba3bc13a03f8bc
24376 Author: Chris Wilson <chris@chris-wilson.co.uk>
24377 Date: Thu Jun 30 17:53:49 2011 +0100
24379 sna: Mappable aperture is region 0 on gen2
24381 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24383 commit c0434ab49035bf278dad6f5f84a541ea58536fb4
24384 Author: Chris Wilson <chris@chris-wilson.co.uk>
24385 Date: Thu Jun 30 16:31:28 2011 +0100
24387 sna: Distinguish 830/845 vs 855/865 using the generation id
24389 Remove the PCI ID device checks by using the simpler check on the
24390 generation id for errata pertaining to 830/845.
24392 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24394 commit 18d08e49d270b7a05f14a309759c9315e5ab9679
24395 Author: Chris Wilson <chris@chris-wilson.co.uk>
24396 Date: Tue Jun 28 20:25:46 2011 +0100
24398 uxa/glyphs: Fallback instead of crashing on large strings
24400 Not ideal, but being slow is a major improvement over losing data.
24402 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36860
24403 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24405 commit 95866bd6bdaeb612732fd43ec1b93e088763086e
24406 Author: Chris Wilson <chris@chris-wilson.co.uk>
24407 Date: Mon Jun 27 21:07:22 2011 +0100
24409 sna/video: Use EXTEND_PAD to avoid mixing in the border color
24411 ...which is 0 and appears green around an unaligned YUV-video.
24413 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38723
24414 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24416 commit 7446b39f04306ae2ffccdfe8dd6c8bda64a27043
24417 Author: Chris Wilson <chris@chris-wilson.co.uk>
24418 Date: Mon Jun 27 15:26:49 2011 +0100
24420 sna: Disable XVideo using the TexturedAdapter if the GPU is wedged
24422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24424 commit 44cd6ebf43aa5b7f23c03629a11cdf83b6e96e3a
24425 Author: Chris Wilson <chris@chris-wilson.co.uk>
24426 Date: Sun Jun 26 18:41:20 2011 +0100
24428 sna: Typo in debug assertion.
24430 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24432 commit 5cd7f5b59857f221bd3aa9965b43b938ff7824ed
24433 Author: Chris Wilson <chris@chris-wilson.co.uk>
24434 Date: Sun Jun 26 18:36:05 2011 +0100
24436 sna/dri: Mark the GPU bo as damaged upon a page-flip
24438 ...so that CPU pixmap is correctly invalidated for the next readback.
24439 For instance, if you were to take a screenshot on a composited destkop.
24441 Reported-by: Sitosfe Wheeler <sitsofe@yahoo.com>
24442 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24444 commit c4f7959266571f14c43c0f0cea69134bb8556875
24445 Author: Chris Wilson <chris@chris-wilson.co.uk>
24446 Date: Sun Jun 26 14:09:11 2011 +0100
24448 sna: Only create bo up to half the size of the mappable aperture
24450 As we use GTT mappings if writing directly into the tiled buffer and the
24451 available aperture is reported by the kernel as the total GTT and not
24452 limited to the fenceable/mappable region, we need to manually probe this
24453 value and ensure that our creation and fenced routines observe this
24456 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24458 commit 130350111720a0e4ae825bb9d21a8c481cd2b548
24459 Author: Chris Wilson <chris@chris-wilson.co.uk>
24460 Date: Sat Jun 25 23:21:20 2011 +0100
24462 sna/dri: Reorganise schedule_swap to not confuse immediate and wait paths
24464 Eeek, the wait-for-target-msc was using the immediate swap path, meaning
24465 that for copy-swaps the copy was submitting immediately but the client
24466 throttled waiting upon the target vblank. What is actually intended is
24467 for the presentation to be delayed until the target_msc.
24469 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24471 commit b460b9645451af84136c5daebbc00c7545de67f4
24472 Author: Chris Wilson <chris@chris-wilson.co.uk>
24473 Date: Sat Jun 25 16:34:15 2011 +0100
24475 sna/dri: Fix composited copy-swaps
24477 The secret is not to cheat and render directly to the front buffer, but
24478 remember to mark the Window as damaged.
24480 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24482 commit 28fffbd1d07890c8b4d697369159f6a30f267675
24483 Author: Chris Wilson <chris@chris-wilson.co.uk>
24484 Date: Sat Jun 25 16:33:31 2011 +0100
24486 sna/display: Protect against drmModeGetCrtc returning NULL
24488 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24490 commit 2c73b4651a04390668c3ed2aa10699663e44fd8a
24491 Author: Chris Wilson <chris@chris-wilson.co.uk>
24492 Date: Sat Jun 25 16:32:30 2011 +0100
24494 sna/gen4+: Use the drawable rectangle offset for copy boxes
24496 Saves a little bit of work whilst emitting the rectangles.
24498 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24500 commit 3833ff967766b0b99f1d636c6453de1783a90586
24501 Author: Chris Wilson <chris@chris-wilson.co.uk>
24502 Date: Sat Jun 25 14:02:50 2011 +0100
24504 sna: Correct typo in computing damage of PolyPoint
24506 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38614
24507 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24509 commit 471115a9806162bf0cafe27947f49d94fdd63d87
24510 Author: Chris Wilson <chris@chris-wilson.co.uk>
24511 Date: Fri Jun 24 13:32:45 2011 +0100
24513 sna: Also allow BLT copies to discard the alpha channel
24515 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24517 commit c3b1a0d7046a83b6daec03e5a562116e3adf3c71
24518 Author: Chris Wilson <chris@chris-wilson.co.uk>
24519 Date: Fri Jun 24 12:19:09 2011 +0100
24521 sna/dri: Copy to real front upon swap
24523 The front-buffer of a DRI2 drawable, may not in fact be pointing to the
24524 scanout pixmap. So override the destination for swapbuffers to update
24527 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24529 commit 669378cd8db8af89d12f1b95a07fc1ead0b1569c
24530 Author: Chris Wilson <chris@chris-wilson.co.uk>
24531 Date: Fri Jun 24 11:53:17 2011 +0100
24533 sna: Wrap the fbcon in a scratch pixmap for render-copy across depth changes
24535 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24537 commit 33b2ea0392fe944b210ef744398aa22989bfdf33
24538 Author: Chris Wilson <chris@chris-wilson.co.uk>
24539 Date: Fri Jun 24 11:43:07 2011 +0100
24541 sna: Avoid using the BLT to copy between mismatching depths
24543 We either conflated bpp (which fails given a mixture of depth-24 and
24544 depth-30 pixmaps) or neglected to check at all.
24546 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24548 commit 28f6dc24d3c50aa399dd098a9ba6a18a95837800
24549 Author: Chris Wilson <chris@chris-wilson.co.uk>
24550 Date: Fri Jun 24 11:24:59 2011 +0100
24552 sna: Support depth-30 and some more logging to show the depth
24554 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24556 commit 312460ea69792594fe9acca8591ead55baea7d99
24557 Author: Xiang, Haihao <haihao.xiang@intel.com>
24558 Date: Thu Jun 23 00:20:03 2011 +0800
24560 Xv: set up pipeline for Xv on Ivybridge
24562 The configuration is same as that on Sandybridge, but
24563 many state commands are changed
24565 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
24567 commit 90a0800588f83d78254136b94bca656c4c24215d
24568 Author: Xiang, Haihao <haihao.xiang@intel.com>
24569 Date: Tue Jun 21 14:14:38 2011 +0800
24571 Xv: upload new shaders to GEM objects for Xv on Ivybridge
24573 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
24575 commit 00614f712b243a7724cf919f94d204a123d2789a
24576 Author: Xiang, Haihao <haihao.xiang@intel.com>
24577 Date: Thu Jun 23 00:09:21 2011 +0800
24579 Xv: update SURFACE_STATE & SAMPLER_STATE for Xv on Ivybridge
24581 SURFACE_STATE & SAMPLER_STATE are changed since Ivybridge
24583 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
24585 commit 70f884772a3adc5dbf56572499c5fb2e080aa84a
24586 Author: Xiang, Haihao <haihao.xiang@intel.com>
24587 Date: Thu Jun 23 00:02:38 2011 +0800
24589 Xv: New shaders for Xv on Ivybridge
24591 Redefine some M4 macros, also update the check for
24592 intel-gen4asm to support Ivybridge
24594 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
24596 commit 3cf423bd3a9483181e59ee87376a9487fa1f923d
24597 Author: Xiang, Haihao <haihao.xiang@intel.com>
24598 Date: Wed Jun 22 23:33:08 2011 +0800
24600 Xv: separate fragments from M4 macros
24602 It is to prepare for Xv on Ivybridge. The difference from Sandybridge
24603 is that all message payload must be in GRF registers instead of MRF registers
24604 on Ivybridge. We will only redefine some M4 macros for Ivybridge
24606 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
24608 commit 58d7a89b93ba4022f45465e479d2799b8903137a
24609 Author: Chris Wilson <chris@chris-wilson.co.uk>
24610 Date: Fri Jun 24 00:35:30 2011 +0100
24612 sna: Don't render to invalid surfaces
24614 Fixes a regression from d0362a. In bypassing the is_wedged checked, we
24615 also ended up bypassing the checks that we could indeed render to the
24616 target bo. With the result that we were creating GPU buffers for SHM
24617 surfaces, something that requires Xserver fixes before we can actually
24620 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24622 commit 3c1d5bc35e02dfd8234c029dcab5661e1656ca8e
24623 Author: Chris Wilson <chris@chris-wilson.co.uk>
24624 Date: Fri Jun 24 00:34:11 2011 +0100
24626 sna: Clip the fbcon to the frontbuffer
24628 ...both to correct the placement of the fbcon into the smaller scanout and
24629 to ensure that we correctly clip the boxes to be copied.
24631 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24633 commit e3b4c10f239f69e5282bde1d3b0c3685b24cf0aa
24634 Author: Chris Wilson <chris@chris-wilson.co.uk>
24635 Date: Thu Jun 23 20:54:28 2011 +0100
24637 sna: Relent and workaround the lack of O_NONBLOCK on /dev/dri/card0
24639 So in order not to block server shutdown, check that the fd is readable
24640 before attempting to read from it.
24642 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24644 commit 33d3077266942b23c5bc05ef4a9a6cebe357272a
24645 Author: Chris Wilson <chris@chris-wilson.co.uk>
24646 Date: Thu Jun 23 15:53:45 2011 +0100
24648 sna: Debug compile fix, and some extra comments
24650 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24652 commit 03e6aee8fd2cc7486fbd2ec55a537c712786486b
24653 Author: Chris Wilson <chris@chris-wilson.co.uk>
24654 Date: Thu Jun 23 15:51:59 2011 +0100
24656 sna: For an unbounded op, we need to convert the whole surface
24658 Otherwise if we fail to check, then we create a 0x0 surface to sample
24659 with the operation -- net result is that we end up using a clear source
24660 instead of the desired mask.
24662 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24664 commit 356ded1308e8fe5f8dbec217b219228a991356d5
24665 Author: Adam Jackson <ajax@redhat.com>
24666 Date: Wed Jun 22 17:17:15 2011 -0400
24668 Don't include xf86Priv.h
24670 I'd like it to not be in the SDK anymore, and we're not using anything
24673 Signed-off-by: Adam Jackson <ajax@redhat.com>
24674 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24676 commit 8fb98e22a63a4354b049ace7870abb6406541c28
24677 Author: Jian Zhao <jian.j.zhao@intel.com>
24678 Date: Wed Jun 22 08:54:59 2011 +0800
24680 configure: Define required_xorg_xserver_version later so it can make successfully.
24682 commit ea71133da78632d4cfee5b0b4c96e8dddd6cdf44
24683 Author: Chris Wilson <chris@chris-wilson.co.uk>
24684 Date: Tue Jun 21 12:28:13 2011 +0100
24686 sna/video: Use pwrite for upload of unclipped, unrotated frames
24688 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24690 commit 2f6afb5b1f02cc448da1b342627108ceddda4f0d
24691 Author: Chris Wilson <chris@chris-wilson.co.uk>
24692 Date: Mon Jun 20 17:29:17 2011 +0100
24694 sna: Only reduce gpu damage after migration
24696 Otherwise we end up doing work for no reason, ala the regression in
24697 grads-heat-map. However, it is important that the damage is reduced at
24698 some point or else it may grow unbounded. Hopefully normal usage will
24699 never hit the pathologocial case...
24701 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24703 commit 88d840b53d64e67cfee14cc4cef8da46d716c401
24704 Author: Chris Wilson <chris@chris-wilson.co.uk>
24705 Date: Mon Jun 20 15:22:57 2011 +0100
24707 sna/gen4: Check for sufficient batch space before emitting pipelined pointers
24709 For the broken shader workaround, we need to reset the pipeline state
24710 before every blt as the state is changed by the magic CA pass. So we
24711 need to check that we actually have sufficient space to emit the
24712 pipelined pointers before doing so.
24714 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24716 commit 84bd6e51199bf8b49e6f6e0c29a1d71c89f6996f
24717 Author: Chris Wilson <chris@chris-wilson.co.uk>
24718 Date: Mon Jun 20 15:22:00 2011 +0100
24720 sna/gen4: Correct the number of dwods required to emit the ca pass
24722 We need 7 for pipelined pointers, 6 for the primitive and 1 for the
24725 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24727 commit 22255335e428cb1f349fe4198d36e0ab3fe06948
24728 Author: Chris Wilson <chris@chris-wilson.co.uk>
24729 Date: Mon Jun 20 13:51:12 2011 +0100
24731 sna: Don't perform a GPU copy of the scanout if it is wedged.
24733 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24735 commit e2684498a4905b4ac157a1c752c1c4f75301e7ad
24736 Author: Chris Wilson <chris@chris-wilson.co.uk>
24737 Date: Mon Jun 20 13:50:43 2011 +0100
24739 sna: Add some debugging asserts for handling wedged GPUs
24741 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24743 commit d0362a20bb3ead2fab9401b63209f917bdd72d55
24744 Author: Chris Wilson <chris@chris-wilson.co.uk>
24745 Date: Sun Jun 19 19:33:01 2011 +0100
24747 sna: Bypass the gpu-is-wedged? check for forced bo creation
24749 In order to attach a scanout (and use swrast) on a wedged GPU, we need to
24750 bypass the safety checks for normal pixmap creation.
24752 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24754 commit 585667c2f9f88554ed89ff21ae38600f761d964c
24755 Author: Chris Wilson <chris@chris-wilson.co.uk>
24756 Date: Sat Jun 18 15:52:22 2011 +0100
24758 sna: Bump the required xserver version to 1.10
24760 SNA requires some pending bug fixes to the xserver so it makes little
24761 sense to conditionalise the code and deliberately cause broken
24764 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=3843
24765 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24767 commit de0286ed9f7814f903d56e488c7a4b3eb0ca56b0
24768 Author: Chris Wilson <chris@chris-wilson.co.uk>
24769 Date: Fri Jun 17 09:40:53 2011 +0100
24771 sna/display: After copying the fbcon, tell the server that we have a background
24773 ... so that the core knows to skip the clear.
24775 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24777 commit f3869d60cc3bc349d1efba642baa011378f1c09b
24778 Author: Chris Wilson <chris@chris-wilson.co.uk>
24779 Date: Fri Jun 17 09:37:15 2011 +0100
24781 sna/display: Apply damage for the fbcon copy
24783 ... so that any immediate shadow usage will read back the fbcon
24786 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24788 commit 39f0b0ae17d00402971bb62555d3807bdf39ee93
24789 Author: Chris Wilson <chris@chris-wilson.co.uk>
24790 Date: Thu Jun 16 23:49:06 2011 +0100
24792 sna: Copy the fbcon contents onto the front buffer upon X startup
24794 This patch has been carried by the distributions every since they
24795 started doing graphical boot splashes. Time to integrate it and give it
24798 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24800 commit 37ba33f502af0878f6cb75f890c9fe288d0f7e41
24801 Author: Chris Wilson <chris@chris-wilson.co.uk>
24802 Date: Fri Jun 17 07:03:25 2011 +0100
24804 sna: Fix kgem_create_from_name error paths and to mark the bo as non-reusable
24806 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24808 commit 665bc1735748f266fad5255f82c40edc59ef98e5
24809 Author: Chris Wilson <chris@chris-wilson.co.uk>
24810 Date: Fri Jun 17 07:00:54 2011 +0100
24812 sna/display: Remove the caching of the drmModeCrtc
24814 We only use it for the id. Everything else stored on it, like the
24815 buffer_id, is not permanent and we need to query the current status as
24818 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24820 commit ad6235cfb11a5dfdd27106ac45106aaacba73bde
24821 Author: Chris Wilson <chris@chris-wilson.co.uk>
24822 Date: Thu Jun 16 19:02:36 2011 +0100
24824 sna/accel: convert BOX_ADD_RECT to an inline function
24826 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24828 commit a86f43988c4ded10cc3b1b964668300c9aa3af59
24829 Author: Chris Wilson <chris@chris-wilson.co.uk>
24830 Date: Thu Jun 16 19:00:09 2011 +0100
24832 sna/accel: convert BOX_ADD_PT to an inline function
24834 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24836 commit f67a26590aadac8b9b9870ae3b03d23eb7eda1a2
24837 Author: Chris Wilson <chris@chris-wilson.co.uk>
24838 Date: Thu Jun 16 18:58:38 2011 +0100
24840 sna/accel: convert TRIM_AND_TRANSLATE_BOX to an inline function
24842 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24844 commit df1b117dedfdda425fbc3e4ee394169902df4716
24845 Author: Chris Wilson <chris@chris-wilson.co.uk>
24846 Date: Thu Jun 16 18:57:38 2011 +0100
24848 sna/accel: convert TRANSLATE_BOX to an inline function
24850 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24852 commit 76ff3c715cf6059316d08217e7e7783ca0388393
24853 Author: Chris Wilson <chris@chris-wilson.co.uk>
24854 Date: Thu Jun 16 18:56:24 2011 +0100
24856 sna/accel: convert TRIM_BOX to an inline
24858 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24860 commit e4a66368e71df288323fce66b27237f1d9f93c3a
24861 Author: Chris Wilson <chris@chris-wilson.co.uk>
24862 Date: Thu Jun 16 18:55:03 2011 +0100
24864 sna/accel: convert CLIP_BOX to an inline function
24866 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24868 commit d7b51cc5d6e624a72b2f2fb92b4488e49a1c749c
24869 Author: Chris Wilson <chris@chris-wilson.co.uk>
24870 Date: Thu Jun 16 18:52:56 2011 +0100
24872 sna/accel: convert BOX_EMPTY macro to an inline function
24874 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24876 commit 1c3c08e80f08adb12a03277a60c4a56115281743
24877 Author: Chris Wilson <chris@chris-wilson.co.uk>
24878 Date: Thu Jun 16 18:37:43 2011 +0100
24880 sna: Fix the typo in BOX_EMPTY
24882 BOX_EMPTY(&box) surprisingly compiles and generates interesting code that
24883 evaluates to if (0). Oops.
24885 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24887 commit 17bf0019b6962c06c87a08e4b41799481eaccd26
24888 Author: Chris Wilson <chris@chris-wilson.co.uk>
24889 Date: Thu Jun 16 14:26:06 2011 +0100
24891 sna: Only retire requests after IO if the bo was active
24893 We only want to process retirements if we know we caused a stall waiting
24894 for the GPU to release the bo before performing IO.
24896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24898 commit b707f18d17564653a2b45adf46a4e8eafcc6ab2c
24899 Author: Chris Wilson <chris@chris-wilson.co.uk>
24900 Date: Thu Jun 16 14:16:06 2011 +0100
24902 sna: Purge the bo cache only if we detect aperture thrashing
24904 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24906 commit 855ced5a4eb67e287001ae22cb6489639c548cfa
24907 Author: Chris Wilson <chris@chris-wilson.co.uk>
24908 Date: Thu Jun 16 13:24:14 2011 +0100
24910 uxa: Remove the attempt to use the 3D pipeline for core rendering
24912 The attempt was still ridden with bugs, such as
24914 http://bugs.freedesktop.org/show_bug.cgi?id=28768
24915 http://bugs.freedesktop.org/show_bug.cgi?id=28798
24916 http://bugs.freedesktop.org/show_bug.cgi?id=28908
24917 http://bugs.freedesktop.org/show_bug.cgi?id=29401
24919 A fresh approach was taken with SNA, but in the mean time before that
24920 can be enabled downstream, restore correct behaviour.
24922 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24924 commit dd4ad11e0d432f3a8eff4574a241c16f8463278d
24925 Author: Chris Wilson <chris@chris-wilson.co.uk>
24926 Date: Thu Jun 16 12:53:22 2011 +0100
24930 A pox on all those who write trivial patches, attempt to compile on a
24931 system not having dri2proto and just assume that all is good.
24933 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24935 commit 2ff36af7e1841a82fd43a9317d07b3229a309df5
24936 Author: Chris Wilson <chris@chris-wilson.co.uk>
24937 Date: Thu Jun 16 11:23:16 2011 +0100
24939 sna/accel: Prevent initialising a region with an invalid box
24941 pixman complains if you attempt to use a region created with an invalid
24942 box, so don't do that.
24944 Reported-by: Reinhard Karcher <reinhard.karcher@gmx.net>
24945 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24947 commit 3c02110052ecf69d43577ed8537a8850227da560
24948 Author: Chris Wilson <chris@chris-wilson.co.uk>
24949 Date: Thu Jun 16 10:30:58 2011 +0100
24951 sna/dri: Add some comments to explain the choice of flags upon the bo
24953 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24955 commit 881f961a1c78e554e566ef624d6edcf28fee8fdf
24956 Author: Chris Wilson <chris@chris-wilson.co.uk>
24957 Date: Wed Jun 15 19:11:05 2011 +0100
24959 sna: Clear the gpu flag once is not busy and no flush is pending
24961 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24963 commit a1ee4b930846d4ba9274028c08800b882fc926f1
24964 Author: Chris Wilson <chris@chris-wilson.co.uk>
24965 Date: Tue Jun 14 23:46:15 2011 +0100
24967 sna/dri: And fix the blit swap paths
24969 Following on the recent successes with the pageflip paths.
24971 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24973 commit 2f675cf402a6d158448a9300779829a10ca991fd
24974 Author: Chris Wilson <chris@chris-wilson.co.uk>
24975 Date: Tue Jun 14 08:13:58 2011 +0100
24977 sna: Split zaphod across the crtcs
24979 Since we have no global resource allocator for zaphod mode, that's what
24980 RandR-1.4 solves, we have to further constrain zaphod mode to only use
24981 one crtc per screen. This also means that you must match the output
24982 restrictions within the Screen definitions, noting that the crtc pipe id
24983 corresponds with the screen number.
24985 Reportede-by: Phillp Haddad <phillip.haddad@gmail.com>
24986 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24988 commit 49f265915a4bca52724c345f04cddfeaca972655
24989 Author: Chris Wilson <chris@chris-wilson.co.uk>
24990 Date: Tue Jun 14 07:59:50 2011 +0100
24992 sna: Compile fix for debugging enabled
24994 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
24996 commit 12b8a62652e219de255fdd8b3c801c23274130e9
24997 Author: Chris Wilson <chris@chris-wilson.co.uk>
24998 Date: Mon Jun 13 20:51:56 2011 +0100
25000 test: Exclude the long running stress tests from make distcheck
25002 tinderbox complains when it is left running for days on end...
25004 Reported-by: Jeremy Huddleston <jeremyhu@apple.com>
25005 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25007 commit 4173ca344bb7d3fb7c03ebdce0e31aa82dba1192
25008 Author: Chris Wilson <chris@chris-wilson.co.uk>
25009 Date: Mon Jun 13 22:31:55 2011 +0100
25011 sna: Invalidate the mode if the front pixmap was swapped whilst blanked
25013 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25015 commit 2158534421f26bcd1eca9ad74ab7a30f27e47a75
25016 Author: Chris Wilson <chris@chris-wilson.co.uk>
25017 Date: Mon Jun 13 15:27:23 2011 +0100
25019 sna/dri: Accurately track front and pending front for async flips
25021 By not tracking the front buffer correctly, i.e. performing the exchange
25022 on every swap, GL_FRONT was no longer pointing at the updated buffer and
25023 neither was the root pixmap. So both X and GL would read the wrong
25024 buffer was the flip was pending.
25026 The other issue was that we would feed the old front buffer back to the
25027 application as a future back buffer (due to buffer caching) and so the
25028 kernel would duly insert a WAIT_EVENT for the pending flip to complete
25029 before allowing rendering to affect it.
25031 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25033 commit 2facaa910798c0e7adb34345e6d6913849b7d8a3
25034 Author: Chris Wilson <chris@chris-wilson.co.uk>
25035 Date: Mon Jun 13 11:46:04 2011 +0100
25037 sna/dri: Complete the pending async flip even after DrawableGone
25039 If the client exits before the async page-flip completes then we still
25040 need to decrement the pending flip count in order to correctly flip the
25043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25045 commit 86888723901030c5cb9da16cbb03508d4b13332f
25046 Author: Chris Wilson <chris@chris-wilson.co.uk>
25047 Date: Sun Jun 12 22:38:01 2011 +0100
25049 sna: Just do a pointer exchange when flipping with no scanout
25051 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25053 commit b881d1b16a8beec024ac18eca8308418e34cb05d
25054 Author: Chris Wilson <chris@chris-wilson.co.uk>
25055 Date: Sun Jun 12 20:13:26 2011 +0100
25057 sna: Check that the scanout is still attached before waiting for scanline
25059 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25061 commit bb08e73c557c379ef131225ffea5aa729ca781d6
25062 Author: Chris Wilson <chris@chris-wilson.co.uk>
25063 Date: Sun Jun 12 20:07:20 2011 +0100
25065 sna: Prevent tracking scratch pixmaps on older xservers
25067 Before the xserver is patched to send notification of release of the
25068 scratch pixmaps, we cannot associated GPU resources with them. In
25069 practice, the only advantage for doing so is for SHM pixmaps...
25071 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25073 commit ff262aca9c46b7616c59d8e8f04d33a5bbc6e324
25074 Author: Chris Wilson <chris@chris-wilson.co.uk>
25075 Date: Sat Jun 11 10:15:14 2011 +0100
25077 sna/dri: Fix tripple-buffering for vblank_mode=0
25079 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25081 commit b199bc2b415fd0f06e456e0b1f44dd25ba8dbe02
25082 Author: Chris Wilson <chris@chris-wilson.co.uk>
25083 Date: Sat Jun 11 09:57:46 2011 +0100
25085 sna/dri: Skip querying current msc for immediate swaps
25087 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25089 commit 5f4dbcc8bde3690668bde739ac61178037612475
25090 Author: Chris Wilson <chris@chris-wilson.co.uk>
25091 Date: Sat Jun 11 09:06:50 2011 +0100
25093 sna/dri: Queue page-flips immediately for next vblank
25095 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25097 commit 00e8b4bec6dc5bb46e6f073af9c1fa34f4a2c2c9
25098 Author: Chris Wilson <chris@chris-wilson.co.uk>
25099 Date: Sat Jun 11 09:00:16 2011 +0100
25101 sna/dri: For offscreen, swaps just exchange the buffers
25103 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25105 commit e9397b28eff5be4cd5e8a2ceca80333424f24391
25106 Author: Chris Wilson <chris@chris-wilson.co.uk>
25107 Date: Sat Jun 11 08:54:04 2011 +0100
25109 sna: Process dri wakeups first
25111 The goal is to minimise the latency in receiving the event from the
25112 kernel and acting upon it.
25114 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25116 commit bcad5b21fef4573f3144608364cf5b1ea6d241e3
25117 Author: Chris Wilson <chris@chris-wilson.co.uk>
25118 Date: Fri Jun 10 21:27:38 2011 +0100
25120 sna: Unbreak configure after last commit
25122 I went a step too far... I still need some define in order to switch
25123 between uxa/sna at compile time.
25125 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25127 commit d0d65940b4f4d4993eccb8c66ec7e8b633b331da
25128 Author: Chris Wilson <chris@chris-wilson.co.uk>
25129 Date: Fri Jun 10 19:51:12 2011 +0100
25131 sna: Remove the ability to disable chipset specific code
25133 This was a fun little, but pointless, exercise.
25135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25137 commit 6f035e80b98e478ff4c59427e50e57ecd710da62
25138 Author: Chris Wilson <chris@chris-wilson.co.uk>
25139 Date: Fri Jun 10 18:41:56 2011 +0100
25141 sna/dri: Keep reference on swap buffers until the end of the pageflip
25143 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25145 commit cf6c7901825bb831443615ef21555c5843184bdb
25146 Author: Chris Wilson <chris@chris-wilson.co.uk>
25147 Date: Fri Jun 10 17:49:13 2011 +0100
25149 sna: Use the ScreenPixmap->serialNumber as a generation count
25151 DRI2 presumes that the pixmap->serialNumber can be used as unique id.
25152 If it changes revokes *all* the buffers, it presumes a new pixmap has
25153 been attached to the window, for example after a reconfiguration event
25154 (resizing of a window, or a mode switch). However, as we updated the
25155 root pixmap upon a pageflip, we were triggering revocations everytime,
25156 causing further revocations and massive aperture thrashing.
25158 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25160 commit 3f19270d6795738b58d9ce1721896b3067b38c9f
25161 Author: Chris Wilson <chris@chris-wilson.co.uk>
25162 Date: Fri Jun 10 15:29:57 2011 +0100
25164 sna/dri: Actually restrict pageflipping to properly aligned fs windows
25166 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25168 commit 9e9c628708ae838e5df3df89331243f84b297e8e
25169 Author: Chris Wilson <chris@chris-wilson.co.uk>
25170 Date: Fri Jun 10 15:17:18 2011 +0100
25172 sna/dri: Create pixmap suitable for the framebuffer for fullscreen buffers
25174 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25176 commit b22de0b01453884d5018e7e0faef439421fa63a0
25177 Author: Chris Wilson <chris@chris-wilson.co.uk>
25178 Date: Fri Jun 10 13:30:54 2011 +0100
25180 sna/blt: Suppress overwriting boxes through fill_boxes()
25182 As the introduction of the unrolled fill_boxes() was bypassing the
25183 effective optimisation in fill_one().
25185 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25187 commit ebeebc2044610e6e31d9cef746c768d2e435f32c
25188 Author: Chris Wilson <chris@chris-wilson.co.uk>
25189 Date: Fri Jun 10 13:07:13 2011 +0100
25191 sna: Use temporary for storing the current crtc box when computing best crtc
25193 ... as the caller may be reusing an input parameter for the result.
25195 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25197 commit b3949d0d7424242703c9e60e2e33908499aa6a43
25198 Author: Chris Wilson <chris@chris-wilson.co.uk>
25199 Date: Fri Jun 10 12:45:24 2011 +0100
25201 sna/kgem: remove erroneous assert during submit
25203 This was only valid for gen6+ with the multiple rings.
25205 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25207 commit 33abb68da2c49081f28898f93cc4e14702b40d70
25208 Author: Chris Wilson <chris@chris-wilson.co.uk>
25209 Date: Fri Jun 10 12:44:56 2011 +0100
25211 sna/kgem: remove redundant clears upon initialisation
25213 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25215 commit cfde920a2e449efaa387850173e46f87d928023b
25216 Author: Chris Wilson <chris@chris-wilson.co.uk>
25217 Date: Fri Jun 10 11:27:04 2011 +0100
25219 sna: Remember to decouple the fb on closing
25221 ... so that we actually attach a new one after regen!
25223 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25225 commit b6192ce95af8aa8be6d9a68a500393f2a73c2f4c
25226 Author: Chris Wilson <chris@chris-wilson.co.uk>
25227 Date: Fri Jun 10 11:05:41 2011 +0100
25229 sna/gen4: Tweak calculation of required batch space for CA
25231 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25233 commit c9b89fc87f1d43a715cd9c5915bfed957175986d
25234 Author: Chris Wilson <chris@chris-wilson.co.uk>
25235 Date: Fri Jun 10 10:53:29 2011 +0100
25237 sna: Drain the event queues before CloseScreen
25239 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25241 commit ea560e25dfcce3750339b477a1a3c05f504cf91e
25242 Author: Chris Wilson <chris@chris-wilson.co.uk>
25243 Date: Fri Jun 10 10:44:27 2011 +0100
25245 sna: Workaround the random free of the ScreenPixmap in fbCloseScreen
25247 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25249 commit 1f364c6d241e8971e8fc1447871c718240e4563e
25250 Author: Chris Wilson <chris@chris-wilson.co.uk>
25251 Date: Fri Jun 10 10:03:01 2011 +0100
25253 sna: Reset the kgem state on server regen
25255 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25257 commit 95ae39a93467655b2107b82bd7970f732c8de066
25258 Author: Chris Wilson <chris@chris-wilson.co.uk>
25259 Date: Fri Jun 10 09:57:22 2011 +0100
25261 sna: Add some debug statements to the core DDX interface
25263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25265 commit 0594724dc4e9e459240b8b290ddf42ed711be22b
25266 Author: Chris Wilson <chris@chris-wilson.co.uk>
25267 Date: Fri Jun 10 08:18:07 2011 +0100
25269 sna/gen4: Restore normal state after CA pass and FLUSH_EVERY_VERTEX
25271 By working around the broken shaders, we emitted the CA rectangle in the
25272 middle of a sequence of glyphs and left the state setup for CA. So we
25273 need to reset the pipeline state at the start of every composite blt.
25275 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25277 commit 2dff681efe1b89ffe59e5d33979c6e059b72a786
25278 Author: Chris Wilson <chris@chris-wilson.co.uk>
25279 Date: Fri Jun 10 08:23:19 2011 +0100
25281 sna/blt: Remove a couple of unused members from its op-state
25283 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25285 commit fbbd1c9dde3b2fafd8aaebb88daf3c546d86e5a4
25286 Author: Chris Wilson <chris@chris-wilson.co.uk>
25287 Date: Fri Jun 10 08:22:00 2011 +0100
25289 sna/glyphs: Fix glyphs to dst fallback path
25291 commit 3f46c34d534e389f541b5c7987b78981c556d868
25292 Author: Chris Wilson <chris@chris-wilson.co.uk>
25293 Date: Fri Jun 10 08:20:12 2011 +0100
25295 sna/composite: Tweak setup for the minimal composite extents
25297 Clip the operation extents before passing to the backend to setup the
25298 operation, so that we only need to upload the minimum amount of data
25299 necessary, or avoid tiling etc.
25301 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25303 commit a34caf1e3ded03dca53bbea3432dcd45101460fd
25304 Author: Chris Wilson <chris@chris-wilson.co.uk>
25305 Date: Thu Jun 9 20:45:18 2011 +0100
25307 sna/gradient: Check solid busy status not current domain
25309 We can write into the bo if it has no outstanding requests, whereas we
25310 were checking to see it was last on the gpu instead.
25312 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25314 commit 1638a0a6c0b335dd31def93a9199f7407ae51170
25315 Author: Chris Wilson <chris@chris-wilson.co.uk>
25316 Date: Thu Jun 9 16:00:30 2011 +0100
25318 sna/dri: Add a comment to warn about a potential race between DRI clients
25320 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25322 commit 4b54c89fbddc7aa72e5f2c5eded44e84bce70b77
25323 Author: Chris Wilson <chris@chris-wilson.co.uk>
25324 Date: Thu Jun 9 15:53:11 2011 +0100
25326 sna: Place the pixmap on the DRI dirty list after damaging the shadow copy
25328 In order that we remember to flush the contents back to the GPU
25329 before we wake up the DRI clients, we need to add those dirty
25330 pixmaps to the flushing list. I caught the obvious place, the
25331 central move-to-cpu, but I missed a couple of paths were we
25332 optimise the copy onto the shadow.
25334 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25336 commit 190a6786563aadd69f63bb7b509f3edf712e81ef
25337 Author: Chris Wilson <chris@chris-wilson.co.uk>
25338 Date: Thu Jun 9 15:13:17 2011 +0100
25340 sna: Periodically reduce the gpu damage (after migrating any cpu damage)
25342 This will prevent the gpu damage list from growing unbounded.
25344 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25346 commit 38d084ee6241527193ba08c6600b6734e1833118
25347 Author: Chris Wilson <chris@chris-wilson.co.uk>
25348 Date: Thu Jun 9 14:53:34 2011 +0100
25350 sna: Move !kgem->wedged check down into kgem_can_create_2d
25352 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25354 commit d1a3bb1467429310f5d17bf6afbf2eeb043ae64d
25355 Author: Chris Wilson <chris@chris-wilson.co.uk>
25356 Date: Thu Jun 9 14:51:49 2011 +0100
25358 sna: Prefer inactive buffers when creating a bo for a dirty pixmap
25360 As we will immediately attempt to replace it with an inactive when
25361 moving the data to the GPU, short-circuit that replacement.
25363 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25365 commit 8ee53cc7449de7df294f49d089ff1d6b2e944323
25366 Author: Chris Wilson <chris@chris-wilson.co.uk>
25367 Date: Thu Jun 9 13:05:19 2011 +0100
25369 sna: Debugging compile fix
25371 Still not enough coffee this morning.
25373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25375 commit 6f59d7d19053fed4e9d6402a5f03bc8a62cb56af
25376 Author: Chris Wilson <chris@chris-wilson.co.uk>
25377 Date: Thu Jun 9 12:59:59 2011 +0100
25379 sna: Use --enable-debug=full to enable extremely verbose debugging
25381 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25383 commit 021209d5d3add8b28143611cfad4c5481a2945a3
25384 Author: Chris Wilson <chris@chris-wilson.co.uk>
25385 Date: Thu Jun 9 10:27:12 2011 +0100
25387 sna: Remove the stubs from sna_render.c
25389 These only existed to work around an include order problem, when kgem
25390 was intended to be entirely separable from sna. Moving the function
25391 pointer into kgem simplifies matters.
25393 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25395 commit 6141b1aea159759e7e9dcf2561deb9d8c02bd0a2
25396 Author: Chris Wilson <chris@chris-wilson.co.uk>
25397 Date: Thu Jun 9 10:06:08 2011 +0100
25399 sna: Warn after detecting a hung GPU
25401 ...and include the instructional error message from uxa.
25403 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25405 commit 1786d2ee03a190ebe242ac7b58bed7cc31d66b16
25406 Author: Chris Wilson <chris@chris-wilson.co.uk>
25407 Date: Thu Jun 9 09:48:57 2011 +0100
25409 sna/accel: Add debug messages for falling back due to a wedged GPU
25411 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25413 commit 4d509d501b09f565fea232947e6f53d54a08749f
25414 Author: Chris Wilson <chris@chris-wilson.co.uk>
25415 Date: Thu Jun 9 01:03:16 2011 +0100
25417 sna: Tweak retiring old bo
25419 As we check for retirement everytime we wakeup, it is seldom useful to
25420 check again until we know we have invoked an operation that may block.
25421 But when we do check, we do not want to scan the entire active list
25422 looking for flushing candidates, so track those on a separate list.
25424 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25426 commit 17be5e2eb41b2ada94954b87c855961003c2fc7c
25427 Author: Chris Wilson <chris@chris-wilson.co.uk>
25428 Date: Thu Jun 9 00:38:29 2011 +0100
25430 sna: Reduce the frequency of the timer interrupts
25432 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25434 commit d0b09f10242e09adfddbbe6ee1ea39dbf5d1b32d
25435 Author: Chris Wilson <chris@chris-wilson.co.uk>
25436 Date: Wed Jun 8 20:34:57 2011 +0100
25438 sna: Make the device fd non-blocking
25440 So that we can simply query it from each of the Zaphod instances without
25441 blocking. Requires a fixed kernel...
25443 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25445 commit 67ed14c5369cfd967f90e5a3e6053c95a48dd16e
25446 Author: Chris Wilson <chris@chris-wilson.co.uk>
25447 Date: Wed Jun 8 19:42:02 2011 +0100
25449 sna/dri: Set bo=NULL to actually detect unknown dri attachment points.
25451 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25453 commit 68755bc5ff830dc5ed6a34b18a1217b7d1003ce4
25454 Author: Chad Versace <chad@chad-versace.us>
25455 Date: Wed Jun 8 18:34:02 2011 +0100
25457 sna/dri: Add support for DRI2BufferStencil and DRI2BufferHiz
25459 Based on Chad's commits 2662db5b and 3daba7e6:
25461 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25463 commit 0030c3aab6172cdb83ca8f1b7b07c7db5915990d
25464 Author: Chris Wilson <chris@chris-wilson.co.uk>
25465 Date: Wed Jun 8 18:31:12 2011 +0100
25467 sna/dri: Drop support for old dri2 versions with incompatible ABI
25469 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25471 commit 715d466ad44e82b740f5454c41db944863420596
25472 Author: Chris Wilson <chris@chris-wilson.co.uk>
25473 Date: Wed Jun 8 13:32:59 2011 +0100
25475 sna/dri: valgrindify
25477 Lots of scary warnings found by valgrind.
25479 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25481 commit a62db5b050dee10246c02c72385358acb5e72b56
25482 Author: Chad Versace <chad@chad-versace.us>
25483 Date: Mon May 23 14:23:54 2011 -0700
25485 dri: Add support for DRI2BufferStencil and DRI2BufferHiz
25487 And bump configure.ac to require dri2proto >= 2.6, because
25488 DRI2BufferStencil and DRI2BufferHiz were introduced in that version.
25490 When a client requests DRI2BufferHiz or DRI2BufferStencil,
25491 I830DRI2CreateBuffer() now returns a Y-tiled buffer. The stencil buffer is
25492 handled as a special case due its quirky pitch requirements.
25494 CC: Eric Anholt <eric@anholt.net>
25495 CC: Ian Romanick <idr@freedesktop.org>
25496 CC: Kristian Høgsberg <krh@bitplanet.net
25497 Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
25498 Signed-off-by: Chad Versace <chad@chad-versace.us>
25500 commit 263daba7e6afb37bd471fdc8dd8e4458da0a45ff
25501 Author: Chad Versace <chad@chad-versace.us>
25502 Date: Sun Jun 5 19:48:19 2011 -0700
25504 dri: Do not create DRI2 buffers for unrecognized DRI2 buffer tokens
25506 Before this commit, if a client were to request an unrecognized DRI2
25507 buffer, such as DRI2BufferStencil, then I830DRI2CreateBuffer() allocated
25508 and returned an X-tiled buffer by accident. The problem was that
25509 unrecognized tokens were caught by the default case of a switch statement.
25511 Now, when given unrecognized DRI2 tokens, I830DRI2CreateBuffers() returns
25514 This shouldn't break older Mesa versions, because they never query (via
25515 DRI2GetBuffersWithFormat) for the drawable's DRI2BufferStencil.
25517 CC: Eric Anholt <eric@anholt.net>
25518 CC: Ian Romanick <idr@freedesktop.org>
25519 CC: Kenneth Graunke <kenneth@whitecape.org>
25520 CC: Kristian Høgsberg <krh@bitplanet.net
25521 Signed-off-by: Chad Versace <chad@chad-versace.us>
25523 commit 038f29c259c1463c0bff12f5f8f0b228c02e45b3
25524 Author: Chris Wilson <chris@chris-wilson.co.uk>
25525 Date: Tue Jun 7 19:16:01 2011 +0100
25527 sna/blt: Flush batch before overwrite surface data
25529 A left-over from before the surface was embedded into the tail of the
25530 batch, we were only checking for room against the total size of the
25531 batch buffer. So under the wrong set of circumstances we ended up
25532 overwriting surface data with batch and triggering a GPU hang on gen4+.
25534 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25536 commit 265d94e0aa46b30a3198893544dd3619cc9145de
25537 Author: Chris Wilson <chris@chris-wilson.co.uk>
25538 Date: Tue Jun 7 14:23:51 2011 +0100
25540 sna: Add zaphod support
25542 Zaphod support is a rudimentary method for creating an Xserver with
25543 multiple screens from a single device. The Device is instantiated, with
25544 a duplication of its resources, as many as required up to a maximum of
25545 the number of its outputs, and each instance is attached to a Screen
25546 and added to the ServerLayout. A Device can be bound to a selection of
25547 outputs using a comma separated list of RandR names.
25549 Note: in general, this is not the preferred solution! And will be
25550 superseded by per-crtc-pixmaps in RandR-1.4.
25552 For example, the following xorg.conf fragment creates an XServer with
25553 two screens, one attached to the LVDS panel on the laptop, and the other
25554 to any external output:
25557 Identifier "Intel0"
25560 Option "ZaphodHeads" "LVDS1"
25565 Identifier "Intel1"
25568 Option "ZaphodHeads" "DVI1,VGA1"
25573 Identifier "Screen0"
25578 Identifier "Screen1"
25582 Section "ServerLayout"
25583 Identifier "default"
25588 Based on a patch by Ben Skegs <bskeggs@redhat.com>
25590 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25592 commit ad5ead8257b7dda84a886461873d87831d43e663
25593 Author: Chris Wilson <chris@chris-wilson.co.uk>
25594 Date: Tue Jun 7 11:06:52 2011 +0100
25596 sna/gen2: Support covered xrgb sources on 830/845
25598 830/845 cannot directly sample from an x8r8g8b8 source, but if we know
25599 that we are only sampling from within the confines of the source then we
25600 force the alpha channel to one. (Outside of the source we require the
25601 sampler to return a==0.)
25603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25605 commit 790f90a277ff2a6fbb615b8cd3c06faa0d549c41
25606 Author: Chris Wilson <chris@chris-wilson.co.uk>
25607 Date: Tue Jun 7 08:18:30 2011 +0100
25609 sna/gen6: Initialise a couple more composite op members for copy_boxes
25611 Valgrind detected that I missed initialised a couple of fields for
25612 use with the generic state emission paths:
25614 ==28683== Conditional jump or move depends on uninitialised value(s)
25615 ==28683== at 0x83BE646: gen6_get_blend (gen6_render.c:251)
25616 ==28683== by 0x83BF769: gen6_emit_state (gen6_render.c:818)
25617 ==28683== by 0x83C38ED: gen6_emit_copy_state (gen6_render.c:2280)
25618 ==28683== by 0x83C3C89: gen6_render_copy_boxes (gen6_render.c:2356)
25619 ==28683== Conditional jump or move depends on uninitialised value(s)
25620 ==28683== at 0x83C15C3: gen6_rectangle_begin (gen6_render.c:1458)
25621 ==28683== by 0x83C177D: gen6_get_rectangles (gen6_render.c:1502)
25622 ==28683== by 0x83C3D16: gen6_render_copy_boxes (gen6_render.c:2363)
25624 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25626 commit 4e443cbef589e016d085e8023677f47c1f01a858
25627 Author: Chris Wilson <chris@chris-wilson.co.uk>
25628 Date: Tue Jun 7 00:24:17 2011 +0100
25630 sna: Add some more debug commentary to render picture source migration
25632 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25634 commit 367298c5109b47fa4961a60fa3cb454c800f02c3
25635 Author: Chris Wilson <chris@chris-wilson.co.uk>
25636 Date: Mon Jun 6 23:21:09 2011 +0100
25638 sna: Subtract the Solid RenderFillRectangles from CPU damage
25640 ... and so avoid having to move it the GPU, as seen in the wild. It
25641 looks like I will actually need to handle mixed Render/Core operations
25642 on the frontbuffer.
25644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25646 commit 62e4266b4dc9b5c3d165aca980738fa777a0f259
25647 Author: Chris Wilson <chris@chris-wilson.co.uk>
25648 Date: Mon Jun 6 16:20:38 2011 +0100
25650 sna/gen5: Only emit the non-pipelined op after BLT commands
25652 We were always terminating the batch with the non-pipelined op, and not
25653 just at the end of a BLT sequence.
25655 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25657 commit 8f97157d2e807d6c9acb521ec59757f419e6518b
25658 Author: Chris Wilson <chris@chris-wilson.co.uk>
25659 Date: Mon Jun 6 10:21:17 2011 +0100
25661 sna/gen2: Replicate alpha for non-CA masks
25663 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25665 commit c8a2fa3a2e3f22e12f363e8ea060c70585156190
25666 Author: Chris Wilson <chris@chris-wilson.co.uk>
25667 Date: Mon Jun 6 08:43:34 2011 +0100
25669 sna/gen2: Correct command length for CA LOAD_IMMEDIATE_STATE_1
25671 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25673 commit a89fc7181b6ccec8ccca2cd4ccab4f180f663656
25674 Author: Chris Wilson <chris@chris-wilson.co.uk>
25675 Date: Mon Jun 6 08:39:17 2011 +0100
25677 sna/gen2: Only emit the mask texcoord if there is a mask
25679 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25681 commit 4fb7784e1e1b2b999c0c2420583e9c48eb1127ac
25682 Author: Chris Wilson <chris@chris-wilson.co.uk>
25683 Date: Mon Jun 6 08:27:17 2011 +0100
25685 sna/gen3: Non-affine texcoords require space for 4 floats not 3.
25687 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25689 commit d9344ab8d04fe4795a3a532c5c0e85dc11ecf968
25690 Author: Chris Wilson <chris@chris-wilson.co.uk>
25691 Date: Mon Jun 6 08:25:53 2011 +0100
25693 sna/gen2: Set op->floats_per_vertex
25695 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25697 commit 741c1101f17bfa379a15178e4760211737fa18cf
25698 Author: Chris Wilson <chris@chris-wilson.co.uk>
25699 Date: Mon Jun 6 08:22:45 2011 +0100
25701 sna/gen2: Enable selection of gen2 only
25703 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25705 commit c76ec69660bf40de052159f60c16de8fd87341a8
25706 Author: Chris Wilson <chris@chris-wilson.co.uk>
25707 Date: Mon Jun 6 08:02:09 2011 +0100
25709 sna/gen2: The inline primitive takes a length, not a vertex count
25711 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25713 commit 91424d49373d8117e71441fa3221bd1801171692
25714 Author: Eric Anholt <eric@anholt.net>
25715 Date: Tue May 31 23:29:43 2011 -0700
25717 uxa: Simplify uxa_poly_fill_rect by only clipping once.
25719 Reviewed-by: Keith Packard <keithp@keithp.com>
25721 commit e0066e77e026b0dd0daa0c3765473c7d63aa6753
25722 Author: Eric Anholt <eric@anholt.net>
25723 Date: Tue May 31 23:17:16 2011 -0700
25725 uxa: Simplify Composite solid acceleration for spans by only clipping once.
25727 Unlike the previous commit removing this style of code, the code in
25728 this one was originally wrong, and would fail to clip in the second
25729 pass of clipping when y was > pbox->y2.
25731 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=37233
25732 Reviewed-by: Keith Packard <keithp@keithp.com>
25734 commit ace324e4aa27effdd621156eec03f3f87b610732
25735 Author: Eric Anholt <eric@anholt.net>
25736 Date: Tue May 31 23:13:18 2011 -0700
25738 uxa: Simplify BLT solid acceleration for spans filling by only clipping once.
25740 We were clipping each span against the bounds of the clip, throwing
25741 out the span early if it was all clipped, and then walked the clip box
25742 clipping against each of the cliprects. We would expect spans to
25743 typically be clipped against one box, and not thrown out, so we were
25744 not saving any work there. For multiple cliprects, we were adding
25745 work. Only for many spans clipped entirely out of a complicated clip
25746 region would it have saved work, and it clearly didn't save bugs as
25747 evidenced by the many fix attempts here.
25749 Reviewed-by: Keith Packard <keithp@keithp.com>
25751 commit bdb396a44bf64f823a44aafee19b5ec692f1bf5f
25752 Author: Chris Wilson <chris@chris-wilson.co.uk>
25753 Date: Sun Jun 5 19:41:25 2011 +0100
25755 sna: PutImage: Copy straight to GTT if the bo is idle
25757 This saves a copy in the typical PutImage to frontbuffer favoured by
25758 flash. And we also happen to fix a bug if we should be requested to
25759 PutImage outside of the clip region...
25761 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25763 commit 407257570fea1c6456fda47c9fab36c39f169c2b
25764 Author: Chris Wilson <chris@chris-wilson.co.uk>
25765 Date: Sun Jun 5 15:33:27 2011 +0100
25767 sna/gen6: Flush the pipeline before effecting a change of blend modes
25769 ... also make sure that we flush if we change the blend mode for the CA pass.
25771 Reported-by: Ivan Bulatovic <combuster@archlinux.us>
25772 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=37946
25773 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25775 commit 73167711226f430cc9ef4c27f267655d0edf4622
25776 Author: Chris Wilson <chris@chris-wilson.co.uk>
25777 Date: Sun Jun 5 14:56:32 2011 +0100
25779 sna: 915gm does not have 128-byte Y-tiling
25781 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25783 commit 0260c4ce32fa4d7ae8c6ee79d5488b0114e96251
25784 Author: Chris Wilson <chris@chris-wilson.co.uk>
25785 Date: Sun Jun 5 14:39:34 2011 +0100
25787 sna: Fallback if presented with mask under NO_COMPOSITE
25789 Just making sure that the debug paths actually work...
25791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25793 commit fcbe2d9ee79c936cad3ee86836286dcb280f0047
25794 Author: Chris Wilson <chris@chris-wilson.co.uk>
25795 Date: Sat Jun 4 18:11:54 2011 +0100
25797 sna/gen4: Flush every vertex for the magic CA pass
25799 gen4 dies hard if it has two rectangles in the pipeline, and despite the
25800 stringent and crippling efforts to prevent us from efficiently using the
25801 GPU, I missed a flush before submitting the CA rectangle.
25803 Reported-and-tested-by: Fryderyk Dziarmagowski <fdziarmagowski@gmail.com>
25804 References: https://bugs.freedesktop.org/show_bug.cgi?id=28768
25805 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25807 commit bcef98af561939aa48d9236b2dfa2c5626adf4cb
25808 Author: Chris Wilson <chris@chris-wilson.co.uk>
25809 Date: Fri Apr 8 07:17:14 2011 +0100
25811 sna: Introduce a new acceleration model.
25813 The premise is that switching between rings (i.e. the BLT and
25814 RENDER rings) on SandyBridge imposes a large latency overhead whilst
25815 rendering. The cause is that in order to switch rings, we need to split
25816 the batch earlier than is desired and to add serialisation between the
25817 rings. Both of which incur large overhead.
25819 By switching to using a pure 3D blit engine (ok, not so pure as the BLT
25820 engine still has uses for the core drawing model which can not be easily
25821 represented without a combinatorial explosion of shaders) we can take
25822 advantage of additional efficiencies, such as relative relocations, that
25823 have been incorporated into recent hardware advances. However, even
25824 older hardware performs better from avoiding the implicit context
25825 switches and from the batching efficiency of the 3D pipeline...
25827 But this is X, and PolyGlyphBlt still exists and remains in use. So for
25828 the operations that are not worth accelerating in hardware, we introduce a
25829 shadow buffer mechanism through out and reintroduce pixmap migration.
25830 Doing this efficiently is the cornerstone of ensuring that we do exploit
25831 the increased potential of recent hardware for running old applications and
25832 environments (i.e. so that the latest and greatest chip is actually faster
25835 For the curious, sna is SandyBridge's New Acceleration. If you are
25836 running older chipsets and welcome the performance increase offered by
25837 this patch, then you may choose to call it Snazzy instead.
25841 gen3 firefox-fishtank 1203584.56 (1203842.75 0.01%) -> 85561.71 (125146.44 14.87%): 14.07x speedup
25842 gen5 grads-heat-map 3385.42 (3489.73 1.44%) -> 350.29 (350.75 0.18%): 9.66x speedup
25843 gen3 xfce4-terminal-a1 4179.02 (4180.09 0.06%) -> 503.90 (531.88 4.48%): 8.29x speedup
25844 gen4 grads-heat-map 2458.66 (2826.34 4.64%) -> 348.82 (349.20 0.29%): 7.05x speedup
25845 gen3 grads-heat-map 1443.33 (1445.32 0.09%) -> 298.55 (298.76 0.05%): 4.83x speedup
25846 gen3 swfdec-youtube 3836.14 (3894.14 0.95%) -> 889.84 (979.56 5.99%): 4.31x speedup
25847 gen6 grads-heat-map 742.11 (744.44 0.15%) -> 172.51 (172.93 0.20%): 4.30x speedup
25848 gen3 firefox-talos-svg 71740.44 (72370.13 0.59%) -> 21959.29 (21995.09 0.68%): 3.27x speedup
25849 gen5 gvim 8045.51 (8071.47 0.17%) -> 2589.38 (3246.78 10.74%): 3.11x speedup
25850 gen6 poppler 3800.78 (3817.92 0.24%) -> 1227.36 (1230.12 0.30%): 3.10x speedup
25851 gen6 gnome-terminal-vim 9106.84 (9111.56 0.03%) -> 3459.49 (3478.52 0.25%): 2.63x speedup
25852 gen5 midori-zoomed 9564.53 (9586.58 0.17%) -> 3677.73 (3837.02 2.02%): 2.60x speedup
25853 gen5 gnome-terminal-vim 38167.25 (38215.82 0.08%) -> 14901.09 (14902.28 0.01%): 2.56x speedup
25854 gen5 poppler 13575.66 (13605.04 0.16%) -> 5554.27 (5555.84 0.01%): 2.44x speedup
25855 gen5 swfdec-giant-steps 8941.61 (8988.72 0.52%) -> 3851.98 (3871.01 0.93%): 2.32x speedup
25856 gen5 xfce4-terminal-a1 18956.60 (18986.90 0.07%) -> 8362.75 (8365.70 0.01%): 2.27x speedup
25857 gen5 firefox-fishtank 88750.31 (88858.23 0.14%) -> 39164.57 (39835.54 0.80%): 2.27x speedup
25858 gen3 midori-zoomed 2392.13 (2397.82 0.14%) -> 1109.96 (1303.10 30.35%): 2.16x speedup
25859 gen6 gvim 2510.34 (2513.34 0.20%) -> 1200.76 (1204.30 0.22%): 2.09x speedup
25860 gen5 firefox-planet-gnome 40478.16 (40565.68 0.09%) -> 19606.22 (19648.79 0.16%): 2.06x speedup
25861 gen5 gnome-system-monitor 10344.47 (10385.62 0.29%) -> 5136.69 (5256.85 1.15%): 2.01x speedup
25862 gen3 poppler 2595.23 (2603.10 0.17%) -> 1297.56 (1302.42 0.61%): 2.00x speedup
25863 gen6 firefox-talos-gfx 7184.03 (7194.97 0.13%) -> 3806.31 (3811.66 0.06%): 1.89x speedup
25864 gen5 evolution 8739.25 (8766.12 0.27%) -> 4817.54 (5050.96 1.54%): 1.81x speedup
25865 gen3 evolution 1684.06 (1696.88 0.35%) -> 1004.99 (1008.55 0.85%): 1.68x speedup
25866 gen3 gnome-terminal-vim 4285.13 (4287.68 0.04%) -> 2715.97 (3202.17 13.52%): 1.58x speedup
25867 gen5 swfdec-youtube 5843.94 (5951.07 0.91%) -> 3810.86 (3826.04 1.32%): 1.53x speedup
25868 gen4 poppler 7496.72 (7558.83 0.58%) -> 5125.08 (5247.65 1.44%): 1.46x speedup
25869 gen4 gnome-terminal-vim 21126.24 (21292.08 0.85%) -> 14590.25 (15066.33 1.80%): 1.45x speedup
25870 gen5 firefox-talos-svg 99873.69 (100300.95 0.37%) -> 70745.66 (70818.86 0.05%): 1.41x speedup
25871 gen4 firefox-planet-gnome 28205.10 (28304.45 0.27%) -> 19996.11 (20081.44 0.56%): 1.41x speedup
25872 gen5 firefox-talos-gfx 93070.85 (93194.72 0.10%) -> 67687.93 (70374.37 1.30%): 1.37x speedup
25873 gen4 evolution 6696.25 (6854.14 0.85%) -> 4958.62 (5027.73 0.85%): 1.35x speedup
25874 gen3 swfdec-giant-steps 2538.03 (2539.30 0.04%) -> 1895.71 (2050.62 62.43%): 1.34x speedup
25875 gen4 gvim 4356.18 (4422.78 0.70%) -> 3276.31 (3281.69 0.13%): 1.33x speedup
25876 gen6 evolution 1242.13 (1245.44 0.72%) -> 953.76 (954.54 0.07%): 1.30x speedup
25877 gen6 firefox-planet-gnome 4554.23 (4560.69 0.08%) -> 3758.76 (3768.97 0.28%): 1.21x speedup
25878 gen3 firefox-talos-gfx 6264.13 (6284.65 0.30%) -> 5261.56 (5370.87 1.28%): 1.19x speedup
25879 gen4 midori-zoomed 4771.13 (4809.90 0.73%) -> 4037.03 (4118.93 0.85%): 1.18x speedup
25880 gen6 swfdec-giant-steps 1557.06 (1560.13 0.12%) -> 1336.34 (1341.29 0.32%): 1.17x speedup
25881 gen4 firefox-talos-gfx 80767.28 (80986.31 0.17%) -> 69629.08 (69721.71 0.06%): 1.16x speedup
25882 gen6 midori-zoomed 1463.70 (1463.76 0.08%) -> 1331.45 (1336.56 0.22%): 1.10x speedup
25885 gen6 xfce4-terminal-a1 2030.25 (2036.23 0.25%) -> 2144.60 (2240.31 4.29%): 1.06x slowdown
25886 gen4 swfdec-youtube 3580.00 (3597.23 3.92%) -> 3826.90 (3862.24 0.91%): 1.07x slowdown
25887 gen4 firefox-talos-svg 66112.25 (66256.51 0.11%) -> 71433.40 (71584.31 0.14%): 1.08x slowdown
25888 gen4 gnome-system-monitor 5691.60 (5724.03 0.56%) -> 6707.56 (6747.83 0.33%): 1.18x slowdown
25889 gen3 ocitysmap 3494.05 (3502.44 0.20%) -> 4321.99 (4524.42 2.78%): 1.24x slowdown
25890 gen4 ocitysmap 3628.42 (3641.66 9.37%) -> 5177.16 (5828.74 8.38%): 1.43x slowdown
25891 gen5 ocitysmap 4027.77 (4068.11 0.80%) -> 5748.26 (6282.25 7.38%): 1.43x slowdown
25892 gen6 ocitysmap 1401.61 (1402.24 0.40%) -> 2365.74 (2379.14 4.12%): 1.69x slowdown
25894 [Note the performance regression for ocitysmap comes from that we now
25895 attempt to support rendering to and (more importantly) from large
25896 surfaces. By enabling such operations is the only way to one day be
25897 faster than purely using the CPU, in the meantime we suffer regression
25898 due to the increased migration and aperture thrashing. The other couple
25899 of regressions will be eliminated with improved span and shader support,
25900 now that the framework for such is in place.]
25902 The performance increase for Cairo completely overlooks the other
25903 critical aspects of the architecture:
25906 gen3 (800x600): 57.5 -> 96.2
25907 gen4 (800x600): 47.8 -> 74.6
25908 gen6 (1366x768): 100.4 -> 140.3 [F15]
25909 144.3 -> 146.4 [drm-intel-next]
25912 aa10text: 3.47 -> 14.3 Mglyphs/s [unthrottled!]
25913 copywinwin10: 1.66 -> 1.99 Mops/s
25914 copywinpix10: 2.28 -> 2.98 Mops/s
25916 And we do not have a good measure for how much improvement the reworking
25917 of the fallback paths give, except that xterm is now over 4x faster...
25919 PS: This depends upon the Xorg patchset "Remove the cacheing of the last
25920 scratch PixmapRec" for correct invalidations of scratch Pixmaps (used by
25921 the dix to implement SHM operations, used by chromium and gtk+ pixbufs.
25923 PPS: ./configure --enable-sna
25925 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25927 commit 340cfb7f5271fd1df4c8948e5c9336f5b69a6e6c
25928 Author: Diego Elio Pettenò <flameeyes@gmail.com>
25929 Date: Sun May 15 02:17:32 2011 +0200
25931 build: do not use AC_CHECK_FILE to find the header files.
25933 Using AC_CHECK_FILE will cause cross-builds to fail picking the right file;
25934 instead use compile/preprocessor checks properly, and check for
25935 xf86driproto earlier.
25937 Reviewed-by: Rémi Cardona <remi@gentoo.org>
25938 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25940 commit 9d6e02a135efdea1d169d1938359ab2b553e941c
25941 Author: Adam Jackson <ajax@redhat.com>
25942 Date: Tue May 10 15:27:45 2011 -0400
25944 Remove the memory of Option "AccelMethod"
25946 Signed-off-by: Adam Jackson <ajax@redhat.com>
25947 Reviewed-by: Eric Anholt <eric@anholt.net>
25948 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25950 commit 895a46e8ff70195c1a4bdccbeb652e330376f64a
25951 Author: Chris Wilson <chris@chris-wilson.co.uk>
25952 Date: Tue May 10 20:38:25 2011 +0100
25954 dri: Flush the batch after a DRI swap/copy event
25956 To minimise lag in those every so critical games, we want to ensure that
25957 the copy happens as soon as it is received, so we need to flush the
25958 batch after processing a swap event and before we go to sleep.
25960 References: https://bugs.freedesktop.org/show_bug.cgi?id=37068
25961 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25963 commit 0b4ca9313cc7eb4845cf7f4e87c869c0c6d6ff0d
25964 Author: Chris Wilson <chris@chris-wilson.co.uk>
25965 Date: Tue May 10 20:36:10 2011 +0100
25967 video: Flush the batch on the next blockhandler after queuing
25969 In order to avoid video lag and jerky playback we need to ensure that
25970 any queued video is flushed before we go to sleep.
25972 Fixes regression from 6f104189bb.
25974 Reported-and-tested-by: Edward Sheldrake <ejsheldrake@gmail.com>
25975 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=37068
25976 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25978 commit bb8bf2a28bcb5e838bf4bb7e80ce220d28540719
25979 Author: Chris Wilson <chris@chris-wilson.co.uk>
25980 Date: Tue May 10 14:35:02 2011 +0100
25982 Correct chipset detection for Q33, Q35, B43_G1
25984 Everytime we update these tables we trip over this bit of marketing
25987 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
25989 commit fd1ebd44fb72e7bdf57d00f8941cd6110a529cac
25990 Author: Chris Wilson <chris@chris-wilson.co.uk>
25991 Date: Tue May 10 07:30:58 2011 +0100
25993 module: Adopt IVB's more detailed naming convention for SNB
25995 This should fix the seven-fold repetition of "SandyBridge" in the list
25996 of supported chipsets during start-up... And be more useful in bug
25999 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26001 commit e9811bb777dfc51af19836175645400489f7d991
26002 Author: Chris Wilson <chris@chris-wilson.co.uk>
26003 Date: Tue May 10 07:28:23 2011 +0100
26005 Whitespacing cleanup for intel_module.c
26007 Bring intel_module.c into line with the kernel whitespacing rules abided
26008 by everywhere else in the tree.
26010 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26012 commit 79e59fb2a047b1e733a7b0dee608db3311391725
26013 Author: Eric Anholt <eric@anholt.net>
26014 Date: Fri May 6 12:40:12 2011 -0700
26016 Add support for Ivybridge chipset.
26018 This gets display and 2D blit acceleration up and running. No Render
26019 acceleration is provided yet.
26021 commit 792738adfc5164d30358f045875dfc9b199a46da
26022 Author: Eric Anholt <eric@anholt.net>
26023 Date: Fri May 6 13:14:25 2011 -0700
26025 Remove the static list of PciChipset and construct it from SymTabRec instead.
26027 This is one less place the new hardware enabler has to spam the
26028 chipset in. The PciChipset is just a match structure from PciId to
26029 the SymTabRec entry token, and our SymTabRec entry tokens are just the
26030 PciId, so it's trivial to construct.
26032 Acked-by: Kenneth Graunke <kenneth@whitecape.org>
26034 commit 583e80dfa12d6c73fc677c81cb605a07b2768979
26035 Author: Eric Anholt <eric@anholt.net>
26036 Date: Fri May 6 13:00:53 2011 -0700
26038 Use the existing deviceID -> name mapping in SymTabRec instead of duping it.
26040 We need to have this array anyway for the xf86 interfaces, apparently,
26041 so just store the name in one location. This drops the i852/i855
26042 subdevice distinction in the name printed, but I haven't seen us ever
26045 Acked-by: Kenneth Graunke <kenneth@whitecape.org>
26047 commit adf7bbd3a8758de6cdecbace42e399dd27188497
26048 Author: Eric Anholt <eric@anholt.net>
26049 Date: Fri May 6 12:56:53 2011 -0700
26051 Store the chipset info struct in the PCI match struct, instead of a switch().
26053 Acked-by: Kenneth Graunke <kenneth@whitecape.org>
26055 commit 3145530feed879082bcfab11ffc8e7fd0911c920
26056 Author: Chris Wilson <chris@chris-wilson.co.uk>
26057 Date: Sat May 7 19:51:04 2011 +0100
26059 Ensure that the partial batch is flushed upon the blockhandler
26061 Currently, we require that a batch containing a dirty bo be submitted
26062 before we mark the device as requiring a flush. So if we never submit a
26063 batch between block handlers, we can end up sleeping without ever
26064 flushing either the partial batch or the rendering to the scanout.
26066 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36776
26067 Tested-by: Vasily Khoruzhick <anarsoul@gmail.com>
26068 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26070 commit 67e5a74e997f199327f9115c7ba867df3c49da8d
26071 Author: Paul Menzel <paulepanter@users.sourceforge.net>
26072 Date: Fri Apr 22 15:09:12 2011 +0200
26074 NEWS: fix typo (s/2.14/2.15/) to match corresponding release
26076 Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
26077 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26079 commit 0944e2d5749f5122cd9802c4a2421106fa829ea4
26080 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
26081 Date: Wed Apr 20 11:39:55 2011 -0700
26083 Add basic 30 bit depth support
26085 Still need to handle video and gamma correction, but this gets the
26086 display up and running at 30 bit depth if the kernel and display support
26089 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
26091 commit 1d102cc6ed21d1c4afa47800eecd24b9d663f689
26092 Author: Chris Wilson <chris@chris-wilson.co.uk>
26093 Date: Tue Apr 12 13:48:38 2011 +0100
26095 Use SwapbuffersWait config option to control waiting on fullscreen swaps
26097 As fullscreen swaps were going via a different path to the swapping of
26098 ordinary windows, we were no longer honouring the xorg.conf option to
26099 disable swapbuffer waiting.
26101 This changes the code to only use pageflipping if the Option
26102 "SwapbuffersWait" is set to "TRUE" (default).
26104 Jesse's comment was that this should be superseded by actually
26105 supporting asynchronous page flips. As we are missing kernel and dix level
26106 support for that, in the meantime honour the config option.
26108 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26109 Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
26110 Cc: Kristian Høgsberg <krh@bitplanet.net>
26112 commit c9fb69cb2502917dfb2828c90802de7766072899
26113 Author: Chris Wilson <chris@chris-wilson.co.uk>
26114 Date: Sun Apr 17 10:42:05 2011 +0100
26116 i965/video: We need 150 dwords of space for video state emission
26118 (Actually around 131, with additional 10% just for safety.)
26120 Reported-by: Modestas Vainius <geromanas@mailas.com>
26121 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36319
26122 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26124 commit a51cd83d25f2f9f2107219d5671194f931601244
26125 Author: Chris Wilson <chris@chris-wilson.co.uk>
26126 Date: Sun Apr 17 10:36:26 2011 +0100
26128 intel: Beware the unsigned promotion when checking for batch overflows
26130 Reported-by: Modestas Vainius <geromanas@mailas.com>
26131 References: https://bugs.freedesktop.org/show_bug.cgi?id=36319
26132 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26134 commit 030aa3d13643a0acb7126fa1fa6c5a8eac134639
26135 Author: Chris Wilson <chris@chris-wilson.co.uk>
26136 Date: Thu Apr 14 10:25:21 2011 +0100
26140 Spotted too late...
26142 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26144 commit 0e425b30e12b11dbf51eea1d3366e3372f8b93af
26145 Author: Chris Wilson <chris@chris-wilson.co.uk>
26146 Date: Thu Apr 14 10:16:03 2011 +0100
26148 configure,NEWS: 2.15.0 release
26150 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26152 commit 686018f283f1d131073ef5917213e6a8ac013f26
26153 Author: Chris Wilson <chris@chris-wilson.co.uk>
26154 Date: Tue Apr 12 08:23:04 2011 +0100
26156 Turn relaxed-fencing off by default for older (pre-G33) chipsets
26158 There are still too many unresolved bugs, typically GPU hangs, that are
26159 related to using relaxed fencing (i.e. only allocating the minimal
26160 amount of memory required for a buffer) on older hardware, so turn off
26161 the feature by default for the release.
26163 Reported-and-tested-by: Knut Petersen <Knut_Petersen@t-online.de>
26164 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36147
26165 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26166 Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
26168 commit 3d2b79098ccb3fed86be66f619f4ed338741d454
26169 Author: Chris Wilson <chris@chris-wilson.co.uk>
26170 Date: Mon Apr 11 15:20:16 2011 +0100
26172 dri: Rearrange code to compile against xorg-server-1.7
26174 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26176 commit 4fa35dd5e13a58070220f787fc0678f1c679808d
26177 Author: Chris Wilson <chris@chris-wilson.co.uk>
26178 Date: Mon Apr 11 10:44:55 2011 +0100
26180 NEWS: version bump for 2.14.903 snapshot
26182 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26184 commit 97e9557619e58ef769eb7cbf1a03fbd52be7f2ed
26185 Author: Chris Wilson <chris@chris-wilson.co.uk>
26186 Date: Fri Apr 8 13:38:48 2011 +0100
26188 intel: Restore manual flush for old kernels
26190 Daniel Vetter pointed out that the automagic flush by the kernel for the
26191 busy-ioctl was only introduced upstream in 2.6.37. So we still need to
26192 manually emit a flush on old kernels.
26194 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26196 commit fb40bf2b33a6d26f0e6a4e5798d10c905faa8aad
26197 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
26198 Date: Fri Apr 8 12:37:35 2011 +0200
26200 Tell users to grab i915_error_state on gpu hangs
26202 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
26203 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
26205 commit 59ed6b05db99d7e5144f4f8499c67cee9b833673
26206 Author: Chris Wilson <chris@chris-wilson.co.uk>
26207 Date: Thu Apr 7 15:28:46 2011 +0100
26209 Revert "i965: Convert to relative relocations for state"
26211 This reverts commit d2106384be6f9df498392127c3ff64d0a2b17457.
26213 Breaks compiz (but not mutter/gnome-shell) on gen6. Not sure if this is
26214 not seem deep interaction issue with multiple clients sharing the GPU or
26215 just with compiz, but for now we have to revert and suffer the inane
26216 performance hit. It looks suspiciously like another deferred damage
26219 Bugzilla: 51a27e88b073cff229fff4362cb6ac22835c4044
26220 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26222 commit 25521900df11bc71020ee80db2223f979bec5ec6
26223 Author: Chris Wilson <chris@chris-wilson.co.uk>
26224 Date: Thu Apr 7 15:09:30 2011 +0100
26226 gen6: Invalidate texture cache
26228 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26230 commit ad22003033eb502474ae538a97e3b42cf8f83880
26231 Author: Chris Wilson <chris@chris-wilson.co.uk>
26232 Date: Thu Apr 7 10:29:40 2011 +0100
26234 i965: Avoid transform overheads for vertex emit where possible
26236 Minor improvement as the bottlenecks lie elsewhere. But it was annoying me.
26238 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26240 commit 007c2f86cbb386861a1f711786523657f92e12cb
26241 Author: Chris Wilson <chris@chris-wilson.co.uk>
26242 Date: Thu Apr 7 09:53:34 2011 +0100
26244 i965: Refactor to use constant sampler_state offsets
26246 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26248 commit 8dc99b305a514dcd42c4260698e685a66dc95518
26249 Author: Chris Wilson <chris@chris-wilson.co.uk>
26250 Date: Mon Apr 4 22:30:29 2011 +0100
26252 i965: Reset vertex_id after every batch
26254 So that we always remember to re-emit the initial vertex elements state.
26256 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26258 commit 5982ed4da12a964351029c6f786348a11e33dc81
26259 Author: Chris Wilson <chris@chris-wilson.co.uk>
26260 Date: Mon Apr 4 19:32:43 2011 +0100
26262 i965: Always update last_floats_per_vertex
26264 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26266 commit 6f104189bb9439ab0e05f04d4be020813eb04bf9
26267 Author: Chris Wilson <chris@chris-wilson.co.uk>
26268 Date: Mon Apr 4 19:21:10 2011 +0100
26270 Take advantage of the kernel flush for dirty bo in the busy ioctl
26272 Rather than just creating and submitting a batch that simply contains a
26273 flush in order to periodically ensure that rendering reaches the
26274 scanout, we can simply ask the kernel whether the scanout is busy. The
26275 kernel will then submit a flush on our behalf if it is dirty, which
26276 takes advantage of the kernel's dirty state tracking.
26278 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26280 commit 314439860e595c473d168c0cafb2b5b49b7fef30
26281 Author: Chris Wilson <chris@chris-wilson.co.uk>
26282 Date: Mon Apr 4 17:20:04 2011 +0100
26284 Remove unused function 'intel_bo_alloc_for_data'
26286 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26288 commit ced747cefb5e697e6caa65296dff728904f52b93
26289 Author: Chris Wilson <chris@chris-wilson.co.uk>
26290 Date: Mon Apr 4 17:13:50 2011 +0100
26292 Remove the unnecessary MI_FLUSH from the flush handler
26294 The kernel will emit any required flushes between the dri client and the
26295 ddx, and we do not rely on the MI_FLUSH here for scanout.
26297 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26299 commit 79444291a39c42039192a5baa3a71d52300cf4ee
26300 Author: Chris Wilson <chris@chris-wilson.co.uk>
26301 Date: Mon Apr 4 16:34:58 2011 +0100
26303 i965: segregate each vertex element into its own buffer
26305 Reduce the number of relocations emitted by only emitting one relocation
26306 per vertex element per vertex buffer.
26308 References: https://bugs.freedesktop.org/show_bug.cgi?id=35733
26309 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26311 commit d2106384be6f9df498392127c3ff64d0a2b17457
26312 Author: Chris Wilson <chris@chris-wilson.co.uk>
26313 Date: Mon Apr 4 12:33:04 2011 +0100
26315 i965: Convert to relative relocations for state
26317 References: https://bugs.freedesktop.org/show_bug.cgi?id=35733
26318 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26320 commit 972569f6fd1e14519f46e9f50d2509faf1d0aa55
26321 Author: Chris Wilson <chris@chris-wilson.co.uk>
26322 Date: Fri Mar 25 10:46:14 2011 +0000
26324 MI_LOAD_SCAN_LINES_INCL are inclusive and range [0, display height-1]
26326 We have seen GPU hangs with:
26328 batchbuffer at 0x0f9b4000:
26329 0x0f9b4000: 0x09000000: MI_LOAD_SCAN_LINES_INCL
26330 0x0f9b4004: 0x00000300: dword 1
26331 0x0f9b4008: 0x09000000: MI_LOAD_SCAN_LINES_INCL
26332 0x0f9b400c: 0x00000300: dword 1
26333 0x0f9b4010: 0x01820000: MI_WAIT_FOR_EVENT
26334 0x0f9b4014: HEAD 0x02000006: MI_FLUSH
26336 on a 1366x768 display. That according to the specs an invalid command
26339 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=35576
26340 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26342 commit 537a836dd6db384d53b52eb457a7d257c440217f
26343 Author: Chris Wilson <chris@chris-wilson.co.uk>
26344 Date: Fri Apr 1 07:10:25 2011 +0100
26348 Sales & Marketing score another victory in confusing me.
26350 Bugzila: https://bugs.freedesktop.org/show_bug.cgi?id=35854
26351 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26353 commit 630d77bf10ba6234bb9c04538636f7d8aa319aea
26354 Author: Chris Wilson <chris@chris-wilson.co.uk>
26355 Date: Wed Mar 30 07:24:42 2011 +0100
26357 Add a DRI2SwapEventPtr typedef for retro xservers
26359 Because people are still trying to build upon our solid historical
26362 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26364 commit f660df2cb44b310740ed850037ff37891b671f28
26365 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
26366 Date: Tue Mar 29 18:02:24 2011 +0200
26368 Cleanup gen2 tiling confusion
26370 A tile on gen2 has a size of 2kb, stride of 128 bytes and 16 rows.
26372 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
26373 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
26375 commit a2c8442fcb29ef2d0df41518952a2d8df1671070
26376 Author: Chris Wilson <chris@chris-wilson.co.uk>
26377 Date: Tue Mar 29 11:38:45 2011 +0100
26379 configure: version bump for 2.14.902 snapshot
26381 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26383 commit 2f2680793952876c89796174bed53c824fe6d1cb
26384 Author: Chris Wilson <chris@chris-wilson.co.uk>
26385 Date: Tue Mar 29 11:38:18 2011 +0100
26387 NEWS: Entry for 2.14.902 snapshot
26389 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26391 commit 7ccbec801e9ee32fc110db730dfec674a94dea21
26392 Author: Keith Packard <keithp@keithp.com>
26393 Date: Thu Mar 24 11:06:57 2011 -0700
26395 Recover from i830_dri2_add_frame_event out-of-memory condition
26397 If adding either the frame or client resources fails, we need to clean
26398 up afterwards properly.
26400 First, add_frame_event needs to internally clean up after itself by
26401 undoing any partial execution. Second, the callers need to look at the
26402 return value and free the swap/flip info structure when necessary.
26404 Signed-off-by: Keith Packard <keithp@keithp.com>
26405 Reviewed-by: Julien Cristau <jcristau@debian.org>
26407 commit ec133abc4bd8caba15ed54e18621b816afb06981
26408 Author: Chris Wilson <chris@chris-wilson.co.uk>
26409 Date: Thu Mar 24 15:32:50 2011 +0000
26411 Use the per-generation batch context switch for atomic sequences
26413 A modest boost to throughput and reduction in CPU overhead from
26414 not flushing the batch on every transition from BLT to RENDER.
26416 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26418 commit 86f23f21ab57fcbc031bcd2b8f432a08ff4cc320
26419 Author: Keith Packard <keithp@keithp.com>
26420 Date: Wed Mar 23 17:07:49 2011 -0700
26422 Skip client and drawable resource delete calls when deleting frame event
26424 As the frame_event is about to be freed, there's no point in cleaning
26425 up references to the drawable and client.
26427 Signed-off-by: Keith Packard <keithp@keithp.com>
26429 commit e1ff5182304e00c0d392092069422cae7626cf8d
26430 Author: Keith Packard <keithp@keithp.com>
26431 Date: Wed Mar 9 17:00:41 2011 -0800
26433 Handle drawable/client destruction in pending swaps/flips
26435 A pending swap or flip holds references to drawables and clients which
26436 become invalid when destroyed. Add suitable resources to the database
26437 to track those lifetimes and clean up the pending data structure then.
26439 Later, when the pending swap or flip occurs, handle a missing drawable
26440 by just discarding the flip or swap. Handle a missing client by not
26441 sending an event or reply.
26443 Signed-off-by: Keith Packard <keithp@keithp.com>
26445 commit aa0328f342177e33f9c863531712eaa4577b5f2b
26446 Author: Keith Packard <keithp@keithp.com>
26447 Date: Wed Mar 9 16:23:22 2011 -0800
26449 dri2: Make DRI2FrameEvent public and use instead of void *
26451 Instead of using void * for all of the flip_info and swap_info
26452 pointers, just make the underlying structure a public data type and
26455 Signed-off-by: Keith Packard <keithp@keithp.com>
26457 commit ee740778f5d5355c04f6fc4564f598993b106d62
26458 Author: Chris Wilson <chris@chris-wilson.co.uk>
26459 Date: Wed Mar 16 08:37:24 2011 +0000
26461 i915/video: Clip indirect Xv output
26463 In the case of a singular clip box and rendering onto the front buffer
26464 larger than 2048, the clip extents were being ignored. Here we can
26465 simply reduce the size of the indirect pixmap to the clip extents.
26467 Reported-by: Michael Chang <mchang@novell.com>
26468 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=35346
26469 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26471 commit 014fc7abb7b2cc2110e3ab9a0bd6f7cff2c64c05
26472 Author: Chris Wilson <chris@chris-wilson.co.uk>
26473 Date: Wed Mar 9 17:10:50 2011 +0000
26475 dri: Disable page-flip between a tiled buffer and a linear scanout
26477 Keith Packard pointed out a loophole that could cause the DDX to end up
26478 with a tiled scanout even if the user required a linear framebuffer;
26479 that is by using page-flipping we could replace the scanout pixmap with
26480 another of our choosing, and not necessarily tiled.
26482 Close that loophole by only allowing an exchange of buffers between
26483 identical tiling modes. For the common case, this is fine since they
26484 will indeed be allocated with the same tiling. For the linear
26485 framebuffer case with mesa using a tiled pixmap, we force it to blit
26486 onto the scanout instead.
26488 Reported-by: Keith Packard <keith.packard@intel.com>
26489 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26491 commit 049ce4397ddf7fd088ce364cbb53cacf5133176f
26492 Author: Chris Wilson <chris@chris-wilson.co.uk>
26493 Date: Wed Mar 9 08:27:02 2011 +0000
26495 Give each user of tiling separate xorg.conf options
26497 So that you can indeed allocate a linear framebuffer if you so desire
26498 without breaking mesa.
26503 Option "LinearFramebuffer" "False|True" # default false
26508 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26510 commit 0bb1a5f19e09dc553761ddd90bf6319eab94a597
26511 Author: Chris Wilson <chris@chris-wilson.co.uk>
26512 Date: Wed Mar 9 08:26:44 2011 +0000
26514 Update priv->stride after bo reallocation
26516 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26518 commit 34f9a3335f96b16ef9df6213eb0a586c94f8f05b
26519 Author: Chris Wilson <chris@chris-wilson.co.uk>
26520 Date: Wed Mar 2 17:46:19 2011 +0000
26522 configure: Bump for 2.14.901 snapshot
26524 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26526 commit 057783937d7a42a766d0e45174c4c2dc0ffacd40
26527 Author: Chris Wilson <chris@chris-wilson.co.uk>
26528 Date: Wed Mar 2 17:45:43 2011 +0000
26530 NEWS: Add entry for 2.14.901 snapshot
26532 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26534 commit e889d3a709b55a0731ab098b17a3364b9bf39387
26535 Author: Chris Wilson <chris@chris-wilson.co.uk>
26536 Date: Sun Feb 27 10:51:50 2011 +0000
26538 dri: Protect against destroying a foreign DRI drawable
26540 I have no clue as to how such an alien drawable reached us, but we have
26541 the evidence of a segfault to say it can happen.
26543 Reported-by: Bernie Innocenti <bernie@codewiz.org>
26544 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=34787
26545 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26547 commit d21d781466785c317131a8a57606925867265dc8
26548 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
26549 Date: Tue Feb 22 18:31:44 2011 +0100
26551 Fix relaxed tiling on gen2
26553 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
26555 commit 9599fde65a0d8b7e7c85199346f7b620bdd8388d
26556 Author: Bryce Harrington <bryce@canonical.com>
26557 Date: Tue Feb 15 22:30:18 2011 -0800
26559 Quell excessively verbose vblank counter failed error messages
26561 Certain error situations can result in the following printed to
26562 Xorg.0.log at a high enough rate to make log file size a problem.
26564 (WW) intel(0): I830DRI2GetMSC:1062 get vblank counter failed: Invalid argument
26565 (WW) intel(0): I830DRI2ScheduleWaitMSC:1118 get vblank counter failed: Invalid argument
26567 Following in the tradition of commit 0ad6d6e1, limit the warnings to be
26568 output 5 times, then quell the remainder.
26570 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=34322
26571 Ref.: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/710594
26573 Signed-off-by: Bryce Harrington <bryce@canonical.com>
26574 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26576 commit 0ca595e9d533019b241666d29b421c7b36f9647a
26577 Author: Adam Jackson <ajax@redhat.com>
26578 Date: Thu Feb 17 15:26:35 2011 -0500
26580 Fix IGD and IGDNG constants to be comprehensible
26582 Since, with GPU-on-package, it's hard to talk about a model number for
26583 a specific chipset like 855GM, just use the platform names.
26585 Signed-off-by: Adam Jackson <ajax@redhat.com>
26586 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26588 commit 4c66b28870b050493ad96f7b0fe2d70d7ee539c7
26589 Author: Chris Wilson <chris@chris-wilson.co.uk>
26590 Date: Thu Feb 17 16:48:24 2011 +0000
26592 uxa: Fallback if the temporary is too large
26594 If the render operation requires a temporary source Picture and the
26595 operation is large, larger than the maximum permitted bo, then we will
26596 fail to allocate the bo. In this case, we need to fallback and perform
26597 the operation on the CPU rather than dereference a NULL bo.
26599 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=34399
26600 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26602 commit 23f9b14df7c102c1036134835dd5d1a508059858
26603 Author: Chris Wilson <chris@chris-wilson.co.uk>
26604 Date: Sat Feb 12 10:42:34 2011 +0000
26606 i965: Remove broken maximum base addresses from video
26610 The hardware was never limited to 0x1000000 and the kernel can quite
26611 rightly place objects above that limit. Specifying such had no relation
26612 to reality, so why did we do it? TWICE!
26614 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=34017
26615 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26617 commit 9184af921bc2f332fcb6c9b47001414378eab8e2
26618 Author: Javier Jardón <jjardon@gnome.org>
26619 Date: Wed Feb 9 00:36:37 2011 +0000
26621 Update autotools configuration
26623 Use new libtool syntax and silent-rules to silent
26624 the build output a bit (linux-like)
26626 commit 6e721e098b9181e8e77e314f966729d28e705582
26627 Author: Bryce Harrington <bryce@canonical.com>
26628 Date: Fri Feb 4 00:15:13 2011 -0800
26630 Check return value of uxa_acquire_solid() since it can return NULL
26632 uxa_acquire_solid returns NULL under OOM. Thus the value of solid
26633 must be checked before dereferencing it in the uxa_get_offscreen()
26636 Signed-off-by: Bryce Harrington <bryce@canonical.com>
26637 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26639 commit da990536eca09c6de74627541cd56ecfad925eda
26640 Author: Chris Wilson <chris@chris-wilson.co.uk>
26641 Date: Thu Feb 3 09:41:48 2011 +0000
26643 uxa: Undo damage translation before appending
26645 The region is used to paint onto the backing pixmap (and thus
26646 translated) prior to being passed to the damage layer (wrt to the
26647 drawable). So the local translation needs to be undone first.
26649 Identified by Christopher James Halse Rogers.
26651 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=33650
26652 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26654 commit ce1e0969058f8c70b624bc85bb8d6698a35794d3
26655 Author: Chris Wilson <chris@chris-wilson.co.uk>
26656 Date: Mon Jan 31 20:27:26 2011 +0000
26658 Rename 'intel' backlight to match upstream name in 2.6.38
26660 (Just waiting for libbacklight to bring sanity...)
26662 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26664 commit 80f59c8a4f51ccc27653a2de2dee2563b1e911f1
26665 Author: Cyril Brulebois <kibi@debian.org>
26666 Date: Mon Jan 31 06:22:39 2011 +0100
26668 xvmc: Stop using uninitialized variable.
26670 The actual code was removed in 5f64122551, where the get_surface_status
26671 driver callback was dropped. Just return Success if there were no issues
26672 with the parameter.
26674 Get rid of the following with CFLAGS="-Wall -Werror":
26676 | cc1: warnings being treated as errors
26677 | intel_xvmc.c: In function ‘XvMCSyncSurface’:
26678 | intel_xvmc.c:677: error: ‘ret’ may be used uninitialized in this function
26679 | intel_xvmc.c:672: note: ‘ret’ was declared here
26681 Signed-off-by: Cyril Brulebois <kibi@debian.org>
26682 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26684 commit 86284abafbe1364f82509d2656c82ba11cdce6cc
26685 Author: Cyril Brulebois <kibi@debian.org>
26686 Date: Mon Jan 31 06:22:38 2011 +0100
26688 xvmc: Silence gcc, adding a cast to work around signedness issues.
26690 Both exported XvMCPutSlice/XvMCPutSlice2 functions have a prototype with a
26691 parameter whose signedness differs from the one in the put_slice/put_slice2
26692 functions they call. Make it unsigned through a cast.
26694 Get rid of this accordingly with CFLAGS="-Wall -Werror":
26696 | cc1: warnings being treated as errors
26697 | intel_xvmc.c: In function ‘XvMCSyncSurface’:
26698 | intel_xvmc.c:677: error: ‘ret’ may be used uninitialized in this function
26699 | intel_xvmc.c:672: note: ‘ret’ was declared here
26701 Signed-off-by: Cyril Brulebois <kibi@debian.org>
26702 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26704 commit fc2eb64d7a184770ccd708a5b2337a538ca2b9db
26705 Author: Cyril Brulebois <kibi@debian.org>
26706 Date: Mon Jan 31 09:59:04 2011 +0000
26708 xvmc: Silence gcc, add parentheses around arithmetic.
26710 Get rid of those accordingly with CFLAGS="-Wall -Werror":
26712 | cc1: warnings being treated as errors
26713 | i915_xvmc.c: In function ‘i915_mc_one_time_state_emit’:
26714 | i915_xvmc.c:369: error: suggest parentheses around arithmetic in operand of ‘|’
26715 | i915_xvmc.c:374: error: suggest parentheses around arithmetic in operand of ‘|’
26716 | i915_xvmc.c:379: error: suggest parentheses around arithmetic in operand of ‘|’
26717 | i915_xvmc.c: In function ‘i915_mc_load_indirect_render_emit’:
26718 | i915_xvmc.c:708: error: suggest parentheses around arithmetic in operand of ‘|’
26719 | i915_xvmc.c:713: error: suggest parentheses around arithmetic in operand of ‘|’
26721 Signed-off-by: Cyril Brulebois <kibi@debian.org>
26722 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26724 commit 1b9349bceb9172d9bf01638b2df10b9b79e883b4
26725 Author: Cyril Brulebois <kibi@debian.org>
26726 Date: Mon Jan 31 06:22:36 2011 +0100
26728 xvmc: Silence gcc, add parentheses around assignment.
26730 Get rid of the following with CFLAGS="-Wall -Werror":
26731 | CC intel_xvmc_dump.lo
26732 | cc1: warnings being treated as errors
26733 | intel_xvmc_dump.c: In function ‘intel_xvmc_dump_open’:
26734 | intel_xvmc_dump.c:41: error: suggest parentheses around assignment used as truth value
26736 Signed-off-by: Cyril Brulebois <kibi@debian.org>
26737 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26739 commit ae8877e30707cae586565b4deee4c7fd9ac7d3ed
26740 Author: Cyril Brulebois <kibi@debian.org>
26741 Date: Mon Jan 31 06:22:35 2011 +0100
26743 xvmc: Get rid of unused variables.
26745 Signed-off-by: Cyril Brulebois <kibi@debian.org>
26746 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26748 commit 8bd6a7149fcb5f9ed6580dc2aa0180874938a1ea
26749 Author: Chris Wilson <chris@chris-wilson.co.uk>
26750 Date: Sat Jan 29 12:04:09 2011 +0000
26752 Enable conditional compilation of asserts
26754 With '--disable-debug' we define NDEBUG and so disable the asserts, that
26755 is we continue to default to compiling asserts into the code.
26757 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26759 commit a5a1ab7bbc58ec884cd0b90fe7139f402291b92f
26760 Author: Chris Wilson <chris@chris-wilson.co.uk>
26761 Date: Fri Jan 28 17:20:06 2011 +0000
26763 i915: Remove unused 'w' and 'h'
26765 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26767 commit 2c9b3225d8c7b53e2d921705dbb069710abcea5d
26768 Author: Chris Wilson <chris@chris-wilson.co.uk>
26769 Date: Fri Jan 28 17:19:39 2011 +0000
26771 i915: Remove unused 'num_floats' variable
26773 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26775 commit c6dc27562abbc8ca9e873ad502ca49ae010461d2
26776 Author: Chris Wilson <chris@chris-wilson.co.uk>
26777 Date: Mon Jan 24 20:25:27 2011 +0000
26779 uxa: Only recreate the glyph cache on *generational* updates
26781 The screen resources are recreated when the screen is rotated as well,
26782 without being finalized. In this case, we do not need to reconstuct the
26783 cache (or if we did, we would need to tear it down first).
26785 Reported-by: Till Matthiesen <entropy@everymail.net>
26786 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=33412
26787 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26789 commit 5baa63c634990810a66c3150b4f2b76fcee2df38
26790 Author: Xiang, Haihao <haihao.xiang@intel.com>
26791 Date: Fri Jan 21 08:45:52 2011 +0800
26793 Fix an error in 4c4ad555564a80311df1a4b762eb1e119c6d95fb.
26795 See the original patch at https://bugs.freedesktop.org/show_bug.cgi?id=24767
26797 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
26799 commit 3e28a0c0b43823d3104fe8fc50b5994b41fc0cc1
26800 Author: Chris Wilson <chris@chris-wilson.co.uk>
26801 Date: Thu Jan 20 15:07:26 2011 +0000
26803 Create the UXA generational resources during screen create
26805 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26807 commit 4c4ad555564a80311df1a4b762eb1e119c6d95fb
26808 Author: Xiang, Haihao <haihao.xiang@intel.com>
26809 Date: Wed Jan 19 15:25:06 2011 +0000
26811 Correct offset of planes within clipped window
26813 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=24767
26814 Reported-and-tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
26815 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26817 commit 5a22bc999de8555f26c639eaffbe73b680746174
26818 Author: Eric Anholt <eric@anholt.net>
26819 Date: Mon Jan 17 11:32:37 2011 -0800
26821 Quiet compiler warning about is_affine_src same way we do is_affine_mask.
26823 commit 3a2a4b0784f0fc96a5457b18931471f15ad745fc
26824 Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
26825 Date: Mon Jan 17 17:38:23 2011 +0000
26827 Fix textured video when destination is larger than screen
26829 In our application, the screen is never rotated from the point of view
26830 of the driver; instead, the compositor applies a suitable rotation as
26831 it composites the display. This works fine on 945, but on 965, videos
26832 are limited in height to the actual height of the screen.
26834 Change various bits of code so that we use the width and height of the
26835 destination pixmap instead of the width and height of the virtual
26836 screen. This works correctly both for XVideo to offscreen storage
26837 (CompositeRedirect) and for XVideo to the screen (no compositor).
26839 commit fd9235ebe03a01982238cdd6e8b55f613e14b6af
26840 Author: Chris Wilson <chris@chris-wilson.co.uk>
26841 Date: Fri Dec 17 00:58:36 2010 +0000
26843 Delete unused memory allocation flags.
26845 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26847 commit 03248a798409ca4dd1caf5e25a3d5f5a80173260
26848 Author: Chris Wilson <chris@chris-wilson.co.uk>
26849 Date: Fri Dec 17 00:57:01 2010 +0000
26851 Cache the fixed crtc<->pipe relationship
26853 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26855 commit 00a2aee38dc5cbbd4907a593fa8d617f5a20292e
26856 Author: Chris Wilson <chris@chris-wilson.co.uk>
26857 Date: Fri Dec 17 00:50:33 2010 +0000
26859 Remove unused I830Output
26861 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26863 commit a851879695c48e88d421b053a78f867d9902273f
26864 Author: Chris Wilson <chris@chris-wilson.co.uk>
26865 Date: Fri Dec 17 00:48:26 2010 +0000
26867 Remove bitrotted, but fortunately unused, I830CrtcPrivate
26869 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26871 commit f8a5d7737e014e99943b947d088a4f9f1b465d46
26872 Author: Chris Wilson <chris@chris-wilson.co.uk>
26873 Date: Fri Dec 17 00:37:47 2010 +0000
26875 Delete some dead code
26877 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26879 commit 968151898b807b06953b2f4170ac02889f16c8dc
26880 Author: Chris Wilson <chris@chris-wilson.co.uk>
26881 Date: Fri Dec 17 00:18:25 2010 +0000
26883 Remove unused GTT/Map sizes and addresses
26885 These have been made obsolete by KMS and GEM.
26887 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26889 commit 19b5817a54d4a17161340ada9bf20027644ff074
26890 Author: Chris Wilson <chris@chris-wilson.co.uk>
26891 Date: Fri Dec 17 00:15:31 2010 +0000
26893 Remove unused struct _intel_memory definition
26895 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26897 commit 40a7bde9140c0ae6412b8f916b1455d44e2af6cf
26898 Author: Chris Wilson <chris@chris-wilson.co.uk>
26899 Date: Fri Jan 7 23:46:07 2011 +0000
26901 configure: version bump for 2.14.0
26903 commit 8ecbba2aca657da3894eb77e856ac2d4e9efc0cc
26904 Author: Chris Wilson <chris@chris-wilson.co.uk>
26905 Date: Fri Jan 7 23:45:06 2011 +0000
26907 NEWS: Release notes for 2.14.0
26909 commit 22d7b61791c382088a6c0df5dce3a15405d6c495
26910 Author: Chris Wilson <chris@chris-wilson.co.uk>
26911 Date: Tue Jan 4 12:32:01 2011 +0000
26913 i965: Fix off-by-one in assert
26915 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26917 commit a7c7a9108f76aa312f3d5efa466052b914c81484
26918 Author: Chris Wilson <chris@chris-wilson.co.uk>
26919 Date: Tue Jan 4 10:53:30 2011 +0000
26921 configure: version bump for 2.13.903 snapshot
26923 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26925 commit 3fbecf34630c3588618900afdfbd9a4266856178
26926 Author: Chris Wilson <chris@chris-wilson.co.uk>
26927 Date: Tue Jan 4 10:52:44 2011 +0000
26929 NEWS: Add entry for 2.13.903
26931 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26933 commit 90538ecfff9c324d61419f451d41d26ab026f232
26934 Author: U. Artie Eoff <ullysses.a.eoff@intel.com>
26935 Date: Tue Dec 21 17:42:38 2010 -0700
26937 configure: suggest upstream to find macros in case they're missing.
26939 added upstream suggestion to error message in case xorg macros or
26940 xserver macros are missing. removed package manager references in
26941 suggestion since these are 'user' and/or 'distribution' preferences.
26943 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26945 commit 6d50f5a140dbefdf86eb25d18d08473aa7e42acc
26946 Author: Chris Wilson <chris@chris-wilson.co.uk>
26947 Date: Mon Jan 3 23:06:14 2011 +0000
26949 If the crtc is not enabled, then it can't be on
26951 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26953 commit 145c42a101c3150fb1e2cae8a1780b26e8b56f42
26954 Author: Adam Jackson <ajax@redhat.com>
26955 Date: Mon Jan 3 17:52:27 2011 -0500
26957 dri2: Fix interlace computation
26959 scrn->currentMode is a hack for xf86vidmode and in general is wrong for
26960 RANDRful drivers. Use the mode on the associated CRTC instead.
26962 Signed-off-by: Adam Jackson <ajax@redhat.com>
26963 [ickle: crtc->mode is a ModeRec not Ptr]
26964 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26966 commit b0d10ff04a24a951256a79710ba22b325b549ba4
26967 Author: Adam Jackson <ajax@redhat.com>
26968 Date: Mon Jan 3 17:52:26 2011 -0500
26970 xv: Fix interlace computation
26972 scrn->currentMode is a hack for xf86vidmode and in general is wrong for
26973 RANDRful drivers. Use the mode on the associated CRTC instead.
26975 Signed-off-by: Adam Jackson <ajax@redhat.com>
26976 [ickle: crtc->mode is a ModeRec]
26977 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26979 commit 0ad6d6e1a678de7d8c2d4cd7c734887e1deb2f9b
26980 Author: Chris Wilson <chris@chris-wilson.co.uk>
26981 Date: Mon Jan 3 16:52:07 2011 +0000
26983 dri: Only issue a warning for an "impossible" flip return 5 times
26985 As it appears that some kernels do indeed return the "wrong" value,
26986 issuing a warning 60 times a second is a cruel and unusual punishment.
26988 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32680
26989 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
26991 commit d729ef02f2955f7476df4c65403bc1f8e705b780
26992 Author: Chris Wilson <chris@chris-wilson.co.uk>
26993 Date: Sun Jan 2 09:11:10 2011 +0000
26995 dri: Don't wait upon a NULL current mode
26997 There is a race condition between the dri swapbuffers code and
26998 hotplugging whereby we might attempt to execute a wait upon a
26999 non-existent output. This causes a NULL dereference and a loud crash.
27001 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32770
27002 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27004 commit 537fa55ed2449e91f3dd1e04abc720c6818d7227
27005 Author: Chris Wilson <chris@chris-wilson.co.uk>
27006 Date: Sun Jan 2 09:06:28 2011 +0000
27008 dri: Fix the use of the uninitialised bo for flink
27010 Reported-by: Jeff Chua <jeff.chua.linux@gmail.com>
27011 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27013 commit dbc542b9d38ff9781efcdf1d5308e97f04499138
27014 Author: Chris Wilson <chris@chris-wilson.co.uk>
27015 Date: Fri Dec 31 23:29:05 2010 +0000
27017 dri: Protect against using dri with an non-gem pixmap
27019 Reported-by: Jeff Chua <jeff.chua.linux@gmail.com>
27020 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27022 commit 53fbc9f1760ee481cba1f6dceb9e7c97282a2976
27023 Author: Chris Wilson <chris@chris-wilson.co.uk>
27024 Date: Thu Dec 30 15:32:40 2010 +0000
27026 Don't replace the scanout bo through PutImage
27028 As the bo may be pinned for either use by the scanout or through sharing
27029 with another application, under those circumstances we cannot replace
27030 the bo itself but must force the blit for PutImage.
27032 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31367
27033 Reported-and-tested-by: Bjørn Mork <bjorn@mork.no>
27034 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27036 commit 7667ad8432c032aec3a2aa004fc4dfc1877971b3
27037 Author: Chris Wilson <chris@chris-wilson.co.uk>
27038 Date: Fri Dec 24 09:42:29 2010 +0000
27040 dri: Differentiate identical "get vblank failed" messages with line no
27042 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27044 commit 875d4828352c501ec5d0169735fb078807e76f2e
27045 Author: Chris Wilson <chris@chris-wilson.co.uk>
27046 Date: Thu Dec 23 19:36:12 2010 +0000
27048 i830: amalgamate consecutive composites into a single primitive
27050 Improve aa10text on i845 from 218kglyphs/s to 234kglyphs/s
27052 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27054 commit 58d9912e447cf6e29b20747b5ed6ee595687ce8d
27055 Author: Chris Wilson <chris@chris-wilson.co.uk>
27056 Date: Wed Dec 22 13:50:53 2010 +0000
27058 Remove the deprecated function 'XNFprintf'
27060 As we know the maximum length of the string, we can replace our single
27061 usage of XNFprintf with snprintf.
27063 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27065 commit e8db77e019e19ace48592b2f6d96af6baa9a228e
27066 Author: Chris Wilson <chris@chris-wilson.co.uk>
27067 Date: Tue Dec 21 23:03:21 2010 +0000
27069 Revert "Suggest where to find xorg-macros in case it's missing"
27071 I failed my reading comprehension and didn't spot that the help strings
27072 pointed to two separate downstream packages.
27074 Thanks to Carl Worth for pointing out my mistake.
27076 This reverts commit ebb7aca66785730737bc739624006e55d43dfe1b.
27078 commit ebb7aca66785730737bc739624006e55d43dfe1b
27079 Author: Chris Wilson <chris@chris-wilson.co.uk>
27080 Date: Tue Dec 21 22:23:24 2010 +0000
27082 Suggest where to find xorg-macros in case it's missing
27084 [Modified version of U. Artie Eoff's commit to remove the duplicated
27085 string. The paraphrased commit message is repeated below for clarity.]
27087 xorg-macros is often package under an unusual name, so provide some
27088 clues as to what name to try on the user's distribution. Installing
27089 build-deps would be just too easy...
27091 Also include a check for the existence of the XORG_DRIVER_CHECK_EXT
27092 macro, otherwise configure might continue in spite of the missing macro
27093 and cause obtuse syntax errors.
27095 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27097 commit 90576f95f35e528fa2c667df162c5dc7ec98975d
27098 Author: U. Artie Eoff <ullysses.a.eoff@intel.com>
27099 Date: Tue Dec 21 15:10:41 2010 -0700
27101 configure: updated m4 macro check in configure.ac
27103 Added m4 check for XORG_DRIVER_CHECK_EXT macro definition. Updated m4 fatal messages to give better hint on how to resolve error when xorg macros are missing. Previously, configure would continue in spite of the missing macros and the build would fail for syntax errors.
27105 commit 1ac2e04023f84dbf1f3db2ecad1cadd159aa614d
27106 Author: Chris Wilson <chris@chris-wilson.co.uk>
27107 Date: Mon Dec 20 10:55:11 2010 +0000
27109 Undo: Disable BLT for i830 and 845G
27111 Reported-by: György Balló <ballogy@freestart.hu>
27112 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32482
27113 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27115 commit 6f21405454487adf7623cf22f0fdd9f127099afd
27116 Author: Chris Wilson <chris@chris-wilson.co.uk>
27117 Date: Fri Dec 17 16:37:55 2010 +0000
27119 G35 is gen4 and not gen3
27121 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32478
27122 Reported-and-tested-by: Michal Marek <mmarek@suse.cz>
27123 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27125 commit 5743c223877fbff710cdd5114cff6d3ee3108309
27126 Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
27127 Date: Thu Dec 9 03:27:59 2010 +0100
27129 Check consistency of pageflip completion vblank count.
27131 Implements a consistency check on returned vblank
27132 count values of pageflip completion. Impossible
27133 values are detected, a x-warning is logged and
27134 returned (msc,ust) values are marked invalid,
27135 so clients could perform error handling. Such
27136 a warning would indicate bugs in the pageflip
27137 completion routine of future kms drivers or the
27138 ddx and thereby aid driver debugging.
27140 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
27141 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27143 commit 2177e6032146f4720f244e98f7c0d6df1b925784
27144 Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
27145 Date: Thu Dec 9 03:12:35 2010 +0100
27147 Fix reporting of pageflip completion events on multi-head.
27149 When a drawable is page-flipped on multiple crtc's (fullscreen
27150 drawable on mirror-mode or multi-head x-screen), only one pageflip
27151 event is finally delivered, after the last participating crtc signals
27152 flip completion, this to avoid visual corruption.
27154 Old code returned vblank count and timestamps of flip completion
27155 of this last crtc, instead of the values of the "master crtc", the
27156 one that was used for initially scheduling/triggering the pagflip
27157 via vblank events. (master = I830DRI2DrawablePipe(drawable))
27159 This patch makes sure that the pageflip completion values of the
27160 "master" crtc are returned, otherwise client applications will
27161 get confused by the random (msc, ust) values returned by whichever
27162 crtc was the last to complete its flip. Without this, the returned
27163 values change randomly and jump forward and backward in time and
27166 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
27167 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27169 commit 71af40a75fbdd1054b1111e8cbe67ad1f97e6613
27170 Author: Chris Wilson <chris@chris-wilson.co.uk>
27171 Date: Fri Dec 10 15:55:09 2010 +0000
27173 NEWS: 2.14, I meant the upcoming 2.14 release!
27175 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27177 commit 556331a22101d08babbd276236f74ffcc380171c
27178 Author: Chris Wilson <chris@chris-wilson.co.uk>
27179 Date: Fri Dec 10 15:28:37 2010 +0000
27181 configure: version bump for 2.13.902
27183 This is actually the first release candidate for 2.14.
27185 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27187 commit e01ea4a45c1cf8953a6d5ff6c600d76dbbe4760d
27188 Author: Chris Wilson <chris@chris-wilson.co.uk>
27189 Date: Fri Dec 10 15:28:00 2010 +0000
27191 NEWS: Add entry for the 2.13.902 snapshot
27193 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27195 commit 8df6f0a6082ad762ec69be7789a961db0c74154a
27196 Author: Chris Wilson <chris@chris-wilson.co.uk>
27197 Date: Fri Dec 10 12:26:16 2010 +0000
27199 configure: Bump required libdrm to 2.4.23
27201 To pull in the required headers for BLT and relaxed fencing support.
27203 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27205 commit 9b967807c2d240488a715509649663aac3583532
27206 Author: Chris Wilson <chris@chris-wilson.co.uk>
27207 Date: Tue Dec 7 20:45:16 2010 +0000
27209 Revert "i965: The RenderCache flush after every glyph is required for compiz"
27211 This reverts commit 03e8351179b1c25d219842ef3e01ee8e176f594f.
27215 This was only meant to be a temporary debugging hack, not for public
27216 consumption (or embarrassment).
27218 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27220 commit 27e33928f87ecb806d1f8c459d4a4ada76b0e30e
27221 Author: Chris Wilson <chris@chris-wilson.co.uk>
27222 Date: Tue Dec 7 20:42:24 2010 +0000
27224 i965: Mark sure we mark reused render targets as dirty
27226 ... or else we may forget to flush them again.
27228 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27230 commit 03e8351179b1c25d219842ef3e01ee8e176f594f
27231 Author: Chris Wilson <chris@chris-wilson.co.uk>
27232 Date: Tue Dec 7 19:27:57 2010 +0000
27234 i965: The RenderCache flush after every glyph is required for compiz
27236 ... now who can explain why.
27238 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27240 commit d90c8f4d0bb6b79ecdcaa82f87291169c18c64b9
27241 Author: Chris Wilson <chris@chris-wilson.co.uk>
27242 Date: Tue Dec 7 19:22:30 2010 +0000
27244 i965: Invalidate pixmap binding location on reuse.
27246 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27248 commit 4a186a612376bdd6f86c026e8b8b442108868a0a
27249 Author: Chris Wilson <chris@chris-wilson.co.uk>
27250 Date: Tue Dec 7 16:56:57 2010 +0000
27252 Always flush the batch before blocking for new X requests
27254 This should prevent any lag when waiting upon user input, for example
27255 whilst logging in with gdm.
27257 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27259 commit 1ba983034b3a70fb541dc359189c020ee497c634
27260 Author: Chris Wilson <chris@chris-wilson.co.uk>
27261 Date: Tue Dec 7 12:27:29 2010 +0000
27263 uxa: Emit the damage after the render for the workaround in uxa_solid_rects
27265 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27267 commit 00c204a7b2df7a6b501db9b57603534a45ecafa9
27268 Author: Chris Wilson <chris@chris-wilson.co.uk>
27269 Date: Mon Dec 6 16:34:52 2010 +0000
27271 snb: Only emit CC and DepthStencil bos once per batch
27273 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27275 commit 084ae43110240c582ae7cd6452966dd9f22350fa
27276 Author: Chris Wilson <chris@chris-wilson.co.uk>
27277 Date: Mon Dec 6 15:57:19 2010 +0000
27279 snb: Restore drawrect, we need the implicit flush
27281 Something is wrong, we should be tracking when to invalidate the caches
27282 as appropriate, yet I can not finding the missing flush to replace the
27283 implicit one of DRAW_RECTANGLE.
27287 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27289 commit f3a47d7f235d18e4529e3898a48673c7c3cbd489
27290 Author: Chris Wilson <chris@chris-wilson.co.uk>
27291 Date: Mon Dec 6 14:11:05 2010 +0000
27293 snb: Cache pixmap binding locations
27295 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27297 commit 4d48fed9aa0c3c7b84e74b4f72f298f580a8973c
27298 Author: Chris Wilson <chris@chris-wilson.co.uk>
27299 Date: Mon Dec 6 13:50:43 2010 +0000
27301 snb: Cache state between composite ops
27303 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27305 commit a58e5a1bdf10be3b96ecaa3d5e3ee288eab1063f
27306 Author: Chris Wilson <chris@chris-wilson.co.uk>
27307 Date: Mon Dec 6 13:34:53 2010 +0000
27309 snb: Emit more invariants only once
27311 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27313 commit e8f41c371024e2f3845c7652b22af8f5900200f2
27314 Author: Chris Wilson <chris@chris-wilson.co.uk>
27315 Date: Sun Dec 5 21:37:23 2010 +0000
27317 uxa: Prevent reading past the last byte on upload/download
27319 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29752
27320 Reported-by: Sergey Samokhin <prikrutil@gmail.com>
27321 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27323 commit c2fac6ca108734dd9566570b15f01cc476fa53dc
27324 Author: Matthias Hopf <mhopf@suse.de>
27325 Date: Fri Nov 19 04:29:57 2010 +0100
27327 Don't use hardware acceleration on Sandybridge rev 07 hardware or earlier.
27329 This is known to lock up the GPU even with the workaround in place.
27331 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31773
27332 Signed-off-by: Matthias Hopf <mhopf@suse.de>
27333 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27335 commit 8abade8a2139b500251488be5d6390f149666fb2
27336 Author: Chris Wilson <chris@chris-wilson.co.uk>
27337 Date: Sun Dec 5 09:55:43 2010 +0000
27339 display: Flush any pending batches before changing modes.
27341 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27343 commit 8450aba6c731b20e5cd7b55f1d747f00a6523291
27344 Author: Chris Wilson <chris@chris-wilson.co.uk>
27345 Date: Sun Dec 5 01:08:51 2010 +0000
27347 i965: Also flush the vertex buffer when restarting the array.
27349 As a corollary to filling one vertex array and beginning a new one is
27350 remembering to emit the old one before overwriting...
27352 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27354 commit f40f8bcf07bd7f496bfd3069eddfcbeb3d1bf8be
27355 Author: Chris Wilson <chris@chris-wilson.co.uk>
27356 Date: Sun Dec 5 00:26:22 2010 +0000
27358 i965: Check for potential vertex array overflow every time
27360 There was a reason why we need to check at the start of every composite
27361 operation to see if we have enough space in the array to fit the
27362 vertices, which I promptly forgot when moving the code around to make
27367 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27369 commit 55c5f1876e2329a938955967f5d45c814e50beb5
27370 Author: Chris Wilson <chris@chris-wilson.co.uk>
27371 Date: Fri Dec 3 02:05:05 2010 +0000
27373 Wait on the current buffer to complete when running synchronously.
27375 And remove the vestigal wait upon changing crtc as this is more properly
27376 done in the kernel.
27378 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27380 commit 3cc74044ce3546cc7dc2e918cbabbb41a77f4026
27381 Author: Chris Wilson <chris@chris-wilson.co.uk>
27382 Date: Thu Dec 2 12:07:44 2010 +0000
27384 i965: Amalgamate surface binding tables
27386 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27388 commit a1fa0dbfdafea32139d4457d81b9d722df955eaf
27389 Author: Chris Wilson <chris@chris-wilson.co.uk>
27390 Date: Thu Dec 2 11:25:05 2010 +0000
27392 i965: Upload an entire vbo in a single pwrite, rather than per-rectangle
27394 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27396 commit 23437fe6769322d48cc3d264660717475b5d0d74
27397 Author: Chris Wilson <chris@chris-wilson.co.uk>
27398 Date: Fri Dec 3 01:16:16 2010 +0000
27400 i965: Use reciprocal scale factors to avoid the divide per-vertex-element
27402 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27404 commit 0bb135c40e5ac1bf7593ec1d68d2815cbf47aa25
27405 Author: Chris Wilson <chris@chris-wilson.co.uk>
27406 Date: Tue Nov 23 22:27:38 2010 +0000
27408 Disable BLT for i830 and 845G
27410 This pair of chipsets seem broken beyond repair, specifically the
27411 erratum that causes the wrong PTE entry to be invalidated, so disable
27412 our incorrect attempts to use the BLT on those devices.
27414 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27416 commit 33c08882c0d551afb28baef643279901dcc65fa9
27417 Author: Keith Packard <keithp@keithp.com>
27418 Date: Wed Nov 17 16:37:53 2010 +0800
27420 Mark outputs as DPMSModeOn and restore backlight at mode set
27422 The kernel always turns monitors on when doing mode setting, and so no
27423 further DPMS action is required. Note this in the mode setting code by
27424 marking the updated DPMS mode and restoring any saved backlight level.
27426 Signed-off-by: Keith Packard <keithp@keithp.com>
27427 Reviewed-by: Eric Anholt <eric@anholt.net>
27429 commit cc930a37612341a1f2457adb339523c215879d82
27430 Author: Chris Wilson <chris@chris-wilson.co.uk>
27431 Date: Sun Nov 14 19:47:00 2010 +0000
27433 uxa: Relax fencing some more for gen3
27435 Allow fenced allocations even for small pixmaps if the kernel supports
27436 relaxing fencing (where only the used pages are allocated).
27438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27440 commit 3c5b1399e29ef577b8b91655b5e1c215d1b6dfbb
27441 Author: Chris Wilson <chris@chris-wilson.co.uk>
27442 Date: Tue Nov 9 20:20:06 2010 +0000
27444 i915: Disable maximum state addresses
27446 As the kernel controls the relocation of state buffers, we should not
27447 hard code the maximum permissible value for them.
27449 Fixes an eventual hang with full-gtt.
27451 Reported-by: Peter Clifton <pcjc2@cam.ac.uk>
27452 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27454 commit 81d355a8dcdfe3d0924bb843bcc0e6546f695391
27455 Author: Chris Wilson <chris@chris-wilson.co.uk>
27456 Date: Tue Nov 9 09:28:54 2010 +0000
27458 uxa: Fix crash after allocation failure
27460 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31487
27461 Reported-by: Thomas Fjellstrom <tfjellstrom@shaw.ca>
27462 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27464 commit 861ee17ae662ffec2269292f92da7833df508b1a
27465 Author: Carl Worth <cworth@cworth.org>
27466 Date: Fri Nov 5 12:55:02 2010 -0700
27468 Update version number to 2.13.901
27470 In preparation for a snapshot.
27472 commit aaa436b437839a80d340cc5ad598ec4b47ed2eff
27473 Author: Carl Worth <cworth@cworth.org>
27474 Date: Fri Nov 5 12:54:16 2010 -0700
27476 NEWS: Add release notes for 2.13.901 snapshot.
27478 An intermediate snapshot to capture recent developments.
27480 commit 6b66b45b1ff207b80b81973745f42a776448f764
27481 Merge: a44a63d 9f232ad
27482 Author: Carl Worth <cworth@cworth.org>
27483 Date: Fri Nov 5 12:51:55 2010 -0700
27485 Merge commit '2.13.0'
27487 This changes the version number and adds the 2.13.0 release notes,
27488 (which were otherwise missing from the master branch).
27490 commit a44a63d2ff6c01c3dc61de6f736dd441ddd25e52
27491 Author: Chris Wilson <chris@chris-wilson.co.uk>
27492 Date: Fri Nov 5 09:58:45 2010 +0000
27494 Wait for any pending rendering before switching modes.
27496 A perennial problem we have is the accursed WAIT_FOR_EVENT hangs, which
27497 occur when we switch the framebuffer before the WAIT_FOR_EVENT completes
27498 and upsets the GPU.
27500 We have tried more subtle approaches to detected these and fix them up in
27501 the kernel, to no avail. What we need to do is to delay the framebuffer
27502 flip until the WAIT completes, which is quite tricky in the kernel
27503 without new ioctls and round-trips. Instead, apply the big hammer from
27504 userspace and synchronise all rendering before changing the framebuffer.
27505 I expect this not to cause noticeable latency on switching modes (far
27506 less than the actual modeswitch) and should stop these hangs once and
27509 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31401 (...)
27510 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27512 commit 52b32436b9e14a3e13818f80102150ff5bc3c002
27513 Author: Chris Wilson <chris@chris-wilson.co.uk>
27514 Date: Wed Nov 3 19:42:26 2010 +0000
27516 Downgrade tiling allocation failure to a warning
27518 We emitted this message as an error even though we fallback and attempt
27519 to allocate a non-tiled framebuffer before failing (with an appropriate
27522 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27524 commit 18839aaec505f8bbdb0690fe694162bf09a87d5c
27525 Author: Chris Wilson <chris@chris-wilson.co.uk>
27526 Date: Wed Nov 3 18:14:29 2010 +0000
27528 Fallback to shadow for Sandybridge if we don't have access to the BLT
27530 If we attempt to emit BLT batches without kernel support, we just end up
27531 with EINVAL and no rendering. Prevent this, and avoid uncached
27532 rendering, by restoring the shadow fallback paths if there is no BLT
27535 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27537 commit 8ff37667bf864b771d16a58fc5041cb48408b6a8
27538 Author: Eric Anholt <eric@anholt.net>
27539 Date: Tue Nov 2 10:36:03 2010 -0700
27541 Remove the intermittent GEM_THROTTLE call.
27543 This is a holdover from early GEM work when we weren't syncing on the
27544 DRI client side. It would keep clients from getting too far ahead and
27545 killing their interactivity, by bringing everyone to a halt when
27546 anyone was too far ahead.
27548 Now, GL clients throttle themselves to avoid the problem, and it turns
27549 out that in the case that they don't (long rendering to buffers with
27550 no swap), this actually reduces X Server interactivity: instead of
27551 lagging of X rendering behind input, you get no response for seconds
27552 at a time, then a burst of rendering, then nothing again.
27554 Reported by ajax. Tested with moving a window while running
27555 cairo-perf-trace on the GL backend (improvement) and X backend (no
27556 significant change in responsiveness).
27558 commit 540c5742186c26c3aeccb7b5d3ff0f374722a20c
27559 Author: Xiang, Haihao <haihao.xiang@intel.com>
27560 Date: Tue Nov 2 11:05:32 2010 +0800
27562 render: use headerless render target write
27564 It is weird that some rendercheck cases only work fine with headerless write.
27565 Need to update intel-gen4asm to support headerless write
27567 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27569 commit 7a3109312e43ce4c2f600a65fe6bc821a8f0ebde
27570 Author: Xiang, Haihao <haihao.xiang@intel.com>
27571 Date: Wed Oct 27 16:51:28 2010 +0800
27573 render: acceleration for composite on Sandybridge
27575 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27577 commit 0ab2c05a291738ee763eb518c46e6dcfe29249a2
27578 Author: Xiang, Haihao <haihao.xiang@intel.com>
27579 Date: Wed Oct 27 16:49:26 2010 +0800
27581 render: fragments for composite on Sandybridge
27583 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27585 commit 21c86548dc2c80632c49ec6ec4da2935a0c99476
27586 Author: Xiang, Haihao <haihao.xiang@intel.com>
27587 Date: Wed Oct 27 16:47:36 2010 +0800
27589 render: fix send instruction used in sampling fragments
27591 To prepare for composite on Sandybridge
27593 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27595 commit 8d008ca89d44aa10aeb032d631971aaebb388675
27596 Author: Xiang, Haihao <haihao.xiang@intel.com>
27597 Date: Mon Oct 25 15:39:39 2010 +0800
27599 render: set the surface state base address
27601 It is the same as commit 73d4c7d7
27603 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27605 commit b5ae596a0746e5a37e86f5b0980f9265e0061142
27606 Author: Chris Wilson <chris@chris-wilson.co.uk>
27607 Date: Mon Nov 1 11:11:09 2010 +0000
27609 Flush BLT batches before starting an atomic RENDER batch
27611 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27613 commit 42363134bd38fb406d90cdb9e5b227836a793912
27614 Author: Zou Nan hai <nanhai.zou@intel.com>
27615 Date: Mon Nov 1 13:23:36 2010 +0800
27617 Support BLT acceleration on gen6
27619 Signed-off-by: Zou Nan hai <nanhai.zou@intel.com>
27620 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27622 commit 5bed685f765671d63642b24e17abd70579c40ddf
27623 Author: Zou Nan hai <nanhai.zou@intel.com>
27624 Date: Mon Nov 1 13:23:35 2010 +0800
27626 add BLT ring support
27628 gen6+ platform has a BLT engine with seperate
27629 command streamer to support BLT commands.
27631 Signed-off-by: Zou Nan hai <nanhai.zou@intel.com>
27632 [ickle: merge trivial conflict]
27633 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27635 commit 6ec3ff134baeac33d9d5821695d5f80843a96c91
27636 Author: Xiang, Haihao <haihao.xiang@intel.com>
27637 Date: Fri Oct 15 11:21:09 2010 +0800
27639 Xv: don't call intel_wait_for_scanline on Sandybridge
27641 MI_LOAD_SCAN_LINE_INCL command is not available on sandybridge.
27643 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27645 commit 9e73cbd7eea61f2fe69692056ca446a4d95e17c2
27646 Author: Xiang, Haihao <haihao.xiang@intel.com>
27647 Date: Fri Oct 15 11:20:22 2010 +0800
27649 Xv: enable TextureAdaptor for Sandybridge
27651 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27653 commit 3213849907bf1b1468872668d10764054c1034fe
27654 Author: Xiang, Haihao <haihao.xiang@intel.com>
27655 Date: Wed Oct 27 09:54:05 2010 +0800
27657 Xv: setup pipeline for Xv on Sandybridge
27659 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27661 commit 9e4dd27aa848acae1b74b77bbbc0a5bb9f6a502a
27662 Author: Xiang, Haihao <haihao.xiang@intel.com>
27663 Date: Thu Oct 21 16:10:14 2010 +0800
27665 Xv: fragments for xv on Sandybridge.
27667 Need to update intel-gen4asm to build these fragments
27669 Signed--off-by: Xiang, Haihao <haihao.xiang@intel.com>
27671 commit e34b3f6ef5811d5e8d156dd6582df8f442e9792e
27672 Author: Xiang, Haihao <haihao.xiang@intel.com>
27673 Date: Thu Oct 21 13:15:45 2010 +0800
27675 Xv: Send instruction doesn't use implied move when sampling YUV surface
27677 The two fragments will be reused for sampling YUV surface
27678 and send doesn't have implied move on Sandybridge
27680 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27682 commit 73d4c7d7b8efc920ac09885dcace5a536f93b70f
27683 Author: Xiang, Haihao <haihao.xiang@intel.com>
27684 Date: Wed Oct 13 09:13:50 2010 +0800
27686 Xv: set the surface state base address
27688 To prepare for Xv on Sandybridge. It is easy to fill the binding
27689 table without relocation and make sure that the pointer to binding
27690 table only uses bits[15:0].
27692 Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
27694 commit 5afc7472b155e8e940f12a38baf80c298dc3b364
27695 Author: Chris Wilson <chris@chris-wilson.co.uk>
27696 Date: Fri Oct 29 12:23:59 2010 +0100
27698 intel: don't pass a dangling pointer to GET_PARAM
27700 I fail at cut'n'paste.
27702 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27704 commit 6b3ce2e8701e7c3fcd30ca07fc13a2cdcf62560d
27705 Author: Chris Wilson <chris@chris-wilson.co.uk>
27706 Date: Fri Oct 29 11:07:50 2010 +0100
27708 uxa: Enable reduced fence sizes for i915
27710 Depends on libdrm 362457715faacd3101929e5f0d8ae250d0ad09df (for
27711 HAS_RELAXED_FENCING define).
27713 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27715 commit b066ddda31e40afa652ae51ffc35025ac3ef6f2a
27716 Author: Adam Jackson <ajax@redhat.com>
27717 Date: Sun Oct 3 14:42:56 2010 -0700
27719 intel: Listen for hotplug uevents (V3)
27721 This connects the kernel uevent indicating monitor hotplugging to the
27722 RandR notification events so that X applications can be notified
27723 automatically when monitors are connected or disconnected.
27725 This also adds a configuration option to disable hotplug events.
27727 V2: missed a #ifdef HAVE_UDEV around some udev-specific declarations
27729 V3: document Hotplug option in man page
27731 Signed-off-by: Keith Packard <keithp@keithp.com>
27733 commit a1c54f69643671ce296c57d132852e9846cc41d3
27734 Author: Eric Anholt <eric@anholt.net>
27735 Date: Mon Oct 18 13:28:44 2010 -0700
27737 Fix violation of gen6 requirements for depthbuffer tiling.
27739 In general, demoting of tiling of DRI2 buffers seems dubious, as we've
27740 got various bits of functionality that won't all work together unless
27741 buffers are tiled as expected. This just covers one instance of the
27742 problem, caught by assertions in Mesa.
27747 glean/readPixSanity
27751 commit 33133a1d38a4dce60d2e135663b545895fc85abc
27752 Author: Chris Wilson <chris@chris-wilson.co.uk>
27753 Date: Sun Oct 17 09:21:06 2010 +0100
27755 display: do not report failure for setting unrecognised properties
27757 We need to accept any changes to properties not handled by ourselves -- we
27758 can't validate the changes ourselves. Denying those changes breaks EDID
27759 reporting, for example.
27761 Reported-by: Elvis Pranskevichus <el@prans.net>
27762 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27764 commit 8507ef50ba08061ac158848a7c4c136e1755d87e
27765 Author: Chris Wilson <chris@chris-wilson.co.uk>
27766 Date: Wed Oct 13 09:45:59 2010 +0100
27768 Fix driverName regression for i830 from 4083197a
27770 Reported-by: Stefan Glasenhardt <stefan@glasen-hardt.de>
27771 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30808
27772 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27774 commit 15aab1ef982fb67eb2ea463089f070afd9a2cfe5
27775 Author: Chris Wilson <chris@chris-wilson.co.uk>
27776 Date: Wed Oct 13 09:50:25 2010 +0100
27778 Move EDID_COMPLETE_RAWDATA define to intel.h to avoid redifinition warning
27780 The compiler was simply warning that we defined the name prior to
27781 including the original definition, so reorder.
27783 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27785 commit d2bc2212f8be746e697a65555d6d83bc66d9ce43
27786 Author: Stefan Dirsch <sndirsch@suse.de>
27787 Date: Fri Oct 8 17:48:38 2010 +0200
27789 Make driver compile for 1.6 Xserver series again. Part deux.
27791 Signed-off-by: Matthias Hopf <mhopf@suse.de>
27792 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
27794 commit 5472359d6860af655a3c286d30558540376c9fdb
27795 Author: Chris Wilson <chris@chris-wilson.co.uk>
27796 Date: Fri Oct 8 15:23:13 2010 +0100
27798 dri: Check for pixmap privates before dereferencing them
27800 It is still possible for the pixmap allocator to return a software only
27801 pixmap (i.e. without an associated GEM buffer or intel_pixmap), so check
27802 before dereferencing.
27804 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30707
27805 Reported-by: Matthias Hopf <mhopf@suse.de>
27806 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27808 commit 4083197a44d1a1a05d33654b3c7d6e96d7472fe7
27809 Author: Chris Wilson <chris@chris-wilson.co.uk>
27810 Date: Thu Oct 7 12:45:02 2010 +0100
27812 Include a chipset generation number to clarify device specific paths.
27814 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27816 commit 1444ea35706992a1f094fe029e6a7d9eec0f93c6
27817 Author: Chris Wilson <chris@chris-wilson.co.uk>
27818 Date: Wed Oct 6 12:36:30 2010 +0100
27820 dri+shadow: Only tweak the acceleration of CopyRegion if using shadow.
27822 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27824 commit f286891d1415fc607f0c1b574e6bce4b3759493e
27825 Author: Chris Wilson <chris@chris-wilson.co.uk>
27826 Date: Wed Oct 6 11:38:48 2010 +0100
27828 dri: Reattach the fake pixmap for the shadow scanout to the drawable.
27830 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27832 commit 23ee926bcd8ddd7a2785786efdf119e4fae75bfd
27833 Author: Chris Wilson <chris@chris-wilson.co.uk>
27834 Date: Wed Oct 6 10:20:49 2010 +0100
27836 uxa: Skip a pixmap lookup if there is no driver finish access function
27838 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27840 commit 707901bf98073fb56179a0b61c806f85ef8f413c
27841 Author: Chris Wilson <chris@chris-wilson.co.uk>
27842 Date: Wed Oct 6 08:47:20 2010 +0100
27844 uxa: Re-enable acceleration.
27846 A side-effect of discriminating offscreen based on the devPrivate.ptr
27847 was that it broke uxa_finish_access and so after any fallback to s/w on
27848 a Pixmap, it remained in software for the reminder of its life.
27850 Introduce an explicit boolean to mark whether or not hardware
27851 acceleration is enabled for a pixmap (with a GEM buffer).
27853 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27855 commit 54f545e0631a287f421ef6a6f20831624b212cf2
27856 Author: Chris Wilson <chris@chris-wilson.co.uk>
27857 Date: Wed Oct 6 08:50:13 2010 +0100
27859 Revert "Clear pixmap->devPrivate.ptr [regression in 7c7294e]"
27861 This reverts commit 48b4e224297fa807be0e2bc7a67bf7e94579e8de.
27863 The better fix is to manually mark the pixmap when acceleration is and
27864 is not permitted. Whilst the devPrivate.ptr are invalid upon creation,
27865 it is not worth carring code that serves no purpose.
27867 commit d1925deedd70c6c4374e42de5aa44989fe328b58
27868 Author: Chris Wilson <chris@chris-wilson.co.uk>
27869 Date: Wed Oct 6 00:17:03 2010 +0100
27871 Clear pixmap->devPrivate.ptr [regression from 7c7294e]
27873 ModifyPixmapHeader(pixdata = NULL) does not clear the
27874 pixmap->devPrivate.ptr, instead the NULL value is interpreted as meaning
27875 to keep the current value. (This is similar to the interpretation of the
27876 other invalid values like depth=-1 which also mean not to change the
27877 current property). However pixadata=NULL is indeed a valid value, and in
27878 7c7294e, devPrivate.ptr == NULL was used to differentiate a bo pixmap
27879 from a system pixmap. Except that we never did clear the ptr as
27880 intended, and so X would immediately crash.
27882 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27884 commit 4bf2182bf9ee64795e2bf524fedf7b6c36f3cd03
27885 Author: Chris Wilson <chris@chris-wilson.co.uk>
27886 Date: Tue Oct 5 09:59:04 2010 +0100
27888 shadow: Use a cacheable shadow for all generations
27890 Always avoid direct rendering to the uncached scanout buffer,
27891 redirecting all 2D access to the shadow instead. Then for the couple of
27892 platforms where either the BLT is not ready or cannot be trusted (i8xx)
27893 perform the front buffer fixup using the uncached writes from the CPU.
27895 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27897 commit b27a521214728052b8b10342aaacbaf1622b8e2c
27898 Author: Chris Wilson <chris@chris-wilson.co.uk>
27899 Date: Mon Oct 4 20:31:44 2010 +0100
27901 shadow: Enable shadow by default on SandyBridge
27903 SandyBridge 2D support is far from complete, so instead of
27904 permanently falling back and always using uncached GTT mapping for
27905 rendering, use the shadow buffer instead.
27907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27909 commit 3e641459f8997f45288cc3cb314e50f2b5a629fd
27910 Author: Chris Wilson <chris@chris-wilson.co.uk>
27911 Date: Mon Oct 4 20:31:01 2010 +0100
27913 shadow: Disable BLT for SandyBridge
27915 The blitting code is incorrect for SandyBridge so disable until
27916 the BLT ring is ready.
27918 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27920 commit 7c7294ec00d6c3a454a17a1b9983d14d0655162c
27921 Author: Chris Wilson <chris@chris-wilson.co.uk>
27922 Date: Mon Oct 4 11:56:27 2010 +0100
27924 shadow+dri2: Allow dri2 to be independently enabled with shadow
27926 To enable DRI we create GEM buffers for the client to render into with
27927 hardware acceleration. In order to maintain coherency between any 2D
27928 render operations with the independent 3D clients (this includes the
27929 reading of 2D rasterisation by the direct rendering client, e.g.
27930 compiz using texture_from_pixmap) we need to replace the shadow pixmap
27931 with the GTT mapping. Therefore 2D rendering to a DRI buffer will be to
27932 uncached memory and thus penalised -- but the direct rendering clients
27933 will have full hardware acceleration.
27935 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27937 commit 16a5d0ee3ca4d41af6a6876a8baf81e3f715781d
27938 Author: Chris Wilson <chris@chris-wilson.co.uk>
27939 Date: Mon Oct 4 11:21:17 2010 +0100
27941 shadow: Map the scanout directly on i8xx
27943 Even with the minimal use of the BLT to copy from the GTT shadow to the
27944 GTT scanout, i830 was still hanging. Just write to the scanout directly.
27946 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27948 commit 516d235c5b3bd51e9533925140d95d3b4533129c
27949 Author: Chris Wilson <chris@chris-wilson.co.uk>
27950 Date: Mon Oct 4 11:10:57 2010 +0100
27952 Split shadow handling routines to their own file.
27954 This is about to get messy, so separate out the shadow from the normal
27957 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27959 commit 455f2939a661764ebb8d1747d44e16a0a8937808
27960 Author: Chris Wilson <chris@chris-wilson.co.uk>
27961 Date: Sat Oct 2 13:32:58 2010 +0100
27963 Do not claim the PCI device if !KMS
27965 By returning FALSE whilst probing if we can't find a KMS driver, we
27966 allow X to fallback to trying the VESA driver -- rather than failing.
27968 The initial idea for this was by Julien Cristau.
27970 Reported-by: Julien Cristau <jcristau@debian.org>
27971 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27973 commit 97fea2fd0c36d78f4e940b124a7886c4a3f3dfc7
27974 Author: Chris Wilson <chris@chris-wilson.co.uk>
27975 Date: Sat Oct 2 10:05:59 2010 +0100
27977 legacy/i810: Remove pVisualConfigs
27979 Completely unused pulling in an undesired glxproto dependency.
27981 Reported-by: Cyril Brulebois <kibi@debian.org>
27982 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30554
27983 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27985 commit 8784c4f5a1524fb979b00c7ce7981cbc1dcf0ec0
27986 Author: Chris Wilson <chris@chris-wilson.co.uk>
27987 Date: Fri Oct 1 21:40:36 2010 +0100
27989 video: Disable TextureAdaptor for Sandybridge
27991 The acceleration code does not yet exist, so we try to use gen4 shaders
27992 which hangs the chip and makes everybody upset. Disable it until such a
27993 time as support is written for Sandybridge XVideo.
27995 Reported-by: Matthias Hopf <mhopf@suse.de>
27996 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30500
27997 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
27999 commit 9f232ad67c72b7869dcd112207bc41dc01d7d104
28000 Author: Carl Worth <cworth@cworth.org>
28001 Date: Thu Sep 30 08:45:49 2010 -0700
28003 Increment version to 2.13.0
28005 For the new, major release.
28007 commit 99ec9bb0f1eaf2e5c04614957a51624a928526e5
28008 Author: Carl Worth <cworth@cworth.org>
28009 Date: Thu Sep 30 08:45:15 2010 -0700
28011 Add release notes for the 2.13.0 release.
28013 Which is functionalliy identical to 2.12.902.
28015 commit 345c963e443ec325f1ff530512a356ddb318ff70
28016 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
28017 Date: Thu Sep 30 10:58:49 2010 +0800
28019 Always use tiling on Sandybridge
28021 Sandybridge requires kind of buffer must be tiling, like depth.
28022 And we would or have all tiling cases handled fine. So not allow
28023 user to turn off tiling on Sandybridge+ may be fine.
28025 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
28027 commit d4c013cb9d92d2959c9a98eca20e43ab14ae22e3
28028 Author: Carl Worth <cworth@cworth.org>
28029 Date: Tue Sep 28 20:49:52 2010 -0700
28031 Increment version to 2.12.902
28033 Just before making the 2.12.902 (2.13-rc2) snapshot.
28035 commit 35e7da7df0b286b5168a85d540cc9fdf55c1cae6
28036 Author: Carl Worth <cworth@cworth.org>
28037 Date: Tue Sep 28 20:49:10 2010 -0700
28039 Add release notes for the 2.12.902 snapshot.
28041 Summarizing the 5 recent changes to the driver.
28043 commit 5704955538f59a2b7a5eff67cc23953d2f27d523
28044 Author: Carl Worth <cworth@cworth.org>
28045 Date: Tue Sep 28 20:44:40 2010 -0700
28047 Fix to depend on the (just-released) libdrm >= 2.4.22
28049 With libdrm < 2.4.22, the compilation of xf86-video-intel would fail
28050 due to an undefined DRM_MODE_CONNECTOR_eDP. We now ensure that a
28051 sufficiently new libdrm is available before compiling.
28053 commit 40c9d180bf43af84ae05d88e993deedf76e263e6
28054 Author: Chris Wilson <chris@chris-wilson.co.uk>
28055 Date: Sun Sep 26 23:09:04 2010 +0100
28057 Retry framebuffer allocation if first attempt fails.
28059 If we are tiled, we may fail the allocation due to an EIO and so not
28060 being able to set tiling. Try again with an untiled request in this
28063 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28065 commit 537e73f3f935b917f2f5f9b51499cb29d65e3889
28066 Author: Chris Wilson <chris@chris-wilson.co.uk>
28067 Date: Fri Sep 24 17:37:41 2010 +0100
28069 Disable dri2 after forcing fallbacks
28071 If we force fallbacks, then we will only create pixmaps in system
28072 memory, preventing DRI2 from passing valid bo names to the clients. In
28073 this case, they will just fallback to swrast. If we disable DRI2 after
28074 forcing fallbacks (e.g. regenerating after a GPU hang or explicitly
28075 disabled with the shadow buffer) then it is simpler just to disable the
28076 extension and allow mesa to use pure swrast.
28078 Reported-by: Julien Cristau <jcristau@debian.org>
28079 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28081 commit b84925b9c0842ba4dfa3481c09d3a80f84db4838
28082 Author: Matthias Hopf <mhopf@suse.de>
28083 Date: Tue Aug 17 18:05:02 2010 +0200
28085 Make driver compile for 1.6 Xserver series again.
28087 Signed-off-by: Matthias Hopf <mhopf@suse.de>
28089 commit 99f8d68d3dbc54fdd7de4a6df546cfa89422a0c2
28090 Author: Chris Wilson <chris@chris-wilson.co.uk>
28091 Date: Wed Sep 22 12:34:00 2010 +0100
28093 display: Refactor is_panel()
28095 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28097 commit 4ba27e1f2198fc10a1dd268ed9f137bac6e80086
28098 Author: Carl Worth <cworth@cworth.org>
28099 Date: Tue Sep 21 17:43:01 2010 -0700
28101 Bump version to 2.12.901
28103 Just prior to pushing out this new snapshot.
28105 commit 334265fa5ef18405bd8c897343fff49ad53f5fce
28106 Author: Carl Worth <cworth@cworth.org>
28107 Date: Tue Sep 21 17:42:21 2010 -0700
28109 NEWS: Add notes for the 2.12.901 snapshot
28111 (Otherwise known as rc1 for 2.13)
28113 commit bfba6d6148b938588ffbefdf04f4e71683c66dfd
28114 Author: Krzysztof Halasa <khc@pm.waw.pl>
28115 Date: Tue Sep 21 22:56:26 2010 +0100
28117 Allow interlaced modes.
28119 This may not the best method, but it should be a good base on which to
28122 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28124 commit f63c7df6866fed61551d39209b1a262e2a0c4b28
28125 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
28126 Date: Fri Sep 17 16:39:12 2010 +0800
28128 Disable swap buffer wait on Sandybridge
28130 MI_LOAD_SCAN_LINE_INCL command is not available on sandybridge.
28131 I haven't got reply on any substitution for it, so turn it off for now.
28133 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
28135 commit 55b5fe888002258dd863d41977bfbfcf1a922179
28136 Author: Chris Wilson <chris@chris-wilson.co.uk>
28137 Date: Thu Sep 16 16:50:13 2010 +0100
28139 Add alternate pci-id for B43
28141 Confirmed by http://en.wikipedia.org/wiki/Intel_GMA
28143 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30221
28144 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28146 commit 08c2caca48323d6d5701dcef3486f850619d7905
28147 Author: Chris Wilson <chris@chris-wilson.co.uk>
28148 Date: Sun Sep 12 12:34:51 2010 +0100
28150 uxa: Apply source clipping to blits
28152 Yes, this should be done in the higher layers. Yes, I have written code
28153 to that. No, it is not ready, hence add the sanity check to the
28156 This isn't the first report that I've seen, but will be the last.
28158 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30120
28159 Reported-by: rezbit.hex@gmail.com
28160 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28162 commit ae160d7fbfc79e78dad8702efcc55d9c0c25ff67
28163 Author: Chris Wilson <chris@chris-wilson.co.uk>
28164 Date: Fri Sep 10 13:19:12 2010 +0100
28166 shadow: Simply modify the Screen pixmap header
28168 This is a slightly less risky strategy than having to remember to update
28169 all pointers to the old Screen pixmap.
28171 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28173 commit 0515256490d5bcd55f85af83b84918d1bfe7f8f8
28174 Author: Chris Wilson <chris@chris-wilson.co.uk>
28175 Date: Fri Sep 10 00:08:58 2010 +0100
28177 display: Free the EDID blob after we copy it to the output, not before.
28179 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28181 commit 2b96c18165d713cd6781dbf217ec33e11cc961bc
28182 Author: Chris Wilson <chris@chris-wilson.co.uk>
28183 Date: Wed Aug 25 12:56:43 2010 +0100
28185 Enable a shadow buffer and disable GPU acceleration.
28187 An attempt to workaround the incoherency in gen2 chipsets, we avoid
28188 using dynamic reallocation as much as possible.
28190 The first step is to disable allocation of pixmaps using GEM and simply
28191 create them in system memory without a backing buffer object. This
28192 forces all rendering to use S/W fallbacks.
28194 The second step is to allocate a shadow front buffer and assign that to
28195 the Screen pixmap. This ensure that the front buffer remains in the GTT
28196 and pinned for scanout. The shadow buffer will be rendered to in the
28197 normal fashion via the Screen pixmap, and be marked dirty. In the block
28198 handler, the dirty shadow buffer is then blitted (using the GPU) over
28199 the front buffer. This should completely avoid having to move pages
28200 around in the GTT and avoid incurring the wrath of those early chipsets.
28201 Secondly, performance should be reasonable as we avoid the ping-pong
28202 caused by the small aperture and weak GPU forcing software fallbacks.
28204 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28206 commit 0fa4321a765126228170ecb9536f32c134886d51
28207 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
28208 Date: Tue Sep 7 13:56:16 2010 -0700
28210 KMS: initialize backlight support for eDP panels too
28212 Since being able to control the backlight is handy.
28214 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
28216 commit f56c839f30ec39ea22bbf6a7fda4cefd7cbfbfc2
28217 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
28218 Date: Tue Sep 7 13:53:47 2010 -0700
28220 KMS: rename LVDS fields to reflect actual usage
28222 These are used for panels, not just LVDS connections, so name them as such.
28224 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
28226 commit 0d26d950fdada1f59dc6cb31fe2f03004825f773
28227 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
28228 Date: Tue Sep 7 13:46:58 2010 -0700
28230 KMS: add fake EDID on eDP too
28232 This gives us a few more standard modes on eDP panels with just a simple
28233 fixed timing in the VBT, just like on older, LVDS attached panels.
28235 Fixes FDO bug https://bugs.freedesktop.org/show_bug.cgi?id=30069.
28237 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
28238 Tested-by: Manoj Iyer <manoj.iyer@canonical.com>
28239 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
28241 commit 273d34fbc43e54a8539fbe7efc1dca66bbb4e6dc
28242 Author: Chris Wilson <chris@chris-wilson.co.uk>
28243 Date: Tue Sep 7 12:57:04 2010 +0100
28245 display: Query current level after finding max value.
28247 The current backlight value is clamped to the valid range [0, max] and
28248 so as we queried the value before setting the max, we forced the current
28249 backlight to 0 and so set it to be zero on initialising the display.
28253 Bug 30063 - start X will modify brightness value to zero
28254 https://bugs.freedesktop.org/show_bug.cgi?id=30063
28256 which is a regression due to 38f940dfea494d3093236f065392c431be06ae6e.
28258 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28260 commit 53767cc0d0a58d36cd445da3a31c65b349eebbba
28261 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
28262 Date: Tue Sep 7 14:17:05 2010 +0800
28264 Add more sandybridge graphics device ids
28266 New ids for GT2 and GT2+ on desktop and mobile sandybridge, and
28267 server sandybridge device ids.
28269 commit 00f6af2c8ec899258ffab60dc2d9d80c4d9daec8
28270 Author: Chris Wilson <chris@chris-wilson.co.uk>
28271 Date: Sat Sep 4 18:45:11 2010 +0100
28273 display: Set MONITOR_EDID_COMPLETE_RAWDATA for large EDIDs
28275 Quoting Adam Jackson:
28277 "But the X driver looks like
28278 it never sets MONITOR_EDID_COMPLETE_RAWDATA, which means the X core
28279 doesn't know that any sections beyond the first are present, so it won't
28280 ever hand back more than 128 bytes to clients. Boo."
28282 This patch is based on his.
28284 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28286 commit 501e78b0095b00a635dc1dc182fc132f14f08f22
28287 Author: Chris Wilson <chris@chris-wilson.co.uk>
28288 Date: Sat Sep 4 12:37:39 2010 +0100
28290 Force use of GTT and fence registers for mapping tiled objects
28292 If the buffer object is tiled, we need to use the fence registers to
28293 perform the appropriate untiling for CPU access. Ensure that we always
28294 take this path for tiled objects, regardless of their size.
28296 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28298 commit b7a8087fbc5cbd944f8345b8a098270c1f63ebab
28299 Author: Chris Wilson <chris@chris-wilson.co.uk>
28300 Date: Wed Aug 25 15:44:20 2010 +0100
28302 Revert "Leave adjustment of backlight to the driver."
28304 This reverts commit 9c3e34703dcdc2155faeb2c1886930def19d9989.
28306 This commit is not ready, as first the driver needs to handle all
28307 controllers, especially those that ignore the BLC and require their own
28308 interface. Fortunately, by moving that discovery into the kernel - where
28309 it just means finding which ACPI device is attached to the video and has a
28310 backlight interface - the userspace code should become much more sane,
28311 and work even with multi-gpu, multi-lid systems.
28313 But that is for tomorrow.
28315 commit 68a5ad497b557ab9e8bff301098ea78c84d385a2
28316 Author: Chris Wilson <chris@chris-wilson.co.uk>
28317 Date: Wed Aug 25 14:59:04 2010 +0100
28319 uxa: Fallback if faced with large A1 glyphs.
28321 Rather than assert, we should fixup the use of large A1 glyphs. However,
28322 the simplest approach is to simply fallback to s/w.
28326 Bug 29430 - [UXA] Crash due assert (uxa_pixmap_is_offscreen(src_pixmap));
28327 https://bugs.freedesktop.org/show_bug.cgi?id=29430
28329 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28331 commit 9c3e34703dcdc2155faeb2c1886930def19d9989
28332 Author: Chris Wilson <chris@chris-wilson.co.uk>
28333 Date: Sun Aug 22 21:27:45 2010 +0100
28335 Leave adjustment of backlight to the driver.
28337 commit 104cd0554bde1d109a54db7a93700d5edfabd914
28338 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
28339 Date: Mon Aug 23 09:48:22 2010 +0800
28341 Add sandybridge D0 support
28343 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
28345 commit 271dda84be2f81b32dbe5947e64f5558a89e385e
28346 Author: Chris Wilson <chris@chris-wilson.co.uk>
28347 Date: Sun Aug 22 19:57:24 2010 +0100
28349 display: Use the native intel backlight controller
28351 If the i915 driver exposes a native ACPI interface to modify the panel
28352 backlight use it in preference to the generic interfaces. On multi-GPU
28353 systems, the panel backlight is meant to be connected via the IGP and
28354 this ensures that we always find the right interface.
28358 Bug 29273 - XORG Intel driver chooses wrong acpi_video to control
28359 brightness in multi-GPU system
28360 https://bugs.freedesktop.org/show_bug.cgi?id=29273
28362 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28364 commit 42312bbd8c9cbadf0adc4556b76987857d105bd9
28365 Author: Chris Wilson <chris@chris-wilson.co.uk>
28366 Date: Sun Aug 22 09:54:18 2010 +0100
28368 Remove accel_pitch_alignment
28370 This has to be 64 on all generations currently, so replace the variable
28373 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28375 commit 7f86e5b5da88201e5443e26619c35a1a9745cf68
28376 Author: Matt Turner <mattst88@gmail.com>
28377 Date: Sat Aug 21 17:32:25 2010 -0400
28379 Replace ROUND_* macros with ALIGN.
28381 Signed-off-by: Matt Turner <mattst88@gmail.com>
28382 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28384 commit b611bced15c30f7bcd03106ce90668b684c1ada6
28385 Author: Matt Turner <mattst88@gmail.com>
28386 Date: Sat Aug 21 17:32:24 2010 -0400
28388 Use ALIGN macro instead of open coding it.
28390 Signed-off-by: Matt Turner <mattst88@gmail.com>
28391 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28393 commit 8b04b350a983b89eb2d741f55baa297a933ac6ea
28394 Author: Chris Wilson <chris@chris-wilson.co.uk>
28395 Date: Thu Aug 19 20:05:08 2010 +0100
28397 Open-code DRICreatePCIBusID()
28399 During -configure we would attempt to query the availablility of KMS
28400 before the DRI module was loaded, thus we were unable to create a valid
28401 bus identifier and so the query failed and we disowned the device.
28405 Bug 29611 - Xorg -configure fails
28406 https://bugs.freedesktop.org/show_bug.cgi?id=29611
28408 Reported-by: Sergey Samokhin <prikrutil@gmail.com>
28409 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28411 commit c882f6a22a862c1664c375e05e5e6fc4bdb04edb
28412 Author: Chris Wilson <chris@chris-wilson.co.uk>
28413 Date: Wed Aug 18 10:21:22 2010 +0100
28415 Move registration of vsync fd from pre-init to screen-init
28417 Marty Jack reported an issue he found where the page-flipping handler
28418 was being lost on server reset. This results in the swap completion
28419 notification being lost, with the sporadic hang of full screen
28420 applications like Compiz, flash and even glxgears!
28424 Bug 29584 - Server in compute loop
28425 https://bugs.freedesktop.org/show_bug.cgi?id=29584
28427 There are also several possibly related bugs with similar symptoms, i.e.
28428 OpenGL applications hanging on missed swap notifications.
28430 Reported-by: Marty Jack <martyj19@comcast.net>
28431 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28432 Cc: Keith Packard <keithp@keithp.com>
28434 commit 19c48d3b3f33582baa87a9b3a9189e320e4cea45
28435 Author: Chris Wilson <chris@chris-wilson.co.uk>
28436 Date: Mon Aug 9 10:13:58 2010 +0100
28438 display: outputs are enabled automatically by KMS
28440 When an output is attached to a crtc and that crtc is enabled, the
28441 output is automatically enabled so we can remove the redundant manual
28444 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28446 commit 6304cb048c745be81dae13f1d936996e04eaa530
28447 Author: Chris Wilson <chris@chris-wilson.co.uk>
28448 Date: Thu Aug 5 18:13:55 2010 +0100
28450 display: Minor cleanup for adding extra LVDS modes
28452 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28454 commit 41ae9564350cba8ac9f040f22bad8740e08e14c4
28455 Author: Chris Wilson <chris@chris-wilson.co.uk>
28456 Date: Thu Aug 5 18:13:30 2010 +0100
28458 display: Refactor EDID attachment to output.
28460 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28462 commit a8919ab2967a8425e1cdc04b73702ce534f6abf0
28463 Author: Chris Wilson <chris@chris-wilson.co.uk>
28464 Date: Thu Aug 5 09:59:22 2010 +0100
28466 Revert "display: Cache whether we have probed for an EDID"
28468 Dave Airlie advised that hotplug detection can be unreliable and that
28469 mode caching, in general, should be done in the kernel in any case.
28471 This reverts commit 622e600069ab0efd22586c7a71eecbd4baf21c40.
28473 commit 622e600069ab0efd22586c7a71eecbd4baf21c40
28474 Author: Chris Wilson <chris@chris-wilson.co.uk>
28475 Date: Wed Aug 4 09:57:12 2010 +0100
28477 display: Cache whether we have probed for an EDID
28479 Remember for the detection cycle whether we have already probed for the
28480 EDID -- as this can be slow.
28482 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28484 commit a6a707ca13097b85b319283b3a174b1986056ab8
28485 Author: Chris Wilson <chris@chris-wilson.co.uk>
28486 Date: Wed Aug 4 09:54:28 2010 +0100
28488 display: Embed the lvds size into the connector
28490 Remove one very common allocation.
28492 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28494 commit 6c7d105ccae32fba49a07a03b726cc3fe2de3d27
28495 Author: Chris Wilson <chris@chris-wilson.co.uk>
28496 Date: Wed Aug 4 09:51:34 2010 +0100
28498 display: Handle cursor error paths.
28500 Check that the cursor was allocated before freeing.
28502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28504 commit 38f940dfea494d3093236f065392c431be06ae6e
28505 Author: Chris Wilson <chris@chris-wilson.co.uk>
28506 Date: Wed Aug 4 09:50:14 2010 +0100
28508 display: Tidy backlight initialisation
28510 Mostly whitespace and a single error-code fix.
28512 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28514 commit 2b7263b771d94401cb4ea6cbf4dc7a295eeda7c0
28515 Author: Chris Wilson <chris@chris-wilson.co.uk>
28516 Date: Wed Aug 4 09:46:01 2010 +0100
28518 display: Check for buffer overrun in output name lookup.
28520 The kernel may know about more types than we do, so protect ourselves
28521 from reading from beyond the end of the string array.
28523 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28525 commit 56629166915155628e70e6ec8052c0d220a8bb68
28526 Author: Dave Airlie <airlied@redhat.com>
28527 Date: Wed Aug 4 15:35:41 2010 +1000
28529 intel: add output names for later additions to kernel
28531 commit fe7dee7fe1cf12665bc5ba3923cd1a8f74cbbe57
28532 Author: Chris Wilson <chris@chris-wilson.co.uk>
28533 Date: Sun Aug 1 14:15:09 2010 +0100
28535 Remove the final references to the drmmode prefix
28537 In particular fix the compile regression for intel_do_pageflip().
28539 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28541 commit 4f8b279f32cfda0f3b8fa1542b3fe00051520012
28542 Author: Chris Wilson <chris@chris-wilson.co.uk>
28543 Date: Sun Aug 1 12:35:57 2010 +0100
28545 intel_display: Miscellaneous tidy
28547 A mixture of whitespace and closing of leaks on error paths.
28549 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28551 commit db7cd7b9f0ce8e435c121071eaadf8e6dcebf9e8
28552 Author: Chris Wilson <chris@chris-wilson.co.uk>
28553 Date: Sun Aug 1 11:25:42 2010 +0100
28555 Rename drmmode_display to intel_display
28557 And fixup all the drmmode_* functions to have an intel prefix and
28558 categorise those into intel_mode, intel_crtc, intel_output and
28559 intel_property so that the functions are a little more self-descriptive
28560 and, more importantly, are consistent.
28562 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28564 commit 0be3e95c844247746742805830860ace9f546d99
28565 Author: Kristian Høgsberg <krh@bitplanet.net>
28566 Date: Fri Jul 30 09:39:58 2010 -0400
28568 Remove explicit batchbuffer submit in DRI2 copyregion
28570 Now that we submit from the flush callback chain, we know we'll always
28571 submit before the client receives the reply or event that blocks it from
28572 rendering the next frame.
28574 commit 69d65f9184006eac790efcff78a0e425160e95aa
28575 Author: Kristian Høgsberg <krh@bitplanet.net>
28576 Date: Thu Jul 29 18:31:48 2010 -0400
28578 Submit batch buffers from flush callback chain
28580 There are a few cases where the server will flush client output buffers
28581 but our block handler only catches the most common (before going into select).
28582 If the server flushes client buffers before we submit our batch buffer,
28583 the client may receive a damage event for rendering that hasn't happened yet.
28585 Instead, we can hook into the flush callback chain, which the server will
28586 invoke just before flushing output. This lets us submit batch buffers
28587 before sending out events, preserving ordering.
28589 Fixes 28438: [bisected] incorrect character in gnome-terminal under compiz
28590 https://bugs.freedesktop.org/show_bug.cgi?id=28438
28592 Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
28594 commit b68d4fcab5d671946c187947f5d00aedff939852
28595 Author: Chris Wilson <chris@chris-wilson.co.uk>
28596 Date: Wed Jul 28 13:41:41 2010 +0100
28598 drmmode: Only treat a backlight as connected if it has a non-zero max
28600 Optimistically might help
28601 https://bugs.freedesktop.org/show_bug.cgi?id=29273
28603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28605 commit 938ef4eaec0e406e198105d866f6d2a1be5e0ef8
28606 Author: Kristian Høgsberg <krh@bitplanet.net>
28607 Date: Wed Jul 28 07:57:42 2010 -0400
28609 legacy: Remove long gone use of GlxSetVisualConfigs()
28611 This removes the last dependeny on anything GL/GLX in the driver.
28613 commit fba6651a92af76d7a8d034e1acfdad5023cd311c
28614 Author: Kristian Høgsberg <krh@bitplanet.net>
28615 Date: Tue Jul 27 12:59:39 2010 -0400
28617 Drop use of GL types in the driver
28619 Still used in i810 for building the glx visuals.
28621 commit 34b9d45be6cb0c42f031677a0bda8a5013472ae8
28622 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
28623 Date: Tue Jul 27 09:10:04 2010 -0700
28625 Revert "configure.ac: add xi and gl requirements"
28627 This reverts commit d580fa82a1cf339d2d1fd1055d137c0b23cd04f3.
28629 Oops, these should be server or other package dependencies, we don't
28630 actually need them for xf86-video-intel.
28632 commit d580fa82a1cf339d2d1fd1055d137c0b23cd04f3
28633 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
28634 Date: Mon Jul 26 16:44:19 2010 -0700
28636 configure.ac: add xi and gl requirements
28638 Could probably be earlier versions, but having these here makes packaging
28641 commit f46a8dfce59013ce593c15f7166598d0fff9198f
28642 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
28643 Date: Mon Jul 26 23:02:02 2010 +0200
28645 video: kill do { ... } while (ret != -EINTR) loops
28647 Chris Wilson likes to sprinkle these all over, but in this
28648 case it's just misleading. libdrm already does this for us.
28650 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
28652 commit 1cb69b9a77b7afbb4358757556065e10a6b15ea8
28653 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
28654 Date: Mon Jul 26 23:00:20 2010 +0200
28656 video: kernel overlay needs triple buffering
28658 The kernel overlay code does asynchronous overlay flips. So keep
28659 onto two old buffers, for otherwise the rendering of the next
28660 frame might overwrite the contents of the currently still displaying
28661 one. With ~25fps videos and ~50 Hz screens that's rather unlikely,
28664 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
28666 commit 68df6b2790891683ee2e58daaad34ef17ae344f5
28667 Author: Gaetan Nadon <memsize@videotron.ca>
28668 Date: Fri Jul 23 17:39:37 2010 -0400
28670 simplify Makefile as per-target compilation flags are not needed
28672 Per-target compilation flags (libIntelXvMC_la_CFLAGS) are required
28673 when multiple targets which require different compiler flags,
28674 are build in the same makefile.
28676 Automake issues a command with -c and -o flags which not all compilers
28677 support. The object fles are prefixed with libIntelXvMC_la.
28678 The macro AM_PROG_CC_C_O must then be used to provide this feature
28679 on compilers that do not have it. If not, a warning is issued at make time.
28681 This macros checks for compiler support and if missing, uses a "compile"
28682 script it generates in the package root directory.
28684 Currently the driver uses per-target flags but the macro is missing.
28685 Rather than adding the macro, this patch stops using per-target flags
28686 by using the AM_CFLAGS variable for all targets in the makefile, as
28689 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
28690 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28692 commit 0ae30b6d1f4f6c2f2db9443f1027ea41849ad087
28693 Author: Gaetan Nadon <memsize@videotron.ca>
28694 Date: Wed Jul 21 16:49:04 2010 -0400
28696 config: add comments for main statements
28698 commit b12b2219677d94933774955b50545efceff57279
28699 Author: Gaetan Nadon <memsize@videotron.ca>
28700 Date: Wed Jul 21 16:07:00 2010 -0400
28702 config: replace deprecated use of AC_OUTPUT with AC_CONFIG_FILES
28704 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28706 commit 8b2d65da41461735b1138fdd6e0efc96419cd815
28707 Author: Gaetan Nadon <memsize@videotron.ca>
28708 Date: Wed Jul 21 14:37:41 2010 -0400
28710 config: replace deprecated AC_HELP_STRING with AS_HELP_STRING
28712 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28714 commit 3feeef5294b1a2df05a25f4f4eb6400be6cc4ee1
28715 Author: Gaetan Nadon <memsize@videotron.ca>
28716 Date: Wed Jul 21 14:05:22 2010 -0400
28718 config: replace deprecated AM_CONFIG_HEADER with AC_CONFIG_HEADERS
28720 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28722 commit 74a828e286c46473b8edf6de32c1989c052dd82c
28723 Author: Gaetan Nadon <memsize@videotron.ca>
28724 Date: Wed Jul 21 09:27:42 2010 -0400
28726 config: complete AC_INIT m4 quoting
28728 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28730 commit 33ec66d7006e2a7e287fd77c53a31db49220591a
28731 Author: Gaetan Nadon <memsize@videotron.ca>
28732 Date: Tue Jul 20 21:54:11 2010 -0400
28734 config: remove unrequired AC_SUBST([DRI_CFLAGS])
28736 This macro is called by PKG_CHECK_MODULES
28738 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28740 commit b20b8ac937505310fbe68f07e42cafe4ba88da76
28741 Author: Gaetan Nadon <memsize@videotron.ca>
28742 Date: Tue Jul 20 20:24:42 2010 -0400
28744 config: remove unrequired AC_HEADER_STDC
28747 "This macro is obsolescent, as current systems have conforming
28748 header files. New programs need not use this macro".
28750 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28752 commit 9fe5169fffeee10b1262f160a15c11c283e2faa7
28753 Author: Gaetan Nadon <memsize@videotron.ca>
28754 Date: Tue Jul 20 19:41:31 2010 -0400
28756 config: remove AC_PROG_CC as it overrides AC_PROG_C_C99
28758 XORG_STRICT_OPTION from XORG_DEFAULT_OPTIONS calls
28759 AC_PROG_C_C99. This sets gcc with -std=gnu99.
28760 If AC_PROG_CC macro is called afterwards, it resets CC to gcc.
28762 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28764 commit e6b8811e5bb30e2cd4e35b1fe84d69f18a73ee0e
28765 Author: Gaetan Nadon <memsize@videotron.ca>
28766 Date: Tue Jul 20 18:45:19 2010 -0400
28768 config: update AC_PREREQ statement to 2.60
28770 Unrelated to the previous patches, the new value simply reflects
28771 the reality that the minimum level for autoconf to configure
28772 all x.org modules is 2.60 dated June 2006.
28774 ftp://ftp.gnu.org/gnu/autoconf/autoconf-2.60.tar.gz
28776 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
28778 commit 2d85efaa36791bd441f069b612ee86387353d1ed
28779 Author: Gaetan Nadon <memsize@videotron.ca>
28780 Date: Tue Jul 20 16:15:29 2010 -0400
28782 config: upgrade to util-macros 1.8 for additional man page support
28784 Use MAN_SUBST now supplied in XORG_MANPAGE_SECTIONS
28785 The value of MAN_SUBST is the same for all X.Org packages.
28787 commit 142ffa287236adbc3c5f6ead46c63c4c541ace0c
28788 Author: Chris Wilson <chris@chris-wilson.co.uk>
28789 Date: Thu Jul 22 16:42:26 2010 +0100
28791 video/i915: ValidateGC after setting clip.
28793 Order is important. And ensure that the scratch GC is performing clip by
28798 Bug 29213 - video artifacts if used dualscreen mode
28799 https://bugs.freedesktop.org/show_bug.cgi?id=29213
28801 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28803 commit c57840b272ba88fddf22484929608431879b0fab
28804 Author: Chris Wilson <chris@chris-wilson.co.uk>
28805 Date: Wed Jul 21 09:06:06 2010 +0100
28807 uxa: Check for failed pixmap allocation
28811 Bug 29187 - crash in intel_drv
28812 https://bugs.freedesktop.org/show_bug.cgi?id=29187
28815 0: /usr/bin/X (xorg_backtrace+0x28) [0x466808]
28816 1: /usr/bin/X (0x400000+0x67c79) [0x467c79]
28817 2: /lib/libpthread.so.0 (0x7ff19b297000+0xef60) [0x7ff19b2a5f60]
28818 3: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7ff197986000+0x34684) => uxa/uxa-render.c:841
28820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28822 commit 7a4bfaf42412f373bbc8ebe1536018dcacde7436
28823 Author: Dave Airlie <airlied@redhat.com>
28824 Date: Tue Jul 20 11:30:44 2010 +1000
28826 intel: respect tiling disable.
28828 For testing purposes its nice to know tiling isn't being used anywhere.
28830 Signed-off-by: Dave Airlie <airlied@redhat.com>
28832 commit d48d584a822ae5adae64f6d5f09faa4ac9b8de73
28833 Author: Chris Wilson <chris@chris-wilson.co.uk>
28834 Date: Thu Jul 15 14:37:59 2010 +0100
28836 video: Free the buffers immediately after turning off.
28838 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28840 commit 24bdfe0d5eb4e890e9c63bbb4617efaa0768ab7f
28841 Author: Chris Wilson <chris@chris-wilson.co.uk>
28842 Date: Thu Jul 15 13:54:04 2010 +0100
28844 video: Reuse the old buffers.
28846 After passing the new buffer to the kernel, the old buffer is unpinned
28847 and becomes available for re-use. So keep hold of the old buffer and
28848 swap after a PutImage. This greatly reduces the amount of CPU time
28849 consumed by the kernel on behalf of the video overlay -- by only
28850 allocating two buffers for an entire sequence, we avoid clflushing and
28851 page allocation on every frame.
28853 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28855 commit 2267e5928bd90a6997970584462bd97df447e819
28856 Author: Chris Wilson <chris@chris-wilson.co.uk>
28857 Date: Tue Jul 13 10:36:34 2010 +0100
28859 Workaround a broken container_of define in list.h
28861 commit 798c3a5fc6e3d44c9b94e6bc4e72385a10939f52
28862 Author: Chris Wilson <chris@chris-wilson.co.uk>
28863 Date: Tue Jul 13 10:30:33 2010 +0100
28865 Teardown the bufmgr on shutdown as well.
28867 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28869 commit b2e98227d11a60e90b3ea518f14962f2764ffa68
28870 Author: Chris Wilson <chris@chris-wilson.co.uk>
28871 Date: Tue Jul 13 10:15:34 2010 +0100
28873 Remove the duplicate drmmode prototypes.
28875 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28877 commit 5de1b74d64f807b59c730871d4cb171484db9780
28878 Author: Chris Wilson <chris@chris-wilson.co.uk>
28879 Date: Tue Jul 13 09:34:53 2010 +0100
28881 modes: There may be more than one crtc and output... DESTROY THEM ALL!
28883 In order to cleanup all CRTCs and outputs on shutdown, we need to keep a
28884 list of the individual structures and iterate over that list on
28887 Also, the output and crtcs are configured just once and not for each
28888 screen generation so move the shutdown to the termination and not on
28891 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28893 commit 3a7c25ff8ddd45c9d9eca5cc2228552847ca9e7d
28894 Author: Chris Wilson <chris@chris-wilson.co.uk>
28895 Date: Mon Jul 12 19:47:46 2010 +0100
28897 video: Apply overlay stride errata for i830 and i845
28899 Due to an erratum on these chipsets, the overlay stride must be a
28900 multiple of 256 bytes.
28902 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28904 commit 56e5816252afbc7dd555c308ca678fce7ee584cc
28905 Author: Chris Wilson <chris@chris-wilson.co.uk>
28906 Date: Mon Jul 12 17:26:37 2010 +0100
28908 video: Copy DummyEncoding into each adapter.
28910 As we use the static DummyEncoding and may attempt to modify it for each
28911 adaptor (on each device), we should use copies instead.
28913 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28915 commit e30f0338fbf894fa3fdab20f4f9a97495cd341da
28916 Author: Keith Packard <keithp@keithp.com>
28917 Date: Thu Jul 1 09:56:40 2010 -0400
28919 Destroy screen pixmap on screen close.
28921 This avoids a memory leak on server reset.
28923 Signed-off-by: Keith Packard <keithp@keithp.com>
28924 [ickle: Added comments from Keith that explain the necessity of
28925 destroying the pixmap ourselves and why chaining up in this instance is
28926 not the correct approach.]
28927 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
28929 commit 6e08b0f48f0d6ddaca963de6f4fd3b56f32512c6
28930 Author: Fernando Carrijo <fcarrijo@yahoo.com.br>
28931 Date: Thu Jul 1 07:15:22 2010 -0300
28933 Purge macro NEED_EVENTS
28935 Signed-off-by: Fernando Carrijo <fcarrijo@yahoo.com.br>
28936 Acked-by: Tiago Vignatti <tiago.vignatti@nokia.com>
28937 Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
28939 commit 5ce3f536b7169a1ee9eadf830e61bab230dd7239
28940 Author: Chris Wilson <chris@chris-wilson.co.uk>
28941 Date: Fri Jul 9 14:06:43 2010 +0100
28943 drmmode: Destroy the output on shutdown
28945 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28947 commit 6ff369cd2602497ee11273b50f76d1a7bf25991d
28948 Author: Chris Wilson <chris@chris-wilson.co.uk>
28949 Date: Fri Jul 9 14:00:31 2010 +0100
28951 drmmode: Destroy Crtc on screen shutdown
28955 Bug 26946 - CRTC cursor BO leak in 2D
28956 https://bugs.freedesktop.org/show_bug.cgi?id=26946
28958 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28960 commit 6fba8c449f61341a98a5ae8d97effa6fd0610fd4
28961 Author: Chris Wilson <chris@chris-wilson.co.uk>
28962 Date: Fri Jul 9 12:12:13 2010 +0100
28964 Add support for I854.
28966 I spotted that the kernel knew of the I854, but the pci-id was never
28969 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28971 commit 141e88c8730a099a6ca5eab1350c2e53a680cb0d
28972 Author: Chris Wilson <chris@chris-wilson.co.uk>
28973 Date: Fri Jul 9 12:11:07 2010 +0100
28975 video: forgotten amendment to previous commit.
28977 An extra sanity check to skip the wait if all clipped...
28979 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
28981 commit 272d1c14a39c32ade39b5a8b080a891f2b3d6e8e
28982 Author: Chris Wilson <chris@chris-wilson.co.uk>
28983 Date: Fri Jul 9 10:41:19 2010 +0100
28985 video: apply the crtc box checks from dri.
28987 The dri code is much more careful in ensuring that the scan lines that
28988 is waits for are valid. Copy this code to video, with a bit of work this
28989 can be refactored, and perhaps even teach dri how to handle rotated
28994 Bug 28964 - [i965gm] GPU infinite MI_WAIT_FOR_EVENT while watching video
28996 https://bugs.freedesktop.org/show_bug.cgi?id=28964
28998 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29000 commit 75850e824bd66b961b9ce21696d9090169f7b28f
29001 Author: Evan McClain <aeroevan@gmail.com>
29002 Date: Thu Feb 25 01:51:42 2010 -0500
29004 Add mbp_backlight support.
29006 Acked-by: Julien Cristau <jcristau@debian.org>
29008 commit a2aa4c23f65b03c389c2ba8eac0298dfb205e67d
29009 Author: Dave Airlie <airlied@redhat.com>
29010 Date: Mon Jul 5 14:02:42 2010 +1000
29012 uxa: oops typo in previous commit
29014 commit feff2ec80eeeba5074d19efcddf4867adf639b4f
29015 Author: Dave Airlie <airlied@redhat.com>
29016 Date: Mon Jul 5 09:07:08 2010 +1000
29018 uxa: don't compare planemask with FB_ALLONES.
29020 planemask is an unsigned long initialised to ~0, on 64-bit this is not equal
29021 to an (unsigned int)-1.
29023 Use the macro provided to do this.
29025 Signed-off-by: Dave Airlie <airlied@redhat.com>
29027 commit 7ad9c0bebf2c71db574e0302264859ce53cecd37
29028 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
29029 Date: Fri Jul 2 12:39:27 2010 -0700
29031 add cscope files to gitignore
29033 commit afcd41820d7481912b49093670e9d7a4a6a2338f
29034 Author: Chris Wilson <chris@chris-wilson.co.uk>
29035 Date: Thu Jul 1 22:14:03 2010 +0100
29037 Reduce front buffer stride prior to rejection
29039 If we reject the front buffer because it has too large a stride, repeat
29040 the allocation using untiled for the cases where we can utilize laxer
29041 hardware restrictions.
29043 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29045 commit f8778b66a98d68e622cdf5b2d1fd959c4f7a2b2b
29046 Author: Chris Wilson <chris@chris-wilson.co.uk>
29047 Date: Thu Jul 1 21:58:43 2010 +0100
29049 drmmode: Add missing newlines at the end of log messages.
29051 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29053 commit 690fbd1a64c1e1fd07f7f8a5e39e05a33e4d0721
29054 Author: Chris Wilson <chris@chris-wilson.co.uk>
29055 Date: Thu Jul 1 21:55:04 2010 +0100
29057 drmmode: Use a copy of the converted mode on resize
29059 Avoid a potential use-after-free of the copied mode string by reusing
29060 the converted kernel mode on resize.
29062 ==19897== Invalid read of size 8
29063 ==19897== at 0x661C330: ??? (strcpy.S:1308)
29064 ==19897== by 0x8618AE7: drmmode_set_mode_major (drmmode_display.c:293)
29065 ==19897== by 0x8618E6F: drmmode_xf86crtc_resize (drmmode_display.c:1299)
29066 ==19897== by 0x529A77: xf86RandR12ScreenSetSize (xf86RandR12.c:708)
29067 ==19897== by 0x4BD528: ProcRRSetScreenSize (rrscreen.c:301)
29068 ==19897== by 0x42B820: Dispatch (dispatch.c:432)
29069 ==19897== by 0x4254C9: main (main.c:289)
29070 ==19897== Address 0x72e91e0 is 0 bytes inside a block of size 9 free'd
29071 ==19897== at 0x4C23DBC: free (vg_replace_malloc.c:325)
29072 ==19897== by 0x48424F: xf86DeleteMode (xf86Mode.c:1921)
29073 ==19897== by 0x4942B7: xf86ProbeOutputModes (xf86Crtc.c:1572)
29074 ==19897== by 0x5290BB: xf86RandR12GetInfo12 (xf86RandR12.c:1551)
29075 ==19897== by 0x5313AE: RRGetInfo (rrinfo.c:202)
29076 ==19897== by 0x4BCCAA: rrGetScreenResources (rrscreen.c:337)
29077 ==19897== by 0x42B820: Dispatch (dispatch.c:432)
29078 ==19897== by 0x4254C9: main (main.c:289)
29080 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29082 commit 772f8236d50725f0b330508616b4f2a9a910662a
29083 Author: Chris Wilson <chris@chris-wilson.co.uk>
29084 Date: Wed Jun 30 13:58:05 2010 +0100
29086 dri: Handle errors during GetBuffers() gracefully.
29088 Unwind the array of Pixmaps already allocated and report failure for the
29089 old dri GetBuffers() path.
29091 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29093 commit 17884af4eda01c05e7429ed6a195f3e4ceeff892
29094 Author: Chris Wilson <chris@chris-wilson.co.uk>
29095 Date: Mon Jun 28 17:44:45 2010 +0100
29097 Repair the damage to 'make distcheck' after splitting out i810
29099 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29101 commit 28c0ca676c47e7e38fabdd9ef24a70bd26701f33
29102 Author: Chris Wilson <chris@chris-wilson.co.uk>
29103 Date: Fri Jun 25 13:40:22 2010 +0100
29105 Remove unused inclusion of <sys/mman.h>
29107 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29109 commit 2306e675c48bc00a8ca1f4545bb88c97d6ef4fa6
29110 Author: Chris Wilson <chris@chris-wilson.co.uk>
29111 Date: Fri Jun 25 13:32:02 2010 +0100
29113 configure: Remove check for unused function, mprotect
29115 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29117 commit 6695296ca28da7c4f551b0e80a8d06b557d59a5e
29118 Author: Chris Wilson <chris@chris-wilson.co.uk>
29119 Date: Fri Jun 25 13:31:11 2010 +0100
29121 configure: Remove unused checks for xext
29123 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29125 commit 5c663ce8442bd3b69c96e29061c57cc565fca91d
29126 Author: Chris Wilson <chris@chris-wilson.co.uk>
29127 Date: Sun Jun 20 15:01:16 2010 +0100
29129 Rename common infrastructure to the intel namespace.
29131 After splitting out the i810 driver into its own legacy directory, we
29132 can identify the common routines not as i830 but as intel. This
29133 clarifies the code which *is* i830 specific.
29135 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29137 commit 797d173a9af12055ba2609293182b616dd673ef4
29138 Author: Chris Wilson <chris@chris-wilson.co.uk>
29139 Date: Sun Jun 20 12:52:10 2010 +0100
29141 i810: Move into a legacy directory.
29143 The driver is still built but is no longer under active development so
29144 move it and supporting files to a new directory.
29146 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29148 commit cd61531a78b9a2cbdd6019199c43258f83e6e2ce
29149 Author: Chris Wilson <chris@chris-wilson.co.uk>
29150 Date: Sun Jun 20 21:03:44 2010 +0100
29152 Remove unused configure option: --enable-video-debug
29154 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29156 commit ca16e3133658504b68a6b5c83bb8712899016544
29157 Author: Carl Worth <cworth@cworth.org>
29158 Date: Thu Jun 24 14:07:15 2010 -0700
29160 Update version to 2.12.0
29162 For the imminent 2.12.0 release.
29164 commit 8caf1564a75d65636d9b7a4da03219bb302bb37b
29165 Author: Carl Worth <cworth@cworth.org>
29166 Date: Thu Jun 24 14:06:43 2010 -0700
29168 NEWS: Add notes for the 2.12.0 release.
29172 commit b58a6a39c1568800938eb0e3ebc7664683b61200
29173 Author: Chris Wilson <chris@chris-wilson.co.uk>
29174 Date: Wed Jun 23 23:56:44 2010 +0100
29176 uxa: Fallback to pixman if source is out-of-bounds
29178 If the source is outside the drawable, then CopyArea will fail to
29179 initialise the source correctly. The simplest fix in this case is to
29180 fallback to pixman to generate the source texture.
29184 Bug 28497 - Graphics corruption after opening a specific website
29185 https://bugs.freedesktop.org/show_bug.cgi?id=28497
29187 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29189 commit 6d33e578de4e23336ac69cc3c5d0935a65d4dda1
29190 Author: Chris Wilson <chris@chris-wilson.co.uk>
29191 Date: Wed Jun 23 21:28:50 2010 +0100
29193 Limit maximum tiled stride to 8k and untiled to 32k.
29195 Tiling on gen 2/3 hardware is only supported for pitches up to 8192
29196 bytes, so above this limit the surface will be untiled and we will no
29197 longer have to comply with the power-of-two pitch alignment. So
29198 disabling tiling for these too wide surface should ~halve the memory
29199 requirement for the full surface.
29201 Also the absolute limit for the 2D blitter is 32,768 bytes. The
29202 documentation says "up to 32,768 bytes" and my PineView box was
29203 malfunction with a surface stride of 32,768 so set the limit to be
29208 Bug 28497 - Graphics corruption after opening a specific website
29209 https://bugs.freedesktop.org/show_bug.cgi?id=28497
29211 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29213 commit 5bf470bd38b1f6a7a540585186a54c9dbbca98f9
29214 Author: Chris Wilson <chris@chris-wilson.co.uk>
29215 Date: Mon Jun 21 22:28:58 2010 +0100
29219 Oops, I spent more time discussing these flushing bugs than I spent
29220 paying attention to what I was actually doing.
29222 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29224 commit 0203cf91b54ee8a7cea8560e559288ee9b6e8554
29225 Author: Chris Wilson <chris@chris-wilson.co.uk>
29226 Date: Mon Jun 21 22:25:08 2010 +0100
29228 Do not clear need_mi_flush within the batch.
29230 This is a situation that should not be possible, need_mi_flush being
29231 true but the list of pending flush pixmaps being clear. However, an
29232 earlier bug in doing just that revealed this minor bug. So for
29233 correctness, be careful not to clear need_mi_flush without emitting a
29236 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29238 commit 5107b6fa26ecfdbdd60b869a86765c9c484db3a2
29239 Author: Chris Wilson <chris@chris-wilson.co.uk>
29240 Date: Mon Jun 21 22:21:58 2010 +0100
29242 i965: Mark the render target as dirty within composite_setup()
29244 The key difference between i965 and earlier, is that the surfaces passed
29245 to the samplers through an indirect table and so the batch and render
29246 target was not being marked dirty by the relocation (since the
29247 relocation only happens within prepare_composite() which may have been
29248 in another batch.) Simply call intel_pixmap_mark_dirty() when binding
29249 the sampler table into the batch to ensure that the dirty is tracked
29252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29254 commit bebd64d82121d2b61893e1d0adebda03f4faea61
29255 Author: Chris Wilson <chris@chris-wilson.co.uk>
29256 Date: Mon Jun 21 22:05:19 2010 +0100
29258 Also submit any pending flush for *this* batch in the BlockHander.
29260 We still need to submit an additional flush if we have further writes
29261 since the last flush.
29263 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29265 commit c4d200517755e6b0f48d6c6a6ba8dd0ddc883d8b
29266 Author: Chris Wilson <chris@chris-wilson.co.uk>
29267 Date: Mon Jun 21 21:57:21 2010 +0100
29269 Only append the pixmap to the flushing list if we are writing to it.
29271 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29273 commit c942585098ac84ae461821cbb8f52dedce7a0da1
29274 Author: Chris Wilson <chris@chris-wilson.co.uk>
29275 Date: Mon Jun 21 21:45:04 2010 +0100
29277 Emit the flush after a potential draw from the BlockHandler.
29279 As the batch submit may not trigger further drawing through flushing the
29280 vertices, pass the requirement to emit the flush down to the submission
29281 routine so that the flush can be appended after the final commands.
29283 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29285 commit e8783869ad55d337601b6f6a51c02f6576c64f38
29286 Author: Chris Wilson <chris@chris-wilson.co.uk>
29287 Date: Sun Jun 20 00:09:20 2010 +0100
29289 uxa: Apply the source offsets to the pixmap source, not target.
29291 A slight confusion in computing the correction image location resulted
29292 in the application of the source offsets to the pixel location in the
29293 target and not in the source as intended.
29295 Fixes the visual corruption of the scrollbar in Chromium, and hopefully
29296 the crash reported by Robert Hooker when starting gdm after plymouth.
29298 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29300 commit 4b7142baa0b3bf6f38843d06aadc579d8624cefc
29301 Author: Chris Wilson <chris@chris-wilson.co.uk>
29302 Date: Sat Jun 19 12:30:23 2010 +0100
29304 uxa: Enable SHM pixmaps
29306 Now with streaming uploads and downloads for composite operations in
29307 place, shared memory pixmaps are no longer that dire performance wise.
29308 With careful use these can in fact be the most efficient means of
29309 transfer between a wholly software renderer in the client and a backing
29310 store. For instance, Chromium renders internally to an ARGB32 image
29311 buffer and uses a shared pixmap to composite dirty regions into the
29312 backing store. Thereby using the GPU to either perform the blit or the
29313 format conversion. Enabling shared pixmaps, reduces our CPU overhead
29314 whilst scrolling by a factor of 5 or so.
29316 And this is achieved simply by deleting obsolete code!
29318 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29320 commit d748f8e6fcebdf5b5b00c90b6c3d519560ee2928
29321 Author: Chris Wilson <chris@chris-wilson.co.uk>
29322 Date: Thu Jun 17 18:29:55 2010 +0100
29324 uxa: Use accelerated get_image for copying to !offscreen
29326 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29328 commit 78ee25f005fac9e9cc85425aa3fca8c396d015ea
29329 Author: Chris Wilson <chris@chris-wilson.co.uk>
29330 Date: Thu Jun 17 18:37:32 2010 +0100
29332 uxa: Match depth 30 to format.
29334 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29336 commit af5c4fc96db56a14bf6347774d8e8f23b38c5967
29337 Author: Chris Wilson <chris@chris-wilson.co.uk>
29338 Date: Sat Jun 19 13:36:20 2010 +0100
29340 uxa: Check for allocation failure.
29342 Check for the NULL Picture prior to passing it to the backends for
29345 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29347 commit be55066c6481b4c5e2cd39ef1c0f3be88cae0c93
29348 Author: Chris Wilson <chris@chris-wilson.co.uk>
29349 Date: Thu Jun 17 14:23:21 2010 +0100
29351 i830: Remove domain tracking from pixmaps.
29353 The 4 integers can be reduced to a single boolean value, so do so.
29355 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29357 commit c187da9a24e98cbdf86f0f0dfbdcd9025c8a4c76
29358 Author: Chris Wilson <chris@chris-wilson.co.uk>
29359 Date: Wed Jun 16 13:16:28 2010 +0100
29361 i830: GetImage acceleration.
29363 The presumption is that we wish to keep the target hot, so
29364 copy to a new bo and move that to the CPU in preference to
29365 causing ping-pong of the original.
29367 Also the gpu is much faster at detiling.
29370 400000 trep @ 0.1128 msec ( 8860.0/sec): GetImage 10x10 square
29371 18000 trep @ 1.3839 msec ( 723.0/sec): GetImage 100x100 square
29372 800 trep @ 30.0987 msec ( 33.2/sec): GetImage 500x500 square
29375 180000 trep @ 0.1478 msec ( 6770.0/sec): GetImage 10x10 square
29376 60000 trep @ 0.4545 msec ( 2200.0/sec): GetImage 100x100 square
29377 4000 trep @ 8.0739 msec ( 124.0/sec): GetImage 500x500 square
29379 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29381 commit 0e0101758438debf98f989d815989b45e78cf5f6
29382 Author: Chris Wilson <chris@chris-wilson.co.uk>
29383 Date: Tue Jun 15 12:48:57 2010 +0100
29385 i830: Tidy i830_uxa_put_image()
29387 Use a single code path to upload the image data after selecting the
29388 right bo, and take advantage of pwrite() when possible.
29392 Bug 28569 - [i965] IGN's flash-based video player crashes X
29393 https://bugs.freedesktop.org/show_bug.cgi?id=28569
29395 Bug 28573 - [i965] Fullscreen flash and windowed SDL games fail to
29397 https://bugs.freedesktop.org/show_bug.cgi?id=28573
29399 Reported-and-tested-by: Brian Rogers <brian@xyzw.org>
29400 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29402 commit 2ff7a2fc9d939554502e9adec89cae5597a9e82c
29403 Author: Chris Wilson <chris@chris-wilson.co.uk>
29404 Date: Thu Jun 17 13:56:30 2010 +0100
29406 i915: Force the emission of BUF_INFO on every composite_setup
29408 We should be able to eliminate these as the drawable remains unchanged.
29409 However, the implicit flush of BUF_INFO fixes the rendering in KDE.
29410 Alternatively, we need an MI_FLUSH | INHIBIT_RENDER_CACHE_FLUSH between
29411 composites. (Note that it is not stale cache data causing the rendering
29412 corruption and that a pipelined flush is not sufficient either.) Also,
29413 having tried varies points at which to flush, the only place where the
29414 flush is effective seems to be between composite operations - that is a
29415 flush after 2D is not sufficient.
29417 Reported-by: Vasily Khoruzhick <anarsoul@gmail.com>
29418 Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
29419 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29421 commit a25573d5c47ebea34c076075e1993233d7db2b4f
29422 Author: Chris Wilson <chris@chris-wilson.co.uk>
29423 Date: Tue Jun 15 20:26:19 2010 +0100
29425 drmmode: Use the tiled stride for the rotated pixmap.
29427 After d41684d5459 we now allocate all framebuffers as tiled bo, and so
29428 we must be careful to use the appropriate stride as returned from the
29429 allocation, instead of assuming that it is just an aligned width.
29433 Bug 28461 - screen rotation results in corrupted output.
29434 https://bugs.freedesktop.org/show_bug.cgi?id=28461
29436 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29437 Reported-by: Till Matthiesen <entropy@everymail.net>
29439 commit 23ddcf45344936da2931a3bf25bd187a9ea28344
29440 Merge: 254bcc4 ffc6897
29441 Author: Julien Cristau <jcristau@debian.org>
29442 Date: Tue Jun 15 17:46:52 2010 +0100
29444 Merge commit '2.11.901'
29446 commit 254bcc40fca6ed9aa0b71f2115c654667f16aec6
29447 Author: Chris Wilson <chris@chris-wilson.co.uk>
29448 Date: Tue Jun 15 10:24:40 2010 +0100
29450 configure: Bump drm requirement to 2.4.21
29452 We depend upon the recent introduction of drm_intel_bo_is_reusable().
29454 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29455 Reported-by: Julien Cristau <jcristau@debian.org>
29457 commit ffc689767fa56f2c7710e6f396441c3d8a24ea18
29458 Author: Carl Worth <cworth@cworth.org>
29459 Date: Mon Jun 14 17:35:20 2010 -0700
29461 Update version to 2.11.901
29463 For the 2.12 RC1 snapshot we're making now.
29465 commit 68c8b3f12482489f2e2e8040fd6728f1a2b9e420
29466 Author: Carl Worth <cworth@cworth.org>
29467 Date: Mon Jun 14 17:34:38 2010 -0700
29469 NEWS: Add notes for 2.11.901
29471 This is the 2.12 RC1 snapshot we're making now.
29473 commit 8899cd03e2e7ee673e8b3b2a9d2873839cdc1bf8
29474 Author: Carl Worth <cworth@cworth.org>
29475 Date: Thu Jun 10 15:38:54 2010 -0700
29477 Fix to require libdrm 2.4.19 or newer.
29479 This is necessary for the drm_intel_bo_emit_reloc_fence function which
29480 did not exist in 2.4.16 which is all we were requiring earlier.
29482 commit 995a4b2b1d09c31672d9258a8ac732dcf9a8fe9f
29483 Author: Chris Wilson <chris@chris-wilson.co.uk>
29484 Date: Mon Jun 14 12:15:34 2010 +0100
29486 i965: Sanity check ComponentAlpha status in prepare_composite
29490 Bug 28446 - Garbled Font with Mathematica 7
29491 https://bugs.freedesktop.org/show_bug.cgi?id=28446
29493 Rewriting the glyphs to render to the destination directly and removing
29494 the more expensive multiple invocations of CompositePicture per picture
29495 was a great performance boost -- except that it needs special handling
29496 in the backend in order to not fallback. Having done so for i915, I
29497 neglected to ensure the sanity checking in i965_prepare_composite() was
29498 sufficient. As it turns out, it was not and so we misrendered CA-glyphs
29499 when rendering directly to the destination. This causes us to fallback
29500 properly, but is a performance regression as we no longer try the 2-pass
29501 magic helper before resorting to s/w. At the moment, I'd rather live
29502 with the temporary regression and fix i965 to do the same magic as i915,
29503 as it critical to fixing the severe performance issues currently
29504 crippling i965, as I believe that this regression only affects the
29505 minority of applications (incorrect, as it turns out, as the glyphs are
29506 overlapping) rendering directly to the destination.
29508 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29510 commit 84d65bace521ad80b9bede4ae9e3e2a8d0668b90
29511 Author: Chris Wilson <chris@chris-wilson.co.uk>
29512 Date: Mon Jun 14 11:31:40 2010 +0100
29514 Compile fix for alternate list.h from xserver-1.9
29516 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29518 commit 00a54eee8f29ced610ef9d65b077e64bc65655e9
29519 Author: Gaetan Nadon <memsize@videotron.ca>
29520 Date: Sat Jun 12 20:24:10 2010 -0400
29522 COPYING: replace stub file with actual Copyright notices.
29524 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
29526 commit 94217ed5f5a3726db3fad6cac2132d99ed45c154
29527 Author: Chris Wilson <chris@chris-wilson.co.uk>
29528 Date: Sat Jun 12 17:21:44 2010 +0100
29530 uxa: Always clip glyphs to destination.
29532 Even if there is only a single clip rect, since the clip may be smaller
29533 than the drawing rectangle on the destination we need to actually
29534 compute the clipped glyph rectangle.
29536 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29538 commit 5a0a8a1cf6d9b0616d6a097e783f2aa318b45736
29539 Author: Chris Wilson <chris@chris-wilson.co.uk>
29540 Date: Thu Jun 10 22:56:34 2010 +0100
29542 i830: Limit disabling acceleration following EIO to !i965
29544 Following a conversation with Owain G. Ainsworth, it was decided that
29545 the second best approach to handling a wedged GPU was to hope that the
29546 kernel could successfully reset it, which currently is only possible for
29547 i965 and later chipsets.
29549 The best approach is of course to prevent such hangs from ever occurring
29550 in the first place.
29552 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29554 commit 3bf4ca2cdc2493e209bbb9c597bfa17ed7fcf5dc
29555 Author: Chris Wilson <chris@chris-wilson.co.uk>
29556 Date: Wed Jun 9 12:02:21 2010 +0100
29558 i830: Only emit the disabling GPU error message once.
29560 But emit the warning about rendering corruption every time for the
29561 transient errors like out-of-memory.
29563 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29565 commit 35a12f029005d8f432755fb78f10c9d48ea2f347
29566 Author: Chris Wilson <chris@chris-wilson.co.uk>
29567 Date: Wed Jun 9 11:47:37 2010 +0100
29569 Fallback implementation for trapezoids for hung GPUs.
29571 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29573 commit 8c1a8d22978f143044cc3cd0ff2ade5736ee9d48
29574 Author: Chris Wilson <chris@chris-wilson.co.uk>
29575 Date: Wed Jun 9 10:03:08 2010 +0100
29577 Revert "xp:trapezoids"
29579 This reverts commit f429fb9d872950705e11171d0e7407fb7673c786.
29581 An experimental patch I forgot was on my main branch as I was bugfixing.
29584 commit 994aa1ef571ac2ab05f6588f9068d33b536c7a37
29585 Author: Chris Wilson <chris@chris-wilson.co.uk>
29586 Date: Wed Jun 9 09:59:36 2010 +0100
29588 uxa: Handle all-clipped out case with destination glyphs.
29590 Fixes the crash reported in:
29592 Bug 28446 - Garbled Font with Mathematica 7
29593 https://bugs.freedesktop.org/show_bug.cgi?id=28446
29595 pDst=0x3d663c0, src_x=0, src_y=0, xDst=142, yDst=112, nlist=0,
29596 list=0x7fffea026580, glyphs=0x7fffea025d88, extents=0x0)
29597 at uxa-glyphs.c:809
29605 this_atlas = 0x2456d00
29611 uxa_screen = 0x2443eb0
29612 src_pixmap = 0x37c29e0
29613 dst_pixmap = 0x45ddbf0
29614 localSrc = 0x361a450
29615 glyph_atlas = 0x2456d00
29620 box = {x1 = 23152, y1 = -5630, x2 = 32767, y2 = 0}
29621 __PRETTY_FUNCTION__ = "uxa_glyphs_to_dst"
29623 Though the meat of that bug regarding the incorrect remains unsolved.
29625 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29627 commit f429fb9d872950705e11171d0e7407fb7673c786
29628 Author: Chris Wilson <chris@chris-wilson.co.uk>
29629 Date: Tue Jun 8 10:24:51 2010 +0100
29633 commit 0776a42b70f2de7b7d7d8804046c79442da1cb8a
29634 Author: Chris Wilson <chris@chris-wilson.co.uk>
29635 Date: Tue Jun 8 10:43:24 2010 +0100
29639 commit d41684d54592cf93554a4d6534e7ea74562b1798
29640 Author: Eric Anholt <eric@anholt.net>
29641 Date: Mon Jun 7 11:18:09 2010 -0700
29643 Allocate rotate shadow buffers using the usual framebuffer allocator.
29645 This means we can get tiling on them, which should significantly boost
29646 performance, and also allow for FBC.
29648 commit b5c9de10ba3a811f312e622b97ee5e601bd3b74f
29649 Author: Eric Anholt <eric@anholt.net>
29650 Date: Sun Jun 6 23:41:23 2010 -0700
29652 Allocate a correctly sized framebuffer when tiling by using libdrm's support.
29654 When I made libdrm stop overallocating so much memory for the purpose
29655 of bo caching, things started scribbling on the bottom of my
29656 frontbuffer (and vice versa, leading to GPU hangs). We had the usual
29657 mistake of size = tiled_pitch * height instead of size = tiled_pitch *
29658 tile_aligned_height.
29660 commit e6acbc763229fd5b5b2cc1d65136404d02ac4655
29661 Author: Chris Wilson <chris@chris-wilson.co.uk>
29662 Date: Mon Jun 7 18:04:51 2010 +0100
29664 uxa: Setup acceleration functions prior to the damage layer
29666 We need to install the acceleration functions so that they are wrapped
29667 by the Damage layer. This fixes the corruption under a compositing WM
29668 introduced in commit 8700673157fdd3a87ad5150f2f30823261fec519.
29670 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29671 Reported-and-tested-by: Arkadiusz Miśkiewicz <arekm@maven.pl>
29673 commit 1788b16eb26580624dd3df96ed84f68d40e78ed7
29674 Author: Chris Wilson <chris@chris-wilson.co.uk>
29675 Date: Mon Jun 7 15:31:47 2010 +0100
29677 i915: Fix typo from previous commit.
29679 A trivial change, I thought, having tested it before rebasing, unworthy
29680 even of a perfunctory compile test. How wrong I was.
29682 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29684 commit d9bc36ae03e3c56ab165d24fe4bdc6688fd62b6d
29685 Author: Chris Wilson <chris@chris-wilson.co.uk>
29686 Date: Mon Jun 7 15:28:18 2010 +0100
29688 i915: Remove screen size limit from video setup.
29690 The i915 textured video routine know how to handle drawing on an output
29691 larger than the 3D pipe, so allow them to do so.
29693 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29695 commit 6555ef5fd16c9e4fe733caf9e831f29f97462278
29696 Author: Chris Wilson <chris@chris-wilson.co.uk>
29697 Date: Mon Jun 7 13:16:34 2010 +0100
29699 i915: Replace structure passing with macros for shader generation.
29701 gcc is horribly bad at collapsing the constants:
29703 text data bss dec hex filename
29704 282336 8720 256 291312 471f0 intel_drv.so.old
29705 269280 8720 256 278256 43ef0 intel_drv.so
29707 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29709 commit d56ea7a852d7090360fe080acec268de55ee908d
29710 Author: Chris Wilson <chris@chris-wilson.co.uk>
29711 Date: Mon Jun 7 00:20:35 2010 +0100
29713 Use the direct dixGevPrivate() API when available
29715 This is quicker and smaller than the old indirect function call to
29716 dixLookupPrivate().
29718 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29720 commit 8700673157fdd3a87ad5150f2f30823261fec519
29721 Author: Chris Wilson <chris@chris-wilson.co.uk>
29722 Date: Mon Jun 7 00:16:18 2010 +0100
29724 Adapt glyphs for changes in devPrivates API
29726 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29728 commit 42ddc39430a10513c49a415ddf1a0dc5fa52089e
29729 Author: Keith Packard <keithp@keithp.com>
29730 Date: Fri Apr 30 13:33:54 2010 -0700
29732 Adapt to DevPrivate API changes
29734 This allows the driver to be built against either the old or new
29737 Signed-off-by: Keith Packard <keithp@keithp.com>
29739 commit 2c1fda08e889cad07acb452230da06f9c383d21c
29740 Author: Eric Anholt <eric@anholt.net>
29741 Date: Fri Jun 4 16:04:37 2010 -0700
29743 Use libc instead of deprecated libc wrappers for malloc/calloc/free.
29745 commit b586624d4f2908d2a998ba87fe0ae31c10f46b91
29746 Author: Chris Wilson <chris@chris-wilson.co.uk>
29747 Date: Sun Jun 6 17:11:40 2010 +0100
29749 uxa: Force fallback for copies.
29751 All but uxa_copy_window() perform the preliminary checks for whether
29752 acceleration is available. The simplest method for adding the fallback
29753 for uxa_copy_window() seems to be to add it in the core copy function,
29756 This allows X to survive a little longer once we encounter a GPU hang.
29758 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29760 commit 6db1e5231b7a0e79611f771d4efea686f7849e04
29761 Author: Chris Wilson <chris@chris-wilson.co.uk>
29762 Date: Wed Jun 2 20:43:49 2010 +0100
29764 dri: Protect against NULL dereference following GPU hang.
29768 Bug 28361 - "glresize" causes server segfault with single buffering.
29769 https://bugs.freedesktop.org/show_bug.cgi?id=28361
29771 [ 14528.767] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Input/output error.
29772 [ 14528.767] (EE) intel(0): Disabling acceleration.
29773 [ 14528.788] Backtrace:
29774 [ 14528.858] 0: /usr/bin/X (xorg_backtrace+0x28) [0x491818]
29775 [ 14528.858] 1: /usr/bin/X (0x400000+0x65ca9) [0x465ca9]
29776 [ 14528.858] 2: /lib/libpthread.so.0 (0x7f9df2dc9000+0xedf0) [0x7f9df2dd7df0]
29777 [ 14528.858] 3: /usr/local/lib/libdrm_intel.so.1 (drm_intel_bo_flink+0x0) [0x7f9defd60c60]
29778 [ 14528.858] 4: /usr/local/lib/xorg/modules/drivers/intel_drv.so (0x7f9deff6a000+0x2fdfd) [0x7f9deff99dfd]
29779 [ 14528.858] 5: /usr/lib/xorg/modules/extensions/libdri2.so (0x7f9df01b8000+0x19e7) [0x7f9df01b99e7]
29780 [ 14528.858] 6: /usr/lib/xorg/modules/extensions/libdri2.so (0x7f9df01b8000+0x1fdb) [0x7f9df01b9fdb]
29781 [ 14528.858] 7: /usr/lib/xorg/modules/extensions/libdri2.so (DRI2GetBuffersWithFormat+0x10) [0x7f9df01ba250]
29782 [ 14528.858] 8: /usr/lib/xorg/modules/extensions/libdri2.so (0x7f9df01b8000+0x3834) [0x7f9df01bb834]
29783 [ 14528.858] 9: /usr/bin/X (0x400000+0x2fc2c) [0x42fc2c]
29784 [ 14528.858] 10: /usr/bin/X (0x400000+0x24da5) [0x424da5]
29785 [ 14528.858] 11: /lib/libc.so.6 (__libc_start_main+0xe6) [0x7f9df1d60a26]
29786 [ 14528.858] 12: /usr/bin/X (0x400000+0x24959) [0x424959]
29787 [ 14528.858] Segmentation fault at address 0x20
29788 [ 14528.858] Fatal server error:
29789 [ 14528.858] Caught signal 11 (Segmentation fault). Server aborting
29791 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29793 commit 2989f51caf3134460c2551de597e7e54fe74ee92
29794 Author: Chris Wilson <chris@chris-wilson.co.uk>
29795 Date: Tue Jun 1 23:08:02 2010 +0100
29797 i830: Remove unused coord-adjust.
29799 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29801 commit dc402334f4e9b0de624bc89cd77eae4ec7cf1708
29802 Author: Chris Wilson <chris@chris-wilson.co.uk>
29803 Date: Tue Jun 1 23:03:19 2010 +0100
29805 i915: Centre sampling.
29807 Use centre sampling of textures to match pixman, and remove numerous
29808 off-by-one and visual artefacts when rendering. The classic example for
29809 this is cairo/text/xcomposite-projection where the edge of the rotated
29810 rectangle is jaggy due to the incorrect sample position.
29814 Bug 16917 - [i915] Blur on y-axis also when only x-axis is scaled
29816 https://bugs.freedesktop.org/show_bug.cgi?id=16917
29818 And about 15 tests from the Cairo test suite.
29820 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29822 commit f74b3f82bab11463a0f4de9f614fc6aa1492ef24
29823 Author: Chris Wilson <chris@chris-wilson.co.uk>
29824 Date: Tue Jun 1 22:31:35 2010 +0100
29826 i915; Avoid the implicit flush on changing BUF_INFO
29828 3DSTATE_BUF_INFO is an implicit flush of the piepline, so avoid emitting
29829 that and associated state unless the destination pixmap has actually
29830 changed. This is a win of around 3-5% for cairo-perf-trace, notably for
29833 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29835 commit f2272402035574c206a0e3383c55373c440fd928
29836 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
29837 Date: Tue Jun 1 13:46:15 2010 -0700
29839 DRI2: fix new buffer exchange check
29841 Chris's new buffer exchange check is a good one, but we don't want to
29842 hit the immediate blit fallback path if it fails. We still want to
29843 schedule a blit for sometime in the future, and we need to use it
29844 wherever an exchange might occur (like the secondary flip check or the
29845 currently disabled CanExchange check).
29847 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=28252.
29849 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
29851 commit a386a003e7b0ff7100434dccd51fedd822454a6c
29852 Author: Chris Wilson <chris@chris-wilson.co.uk>
29853 Date: Mon May 31 22:18:43 2010 +0100
29855 uxa: Spans, try again to get the early break correct.
29857 Trigger happy bug fixing. The sign *was* right, the endpoint was wrong.
29859 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29861 commit 1672ee0421a99b745d65d72a3818de01d8100b6b
29862 Author: Chris Wilson <chris@chris-wilson.co.uk>
29863 Date: Mon May 31 22:08:43 2010 +0100
29865 uxa: Sign reversal on early break from spans passing the YXband
29867 Introduced with e5c971e7639095d38da3518a5dc404b708d45cfb.
29869 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29871 commit cd38b705be84901a38380dc2a1b90889b553d565
29872 Author: Chris Wilson <chris@chris-wilson.co.uk>
29873 Date: Mon May 31 17:23:36 2010 +0100
29875 Disable acceleration if we detect a hardware error.
29877 This is wildly optimistic, but it should work in a surprising number of
29878 error situations and some output in those cases will be hopefully be
29879 better than none...
29881 If we submit a batchbuffer and the kernel reports the GPU is hung (which
29882 will be caused by an earlier execbuffer, and so the kernel should have
29883 had enough time to determine whether or not it could reset the GPU) then
29884 disable any further attempt to accelerate gfx and force fallbacks to map
29885 the buffers and use the CPU. We cannot normally map any more buffers if
29886 the GPU is hung, so only those already mapped prior to the hang can be
29887 written to, or those allocated in system memory. However, we can expect
29888 that the framebuffer is already mapped, and so have a reasonable
29889 expectation to continue to see the display update.
29891 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29893 commit 5fff430046db2030f89e49beb66c6476dc3fcd4c
29894 Author: Chris Wilson <chris@chris-wilson.co.uk>
29895 Date: Sat May 29 17:35:00 2010 +0100
29899 Rewrite glyph rendering to avoid the intermediate buffer, accumulating
29900 the glyph rectangles directly in the backend composite routines. And
29901 modify the glyph cache routines to fully utilise the allocated size of
29902 the tiled buffer on older hardware. To do this we alias all glyph sizes
29903 into the same texture using a technique suggested by Keith Packard.
29906 885/856-> 1150/1110 kglyph/s (aa/rgb)
29908 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29910 commit d31abccd41c417338aac7c681e8bc6bd187b1843
29911 Author: Chris Wilson <chris@chris-wilson.co.uk>
29912 Date: Mon May 31 12:14:23 2010 +0100
29914 i915: Support textured video on an extended desktop.
29916 Handle rendering textured video onto an extended desktop (>2048) by
29917 using a temporary pixmap. Note that we still cannot handle rendering to
29918 a greater than 2048 destination region, for that we will need to tile.
29919 Hmm, time to request a 2560x1600, 10bpc monitor...
29921 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29923 commit 2cfd5bc134f0dd86ea714594d61f6d5eb29019ce
29924 Author: Chris Wilson <chris@chris-wilson.co.uk>
29925 Date: Sat May 29 17:55:19 2010 +0100
29927 dri: Compilation fix.
29929 17:53 < arekm> ickle: i830_dri.c:630:28: error: ‘DrawableRec’ has no member named ‘bpp’
29930 17:53 < arekm> ickle: i830_dri.c:630:57: error: ‘DrawableRec’ has no member named ‘bpp’
29932 * sigh. I need to fix this machine to have the right version of the
29935 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29937 commit e2615cdeef078dbd2e834b68c437f098a92b941d
29938 Author: Chris Wilson <chris@chris-wilson.co.uk>
29939 Date: Sat May 29 16:37:12 2010 +0100
29941 dri: Only flip if the front and back pixmaps match.
29943 An unredirected window (thanks Michel for the reminder) is backed by the
29944 Screen pixmap, and so uses a reference of that as its front buffer. The
29945 back buffer is a pixmap appropriately sized for the drawable. When the
29946 application requests to swap its buffers, obviously we cannot simply
29947 exchange the front and back buffer as they do not match, but need to copy
29948 the appropriate region from the back to the front.
29950 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29952 commit 8b2039187fc23f5e9dddf4081f8b1ab6f4c01416
29953 Author: Chris Wilson <chris@chris-wilson.co.uk>
29954 Date: Sat May 29 12:14:55 2010 +0100
29956 Revert "dri: Use size from backing pixmap when creating buffers."
29958 This reverts commit 44d45d3fa56f121ce89ffe5b28beb48be01a95df.
29960 Michel Dänzer pointed out the flaw in using the pixmap size instead of
29963 Using the backing pixmap dimensions for this is not desirable. In
29964 particular, it means that the DRI2 buffers of non-redirected windows
29965 always have the same size as the screen. But even for redirected windows
29966 it wastes some graphics memory with a re-parenting window manager, that
29967 is if it doesn't break in various ways due to the top left corner of the
29968 DRI2 buffers no longer corresponding to the top left corner of the window.
29970 commit 44d45d3fa56f121ce89ffe5b28beb48be01a95df
29971 Author: Chris Wilson <chris@chris-wilson.co.uk>
29972 Date: Sat May 29 10:39:28 2010 +0100
29974 dri: Use size from backing pixmap when creating buffers.
29976 This avoid using the garbage values stored in the Screen drawable,
29977 instead of the true values which are only maintained in its backing
29978 pixmap. The consequence of using the wrong size was to hand a 1x1
29979 pixmap to metacity/mutter and have it believe it was a full screen
29980 drawable; GPU hangs ensued if using page flipping.
29982 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29984 commit 90c74a43149ec72bef64fb7e64b812294bda69a5
29985 Author: Chris Wilson <chris@chris-wilson.co.uk>
29986 Date: Fri May 28 21:49:38 2010 +0100
29988 i915: Don't re-emit vertex size unless it has changed.
29990 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
29992 commit a94ae175d6a69243a4465de0be0a8a3368e0ab28
29993 Author: Eric Anholt <eric@anholt.net>
29994 Date: Fri May 28 12:40:02 2010 -0700
29996 uxa: Fix prepare_solid being called without check_solid first.
29998 Fixes GPU hang on gen6.
30000 commit 66c90158e45e890c6f655195c6707d216c7ac4c5
30001 Author: Chris Wilson <chris@chris-wilson.co.uk>
30002 Date: Fri May 28 17:13:30 2010 +0100
30004 uxa: Skip the redundant miComputeCompositeRects() when adding to the mask
30006 As we are in full control of the destination (the temporary glyph mask)
30007 and the source (the glyph cache) we know that there are no clip regions
30008 on either and so can skip computing the composite rectangles. (We trust
30009 the device clipping to prevent compositing outside the target.)
30011 x11perf on PineView:
30012 701/686 -> 881/856 kglyphs/s [aa/rgb]
30014 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30016 commit 5b2254838eb89bf77eeb893c73eecb8c737822f4
30017 Author: Chris Wilson <chris@chris-wilson.co.uk>
30018 Date: Fri May 28 12:47:26 2010 +0100
30020 uxa: Make the glyph caches' fixed size explicit.
30022 Until we actual resize the glyph cache dynamically, make it obvious to
30023 the reader and the compiler that the size is fixed.
30025 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30027 commit 11581dda99cb2e4ae78fc73be4b02185b3be58ed
30028 Author: Chris Wilson <chris@chris-wilson.co.uk>
30029 Date: Fri May 28 11:38:05 2010 +0100
30031 uxa: Use a glyph private rather than a hash table.
30033 Store the cache position directly on the glyph using a devPrivate rather
30034 than an through auxiliary hash table.
30036 x11perf on PineView:
30037 650/638 kglyphs/s -> 701/686 kglyphs/s [aa/rgb]
30039 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30041 commit 73111cf2a212ee5cc2e03af1c600867df0c55b39
30042 Author: Chris Wilson <chris@chris-wilson.co.uk>
30043 Date: Wed May 26 21:05:27 2010 +0100
30045 Decouple non-reusuable pixmaps from batch lists on unref.
30047 ==7596== Invalid write of size 4
30048 ==7596== at 0x491ACA8: intel_batch_teardown (i830_batchbuffer.c:118)
30049 ==7596== by 0x491C9D6: I830CloseScreen (i830_driver.c:1419)
30050 ==7596== by 0x8103A9C: RRCloseScreen (randr.c:105)
30051 ==7596== by 0x80DE794: xf86CrtcCloseScreen (xf86Crtc.c:759)
30052 ==7596== by 0x80BEBA3: DGACloseScreen (xf86DGA.c:268)
30053 ==7596== by 0x80D044B: DPMSClose (xf86DPMS.c:134)
30054 ==7596== by 0x488B050: XvCloseScreen (xvmain.c:320)
30055 ==7596== by 0x81841B1: VidModeClose (xf86VidMode.c:110)
30056 ==7596== by 0x80EB12F: CursorCloseScreen (cursor.c:191)
30057 ==7596== by 0x810CA17: AnimCurCloseScreen (animcur.c:108)
30058 ==7596== by 0x816937E: compCloseScreen (compinit.c:86)
30059 ==7596== by 0x48D39B9: glxCloseScreen (glxscreens.c:221)
30060 ==7596== Address 0x49c1a50 is 24 bytes inside a block of size 52 free'd
30061 ==7596== at 0x4024866: free (vg_replace_malloc.c:325)
30062 ==7596== by 0x80B023C: Xfree (utils.c:1096)
30063 ==7596== by 0x4927CFD: i830_set_pixmap_bo (i830_uxa.c:647)
30064 ==7596== by 0x491C9B4: I830CloseScreen (i830_driver.c:1413)
30065 ==7596== by 0x8103A9C: RRCloseScreen (randr.c:105)
30066 ==7596== by 0x80DE794: xf86CrtcCloseScreen (xf86Crtc.c:759)
30067 ==7596== by 0x80BEBA3: DGACloseScreen (xf86DGA.c:268)
30068 ==7596== by 0x80D044B: DPMSClose (xf86DPMS.c:134)
30069 ==7596== by 0x488B050: XvCloseScreen (xvmain.c:320)
30070 ==7596== by 0x81841B1: VidModeClose (xf86VidMode.c:110)
30071 ==7596== by 0x80EB12F: CursorCloseScreen (cursor.c:191)
30072 ==7596== by 0x810CA17: AnimCurCloseScreen (animcur.c:108)
30074 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30076 commit a6fb6aa5f9c1c81b8b81b9d1138a526ffbbd5ad3
30077 Author: Chris Wilson <chris@chris-wilson.co.uk>
30078 Date: Wed May 26 20:31:45 2010 +0100
30080 Add vertex bo to the list of buffers to be torn down.
30082 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30084 commit 5dce69002dd48dd6dce4f0e7ad9ecf2c4c17cfe4
30085 Author: Chris Wilson <chris@chris-wilson.co.uk>
30086 Date: Wed May 26 20:26:42 2010 +0100
30088 i965: Remove ATOMIC_BATCH.
30090 This paranoid check is deceased; pining for the fjords.
30092 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30094 commit 06ebb55d308defbf8e01e30e4f994c2dd1071c40
30095 Author: Eric Anholt <eric@anholt.net>
30096 Date: Sun May 23 20:33:47 2010 -0700
30098 Add a workaround for Ironlake errata relating to disabling the clipper.
30100 commit 158a158dad26b6888dd2ebaf13e6020894475fcb
30101 Author: Eric Anholt <eric@anholt.net>
30102 Date: Sun May 23 20:28:24 2010 -0700
30104 Add a workaround for Ironlake errata regarding blits and other engines.
30106 commit 3461f8f4bc2a49fd41c408d225c592623404be91
30107 Author: Eric Anholt <eric@anholt.net>
30108 Date: Sun Apr 18 19:14:13 2010 -0700
30110 Remove remaining REG_DUMPER build stuff.
30112 commit 03bbb4c896ef3cd275312b413a2c85d9f499c032
30113 Author: Chris Wilson <chris@chris-wilson.co.uk>
30114 Date: Wed May 26 10:16:36 2010 +0100
30116 uxa: Perform manual damage for CompositeRects
30118 [xserver-1.8] The damage layer doesn't wrap CompositeRects, so we need to
30119 manually append the damaged region ourselves. This works for
30120 miCompsiteRects since that translates the call into multiple invocations
30121 of either PolyFillRectangle or Composite, which themselves cause damage.
30125 Bug 28120 - Tint2's tooltip borders end up at 0,0 and do not disappear
30126 https://bugs.freedesktop.org/show_bug.cgi?id=28120
30128 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30130 commit b9ada52a3081f54e0ed094ac31188bb240866f81
30131 Author: Chris Wilson <chris@chris-wilson.co.uk>
30132 Date: Wed May 26 10:14:52 2010 +0100
30134 uxa: Force the alpha value to 0xffff when treating Over as Src
30136 Since we have at most 8 bits of alpha, we treat >= 0xff00 as opaque.
30137 However, being paranoid we should set the alpha value to 0xfff in case
30138 something unexpected happens when converting from the xRenderColor to
30141 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30143 commit 3055d40164590147d35b5e7059ebe5f5858c85fa
30144 Author: Chris Wilson <chris@chris-wilson.co.uk>
30145 Date: Wed May 26 10:12:57 2010 +0100
30147 uxa: Use Composite rather than solid blitter for PolyRect
30149 Due to the relocation overhead, using a single composite with many
30150 rectangles outperforms many solid blits.
30152 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30154 commit ec2437f958ec4f5ac5222b37cba4cd403b5c8855
30155 Author: Chris Wilson <chris@chris-wilson.co.uk>
30156 Date: Wed May 26 10:12:18 2010 +0100
30158 uxa: Add PICT format mapping for depth 4 pixmaps.
30160 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30162 commit 309bd3a29943ef7502d40c67be49c19a59b6d1a2
30163 Author: Chris Wilson <chris@chris-wilson.co.uk>
30164 Date: Wed May 26 10:11:32 2010 +0100
30166 i830: Skip an empty fill.
30168 In the extremely unlikely event that the higher layer erroneous gave us
30169 an empty fill, skip it.
30171 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30173 commit b645ec83e0d86f2247b8338ceab60b9502516e70
30174 Author: Chris Wilson <chris@chris-wilson.co.uk>
30175 Date: Tue May 25 09:49:20 2010 +0100
30177 uxa: Apply the drawable offset to the solid rects
30181 Bug 28120 - Tint2's tooltip borders end up at 0,0 and do not disappear
30182 https://bugs.freedesktop.org/show_bug.cgi?id=28120
30184 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30186 commit 9d8ac271404ff219c3d40ae8f829e8b76ac7c359
30187 Merge: 509df27 ea07535
30188 Author: Chris Wilson <chris@chris-wilson.co.uk>
30189 Date: Fri May 21 11:01:30 2010 +0100
30191 Merge branch 'glyphs'
30193 Tweak glyphs to improve x11perf on i915 by about 33%.
30194 PineView, aa10text: 460 -> 617 kglyphs/s.
30195 PineView, rgb10text: 434 -> 610 kglyphs/s.
30199 xcb poppler 18.636 -> 13.958: 1.34x speedup
30200 xlib firefox-talos-gfx 71.905 -> 56.232: 1.28x speedup
30201 xcb firefox-talos-gfx 72.882 -> 57.969: 1.26x speedup
30202 xlib gnome-terminal-vim 38.126 -> 34.472: 1.11x speedup
30203 xcb gnome-terminal-vim 35.164 -> 32.573: 1.08x speedup
30204 xlib poppler 19.634 -> 18.246: 1.08x speedup
30206 Note the lack of significant improvement for firefox-planet-gnome.
30208 commit ea07535240dafc4c6ef55b4b7a2eeaa595febe86
30209 Author: Chris Wilson <chris@chris-wilson.co.uk>
30210 Date: Fri May 21 14:33:18 2010 +0100
30212 i915: Emit CA over using OutReverse + Add passes
30215 578/621 -> 610/617 kglyphs/sec [rgb/aa]
30217 commit 80a9e64f50aeda6004e3aba1fbfdda50bb1f1c82
30218 Author: Chris Wilson <chris@chris-wilson.co.uk>
30219 Date: Fri May 21 09:55:55 2010 +0100
30221 uxa: Use temporary dest when target is too large for compositor
30223 If the destination cannot fit into the 3D pipeline when we need to
30224 composite, we fallback to doing the operation on the CPU. This is very
30225 slow, and quite easy to trigger on i915 by plugging in an external
30228 An alternative is to extract the extents of the operation from the
30229 destination using the blitter which can usually handle much larger
30230 operations. This gives us a temporary target that can fit into the 3D
30231 pipeline and thus be accelerated, before copying back into the larger
30234 For x11perf this boosts glyph rendering on PineView, from 38kglyphs/s to
30235 480kglyphs/s. Just a little shy of the native performance of 601kglyphs/s
30237 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30239 commit 91f560034fc2695680d1208a78fc56d814b0da79
30240 Author: Chris Wilson <chris@chris-wilson.co.uk>
30241 Date: Thu May 20 10:26:59 2010 +0100
30243 uxa: Composite glyphs directly onto dst when possible.
30245 Without using a mask and compositing directly onto the destination,
30246 takes us from 580 kglyphs/s to 850 kglyphs/s on i945 [x11perf -aa10text].
30248 However, the extra intersection check almost entirely cancels out the
30249 speed up and we discover that the glyphs in x11perf are always
30250 overlapping. Nothing is ever easy.
30252 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30254 commit e3ece83f577d3664962edeec6ab5bdc41c5d77cf
30255 Author: Chris Wilson <chris@chris-wilson.co.uk>
30256 Date: Wed May 19 16:52:11 2010 +0100
30258 i915: compute normalized texcoords using a scale factor.
30260 500 -> 580kglyphs/s on i945.
30262 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30264 commit c2abf8d659b8b161a4f9df100b614ee3c8f8e458
30265 Author: Chris Wilson <chris@chris-wilson.co.uk>
30266 Date: Wed May 19 16:18:40 2010 +0100
30268 uxa: translate the region in line for composites
30270 When compositing, we need to convert the box into a rect and so the
30271 advantages of using REGION_TRANSLATE are lost.
30273 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30275 commit 2adf823b80b7b1f6df1bf3422a1219e93321a8fb
30276 Author: Chris Wilson <chris@chris-wilson.co.uk>
30277 Date: Wed May 19 15:02:58 2010 +0100
30279 i915: Add special case primitive emitters for glyphs.
30281 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30283 commit f64ab9e0d97dd9c654b4ae1924e62ef6813d9bb0
30284 Author: Chris Wilson <chris@chris-wilson.co.uk>
30285 Date: Wed May 19 14:50:12 2010 +0100
30287 i915: Move vertices into a vertex buffer object.
30289 In theory this should allow us to pack far more operations into a single
30290 batch buffer, and reduce our overheads.
30292 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30294 commit 2b050f330f78d02e7f476e55be29d760271ac61c
30295 Author: Chris Wilson <chris@chris-wilson.co.uk>
30296 Date: Wed May 19 10:57:46 2010 +0100
30298 Use pwrite to upload the batch buffer
30300 By using pwrite() instead of dri_bo_map() we can write to the batch buffer
30301 through the GTT and not be forced to map it back into the CPU domain and
30302 out again, eliminating a double clflush.
30304 Measing x11perf text performance on PineView:
30307 16000000 trep @ 0.0020 msec (511000.0/sec): Char in 80-char aa line (Charter 10)
30308 16000000 trep @ 0.0021 msec (480000.0/sec): Char in 80-char rgb line (Charter 10)
30310 16000000 trep @ 0.0019 msec (532000.0/sec): Char in 80-char aa line (Charter 10)
30311 16000000 trep @ 0.0020 msec (496000.0/sec): Char in 80-char rgb line (Charter 10)
30313 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30315 commit dcef703a7cdcf360f12312a338361697acffc3e9
30316 Author: Chris Wilson <chris@chris-wilson.co.uk>
30317 Date: Wed May 19 09:30:02 2010 +0100
30319 Kill paranoid assertions on every write into the batchbuffer.
30321 On my PineView box these represent ~5% overhead on x11perf text:
30324 16000000 trep @ 0.0020 msec (495000.0/sec): Char in 80-char aa line (Charter 10)
30325 12000000 trep @ 0.0022 msec (461000.0/sec): Char in 80-char rgb line (Charter 10)
30328 16000000 trep @ 0.0020 msec (511000.0/sec): Char in 80-char aa line (Charter 10)
30329 16000000 trep @ 0.0021 msec (480000.0/sec): Char in 80-char rgb line (Charter 10)
30331 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30333 commit bc41f84e01f18548b05c670e1fd0d641adc28d0f
30334 Author: Chris Wilson <chris@chris-wilson.co.uk>
30335 Date: Tue May 18 23:54:13 2010 +0100
30337 i915: Emit composite primitive with specialised functions.
30339 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30341 commit 4a3476ea094e84887fefb558e0bba023fee34151
30342 Author: Chris Wilson <chris@chris-wilson.co.uk>
30343 Date: Tue May 18 23:25:59 2010 +0100
30345 i915: amalgamate composite into a single primitive list
30347 Combine all the calls to composite between prepare_composite and
30348 done_composite into a single primitive list, rather than a primitive
30349 call per composite().
30351 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30353 commit e5c971e7639095d38da3518a5dc404b708d45cfb
30354 Author: Chris Wilson <chris@chris-wilson.co.uk>
30355 Date: Tue May 18 22:16:17 2010 +0100
30359 Use composite rather than solid blits in order to bring performance on
30360 a par with the CPU when using GEM and relocations.
30362 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30364 commit 509df27c7401e96d3062890da73f6af6629adef4
30365 Author: Kristian Høgsberg <krh@bitplanet.net>
30366 Date: Tue May 18 10:01:52 2010 -0400
30368 dri: Clean up DRI2 API #ifdefs a bit
30370 Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
30372 commit 5e04a81369dfc76bbc1ebbd4db00c23e71eb9e1c
30373 Author: Chris Wilson <chris@chris-wilson.co.uk>
30374 Date: Mon May 17 15:16:25 2010 +0100
30376 i830: Remove vestigal debugging ALWAYS_FLUSH and ALWAYS_SYNC
30378 These are now debugging options exposed in Xorg.conf, and now unused int
30381 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30383 commit 723cc45b2707aa160eac6d5b56a20e19cc42c8a3
30384 Author: Chris Wilson <chris@chris-wilson.co.uk>
30385 Date: Mon May 17 15:14:55 2010 +0100
30387 dri: Check error code from GetScratchGC()
30389 It may fail so be prepared, and do use the right drawable!
30391 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30393 commit 2c00297bc3b4969f28e2eb19bce36ece50ce4ccb
30394 Author: Chris Wilson <chris@chris-wilson.co.uk>
30395 Date: Sun May 16 19:15:37 2010 +0100
30397 uxa: Replace solid planemask [0xffffffff] with FB_ALLONES
30399 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30401 commit 2c69709d8afa6e9c0990efc463df0061536585e1
30402 Author: Chris Wilson <chris@chris-wilson.co.uk>
30403 Date: Sun May 16 18:41:52 2010 +0100
30405 i830: Encode surface bpp into format
30409 Bug 28135 - [855GM] Slowdown/High CPU-Usage after Git-Commit
30410 926fbc7d90ac1d0d49d154f136f9c9ed613c98c2
30411 https://bugs.freedesktop.org/show_bug.cgi?id=28135
30413 The simple answer is that I had assumed that 0 was a reserved value.
30414 However, without the bbp encoded into the format 0 was used for a8r8g8b8
30415 and r5g6b5, which are very common formats!
30417 The other possibility for the slowdown is that gtkperf is using of the
30418 now verboten xrgb formats -- but would in fact be valid if the source
30419 covers the clip and we could fixup the alpha value in the fixed function
30422 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30424 commit 21b5fd427f5ed6c99276349a7ae128eea580dd72
30425 Author: Chris Wilson <chris@chris-wilson.co.uk>
30426 Date: Sun May 16 13:51:41 2010 +0100
30428 uxa: Tidy uxa_solid_rects()
30430 Move the operator reduction after a few fallbacks, closer to its use.
30432 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30434 commit 61835701fd411d3bb550ceee3365e30639e46861
30435 Author: Chris Wilson <chris@chris-wilson.co.uk>
30436 Date: Sun May 16 13:26:07 2010 +0100
30438 uxa: Patterns are acquired at 0,0
30440 Set the correct offset for the gradients patterns after rendering to a
30443 Fixes cairo/test/huge-radial and friends
30445 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30447 commit 89f43f69a9c854fb2b82538a8938002675ca19e4
30448 Author: Chris Wilson <chris@chris-wilson.co.uk>
30449 Date: Sat May 15 18:34:54 2010 +0100
30451 uxa: Force an alpha channel when rendering source fallbacks
30453 As the source may not cover the extents, we need to represent those
30454 areas as transparent in the fallback picture, ergo we need an alpha
30455 channel. We could be smarter and force a format conversion when
30456 necessary, and we could let the backend choose the most appropriate
30459 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30461 commit 524fd2dd0da8f69d4c3f5aebe3a345fe34af609d
30462 Author: Chris Wilson <chris@chris-wilson.co.uk>
30463 Date: Sat May 15 18:23:21 2010 +0100
30465 uxa: Apply clip for solid rectangles.
30469 Bug 28120 - Tint2's tooltip borders end up at 0,0 and do not disappear
30470 https://bugs.freedesktop.org/show_bug.cgi?id=28120
30472 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30474 commit 58b089febceca1e915607bb723ee658aaa9dbed3
30475 Author: Chris Wilson <chris@chris-wilson.co.uk>
30476 Date: Sat May 15 09:11:46 2010 +0100
30478 uxa: Avoid using blits when with PictFilterConvolution
30482 Bug 28098 Compiz renders shadows wrong, garbage line of pixels along left
30483 and top edge of windows
30484 https://bugs.freedesktop.org/show_bug.cgi?id=28098
30486 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30488 commit ef95899f5b21453daeabf81a3015b22456d21fec
30489 Author: Chris Wilson <chris@chris-wilson.co.uk>
30490 Date: Sat May 15 09:02:07 2010 +0100
30492 uxa: Check the w-scaling component is 1 for an translation matrix
30494 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30496 commit 9c3da71349bcfeabae08f1572cf602c357bf7641
30497 Author: Chris Wilson <chris@chris-wilson.co.uk>
30498 Date: Sat May 15 01:05:24 2010 +0100
30500 i830: Remove xrgb conversion to argb, no longer required.
30502 All textures are now properly declared so that the alpha swizzling
30503 occurs in the sampler or not at all. The downside is that for quite a
30504 few composite operations we have to fallback to software on older
30505 hardware. There is scope for more performing the alpha expansion in
30506 shaders or combiners when we know the picture covers the clip - which is
30507 almost all of the time for normal operations especially those
30508 constructed by Cairo.
30510 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30512 commit 926fbc7d90ac1d0d49d154f136f9c9ed613c98c2
30513 Author: Chris Wilson <chris@chris-wilson.co.uk>
30514 Date: Sat May 15 01:01:20 2010 +0100
30516 i830: Remove incorrectly mapped tex formats.
30518 We no longer workaround the lack of alpha expansion for xrgb textures as
30519 this interferes with EXTEND_NONE, though we could if we know the source
30522 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30524 commit 95654cffa8643f61f153218f27e800df75b9bdee
30525 Author: Chris Wilson <chris@chris-wilson.co.uk>
30526 Date: Sat May 15 00:50:42 2010 +0100
30528 uxa: Fix order of conditionals to only run fill_region for SRC or opaque
30530 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30532 commit f67b45965b527699794ee21174809c5d71c03f22
30533 Author: Chris Wilson <chris@chris-wilson.co.uk>
30534 Date: Sat May 15 00:50:20 2010 +0100
30536 uxa: Expand the range of compatible formats to cover all bpp.
30538 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30540 commit 82d07fdf10cac2211af74ebf7d519daacd7084c0
30541 Author: Chris Wilson <chris@chris-wilson.co.uk>
30542 Date: Sat May 15 00:49:39 2010 +0100
30544 uxa: Only use 1x1R as a solid with an opaque format or SRC
30546 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30548 commit 3bca186a7ead84d4f0e71dfd83847ef77a4e390b
30549 Author: Chris Wilson <chris@chris-wilson.co.uk>
30550 Date: Sat May 15 00:48:31 2010 +0100
30552 uxa: Call check_solid before running the solid blitter.
30554 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30556 commit 213816c30b5ca6146e26ded85b2aa31ca16ca9bf
30557 Author: Chris Wilson <chris@chris-wilson.co.uk>
30558 Date: Sat May 15 00:47:55 2010 +0100
30560 i915: Load texture into directly into OC when possible.
30562 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30564 commit 737de9a779d05fe6fb1efe1904fda2cc452ff137
30565 Author: Chris Wilson <chris@chris-wilson.co.uk>
30566 Date: Fri May 14 23:48:27 2010 +0100
30568 uxa: Disable compatible src xrgb and dst argb
30570 I'm seeing garbage alpha for rendercheck blend:
30572 x8r8g8b8a 10x10 SRC ar8g8b8a
30574 so disable blitting until I work out if we can fast-path it.
30576 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30578 commit 271240fd477b11885947da32868bc77151681559
30579 Author: Chris Wilson <chris@chris-wilson.co.uk>
30580 Date: Fri May 14 23:38:02 2010 +0100
30582 i915: Remove a couple of unsupported 16bpp no-alpha tex formats
30584 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30586 commit a7c318d21c9035f6cab48c16e734b9ecb6f4b238
30587 Author: Chris Wilson <chris@chris-wilson.co.uk>
30588 Date: Fri May 14 23:30:21 2010 +0100
30590 uxa: Parse BGRA pixel formats.
30592 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30594 commit f7bbcc492a05f0801c64ee884798cf1a7ebb71c2
30595 Author: Chris Wilson <chris@chris-wilson.co.uk>
30596 Date: Fri May 14 23:29:13 2010 +0100
30598 Split the prepare blitter functions into check + prepare.
30600 Allow us to check whether we can handle the operation using the blitter
30601 prior to doing any work.
30603 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30605 commit 4be8d7eb89e61ffb2ceb19f1f84260e581187692
30606 Author: Chris Wilson <chris@chris-wilson.co.uk>
30607 Date: Fri May 14 21:13:21 2010 +0100
30609 i915: Don't force alpha=1 for RGB drawables in the shader.
30611 I was blindly fixing rendercheck without thinking. We need to force the
30612 alpha value to be in the blend unit and not before -- otherwise we
30613 generate the incorrect result whilst blending. D'oh.
30615 commit b9a5e36f957e42eaf0387ad15f2298d3e9538b05
30616 Author: Chris Wilson <chris@chris-wilson.co.uk>
30617 Date: Fri May 14 20:52:48 2010 +0100
30619 uxa: enable solid rects for backends that require pixmaps
30621 Convert the color into a (cached) pixmap if the backend cannot handle
30622 the SolidFill natively.
30624 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30626 commit a21297d7cc5945704d8cbd28ad0a2a3f6096fa47
30627 Author: Chris Wilson <chris@chris-wilson.co.uk>
30628 Date: Fri May 14 17:51:42 2010 +0100
30630 drm: Remove pin(); unpin() sync
30632 GEM handles serialisation of the new front buffer with respect to page
30633 flipping and rendering and reports back when the flip is complete.
30634 Adding a sync point here is then redundant.
30636 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30638 commit 7ee73d2c6f373f3d6829cb1876b1940db62be452
30639 Author: Chris Wilson <chris@chris-wilson.co.uk>
30640 Date: Fri May 14 17:51:20 2010 +0100
30642 drm: Remove unused old_front parameter from drmmode_do_pageflip.
30644 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30646 commit 030d56279bf14d9ddd42d8fdbeaa66ef3f557b4d
30647 Author: Chris Wilson <chris@chris-wilson.co.uk>
30648 Date: Fri May 14 16:53:40 2010 +0100
30650 drm: don't overwrite the old intel->front_buffer
30652 It's now handled in the common ExchangeBuffers() path.
30654 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30656 commit 5bd022739511104e4978bdfd2f4a1f60b9206723
30657 Author: Chris Wilson <chris@chris-wilson.co.uk>
30658 Date: Fri May 14 15:47:09 2010 +0100
30660 i830: Teardown batch entries on reset.
30662 By not cleaning up the batch entries when resetting the X server, we left
30663 the pointers in an inconsistent state and caused X to crash.
30665 commit 0d2392d44aae95d6b571d98f7ec323cf672a687f
30666 Author: Chris Wilson <chris@chris-wilson.co.uk>
30667 Date: Fri May 14 10:32:12 2010 +0100
30669 dri: Hold reference to buffers across swap
30671 As we schedule swaps for some time in the future and may process a
30672 detachment prior to receiving the vblank notification from the kernel,
30673 we need to hold a reference to the buffers for our swap event handler.
30676 Bug 28080 - "glresize" causes X server segfault with indirect rendering.
30677 https://bugs.freedesktop.org/show_bug.cgi?id=28080
30679 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30681 commit 8de09a0707ee1be1b919b979843711728618ef27
30682 Author: Chris Wilson <chris@chris-wilson.co.uk>
30683 Date: Thu May 13 17:15:28 2010 +0100
30685 uxa: Convert 1x1R back to solid_fill
30687 In the change to prevent blitting between incompatible sources, we also
30688 prevented 1x1R pixmaps from being used for solid fills. Reorder the
30689 sequence of conditions to enable this fast path again.
30691 commit 92e9cf8af784b13c28030e38b15c8decf29e6a32
30692 Author: Chris Wilson <chris@chris-wilson.co.uk>
30693 Date: Thu May 13 15:12:51 2010 +0100
30695 uxa: Only use solid_fill for SRC.
30697 commit d1bd14e8b6754ba9f797642e1b33bf689e19417b
30698 Author: Chris Wilson <chris@chris-wilson.co.uk>
30699 Date: Thu May 13 15:11:16 2010 +0100
30701 uxa: Replace source for CLEAR with a transparent solid
30703 This means that we will hit the faster try_solid_fill path instead.
30705 commit cdab72c405434ecbf7a79e402ff2d65d6a728179
30706 Author: Chris Wilson <chris@chris-wilson.co.uk>
30707 Date: Thu May 13 15:12:32 2010 +0100
30709 uxa: Fallback early if compositing with alphaMaps
30711 commit 25811dc7b7d1ad3fb01c31197d1ae1fe5b498975
30712 Author: Chris Wilson <chris@chris-wilson.co.uk>
30713 Date: Thu May 13 15:09:28 2010 +0100
30715 i915: Force output alpha to 1. if dst has no alpha channel.
30717 Ensure that garbage is not stored in the unused alpha channel so that
30718 we can rely on it being currently initialiased when used as a source or
30719 returning via GetImage.
30721 Partial fix for rendercheck -t blend
30723 commit 0e726b85ca6013ae9dc51391aaa309203352b61e
30724 Author: Chris Wilson <chris@chris-wilson.co.uk>
30725 Date: Thu May 13 09:41:39 2010 +0100
30727 i915: Add a2r10g10b10 format and friends
30729 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30731 commit 9f54107f866a25cf670f81f7c52b8c108728c6a5
30732 Author: Chris Wilson <chris@chris-wilson.co.uk>
30733 Date: Tue May 11 14:55:16 2010 +0100
30735 dri2: Handle reference counting across page flipping
30737 1. Instead of swapping bos, swap the entire private structure.
30739 2. If we update the pixmap bo for the Screen, make sure we update the
30740 reference inside intel->front_buffer so that xrandr still functions.
30744 Bug 27922 - i965: Rapidly resizing OpenGL window causes GPU to hang.
30745 https://bugs.freedesktop.org/show_bug.cgi?id=27922
30747 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30749 commit 6c27f6e4f76b97df71094acf25083b2922966b42
30750 Author: Chris Wilson <chris@chris-wilson.co.uk>
30751 Date: Wed May 12 12:01:46 2010 +0100
30753 uxa: Avoid glyph ping-pong with !offscreen destination
30755 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30757 commit d5383c2073e181b9ec352461c05d6202ad37f9d6
30758 Author: Chris Wilson <chris@chris-wilson.co.uk>
30759 Date: Wed May 12 10:53:14 2010 +0100
30761 uxa: Avoid ping-pong with !offscreen destination and traps
30763 If we are destined to target an !offscreen drawable, then uploading the
30764 trapezoid mask to a bo is the last thing we actually want to do...
30766 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30768 commit 00664b8f9d3da8d0d6aa53471ed3a8a8f6391660
30769 Author: Chris Wilson <chris@chris-wilson.co.uk>
30770 Date: Wed May 12 12:41:51 2010 +0100
30772 uxa: Fallback when compositing to a !offscreen destination
30774 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30776 commit 0c6372a77fd8c051534c1b420a02a8737aa1dd01
30777 Author: Chris Wilson <chris@chris-wilson.co.uk>
30778 Date: Wed May 12 09:57:27 2010 +0100
30780 i830: Prevent allocation of bo larger than half the aperture
30782 We need to prevent overcommitting the aperture, and in particular if we
30783 allocate a buffer larger than available space we will fail to mmap it in
30784 and rendering will fail. Trying to allocate multiple large buffers in
30785 the aperture, often the case when falling back, causes thrashes and
30786 eviction of useful buffers. So from the outset simply do not allocate a
30787 bo if the the required size is more than half the available aperture
30790 Fixes allocation failure in ocitymap.trace for instance.
30792 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30794 commit 244b7cbfffdcda4761948eaa37ed2a30ca81f107
30795 Author: Chris Wilson <chris@chris-wilson.co.uk>
30796 Date: Wed May 12 09:18:30 2010 +0100
30798 uxa: Use accelerated PutImage for uploading pixman images.
30800 Short-circuits the current use of PutImage from CopyArea, bypassing all
30801 the temporary allocations.
30803 commit cb887cfc670bf63993bd313ff33927afb8198eae
30804 Author: Chris Wilson <chris@chris-wilson.co.uk>
30805 Date: Fri Mar 26 09:59:51 2010 +0000
30809 The cost of performing relocations outweigh the advantages of using the
30810 blitter for solids with lots of rectangles.
30814 Bug 22127 - [UXA] 50% performance regression for XRenderFillRectangles
30815 https://bugs.freedesktop.org/show_bug.cgi?id=22127
30817 By using the 3D pipeline we improve our performance by around 4x on
30818 i945, measured by the jxbench microbenchmark, and a factor of 10x by
30819 short-cutting to the 3D pipeline for blended rectangles.
30821 Before, on a i945GME:
30822 19982.412060 Ops/s; rects (!); 15x15
30823 9599.131693 Ops/s; rects (!); 75x75
30824 3803.654743 Ops/s; rects (!); 250x250
30825 6836.743772 Ops/s; rects blended; 15x15
30826 1443.750000 Ops/s; rects blended; 75x75
30827 495.335821 Ops/s; rects blended; 250x250
30828 23247.933884 Ops/s; rects composition (!); 15x15
30829 10993.073048 Ops/s; rects composition (!); 75x75
30830 3595.905172 Ops/s; rects composition (!); 250x250
30833 87271.145975 Ops/s; rects (!); 15x15
30834 32347.744361 Ops/s; rects (!); 75x75
30835 5884.177215 Ops/s; rects (!); 250x250
30836 73500.000000 Ops/s; rects blended; 15x15
30837 33580.882353 Ops/s; rects blended; 75x75
30838 5858.811749 Ops/s; rects blended; 250x250
30839 25582.317073 Ops/s; rects composition (!); 15x15
30840 6664.728682 Ops/s; rects composition (!); 75x75
30841 14965.909091 Ops/s; rects composition (!); 250x250 [suspicious]
30843 This has no impact on Cairo, but I have a suspicion from watching xtrace
30844 that Qt likes to blit thousands of 1x1 rectangles with the same colour.
30845 However, we are still around 2-3x slower than the reported figures for
30848 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30850 commit c8e10f7791fecf0fed7ad606db062d87d8ff263c
30851 Author: Chris Wilson <chris@chris-wilson.co.uk>
30852 Date: Wed May 12 12:41:00 2010 +0100
30854 debug: Add names for operators
30856 Most useful for confirming my worst fears: unwarranted use of
30859 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30861 commit 6ea8ce640fbc7e5054b521a6f70b303945d0e849
30862 Author: Chris Wilson <chris@chris-wilson.co.uk>
30863 Date: Tue May 11 19:35:59 2010 +0100
30865 xvmc: Build fix with -pedantic
30869 Bug 27352 - RPMLINT error causes build breakage
30870 https://bugs.freedesktop.org/show_bug.cgi?id=27352
30872 Reported-by: Johannes Obermayr <johannesobermayr@gmx.de>
30873 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30875 commit e1b7e8bf1d475b3d00adcb96e6002c20c9e70d1c
30876 Author: Chris Wilson <chris@chris-wilson.co.uk>
30877 Date: Tue May 11 15:50:23 2010 +0100
30879 drmmode: Reorder i830_set_pixmap_bo() so that the correct stride is used.
30881 The pitch needs to be set on the pixmap prior to the private
30882 intel_pixmap structure being created so that it can record the correct
30883 value from the pixmap.
30885 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30887 commit dfbaf9aab80cf33d47148f087fa747eacd508b05
30888 Author: Chris Wilson <chris@chris-wilson.co.uk>
30889 Date: Tue May 11 14:54:15 2010 +0100
30891 i830: Never create a bo for depth=1 pixmaps.
30893 As we can not accelerate these either as a destination or a source,
30894 don't bother allocating a buffer object for them.
30896 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30898 commit 5b7efe375ae951ffcdb2c81e2ad8a2ed86ffbdf2
30899 Author: Chris Wilson <chris@chris-wilson.co.uk>
30900 Date: Tue May 11 14:53:29 2010 +0100
30902 i830: Use set_pixmap_bo() instead of open-coding.
30904 The advantage is that this enables in-flight reuse of the old pixmap if
30907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30909 commit ad8af95dd35921118f45d6cff355c80cf0906ff0
30910 Author: Chris Wilson <chris@chris-wilson.co.uk>
30911 Date: Tue May 11 14:52:54 2010 +0100
30913 i830: Do not cache in-flight non-reusable buffers.
30915 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30917 commit f1048e14d5ef34970cb717e2a37d1c6bb4ea3a34
30918 Author: Chris Wilson <chris@chris-wilson.co.uk>
30919 Date: Tue May 11 11:39:46 2010 +0100
30921 i965: Add texformats mapping for additional pixman formats
30923 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30925 commit a35afd4a2df8bab543700d874274228eddccae5b
30926 Author: Chris Wilson <chris@chris-wilson.co.uk>
30927 Date: Tue May 11 11:26:33 2010 +0100
30929 uxa: Recheck texture after acquiring pattern.
30931 As the first step to handling unsupported texture formats, double check
30932 that the converted pattern can be used as a texture by the card.
30934 Fixes: rendercheck -t repeat
30936 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30938 commit d745cab6c45b9fce5e46eacbdd04ceae911fddbb
30939 Author: Keith Packard <keithp@keithp.com>
30940 Date: Mon May 10 22:58:49 2010 -0700
30942 Must call ValidateGC in i830_uxa_put_image for scratch GC
30944 Always need to call ValidateGC or the scratch GC will not get the
30945 right composite clip.
30947 Signed-off-by: Keith Packard <keithp@keithp.com>
30949 commit 3eded4202ea00d42edf5edad831484ebc6c0c826
30950 Author: Chris Wilson <chris@chris-wilson.co.uk>
30951 Date: Mon May 10 23:38:17 2010 +0100
30953 i915: Fix pixmap based masks.
30955 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30957 commit 1ecd89be039b642affb2f9933f4341d186cb4dea
30958 Author: Chris Wilson <chris@chris-wilson.co.uk>
30959 Date: Mon May 10 23:33:52 2010 +0100
30961 uxa: Protect against valid SourcePict in uxa_acquire_mask()
30963 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30965 commit a8761585ef0b37736d302299d6b03e0957ad3e16
30966 Author: Chris Wilson <chris@chris-wilson.co.uk>
30967 Date: Mon May 10 19:38:24 2010 +0100
30969 i830: Minor cleanup
30971 Remove some extraneous prototypes and unused variables.
30973 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30975 commit 9e9b0d85da801f1f95c7bc7e99566fe7056f5813
30976 Author: Chris Wilson <chris@chris-wilson.co.uk>
30977 Date: Mon May 10 18:37:02 2010 +0100
30979 i830: Update stride when swapping bo for PutImage
30981 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30983 commit 0d4dd00aeafdd625ffe902e26e666b63a0842adc
30984 Author: Chris Wilson <chris@chris-wilson.co.uk>
30985 Date: Mon May 10 12:10:09 2010 +0100
30987 uxa,i915: Handle SourcePict through uxa_composite()
30989 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
30991 commit 21c1c3c7f6eb2b5070d2153b15a8fb1afe938bbb
30992 Author: Chris Wilson <chris@chris-wilson.co.uk>
30993 Date: Mon May 10 10:19:28 2010 +0100
30995 i915: Use 1x1R pixmap for solid drawables
30997 x11perf has a regression
30998 https://bugs.freedesktop.org/show_bug.cgi?id=25068
31002 commit e581ceb7381e29ecc1a172597d258824f6a1d2d3
31003 i915: Use the color channels to pass along solid sources and masks.
31005 Do not convert 1x1R pixmaps into a solid color as the readback from the
31006 bo negates all the performances advantages of using a smaller vertex
31007 buffer and fewer samplers.
31010 aa=66800 glyph/s, rgb=28800 glyphs/s
31013 aa=96800 glyphs/s, rgb=48500 glyphs/s
31015 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31017 commit f52b6e832292c02c0010b19882e38e1097beeda0
31018 Author: Chris Wilson <chris@chris-wilson.co.uk>
31019 Date: Mon May 10 09:39:44 2010 +0100
31021 uxa: Rearrange checking and preparing of composite textures.
31023 x11perf regression caused by 2D driver
31024 https://bugs.freedesktop.org/show_bug.cgi?id=28047
31028 commit a7b800513fcc94e063dfd68d2f63b6bab7fae47d
31029 uxa: Extract sub-region from in-memory buffers.
31031 The issue is that as we extract the region prior to checking whether the
31032 composite can in fact be accelerated, we perform expensive surplus
31033 operations. This is particularly noticeable for ComponentAlpha text,
31034 such as rgb10text. The solution here is to rearrange the
31035 check_composite() prior to acquiring the sources, and only extracting
31036 the subregion if the render path can not actually handle the texture.
31038 Performance (on PineView):
31039 a7b800513^: aa=68600 glyphs/s, rgb=29900 glyphs/s
31040 a7b800513: aa=65700 glyphs/s, rgb=13200 glyphs/s
31041 now: aa=66800 glyph/s, rgb=28800 glyphs/s
31043 The residual lossage seems to be from the extra function call and
31044 dixPrivate lookups. Hmm. More warning is the extremely low performance,
31045 however the results are consistent so the improvement looks real...
31047 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31049 commit 848ab66384508c3ad3e5fb4884e4527f3ebd3bde
31050 Author: Chris Wilson <chris@chris-wilson.co.uk>
31051 Date: Sat May 8 18:36:55 2010 +0100
31053 uxa: Transform composites with a simple translation into a blit
31055 We can also convert a composite with an integer translation into a
31056 blit, so long as the sample extents remains within the source.
31058 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31060 commit a7b800513fcc94e063dfd68d2f63b6bab7fae47d
31061 Author: Chris Wilson <chris@chris-wilson.co.uk>
31062 Date: Wed Apr 14 21:14:34 2010 +0100
31064 uxa: Extract sub-region from in-memory buffers.
31066 If the buffer is too large or not suitable for a GPU operation, we
31067 currently fallback and perform the composite on the CPU. An alternative
31068 is too extract the small region out of the source (as usually the
31069 sample extents are much smaller than the actual surface size) and try
31070 the composite with the new surface.
31072 The effect is particularly noticeable on pathological websites that use
31073 very large background images. For example, http://www.woodtv.com/ uses a
31074 1299x15000 pattern that is obscured by another opaque pattern.
31076 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31078 commit 8562b7bc6740eef2602af76b8685388efd2d4d37
31079 Author: Chris Wilson <chris@chris-wilson.co.uk>
31080 Date: Tue Apr 27 10:27:44 2010 +0100
31082 i830: prepare the uxa pixmap for fbCopyArea.
31084 Complete the prepare access for the PutImage fallback via fbCopyArea(),
31085 by remembering to set the private pointer to the GTT mapping.
31087 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31089 commit 9a5cd65b593ea82e56e0c403f7ddcc2420a64dc3
31090 Author: Chris Wilson <chris@chris-wilson.co.uk>
31091 Date: Mon Apr 26 09:07:07 2010 +0100
31093 i830: if pixman_blt() fails fallback to fbCopyArea()
31095 On older versions of pixman, pixman_blt() can return false if the images
31096 are <= 8bpp. If we are being called from CopyArea, then we cannot return
31097 FALSE here as that will trigger an infinite recursion. Instead we must
31098 manually perform the fallback using fbCopyArea().
31100 Reported-by: Peter Clifton <pcjc2@cam.ac.uk>
31101 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31103 commit 86d349aa7b0a596fbb9530e896c19349405bf3ba
31104 Author: Chris Wilson <chris@chris-wilson.co.uk>
31105 Date: Mon Apr 26 09:12:54 2010 +0100
31107 i830: tidy in flight bo reuse.
31109 A left-over cleanup patch for c374c94. *sigh*
31111 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31113 commit 72fd7d191c33c8d0b75b06ab0717d7ca0d4e019a
31114 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31115 Date: Mon Apr 19 21:56:57 2010 +0200
31119 This is some fallout from my xvmc cleanup.
31121 Original-Patch-by: Rico Tzschichholz <ricotz@t-online.de>
31122 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31124 commit 9494f4e91f8c8c7a0f8d61c6883c9bfceb2cec46
31125 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31126 Date: Fri Apr 16 22:24:01 2010 +0200
31128 i810: adjust the pitch for DRI rendering
31130 Current code forgot to adjust the pitch of the frontbuffer.
31132 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31133 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=16729
31135 commit c374c94e41d6e7d677334171e3255778d77cbe18
31136 Author: Chris Wilson <chris@chris-wilson.co.uk>
31137 Date: Wed Mar 31 11:50:27 2010 +0100
31139 uxa: Reuse in-flight bo
31141 When we need to allocate a new bo for use as a gpu target, first check
31142 if we can reuse a pixmap that has already been relocated into the
31143 aperture as a temporary target, for instance a glyph mask or a clip mask.
31146 backend test min(s) median(s) stddev.
31147 xlib firefox-planet-gnome 50.568 50.873 0.30%
31148 xcb firefox-planet-gnome 49.686 53.003 3.92%
31149 xlib evolution 40.115 40.131 0.86%
31150 xcb evolution 28.241 28.285 0.18%
31153 backend test min(s) median(s) stddev.
31154 xlib firefox-planet-gnome 47.759 48.233 0.80%
31155 xcb firefox-planet-gnome 48.611 48.657 0.87%
31156 xlib evolution 38.954 38.991 0.05%
31157 xcb evolution 26.561 26.654 0.19%
31159 And even more dramatic improvements when using a font size larger than
31160 the maximum size of the glyph cache:
31161 xcb firefox-36-20090611: 1.79x speedup
31162 xlib firefox-36-20090611: 1.74x speedup
31163 xcb firefox-36-20090609: 1.62x speedup
31164 xlib firefox-36-20090609: 1.59x speedup
31166 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31168 commit 96aa7a236ac0605324a94f7b7d10413cb219f071
31169 Author: Chris Wilson <chris@chris-wilson.co.uk>
31170 Date: Wed Apr 14 15:04:53 2010 +0100
31172 i830: Allocate bo's for glyphs larger than 32x32.
31174 As we only use the glyph cache for small glyphs, those large than 32x32
31175 will first be copied to a bo and used as a mask in a composite
31176 operation. We can avoid the allocation and upload per use by allocating
31177 a bo for the over-sized glyph from the start. As the glyph is large
31178 anyway, the excess memory allocation is less significant.
31180 Using normal font sizes, firefox shows no change - as expected. However,
31181 using the 36 font size traces, we see around a 10% improvement on g45.
31184 xcb firefox-36-20090609 127.333 127.897 0.22%
31185 xcb firefox-36-20090611 87.456 88.624 0.66%
31186 xcb firefox-20090601 19.522 20.194 1.69%
31187 xlib firefox-36-20090609 201.054 201.780 0.18%
31188 xlib firefox-36-20090611 133.468 133.717 0.09%
31189 xlib firefox-20090601 23.740 23.975 0.49%
31191 With large glyphs in bo:
31192 xcb firefox-36-20090609 117.256 118.254 0.42%
31193 xcb firefox-36-20090611 79.462 79.962 0.31%
31194 xcb firefox-20090601 19.658 20.024 0.92%
31195 xlib firefox-36-20090609 185.645 188.202 0.68%
31196 xlib firefox-36-20090611 123.592 124.940 0.54%
31197 xlib firefox-20090601 23.917 24.098 0.38%
31199 Thanks to Owain G. Ainsworth for the suggestion!
31201 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31203 commit 2d17bd50af367bead84985c22fdd43d264a5f072
31204 Author: Chris Wilson <chris@chris-wilson.co.uk>
31205 Date: Wed Apr 14 14:54:34 2010 +0100
31207 Revert "Revert "uxa: Try using put_image when copying from a memory buffer.""
31209 This reverts commit 6d50553e8f70d8f2142efdfd6c90bc27a599d0bc.
31211 Now we have taught the fallback path not to infinitely recurse,
31212 re-enable the accelerated path for ShmPutImage and friends.
31214 commit 1cc2c2c44ac72460cf1c4e6bdc13c612235809c9
31215 Author: Chris Wilson <chris@chris-wilson.co.uk>
31216 Date: Wed Apr 14 11:18:25 2010 +0100
31218 i830: Use pixman_blt directly for performing the in-memory copy
31220 In order to avoid an infinite recursion after enabling CopyArea to use
31221 the put_image acceleration to either stream a blit or to copy in-place,
31222 we cannot call CopyArea from put_image for the fallback path. Instead,
31223 we can simply call pixman_blt directly, which coincidentally is a tiny
31226 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31228 commit 324a2810da3fbae35637ba9080f31f9383db0868
31229 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31230 Date: Tue Apr 13 08:39:43 2010 +0200
31232 i830 render: check aperture space requirements
31234 No point not doing this.
31236 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31238 commit 804263c10df2fedb2f0debab6cdfaa6a89184a42
31239 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31240 Date: Sun Apr 11 12:56:24 2010 +0200
31242 render: tell the kernel explicitly when fences are needed
31244 This slighlty improves xrender performance on fence reg starved
31247 I've also changed a few function calls to the new names from the
31248 compat ones while looking at the code.
31250 The i915 textured video path is not converted because atm the xv
31251 code does not use tiled surfaces.
31253 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31255 commit a619a7831228dc52f0fef7d92c92f701e5aeaa94
31256 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31257 Date: Thu Apr 8 18:49:01 2010 +0200
31259 i915 render: use tiling bits where possible
31261 This is in preparation to explicit fence allocation with execbuf2.
31263 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31265 commit 55cd36046e61e8d51b5cb754a81cdb54e3eab166
31266 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31267 Date: Thu Apr 8 14:48:35 2010 +0200
31269 i830 render: use tiling bits where possible
31271 This is in preparation to explicit fence allocation with execbuf2.
31273 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31275 commit 6d50553e8f70d8f2142efdfd6c90bc27a599d0bc
31276 Author: Eric Anholt <eric@anholt.net>
31277 Date: Mon Apr 12 13:44:31 2010 -0700
31279 Revert "uxa: Try using put_image when copying from a memory buffer."
31281 This reverts commit 27195d7dba0f3ff08b92f3fd916cdf5113cbef58.
31282 put_image often calls copy_area. Which calls put_image. Exhausting of
31285 commit 28024f6c5f351ce4cb3434998d5dbfbeded7c606
31286 Author: Chris Wilson <chris@chris-wilson.co.uk>
31287 Date: Mon Apr 12 13:44:01 2010 +0100
31289 Revert "uxa: Add fallback warnings for PutImage."
31291 This reverts commit 299b0338d0811192dc4f8eae5d79453e9882c5d1.
31292 A debugging patch, it was never intended to go into master
31294 commit 27195d7dba0f3ff08b92f3fd916cdf5113cbef58
31295 Author: Chris Wilson <chris@chris-wilson.co.uk>
31296 Date: Wed Mar 31 20:30:27 2010 +0100
31298 uxa: Try using put_image when copying from a memory buffer.
31300 Often, for example in the fallback for ShmPutImage, we will attempt to
31301 use uxa_copy_area() copying to a normal pixmap from a memory buffer.
31302 This triggers a fallback, and maps the destination pixmap back into the
31303 GTT. The accelerated put_image path will attempt to stream a blit to the
31304 destination pixmap if it is currently active, avoiding the stall.
31306 commit 385563417d10b5e8a005bed6ae4de9a8fac1b407
31307 Author: Chris Wilson <chris@chris-wilson.co.uk>
31308 Date: Tue Mar 30 21:10:50 2010 +0100
31310 Review i830_pad_drawable_width()
31312 We appear to have a confusion of stride in terms of pixels, pitch in
31313 terms of bytes and the actual width of the surface.
31314 i830_pad_drawable_width() appears to be operating aligning *pixels* to a
31315 64 pixel boundary and has never used the chars-per-pixel causing
31316 considerable confusion in its callers. Remove the parameter and ensure
31317 that the callers are expecting a value in pixels returned, multiplying
31318 by cpp where necessary to get the pitch.
31320 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31322 commit 299b0338d0811192dc4f8eae5d79453e9882c5d1
31323 Author: Chris Wilson <chris@chris-wilson.co.uk>
31324 Date: Thu Mar 25 13:43:26 2010 +0000
31326 uxa: Add fallback warnings for PutImage.
31328 commit 80f52482c7cde000a76b91fe3d8b6c16baf2141f
31329 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31330 Date: Thu Apr 8 22:30:33 2010 +0200
31332 XvMC: fix memory overflow
31334 Caught by a malloc library assert.
31336 Note to self: Don't just copy&paste codelines around :(
31338 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31339 Buzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27540
31340 Tested-by: Nick Bowler <nbowler@draconx.ca>
31341 Tested-by: Calvin Walton <calvin.walton@gmail.com>
31343 commit 3d7723526b181686d62f09f81fba3ccdbf7d2b1a
31344 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31345 Date: Thu Apr 8 22:21:01 2010 +0200
31347 Xv: fixup overlay stride confusion
31349 For some reason I've made a mess out of the overlay stride constrains.
31352 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31353 Tested-by: Calvin Walton <calvin.walton@gmail.com>
31354 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27453
31356 commit bdf0ee509c26ab35e60f3ed02549efa435be032f
31357 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31358 Date: Thu Apr 8 13:29:04 2010 +0200
31360 i965 Xv: fix chroma pitch
31362 In my recent fix for the chroma pitch for i915 xvmc I've forgotten about
31363 i965 class hw. For videos with a non-even sized stride (measured in dwords)
31364 the chroma pitch was internally incosistent and one dword off.
31366 Fix this by using pitch2 for the chroma pitch in i965 textured video like
31369 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31370 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27417
31371 Tested-by: Nick Bowler <nbowler@draconx.ca>
31372 Tested-by: Sven Arvidsson <sa@whiz.se>
31374 commit f6cb28e909c8685474e69e8bd8600f146d3a74e9
31375 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31376 Date: Thu Mar 11 15:37:54 2010 +0100
31378 libXvMC: rip out debug stuff
31380 Almost totatlly unused, but surely totally useless.
31382 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31384 commit 676028ec0cbd4fadded92feefc8371e075abcc12
31385 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31386 Date: Wed Mar 10 19:23:06 2010 +0100
31388 libXvMC: unify CreateSurface
31390 Simply store the desired bo size in intel_xvmc_context and initialize
31391 it in the driver's create_context function.
31393 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31395 commit 3f590a4577f554313e7577205e471757a39129b6
31396 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31397 Date: Wed Mar 10 19:04:51 2010 +0100
31399 libXvMC i915: kill unused context private fields
31401 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31403 commit 8a31dacb067f813a5b0eafc16d265434e48ec5fc
31404 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31405 Date: Wed Mar 10 18:59:29 2010 +0100
31407 libXvMC: derive driver context from struct intel_xvmc_context
31409 ... by putting struct intel_xvmc_surface at the beginning. Also kill
31410 the common context handling code and simply keep a pointer in the
31411 surface private to the context.
31413 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31415 commit 16e5edde4d97e1818e0dbfbc165bd8199f697336
31416 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31417 Date: Wed Mar 10 16:51:10 2010 +0100
31419 libXvMC: unify DestroySurface
31421 Surface bo's can be freed by the common code, then there's nothing
31422 driver specific left.
31424 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31426 commit 5f64122551f2d8d41d745351b2e4e68208aa6084
31427 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31428 Date: Wed Mar 10 16:25:57 2010 +0100
31430 libXvMC: drop get_surface_status driver callback
31432 It's unused. Also drop all related generic code that tries to do
31433 clever stuff with this callback. These are all remnants from a
31436 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31438 commit 6c44ce9e127e531224bac825da7918800a3e6a1a
31439 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31440 Date: Wed Mar 10 16:20:03 2010 +0100
31442 libXvMC i915: s/i915XvMCSurface/struct intel_xvmc_surface/
31444 The private surface struct now adds nothing. Drop it.
31446 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31448 commit 583ec1f2f7ed32bb42c98baa1ca0fc3735b69b11
31449 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31450 Date: Wed Mar 10 15:55:44 2010 +0100
31452 libXvMC i915: rip out unneeded surface fields
31454 All of these are also stored in the context. Also kill the context
31455 reference counting. Doesn't serve a purpose besides occupying a
31456 pointer to the context in the private surface struct.
31458 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31460 commit c9aaaed829731cda22d057d6ac6424c587770973
31461 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31462 Date: Wed Mar 10 15:27:02 2010 +0100
31464 libXvMC: unify PutSurface
31466 Now that the drm bo is in the common structure, do the PutImage
31467 handling in common code, too.
31469 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31471 commit 4edff6c972311df2370a8323e346eeabd49ced35
31472 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31473 Date: Wed Mar 10 15:13:41 2010 +0100
31475 libXvMC: move drm_intel_bo to common intel_xvmc_surface
31477 i965 lost the last field in it's private surface struct. Kill it
31478 and any associated headers.
31480 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31482 commit 9eeabf876d1f70f5525bbac438414557dd2d4335
31483 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31484 Date: Wed Mar 10 15:00:58 2010 +0100
31486 libXvMC: derive driver surfaces from struct intel_xvmc_surface
31488 ... by putting struct intel_xvmc_surface at the beginning. This
31489 will allow to consolidate surface and bo handling.
31491 Also kill some now dead code used to handle the common surface
31494 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31496 commit a0058369bbdfb49f8df5ee1e50671115927d3565
31497 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31498 Date: Wed Mar 10 14:31:40 2010 +0100
31500 XvMC: kill intel_xvmc_command
31502 We only passed around and actually used the gem handle. Don't
31503 need a struct for one field alone ...
31505 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31507 commit 79fe4caafd21f6ca262901eac4a66ce5a4eaf548
31508 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31509 Date: Tue Mar 9 16:20:33 2010 +0100
31511 XvMC: dynamically allocate adaptor
31513 And kill all the static structures. This way it's clearer what's
31514 common and what's specific. And the code is shorter too.
31516 Also clean up src/i830_hwmc.c - kill the nonstandard surface types
31517 for i915 and the associated code.
31519 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31521 commit 00b63ecb192b7eafea9482bf5c5c5d442ce2c6fc
31522 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31523 Date: Tue Mar 9 15:29:15 2010 +0100
31525 XvMC: directly set driver type in context creation
31527 Instead of threading it through a global variable.
31529 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31531 commit 411cdcb6fcb60085191cf712aba1e0977f145256
31532 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31533 Date: Tue Mar 9 15:23:02 2010 +0100
31535 XvMC: init driver in one function call
31537 No point splitting this up. Gets rid of two #ifdef INTEL_XVMC blocks
31540 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31542 commit b666e3e29f22894f43a8e9f7f3d1b950d431c42d
31543 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31544 Date: Tue Mar 9 14:56:44 2010 +0100
31546 XvMC: move i965 specific stuff to common context
31548 Like for the i915. With this create_context is now also shared.
31550 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31552 commit e0068b1e63bc914940dbeb52de46657fcf5da1ea
31553 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31554 Date: Tue Mar 9 14:44:53 2010 +0100
31556 XvMC: move i915 specific stuff to common context
31558 Doing the same with the i965 code will allow us to share the
31559 create_context function.
31561 src/i915_hwmc.h is now almost empty. Move the last #defines to
31562 src/xvmv/i915_xvmc.c where they are actually used and delete the
31565 Also rename the ddx context struct to something sane.
31567 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31569 commit 8cab7b84da6d06c20ad9dbbcd8f3f3d971b438e7
31570 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31571 Date: Tue Mar 9 14:02:23 2010 +0100
31573 XvMC: unify destroy context function
31575 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31577 commit 323464b8e2669423df573a18fd1137cb3cad6700
31578 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31579 Date: Mon Mar 8 23:46:25 2010 +0100
31581 XvMC: clean up i830_hwmc.c
31583 - Drop unnecessary macros
31584 - Denote i965 specific stuff
31586 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31588 commit cfa907adf800e91a3c1f4698ac701b9b83fa600b
31589 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31590 Date: Mon Mar 8 23:41:56 2010 +0100
31592 XvMC: unify surface functions
31594 Like for the subpicture stuff, share the "do-nothing" functions ...
31595 And fix function name spelling, too.
31597 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31599 commit 036cd4bb82131c08b79f596667f00b19a0c8ab79
31600 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31601 Date: Mon Mar 8 17:33:18 2010 +0100
31603 i965 XvMC: don't create any surface state in the ddx
31605 Like for i915. Also drop that now totally superflous limit on the
31606 available surfaces.
31608 Move the surface struct into the userspace library header now that
31609 the ddx doesn't use it anymore.
31611 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31613 commit 69218cc1b51acfee91b471bb973a4419b79cdd85
31614 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31615 Date: Mon Mar 8 17:05:25 2010 +0100
31617 XvMC: unify subpicture functions
31619 The XvMC driver api in the server is insane. Even for optional stuff
31620 like subpicture support it doesn't check for NULL-pointers. So we
31621 have to retain some dummy functions.
31623 Wonder how many copies of these things exist on fdo ...
31625 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31627 commit 5d104e9c5cd5cd2d34c9035dd38794fa9b5578f3
31628 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31629 Date: Mon Mar 8 17:03:36 2010 +0100
31631 XvMC: unify PutImage handling
31633 Both xvmc are handing in the bo in the exact same way. So move the code
31634 to src/i830_video.c and kill this great oeuvre of spaghetti-code.
31636 The xvmc driver ini and fini also lost their last use, kill them, too.
31638 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31640 commit ebc0ffba79cc4bd564a8bd97327f24a1d15b3d1a
31641 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31642 Date: Mon Mar 8 16:36:24 2010 +0100
31644 XvMC: merge *_hwmc.c into i830_hwmc.c
31646 After unifying i915 and i965, not much will be left of these files.
31647 Therefore merge them to make the following changes easier.
31649 This creates some warnings about some redefined macros, but when this
31650 is all cleaned up they'll all be gone.
31652 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31654 commit 8199f8f191f8385ef3a228e0ec014a8bbd44002b
31655 Author: Daniel Vetter <daniel@fliege.ffwll.ch>
31656 Date: Sun Mar 7 23:17:49 2010 +0100
31658 i915 XvMC: kill more subpict support remnants
31660 This things just won't die!
31662 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31664 commit 29ba8a84f7cf5c29a5f38688a1ac0ccf41d8e4ec
31665 Author: Daniel Vetter <daniel@fliege.ffwll.ch>
31666 Date: Sun Mar 7 23:16:28 2010 +0100
31668 XvMC: everyone's using execbuffer!
31670 XvMC was switched over from batchbuffers a few years ago ...
31672 Signed-off-by: Daniel Vetter <daniel@fliege.ffwll.ch>
31674 commit 440b4d207b730112169827d0b90b46596659b504
31675 Author: Carl Worth <cworth@cworth.org>
31676 Date: Mon Mar 29 11:20:20 2010 -0700
31678 Increase version to 2.11.0
31680 In preparation for release.
31682 commit 1119c4523889ddedecf1722f0e9a2e4a7ad326b2
31683 Author: Carl Worth <cworth@cworth.org>
31684 Date: Mon Mar 29 11:19:31 2010 -0700
31686 NEWS: Add release notes for the 2.11.0 release.
31688 Hurrah! We made it.
31690 commit df3b26847914557eb3c9c70732d91169b1936d77
31691 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
31692 Date: Fri Mar 26 10:44:55 2010 -0700
31694 DRI2: release our private front buffer ref when buffer swapping
31696 Pauli pointed out that we take a ref on the front buffer when exchanging
31697 but forget to release it. The ref is necessary since the set functions
31698 will drop refs as necessary, but once we set the front buffer to point
31699 at the back pixmap, we ned to release our private ref again, or we'll
31702 Reported-by: Pauli Nieminen <suokkos@gmail.com>
31703 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
31705 commit 362a49e71fc41541b6dc121660d98e29da4b14e8
31706 Author: Gaetan Nadon <memsize@videotron.ca>
31707 Date: Thu Mar 25 10:07:41 2010 -0400
31709 uxa make: remove unused XORG_INCS and DIX_CFLAGS variables
31711 Most likely copied from xserver makefile.
31713 Acked-by: Dan Nicholson <dbn.lists@gmail.com>
31714 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
31716 commit 0d1ac4da5288a621d0493cb109fcd1e040e6c5a7
31717 Author: Chris Wilson <chris@chris-wilson.co.uk>
31718 Date: Thu Mar 25 09:12:09 2010 +0000
31720 uxa: Perform the xrgb -> argb conversion not inplace
31722 After reports of segmentation faults caused by
31723 d6b7f96fde1add92fd11f5a75869ae6fc688bf77 and vmware, the most obvious
31724 cause would be illegally writing to the src data when performing the alpha
31725 fill inline. So force the image upload to go via a fresh buffer whenever
31726 we need to modify the incoming data.
31728 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31729 Reported-and-tested-by: Jeff Chua <jeff.chua.linux@gmail.com>
31731 commit 0c47195ca805881e3fbd5b9224be5c930feeeb8c
31732 Author: Chris Wilson <chris@chris-wilson.co.uk>
31733 Date: Wed Mar 24 17:37:39 2010 +0000
31735 i830: Clip solid fills to surface.
31737 There is a reasonable surfeit of evidence to support this error,
31738 for instance: http://bugs.freedesktop.org/attachment.cgi?id=34417
31740 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31742 commit 5537079c29a56133446f1874d24d9e6516825edb
31743 Author: Chris Wilson <chris@chris-wilson.co.uk>
31744 Date: Wed Mar 24 14:59:20 2010 +0000
31746 uxa: After filling the alpha channel xrgb src is compatible with argb dst.
31748 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31750 commit 90a971c60769781f53827b469a9be3aab14cf71c
31751 Author: Chris Wilson <chris@chris-wilson.co.uk>
31752 Date: Wed Mar 24 14:50:45 2010 +0000
31754 uxa: Only reduce a composite to a BLT if it is wholly contained
31756 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31758 commit 2eec53d0b9232970fe3d03ce6c8940ebeea44bee
31759 Author: Chris Wilson <chris@chris-wilson.co.uk>
31760 Date: Tue Mar 23 17:28:22 2010 +0000
31762 uxa: Default to using TILING_X for pixmaps.
31764 On memory constrained hardware, tiling is vital for good performance as
31765 it minimizes cache misses. The downside is that for older hardware
31766 (which often suffers from the lack of bandwidth) requires the use of
31767 fences for many operations, which are in short supply and so may cause
31768 shorter batchbuffers. However our batch buffers are typically short and
31769 so this is unlikely to be a concern and not affect the performance wins.
31771 A quick bit of testing suggests the effect is inconclusive on
31774 xcb 205.470 206.219
31775 xcb-render-0.0 404.704 388.413
31776 xlib 166.410 170.805
31778 A secondary effect of the patch is to workaround a G31 specific hang
31779 when attempting to use linear 2048x2048 surfaces. Bonus!
31782 Bug 25375 - Performance issue using texture from pixmap (tfp) glx extension on 945
31783 http://bugs.freedesktop.org/show_bug.cgi?id=25375
31785 Bug 27100 - GPU Hung copying a 2048x1152 pixmap
31786 http://bugs.freedesktop.org/show_bug.cgi?id=27100
31788 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31789 Tested-by: John <jvinla@gmail.com>
31791 commit 9c037f61a490c96f9095f7ff3fecbf41f5efe9f7
31792 Author: Carl Worth <cworth@cworth.org>
31793 Date: Mon Mar 22 15:23:04 2010 -0700
31795 Update version to 2.10.903
31797 For today's snapshot.
31799 commit 11142cffe4bc81e1ce465ea725fcaef024130082
31800 Author: Carl Worth <cworth@cworth.org>
31801 Date: Mon Mar 22 15:22:28 2010 -0700
31803 NEWS: Add notes for 2.10.903
31805 For today's snapshot.
31807 commit f206816154fd44e61b0bdeaab1aa4c5359c3b70a
31808 Author: Carl Worth <cworth@cworth.org>
31809 Date: Mon Mar 22 14:14:46 2010 -0700
31811 man: Update list of available 'sclaing mode' property values.
31813 Include the names from the current kernel driver along with accurate
31814 descriptions of each. Indicate how to use the values with:
31816 xrandr --output output --set property value
31818 and point the user to "xrandr --prop" for an accurate list of
31819 currently available values.
31823 xf86-video-intel manpage needs update for KMS xrandr properties
31824 http://bugs.freedesktop.org/show_bug.cgi?id=25606
31826 commit d1dfab6b1649214177435629d8e85239dbe44d89
31827 Author: Matthias Hopf <mhopf@suse.de>
31828 Date: Wed Mar 17 15:13:51 2010 +0100
31830 Clear drmmode_output->mode_output in drmmode_output_destroy().
31832 commit 10cd04a84bcb6313903fc23b2d7791658ebc6b8e
31833 Author: Li Peng <peng.li@linux.intel.com>
31834 Date: Sat Mar 20 00:21:48 2010 +0800
31836 Initialize flip_count before using it
31838 Otherwise it would be a random value and drmmode_page_flip_handler()
31839 won't have a chance to call I830DRI2FlipEventHandler() and indicate
31840 a full page flip is complete.
31842 Signed-off-by: Li Peng <peng.li@intel.com>
31844 commit 3d4b3f257fbbb69c6f236d9803abe54a90d7d434
31845 Author: Dave Airlie <airlied@redhat.com>
31846 Date: Thu Mar 18 12:48:39 2010 +1000
31848 intel: free bus id in error path after printing it out.
31850 the error message prints out a freed string, spotted during
31851 code reappropriation to radeon driver.
31853 Signed-off-by: Dave Airlie <airlied@redhat.com>
31855 commit 31d5f84bb4416ef92abd97264d52cdab7a184687
31856 Author: Chris Wilson <chris@chris-wilson.co.uk>
31857 Date: Wed Mar 17 09:11:05 2010 +0000
31859 i915: Correct preamble for emit_composite
31862 http://bugs.freedesktop.org/show_bug.cgi?id=27123
31864 Fatal server error:
31865 i915_emit_composite_setup: ADVANCE_BATCH: under-used allocation 100/104
31867 Introduced with commit d6b7f96fde1add92fd11f5a75869ae6fc688bf77.
31869 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31871 commit d6b7f96fde1add92fd11f5a75869ae6fc688bf77
31872 Author: Chris Wilson <chris@chris-wilson.co.uk>
31873 Date: Sat Mar 6 15:49:04 2010 +0000
31875 Fill alpha on xrgb images.
31877 Do not try to fixup the alpha in the ff/shaders as this has the
31878 side-effect of overriding the alpha value of the border color, causing
31879 images to be padded with black rather than transparent. This can
31880 generate large and obnoxious visual artefacts.
31884 Bug 17933 - x8r8g8b8 doesn't sample alpha=0 outside surface bounds
31885 http://bugs.freedesktop.org/show_bug.cgi?id=17933
31887 and many related cairo test suite failures.
31889 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31891 commit 910fd171a00227025abc8bcc286a740f5bae895b
31892 Author: Chris Wilson <chris@chris-wilson.co.uk>
31893 Date: Fri Mar 5 14:32:13 2010 +0000
31895 i830: Remove coord-adjust for nearest centre-sampling.
31897 Fixes a number of cairo test suite failures.
31900 Bug 16917 - Blur on y-axis also when only x-axis is scaled bilinear
31901 http://bugs.freedesktop.org/show_bug.cgi?id=16917
31903 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
31905 commit 753914acc31947df8e9f2fa3b9c39de745098709
31906 Author: Carl Worth <cworth@cworth.org>
31907 Date: Mon Mar 15 17:32:57 2010 -0700
31909 Bump version to 2.10.902
31911 For the imminent 2.10.902 snapshot release.
31913 commit 185196cb010cb8426ca929a9dc3f028ec9b383db
31914 Author: Carl Worth <cworth@cworth.org>
31915 Date: Mon Mar 15 17:32:20 2010 -0700
31917 NEWS: Add notes for the 2.10.902 snapshot.
31919 Which I'm about to push out.
31921 commit 318aa9ed799197810e2039dbe3ec51559dcc888c
31922 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
31923 Date: Mon Mar 8 23:22:53 2010 +0100
31925 i915 XvMC: fixup colors
31927 My cleanup accidently created a inconsistency in the YUV plane ordering.
31928 I think we can safely assume that I'm colorblind ;)
31930 As Carl Worth rightly pointed out, this change deserves a more elaborate
31933 For Xv planar formats, the three planes are stored consecutively in
31934 memory, ordered Y U V. Now for some totally odd reason (= none at all),
31935 i915 xvmc stored it in Y V U order. Right after the release of 2.10, with
31936 commit "Xv: consolidate xmvc passthrough handling" I've inadvertently
31937 broken xvmc support (which started this whole odyssey into xvmc). When
31938 fixing stuff up, I neglected this special plane ordering and simply
31939 assumed it to be the same as Xv and dropped that special case for i915 in
31940 src/i830_video.c. This patch completes the change to standard YUV plane
31941 ordering by making the corresponding change in src/xvmc/i915_xvmc.c.
31943 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
31945 commit 68629b63740fb6e62c258b8a46669e77660fa4e4
31946 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
31947 Date: Mon Mar 8 15:42:39 2010 -0800
31949 DRI2: make WaitMSC error handling more consistent with ScheduleSwap
31951 Just make it mirror ScheduleSwap: complete the wait on any error
31952 condition so as not to crash the client if the kernel is misbehaving.
31954 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
31956 commit 6df74e61afb7831ebf3fbab8782f46ddccbe262b
31957 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
31958 Date: Mon Mar 8 15:33:20 2010 -0800
31960 DRI2: truncate OML values to 32 bits
31962 We can only handle 32 bit values unless we totally virtualize the count,
31963 since the kernel only handles 32 bits itself. Rather than adding all
31964 that overhead, just tolerate the occasional missed event everytime the
31967 Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
31968 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
31970 commit c66d57080dc034aa7877f47612065e388bbc38a2
31971 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
31972 Date: Mon Mar 8 15:26:24 2010 -0800
31974 DRI2: more WaitMSC fixes
31976 A couple more niggles: make sure we return a target_msc that at least
31977 matches the current count; this is a little more friendly to clients
31978 that missed an event. Also check for >= when calculating the remainder
31979 so we'll catch the *next* vblank event when the calculation is
31980 satisfied, rather than the current one as might happen at times.
31982 Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
31983 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
31985 commit 7845c6ade82085488192bd76729d92fb7b534cc0
31986 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
31987 Date: Mon Mar 8 15:19:14 2010 -0800
31989 DRI2: fixup stray curly brace
31991 Got left in the wrong column after the last cleanup.
31993 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
31995 commit 9656d329e4df28cfe138b657d04e5136bc794ce2
31996 Author: Eric Anholt <eric@anholt.net>
31997 Date: Mon Mar 8 14:34:04 2010 -0800
31999 Put back the pitch alignment for new framebuffers.
32001 I confused a dead assignment with dead code, because one of the args
32002 to the function was an outvalue. Fixes corruption under compiz.
32006 commit b71ca26a026a356763c51c763bcdd8024fc4a783
32007 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
32008 Date: Mon Mar 8 11:27:42 2010 -0800
32010 DRI2: fixup ScheduleWaitMSC similarly to ScheduleSwap
32012 My merge of Mario's patch for this was botched. Fix it up so that OML
32013 waits work correctly, and remove a bogus warning from ScheduleSwap.
32015 Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
32016 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
32018 commit 06b54e089eb21736e6d4f6462bbfca987be0c5cc
32019 Author: Matt Turner <mattst88@gmail.com>
32020 Date: Sun Mar 7 14:27:27 2010 -0500
32022 Don't check for Xinerama.
32024 It doesn't seem to be used anywhere, so don't require it.
32026 CC: Eric Anholt <eric@anholt.net>
32027 CC: Jesse Barnes <jbarnes@virtuousgeek.org>
32028 Signed-off-by: Matt Turner <mattst88@gmail.com>
32030 commit 8ece6cf5afa1bb0d8d9328696422f42f3c3adbd6
32031 Author: Robert Hooker <sarvatt@ubuntu.com>
32032 Date: Sat Mar 6 14:09:12 2010 -0500
32034 Fix build against xserver 1.6 branch.
32036 Signed-off-by: Robert Hooker <sarvatt@ubuntu.com>
32037 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
32039 commit 1cd556420277f103c47ade422f3ec8f8efb2d282
32040 Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
32041 Date: Fri Mar 5 12:32:18 2010 -0800
32043 DRI2: handle target_msc, divisor and remainder properly in DRI2ScheduleSwap
32045 The current code in I830DRI2ScheduleSwap() only schedules the correct
32046 vblank events for the case divisor == 0, i.e., the simple
32047 glXSwapBuffers() case.
32049 In a glXSwapBuffersMscOML() request, divisor can be > 0, which would go
32052 This modified code should handle target_msc, divisor, remainder and the
32053 different cases defined in the OML_sync_control extension correctly for
32054 the divisor > 0 case.
32056 It also tries to make sure that the effective framecount of swap
32057 satisfies all constraints, taking the 1 frame delay in pageflipping mode
32058 and possible delays in blitting/exchange mode due to
32059 DRM_VBLANK_NEXTONMISS into account.
32061 The swap_interval logic in the X-Servers DRI2SwapBuffers() call expects
32062 the returned swap_target from the DDX to be reasonably accurate,
32063 otherwise implementation of swap_interval for the glXSwapBuffers() as
32064 defined in the SGI_swap_interval extension may become unreliable.
32066 For non-pageflipped mode, the returned swap_target is always correct due
32067 to the adjustments done by drmWaitVBlank(), as DRM_VBLANK_NEXTONMISS is
32070 In pageflipped mode, DRM_VBLANK_NEXTONMISS can't be used without severe
32071 impact on performance, so the code in I830DRI2ScheduleSwap() must make
32072 manual adjustments to the returned vbl.reply.sequence number.
32074 This patch adds the needed adjustments.
32076 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
32078 commit 13119ffc034a3e9d6c76339d4fedc62bb3b41257
32079 Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
32080 Date: Fri Mar 5 11:33:45 2010 -0800
32082 DRI2: make MSC waits handle specific target_mscs and divisor/remainders
32084 Previous code only handled divisor == 0 case correctly. This should
32085 honor a given target_msc for the divisor > 0 case and handle the
32086 (msc % divisor) == remainder constraint correctly.
32088 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
32090 commit b6e0b92f398823629ba8a1ea8f5e62fbf959e725
32091 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
32092 Date: Thu Mar 4 10:07:26 2010 -0800
32094 DRI2: handle offscreen drawables better at swap time
32096 If a drawable isn't visible due to DPMS or redirection, we'll just blit
32097 it rather than schedule a swap event. However, we didn't reset the
32098 target_msc, so the swap target we receive from the server could get out
32099 of sync with the vblank count of the drawable's display. So at DPMS on
32100 time, the swap target would be the last good vblank count plus some
32101 large number (since the swaps won't have been throttled).
32103 Solve this by zeroing out the swap target like we should when we fall
32104 back to a blit. Also make the kernel error cases more friendly by
32105 making them fall back to blits too.
32107 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
32109 commit 54ac4e2df987b72529a523ffbde357bec27e3658
32110 Author: Chris Wilson <chris@chris-wilson.co.uk>
32111 Date: Thu Mar 4 21:34:52 2010 +0000
32113 Rate limit batch buffer error.
32115 Once we hit this error it's unlikely that we're coming back - so don't
32116 flood the logs with redundant information.
32118 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32120 commit 066d9b64ee243e0d255d7e12c2134951eef4ade7
32121 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32122 Date: Thu Mar 4 20:50:20 2010 +0100
32124 i915 XvMC: kill dead code
32126 This kills one wip remnant from my i830_memory cleanup and the last
32127 remainings of the subpicture support.
32129 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32131 commit 7b7c724da97f358f5d2d4d4dae75cfc85bde2c62
32132 Author: Eric Anholt <eric@anholt.net>
32133 Date: Thu Mar 4 10:29:09 2010 -0800
32135 Remove remaining fbOffset setting.
32137 In the long long ago, fbOffset was used for DGA. The server now has
32138 only one reference to fbOffset, a leftover setting of it in fbdevhw.
32139 We can safely ignore it now, which is good since we weren't updating
32140 it in other places where the front buffer offset could change.
32142 commit 15026d64d3c0d8ad96e0a73fcae3103d1c41fd46
32143 Author: Eric Anholt <eric@anholt.net>
32144 Date: Thu Mar 4 10:28:09 2010 -0800
32146 Remove a piece of fbOffset cruft from non-DRM support.
32148 commit e8e615289213ba1704e402b923c51ce9d196f06f
32149 Author: Eric Anholt <eric@anholt.net>
32150 Date: Thu Mar 4 10:27:51 2010 -0800
32152 Remove i830_allocate_2d_memory() now that it only called one function.
32154 commit e37b562083aa3293e0c009171724a3f122d8a32d
32155 Author: Eric Anholt <eric@anholt.net>
32156 Date: Thu Mar 4 10:23:12 2010 -0800
32158 Init CRTC cursors with CRTC setup instead of i830_memory.
32160 commit a36bdaba6136054ae7c67943ca0215cfd177bc5b
32161 Author: Eric Anholt <eric@anholt.net>
32162 Date: Thu Mar 4 09:47:10 2010 -0800
32164 Remove intel_sync() at teardown time.
32166 The kernel's still running after we're gone. This didn't matter.
32168 commit 4ada6d7a856a941b834871ff8a7c5505ff26ae23
32169 Author: Eric Anholt <eric@anholt.net>
32170 Date: Thu Mar 4 09:32:42 2010 -0800
32172 Remove 3D state clobber on EnterVT, and always clobber at batch start.
32174 We know that it's clobbered at each batchbuffer, anyway. And even if
32175 this server isn't running DRI2, it can still be clobbered at batch
32176 start in the KMS world.
32178 commit d92d42303e5ca9f4208ed97823ad8f691121370f
32179 Author: Eric Anholt <eric@anholt.net>
32180 Date: Thu Mar 4 09:31:56 2010 -0800
32182 Remove pre-2.6.29 error message handling since we require KMS.
32184 commit faecd155c49229499e29815eb6e79662ed33ddd5
32185 Author: Eric Anholt <eric@anholt.net>
32186 Date: Thu Mar 4 09:31:15 2010 -0800
32188 Move batch and 965 render state setup/teardown to screen init/close.
32190 Whether we're VT switched or not shouldn't impact rendering.
32192 commit 74e2b69a317cc728b00e675c18d2976987407aec
32193 Author: Carl Worth <cworth@cworth.org>
32194 Date: Thu Mar 4 09:51:03 2010 -0800
32196 i915_hwmc: Remove dead code.
32198 Daniel recently identified this code as unneeded (with an #if 0).
32199 Here we take the next step and remove it entirely.
32201 commit 1cc35a8ba403ad95b67caff46b803db1edea5ad3
32202 Author: Carl Worth <cworth@cworth.org>
32203 Date: Thu Mar 4 09:46:33 2010 -0800
32205 uxa: Fix type mismatch to avoid compiler warning.
32207 The code was using uint32_t where an XID (currently an unsigned long)
32208 was specified in the prototype. Use XID to avoid both the warning and
32209 any potential problem.
32211 commit 1d6537ec57535183638c62e3099bc79786c682ca
32212 Author: Carl Worth <cworth@cworth.org>
32213 Date: Thu Mar 4 09:39:28 2010 -0800
32215 i830_video: Remove unused variable.
32217 Certainly just a little leftover from the recent rewrites.
32219 commit 57c0043b9a519da5c9af29019ef632716782bda2
32220 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32221 Date: Tue Mar 2 09:22:33 2010 +0100
32223 Xv: fixup relocation in i965_video.c
32225 The previous code made no sense, (multiplying an offset by 4 is
32226 meaningless). It could have onlt worked with the offset being
32229 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32230 Reviewed-by: Carl Worth <cworth@cworth.org>
32232 commit e606be463f6359fd017791a8d4d85059831a5d41
32233 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32234 Date: Tue Mar 2 21:53:28 2010 +0100
32236 i830_memory: rip out the remainings of the old allocator
32240 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32241 Reviewed-by: Carl Worth <cworth@cworth.org>
32243 commit 371be65fb74789250dbb1e332e46416d931da321
32244 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32245 Date: Tue Mar 2 21:53:27 2010 +0100
32247 XvMC: kill the pinned batchbuffer in the ddx code
32249 It's been unused for quite a while.
32251 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32252 Reviewed-by: Carl Worth <cworth@cworth.org>
32254 commit d5a20c81ab76129a1e0bcefecd9c4a10b5040af4
32255 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32256 Date: Tue Mar 2 21:53:26 2010 +0100
32258 i915 XvMC: kill pinned surface buffer in the ddx code
32260 Like with the per context stuff, also drop the now artificial limit
32261 on surfaces. Again, with that gone, a lot of code can be deleted.
32263 Reviewed-by: Carl Worth <cworth@cworth.org>
32265 commit e6eb257ad32b393fb67c506943d8ece98b9a8061
32266 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32267 Date: Tue Mar 2 21:53:25 2010 +0100
32269 i915 XvMC: kill pinned per-context buffers in the ddx code
32271 There's now not a reason anymore to limit the number of active contexts.
32272 So kill this accounting, too.
32274 With that all gone, per-context state in the ddx is nil, so rip out
32275 all associated code.
32277 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32278 Reviewed-by: Carl Worth <cworth@cworth.org>
32280 commit 47ae1181f6cfd2ab64cd5c8535289fdaa2a9fd9c
32281 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32282 Date: Tue Mar 2 21:53:24 2010 +0100
32284 i915 XvMC: don't stall the cpu anymore
32286 Proper bo management ensures that the cpu doesn't step on buffers
32287 used by the gpu. Drop the now unnecessary synchronization.
32289 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32290 Reviewed-by: Carl Worth <cworth@cworth.org>
32292 commit 9bba123c5f0f081562306c152e70221fc5041a3b
32293 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32294 Date: Tue Mar 2 21:53:23 2010 +0100
32296 i915 XvMC: drop superflous MI_FLUSH
32298 Cache coherency is now fully under the control of gem.
32300 For lack of hw documentation, I had to find out the correct cache
32301 placements by trial and error:
32303 Backward and forward surfaces: I915_GEM_DOMAIN_RENDER
32304 Correlation data: I915_GEM_DOMAIN_SAMPLER
32306 Changing any of them leads to visual corruptions, so I think these
32307 are the correct ones.
32309 Reviewed-by: Carl Worth <cworth@cworth.org>
32311 commit b11623f20e303ae1d90d4a6bf0d5d73970b4e9bf
32312 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32313 Date: Tue Mar 2 21:53:22 2010 +0100
32315 i915 XvMC: switch surfaces to drm_intel_bo
32317 Now the last user of the fixed buffers provided by the ddx is gone!
32319 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32320 Reviewed-by: Carl Worth <cworth@cworth.org>
32322 commit b7f79bfd40db3ed17d7a79750d007b57d7ed157a
32323 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32324 Date: Tue Mar 2 21:53:21 2010 +0100
32326 i915 XvMC: switch corrdata buffer to drm_intel_bo
32330 v2: Correlation data needs to be in the render cache!
32332 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32333 Reviewed-by: Carl Worth <cworth@cworth.org>
32335 commit 62846d88d41f1bbbd16a6e700ca611656fa20f61
32336 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32337 Date: Tue Mar 2 21:53:20 2010 +0100
32339 i915 XvMC: switch msb to drm_intel_bo
32341 Like for the static indirect state buffer.
32343 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32344 Reviewed-by: Carl Worth <cworth@cworth.org>
32346 commit d27955c697e1768559957da8d61087644befd285
32347 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32348 Date: Tue Mar 2 21:53:19 2010 +0100
32350 i915 XvMC: switch sis to drm_intel_bo
32352 I've decided to allocate a new buffer for every render command, to
32353 prevent stalling for the gpu. libdrm bo reuse should take care of
32354 not wasting memory in case the buffer is not busy.
32356 Also always emit the full state, it's not worth it to complicate
32357 the code over a few stores to wc memory.
32359 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32360 Reviewed-by: Carl Worth <cworth@cworth.org>
32362 commit 3203c66fbf04c55adba760e97c9fa1feb9e55f8d
32363 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32364 Date: Tue Mar 2 21:53:18 2010 +0100
32366 i915 XvMC: switch load_indirect_render_emit to batchbuffer macros
32368 Like with one_time_state_emit, this preps for relocatable bo's.
32370 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32371 Reviewed-by: Carl Worth <cworth@cworth.org>
32373 commit d30decae6a9db585f2d2589ae2c3d9e1f9acd33d
32374 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32375 Date: Tue Mar 2 21:53:17 2010 +0100
32377 i915 XvMC: switch psc to drm_intel_bo
32379 Like with the sampler state buffer.
32381 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32382 Reviewed-by: Carl Worth <cworth@cworth.org>
32384 commit b543c355f4fdcb64deb198765d9b8b86f7c234c2
32385 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32386 Date: Tue Mar 2 21:53:16 2010 +0100
32388 i915 XvMC: switch psp to drm_intel_bo
32390 Like with the sampler state buffer.
32392 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32393 Reviewed-by: Carl Worth <cworth@cworth.org>
32395 commit 18c364084ce5495ffdfea38c17c1297d555c6c54
32396 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32397 Date: Tue Mar 2 21:53:15 2010 +0100
32399 i915 XvMC: switch ssb to drm_intel_bo
32401 This also starts to kill the last remnants of the support for
32402 physical addresses for the indirect state buffers. With gem this
32403 would need kernel support (in the form of a new reloc type in
32406 This does not change the ABI between ddx and client libIntelXvMC.
32407 I've decided to do this in one swoop when all the buffer rework is
32410 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32411 Reviewed-by: Carl Worth <cworth@cworth.org>
32413 commit fc9e44f01910eb9fd43def9b2ac531b0742d8b4e
32414 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32415 Date: Tue Mar 2 21:53:14 2010 +0100
32417 i915 XvMC: kill last_flip and last_render
32419 Seems to be a remnant from i810 XvMC support. last_flip is always 0,
32420 so serves no real purpose anymore. Kill it and the associated code.
32422 With last_flip gone, last_render also lost its purpose. Kill it, too.
32424 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32425 Reviewed-by: Carl Worth <cworth@cworth.org>
32427 commit 04aa38a639fefcd00acb485f36ab18c2083ed1bb
32428 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32429 Date: Tue Mar 2 21:53:13 2010 +0100
32431 i915 XvMC: switch one_time_state_emit to batchbuffer macros
32433 This is in preparation for real relocatable drm_bo's instead
32434 of memory at a fixed address. By switching to the batchbuffer
32435 macros (like i965 xvmc) we can use the nice OUT_RELOC macro.
32437 Also align the code more with coding-style elsewhere, i.e. bitops
32438 instead of bitfield structures. The bitfield structures are
32439 quite a mess to work with the batchbuffer macros, so they were
32440 getting in the way, anyway.
32442 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32443 Reviewed-by: Carl Worth <cworth@cworth.org>
32445 commit 24d787335a0c319b8d9243ea1f2726575cf73b2b
32446 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32447 Date: Tue Mar 2 21:53:12 2010 +0100
32449 XvMC: kill dead code in i915_xvmc.c
32451 WIP code that hasn't changed for over two years is unlikely to
32452 suddenly start progressing. Drop it. After all, git can easily
32453 resurect it in cases it's needed.
32455 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32456 Reviewed-by: Carl Worth <cworth@cworth.org>
32458 commit 13266b152ae35ed54984844e3282cbdf20dc2e32
32459 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32460 Date: Tue Mar 2 21:53:11 2010 +0100
32462 XvMC: kill i830_memory in 965 class xvmc
32464 Yes, this breaks binary compat of the struct passed around between
32465 X ddx and the client libXvMC. But we always ship both, so they should
32466 not get out of sync.
32468 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32469 Reviewed-by: Carl Worth <cworth@cworth.org>
32471 commit d39d822cf887a861b37cee92c0b59533370ded2f
32472 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32473 Date: Tue Mar 2 09:22:44 2010 +0100
32475 i830_memory: hide as much of the old memory allocator as possible
32477 The only user left of this stuff is the xvmc support.
32479 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32480 Reviewed-by: Eric Anholt <eric@anholt.net>
32482 commit 65267d4bfbf19942beab72858333c6ee3c719223
32483 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32484 Date: Tue Mar 2 09:22:43 2010 +0100
32486 i830_memory: switch frontbuffer to drm_intel_bo
32488 Yet another user of i830_memory gone for good.
32490 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32491 Reviewed-by: Eric Anholt <eric@anholt.net>
32493 commit 06f147dc04629a8a1534703be570e7f25e41cdd9
32494 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32495 Date: Tue Mar 2 09:22:42 2010 +0100
32497 i830_memory: switch cursors to drm_intel_bo
32499 Minus one user of i830_memory, some more to go.
32501 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32502 Reviewed-by: Eric Anholt <eric@anholt.net>
32504 commit 23d12e3b088f38d25e83d6501d553d98be442d49
32505 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32506 Date: Tue Mar 2 09:22:41 2010 +0100
32508 i830_memory: kill field "pitch"
32512 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32513 Reviewed-by: Eric Anholt <eric@anholt.net>
32515 commit 2fb8feeb56fae364380ffd91749eeb69fb461cff
32516 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32517 Date: Tue Mar 2 09:22:40 2010 +0100
32519 i830_memory: rip out field "size"
32521 Use the one in the drm bo instead.
32523 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32524 Reviewed-by: Eric Anholt <eric@anholt.net>
32526 commit 4d4d763b3d96e2cac99a0b7b03ad9bcaf4465c61
32527 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32528 Date: Tue Mar 2 09:22:39 2010 +0100
32530 i830_memory: kill field "tiling_mode"
32534 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32535 Reviewed-by: Eric Anholt <eric@anholt.net>
32537 commit d0800d098739d5e8ea94be9c5ed2f7a8a86c06dc
32538 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32539 Date: Tue Mar 2 09:22:38 2010 +0100
32541 i830_memory: no memory allocations without a bo!
32543 Kill the corresponding !bo path in i830_free_memory.
32545 Also kill another remnant of the pre-kms era in the same file, while I
32546 was looking at the code.
32548 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32549 Reviewed-by: Eric Anholt <eric@anholt.net>
32551 commit 086c0e25cac1d3dd0a37def8b5cb82c1c6279edf
32552 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32553 Date: Tue Mar 2 09:22:37 2010 +0100
32555 i830_memory: rename i830_bind_all_memory to reflect code reality
32557 It doesn't bind anything anymore, but does a few random things.
32558 Give it a hopefully vague enough name to cover all cases ;)
32560 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32561 Reviewed-by: Eric Anholt <eric@anholt.net>
32563 commit 2a989aa057cee74154419fd0a4911ba1e95582cf
32564 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32565 Date: Tue Mar 2 09:22:36 2010 +0100
32567 i830_memory: rip out the old video memory allocator
32569 Besides the debug stuff the went away in the previous patch,
32570 this stuff was totally unused ...
32572 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32573 Reviewed-by: Eric Anholt <eric@anholt.net>
32575 commit bf83b9a10254966cb73b24e08954154d4296dac3
32576 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32577 Date: Tue Mar 2 09:22:35 2010 +0100
32579 i830_memory: kill i830_desribe_allocations
32581 Totally useless debug function from the pre-gem era. No point
32582 to occasionally spam Xorg.log with a bogus "No memory allocations"
32585 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32586 Reviewed-by: Eric Anholt <eric@anholt.net>
32588 commit ff8a1e1cf728eeae848f6284bb4272193dfa918b
32589 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32590 Date: Tue Mar 2 09:22:34 2010 +0100
32592 i830_memory: rip out field "offset"
32594 Use the one in the drm bo instead.
32596 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32597 Reviewed-by: Eric Anholt <eric@anholt.net>
32599 commit e18ffceb142a9e18968900dd526a2d50fab72900
32600 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32601 Date: Tue Mar 2 09:22:32 2010 +0100
32603 i830_memory: rip out field "end"
32605 It's a left-over from the non-gem era and no longer used at all.
32607 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32608 Reviewed-by: Eric Anholt <eric@anholt.net>
32610 commit 5018fd3097d77a5f31af4cb27e39daa37557b64e
32611 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32612 Date: Mon Mar 1 22:57:40 2010 +0100
32614 libIntelXvMC: kill ums leftovers
32616 On i965 class hw, kernel_exec_fencing was 1 always, anyway. And on
32617 i945, this patch kills a memory leak (dunno how, but it does).
32619 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32621 commit cb06aa32d433f54affe87da2cf964f0308d3c258
32622 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32623 Date: Mon Mar 1 22:57:39 2010 +0100
32625 Xv: fixup XvMC on i915
32627 I've accidentally broken i915 xvmc due to alignment constrains that
32628 break my assumption that Y-pitch == UV-pitch*2. Fix this up by consistenly
32629 using dstPitch2 for the Y-pitch. This also unifies the dst pitch
32630 computation slightly, now that the i915 xvmc special case is gone.
32632 Bugzilla: http://bugs.freedesktop.org/show_bug.cgi?id=25949
32634 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32636 (Minor edit to support compilation without INTEL_XVMC defined by
32637 Carl Worth <cworth@cworth.org>)
32639 commit 80f2a1f3b5e477a12899ea2251aebcfa62689943
32640 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
32641 Date: Mon Mar 1 22:57:38 2010 +0100
32643 Xv: fixup YUV plane offset for xvmc case
32645 In my previous cleanup I've inadvertedly dropped the offset adjustment
32646 code for the xvmc passthrough case. Fix this up.
32648 Also reimplement that ugly hack I've accidently killed to keep i915 class
32649 xvmc a tad bit longer on life support.
32651 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
32652 Tested-by: xunx.fang@intel.com
32654 commit a0ee9c3d9c72962c8d513ec8c43dd4a21e316947
32655 Author: Carl Worth <cworth@cworth.org>
32656 Date: Fri Feb 26 17:05:45 2010 -0800
32658 Update version number to 2.10.901
32660 In order to make a new snapshot.
32662 commit 63b8e890a134e5fc8242b655115491a3905ad0ff
32663 Author: Carl Worth <cworth@cworth.org>
32664 Date: Fri Feb 26 17:05:06 2010 -0800
32666 NEWS: Add notes for the 2.10.901 snapshot.
32668 From skimming the git lot since 2.10.0.
32670 commit d5409303128f5c9ba940ab564ff407a82eabc597
32671 Author: Carl Worth <cworth@cworth.org>
32672 Date: Fri Feb 26 16:12:50 2010 -0800
32674 Remove dead code: fill_detailed_lvds_block
32676 All callers of this function were recently removed, so it can join
32677 them in the bit bucket.
32679 commit 7d0e6ff4dadcf243b1006ce6f85bd06c5f4e4908
32680 Author: Adam Jackson <ajax@redhat.com>
32681 Date: Wed Dec 2 12:05:52 2009 -0500
32683 kms: Fix LVDS mode list construction.
32685 Rather than mangle the EDID block and hope the server does the right
32686 thing, just build a sensible mode list up front. Do this for LVDS where
32687 there is no EDID or where it does not claim to be continuous-frequency
32688 (since in the latter case, the server will add reasonable modes for us).
32690 Signed-off-by: Adam Jackson <ajax@redhat.com>
32692 commit 1730af4437530ab7f6fe6ca74b16efb4a843b3ae
32693 Author: Carl Worth <cworth@cworth.org>
32694 Date: Thu Feb 25 17:12:49 2010 -0800
32696 NEWS: Fix typo (we just released 2.10 *not* 1.10)
32698 Sigh. Every version number that must be typed manually will eventually
32699 be typed incorrectly at some release. (I think that's Owne Taylor's
32702 commit 529bf185fbcb9f7705b315a5106054ee25c1c77f
32703 Author: Eric Anholt <eric@anholt.net>
32704 Date: Wed Feb 24 17:54:13 2010 -0800
32706 In frame event handling, track drawable id instead of drawable pointer.
32708 Windows aren't refcounted, so if the event came in after the window
32709 was destroyed, we'd dereference garbage and segfault.
32711 commit 633c7033170b0e9b468dbee444b94922f6c30940
32712 Author: Eric Anholt <eric@anholt.net>
32713 Date: Wed Feb 24 17:40:30 2010 -0800
32715 Fix up a bunch of inconsistent 4-space indentation in i830_dri.c
32717 commit 9291828a569a01ed4a6ef71f530b93f8a54c84aa
32718 Author: Eric Anholt <eric@anholt.net>
32719 Date: Fri Jan 15 12:51:50 2010 -0800
32721 Add new mobile Sandybridge PCI IDs.
32723 commit 3c71f98b9e5262675e61fafb317d0c35e62a873f
32724 Author: Eric Anholt <eric@anholt.net>
32725 Date: Thu Oct 22 16:55:02 2009 -0700
32727 Add initial defines and probing for Sandybridge
32729 commit c2c670ef18755cf5c878edf8a6b7d1617f54fe73
32730 Author: Eric Anholt <eric@anholt.net>
32731 Date: Fri Feb 19 12:54:48 2010 -0500
32733 Remove more DRI1 dead code.
32735 commit b320449f75adf912e06832dacc0b861ce26353bd
32736 Author: Eric Anholt <eric@anholt.net>
32737 Date: Fri Feb 19 12:44:11 2010 -0500
32739 Remove dead i830_ring.h.
32741 commit 6d48b26098691d772a8de001c89a05b2442ef3b5
32742 Author: Eric Anholt <eric@anholt.net>
32743 Date: Fri Feb 19 12:27:55 2010 -0500
32745 Remove dead i830_display.h.
32747 commit 4126f73051aa73defcf342c81aa0fc9062f83f90
32748 Author: Eric Anholt <eric@anholt.net>
32749 Date: Fri Feb 19 12:25:47 2010 -0500
32751 Remove dead i830_bios.h.
32753 commit ca173c1eaf0e28f7fa138b1965f5e768b3add895
32754 Author: Eric Anholt <eric@anholt.net>
32755 Date: Fri Feb 19 12:24:00 2010 -0500
32757 Remove the last tool now that it's been moved to gpu_tools.
32759 commit 6da7cda5830817b0c977ed4767a9432748f12b09
32760 Author: Eric Anholt <eric@anholt.net>
32761 Date: Fri Feb 19 12:21:47 2010 -0500
32763 Remove swf_dumper. We have intel_mmio_read in gpu_tools to do this.
32765 commit 71c613c8475909e58fc491131bf9c10f4b047b16
32766 Author: Eric Anholt <eric@anholt.net>
32767 Date: Fri Feb 19 12:20:33 2010 -0500
32769 Remove bios_reader now that it's moved to gpu_tools.
32771 commit 7ada4eb3836f58a72f1c5b6b8139030fab9b7d63
32772 Author: Eric Anholt <eric@anholt.net>
32773 Date: Fri Feb 19 12:18:23 2010 -0500
32775 Remove intel_hotplug tool now that it won't work with KMS.
32777 We'd want to put something like it in sysfs or something if we wanted
32778 this tool to exist again.
32780 commit 3284aacfe095f6260eee5ed126685d0638626a29
32781 Author: Eric Anholt <eric@anholt.net>
32782 Date: Fri Feb 19 12:14:12 2010 -0500
32784 Remove dead i830_bios.c.
32786 Things that used it are now either in the kernel or stashed off in
32789 commit 761d386a940081688b9b599c7a5b0270abfc5138
32790 Author: Eric Anholt <eric@anholt.net>
32791 Date: Fri Feb 19 12:13:25 2010 -0500
32793 Remove intel_gtt and intel_lid now that they're in gpu_tools.
32795 commit 6199af00cb47df6347ecd72ff7f6c86a564cec57
32796 Author: Eric Anholt <eric@anholt.net>
32797 Date: Fri Feb 19 12:06:18 2010 -0500
32799 Remove xprintf.c from the old "I can't believe it's not an X Server" tools.
32801 commit cba8e3136a100ba274f48c78df5b23fa8960fa16
32802 Author: Eric Anholt <eric@anholt.net>
32803 Date: Fri Feb 19 12:04:51 2010 -0500
32805 Remove intel_statuspage, now that we have /debug/dri/0/i915_gem_hws
32807 commit 40f5f72e30003993bfe298cd634d4f3f56148d4a
32808 Author: Eric Anholt <eric@anholt.net>
32809 Date: Fri Feb 19 12:01:30 2010 -0500
32811 Remove dead main.c from reg_dumper/
32813 commit 0c5e4a65f5d0f8d41c6bcb4c6978054f40064494
32814 Author: Eric Anholt <eric@anholt.net>
32815 Date: Fri Feb 19 12:00:56 2010 -0500
32817 Delete the audio dumping tool now that it lives in gpu_tools.
32819 commit 6bdab841766c8d9febaa23cb01a7bbc9edfae471
32820 Author: Eric Anholt <eric@anholt.net>
32821 Date: Fri Feb 19 11:21:40 2010 -0500
32823 uxa: Skip adjusting mask coordinates when no mask is present.
32825 Quiets clang warnings about garbage variable usage.
32827 commit a86869e6c3131b83a2ad529bc313270a9f45f5bd
32828 Author: Eric Anholt <eric@anholt.net>
32829 Date: Fri Feb 19 08:15:30 2010 -0800
32831 Fix an unused variable warning for !INTEL_XVMC.
32833 commit ec5deb2bcbf828c3c47488eb77461c26cab38ed6
32834 Author: Eric Anholt <eric@anholt.net>
32835 Date: Fri Feb 19 08:11:00 2010 -0800
32837 Remove dead assignments noticed by clang.
32839 commit a4180eabfa00c256494ed09aa336a17605c49a8a
32840 Author: Eric Anholt <eric@anholt.net>
32841 Date: Fri Feb 19 08:10:17 2010 -0800
32843 Remove stale junk in VideoRam setup, noticed by clang.
32845 commit 03657d4a698e0a25ab3863a15ad4c8b9d64dbb9a
32846 Author: Eric Anholt <eric@anholt.net>
32847 Date: Fri Feb 19 08:03:44 2010 -0800
32849 Remove dead UMS CRTC resize code.
32851 commit 1c3aaad09d6ef207fba748ad4ef4575a26ab2e5c
32852 Author: Chris Wilson <chris@chris-wilson.co.uk>
32853 Date: Wed Feb 17 17:48:32 2010 +0000
32855 dri2: Silence the compiler for an unused function with proto < 4
32857 Move the unused function into the #if DRI2INFOREC_VERSION >= 4 block.
32859 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32861 commit 00e7312dc45e54cd4547a943897a524639cb0b38
32862 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
32863 Date: Thu Feb 11 09:46:12 2010 -0800
32865 DRI2: handle full height blits without tearing in CopyRegion
32867 On 965 and up, if we detect a full height blit, we should just wait for
32868 vblank, rather than try to do a scanline wait for the whole display.
32870 On pre-965, doing a scanline wait followed by a blit works, but in the
32871 full height case we need to give the blitter time to start up, so we
32872 wait until the bottom line of the blit minus 2 padding scanlines to
32875 Fixes FDO bug #22475.
32877 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
32879 commit f0d760bfd71e2b2b40d2b250cd84b4626492ba47
32880 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
32881 Date: Wed Feb 10 11:24:46 2010 -0800
32883 Disable bo reuse on shadow framebuffer
32885 This keeps us from trying to set tiling on it while pinned, which also
32886 keeps us from trying to unpin it in the kernel, causing an error.
32888 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
32890 commit 41784e15d39c29af5a87fe8d0982bd1b0be3562d
32891 Author: Gaetan Nadon <memsize@videotron.ca>
32892 Date: Fri Jan 29 15:54:16 2010 -0500
32894 config: remove dead LINUXDOC macro usage
32896 This module does not generate LINUXDOC documentation.
32898 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
32900 commit 6610bcbac51c9ac970128012f0d4566d8cfba000
32901 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
32902 Date: Fri Feb 5 17:18:50 2010 -0800
32904 DRI2: only use version 4 APIs if kernel support exists
32906 Check for page flipping support before enabling flip and vblank event
32907 support needed for the new DRI2 APIs.
32909 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
32911 commit 1a76fa5574e8e8f88ac3518a4e4494e1af301dc1
32912 Author: Keith Packard <keithp@keithp.com>
32913 Date: Fri Jan 29 23:28:46 2010 -0800
32915 Initialize DRI2 info rec version 4 list of driver names
32917 With DRI2 supporting multiple subsystems, the video driver must
32918 initialize the list of driver names instead of just passing the single
32919 driver name used by Mesa. Without this, the X server will fail to
32920 initialize DRI2 as the numDrivers field in this structure will be
32923 Signed-off-by: Keith Packard <keithp@keithp.com>
32925 commit 918151a7955c26174db80b775205f6ffb4f44ab6
32926 Author: Chris Wilson <chris@chris-wilson.co.uk>
32927 Date: Wed Jan 27 20:33:08 2010 +0000
32929 uxa: Fix compatible_formats() for OVER
32931 In separating the boolean logic out into a separate function, dc6522dd,
32932 I reversed the sense of one particular test:
32934 src->format == dst->format
32936 The OVER optimisation is only valid if the src and dst formats match,
32939 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32941 commit 197cb08a2d54cabbfe97454d7db85cfe1f5f27ba
32942 Author: Chris Wilson <chris@chris-wilson.co.uk>
32943 Date: Mon Jan 25 15:15:04 2010 +0000
32945 Extract pixel value for all formats to avoid hitting fallbacks.
32947 On failing to extract the pixel value for an alpha-only solid we
32948 actually triggered a fallback. Since this path is commonly hitting
32949 whilst fading in images, for example cairo_paint_with_alpha(), the
32950 fallback was detected during the Moblin boot sequence where it was
32951 adding a second to the overall boot time.
32954 fallback intel: Moblin startup is hitting a composite fallback, costing
32955 a ton of performance
32956 https://bugs.freedesktop.org/show_bug.cgi?id=26189
32958 Based on the initial patch by Arjan van de Van.
32960 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32962 commit 5f93d019dc6311dd16b6792ffb60dbfc45ef3d08
32963 Author: Chris Wilson <chris@chris-wilson.co.uk>
32964 Date: Mon Jan 25 15:47:11 2010 +0000
32966 uxa: Adjust uxa_get_color_for_pixmap to match prototype
32968 The prototype says this function returns a Bool and not just an int, so
32969 be pedantic and return TRUE/FALSE.
32971 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32973 commit dc6522dd491831f3843606e3ed57cc01fe2c804c
32974 Author: Chris Wilson <chris@chris-wilson.co.uk>
32975 Date: Sun Jan 24 09:37:23 2010 +0000
32977 uxa: Protect against a potential NULL src->Drawable reference
32979 One of the convoluted if branches dereferenced Drawable when it is
32980 potentially NULL. Avoid this by explicitly handling the NULL Drawable
32981 cases earlier, and enabling solid fills for solid sources.
32983 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32985 commit 31bbd7f919f4f1f545dd49861c15e60561c075cb
32986 Author: Chris Wilson <chris@chris-wilson.co.uk>
32987 Date: Sun Jan 24 09:04:16 2010 +0000
32989 uxa/uxa-render: Always remove useless repeats during composite.
32991 I added a jump if there was no src or mask Drawable, but we do actually
32992 need to check for useless src repeats even if we have a source-only
32995 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
32997 commit 326fe00df4160fbc3513ddbedfec90bdb2d7101b
32998 Author: Chris Wilson <chris@chris-wilson.co.uk>
32999 Date: Sun Jan 24 09:02:05 2010 +0000
33001 uxa: Increase amount of composite fallback verbage
33003 The fallback log for http://bugs.freedesktop.org/show_bug.cgi?id=26189
33004 does not actually state the reason why we actually fallback. This is
33005 possibly because we need to fallback for reasons other than the
33006 operation cannot be performed in hardware -- such as using an alpha map
33007 or the screen is swapped out, so add this information to the fallback
33010 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33012 commit 93cd943d41c646c794b8cb5a960d8f0805e15395
33013 Author: Eric Anholt <eric@anholt.net>
33014 Date: Fri Jan 15 12:53:47 2010 -0800
33016 intel: Use the compositing-aware colorkey filler instead of homebrew fail.
33018 Of course, it's still fail since you can't correctly composite
33019 colorkey overlay, but at least this doesn't spam colorkey to the root
33022 Tested-by: Daniel Vetter <daniel@ffwll.ch>
33024 commit 6e61de896b42cffe1af6b0384cb69265d3d121a1
33025 Author: Alan Coopersmith <alan.coopersmith@sun.com>
33026 Date: Fri Jan 15 15:51:28 2010 -0800
33028 Update Sun license notices to current X.Org standard form
33030 Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
33032 commit 96f45c66eed2631eba98ae416c3afdf540fa5c34
33033 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
33034 Date: Wed Jan 13 13:20:43 2010 -0500
33036 DRI2: if the swap condition is satisfied, complete it immediately
33038 If we get to the point where we check the divisor/remainder equation and
33039 it's satisfied, we should complete the swap immediately.
33041 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
33043 commit 51c75906329a4727e37c8d1f64f257ea9602caa2
33044 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
33045 Date: Fri May 1 14:52:26 2009 -0700
33047 DRI2: support new DRI2 APIs
33049 The new interfaces allow for improved buffer swap, and support for the
33050 SGI_swap_control, SGI_video_sync and OML_sync_control GLX extensions.
33052 The Intel implementation allows page flipping to occur for swaps that
33053 are full screen and not rotated.
33055 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
33057 commit 4902f546be19e3d5bb47f6c75e2199dc4856c0f4
33058 Author: Chris Wilson <chris@chris-wilson.co.uk>
33059 Date: Sun Dec 13 10:44:12 2009 +0000
33061 i965: Ensure that URB_FENCE is aligned to 64-bytes
33063 The PRM (Vol 1, p32) specifies that the URB_FENCE command must not cross
33064 a cache-line boundary (64-bytes) in order to workaround a silicon issue.
33065 Ensure that it does not by inserting an alignment point before the atomic
33068 This is a slightly too large hammer, but the easiest method to work with
33069 the current BEGIN_BATCH/ADVANCE_BATCH protections.
33071 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33073 commit 83626aba357ffb4dd7931daaf163c1dd1d08f9d3
33074 Author: Chris Wilson <chris@chris-wilson.co.uk>
33075 Date: Sun Nov 29 21:39:41 2009 +0000
33077 uxa-glyphs: Enable TILING_X on glyph caches.
33079 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33081 commit 50e07da8094c8c8c593b6eb8c41fc42444851d04
33082 Author: Chris Wilson <chris@chris-wilson.co.uk>
33083 Date: Sun Dec 13 09:35:36 2009 +0000
33085 i830: Do not use vtSema when chosing mapping type.
33087 The mapping type to use is determined by the tiling of the underlying
33088 object, not by whether or not not we control the vt. This was a
33089 left-over wart that was intended to mean that we had GEM and so could
33092 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33094 commit 7a2b7cfab5cdef277f0feb838683422d9fcb0db3
33095 Author: Chris Wilson <chris@chris-wilson.co.uk>
33096 Date: Sun Dec 13 09:33:45 2009 +0000
33098 Consolidate determining maximum sizes for use with GEM
33100 Add a small wrapper function so that the callsites need only call the
33101 single function when checking the available aperture size for
33102 determining the maximum viable size for operations. This will allow us
33103 to easily extend this set in the future by only needing to adding the
33104 check to a single location.
33106 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33108 commit 229d23fb18d696fb7ad476ce335be14ec9811bd3
33109 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33110 Date: Tue Dec 1 14:32:34 2009 +0100
33112 Xv: don't enable XVMC port on unsupported configs
33114 This just makes it _really_ clear, what's supported. No other changes.
33116 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33117 Signed-off-by: Eric Anholt <eric@anholt.net>
33119 commit ce7ba18f577cc9aedddaff303dbc9662a276b4cf
33120 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33121 Date: Tue Dec 1 14:32:33 2009 +0100
33123 Xv: consolidate xmvc passthrough handling
33125 It's now all in I830PutImageTextured. Also kill some leftovers
33126 from XVMC-on-overlay support and ums-XVMC-on-i915 support. Plus
33127 a small comment as a reminder for where to add i915 xvmc support
33130 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33131 Signed-off-by: Eric Anholt <eric@anholt.net>
33133 commit 1ac7c94083a3266e2d5dd932709118436074dd00
33134 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33135 Date: Tue Dec 1 14:32:32 2009 +0100
33137 Xv: hide ugly semantics in i830_clip_video_helper
33139 I'm still curious as to why fixed-point semantics are necessary
33140 for this generic XV helper function that's been causing all this.
33141 Can modern X really run on hw without floating-point support?
33143 Anyway, the ugliness is now all nicely under the carpet (in
33144 i830_clip_video_helper).
33146 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33147 Signed-off-by: Eric Anholt <eric@anholt.net>
33149 commit 31f13fa8a0d4cd28067de37d0d31d23e8200d0ff
33150 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33151 Date: Tue Dec 1 14:32:31 2009 +0100
33153 Xv: move users of x1, x2, y1, y2 to PutImage
33155 After this there are no other external users of these strange variables,
33156 so we can nicely hide them somewhere in the next changeset.
33158 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33159 Signed-off-by: Eric Anholt <eric@anholt.net>
33161 commit ce6526b9b4df304ccd83a0a02a95621300dbaed3
33162 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33163 Date: Tue Dec 1 14:32:30 2009 +0100
33165 Xv: kill unnecessary parameters for hw PutImage functions
33167 This is the first part of my small crusade to rip out x1, x2, y1, y2
33168 from I830PutImage*. These variables have strange semantics (they
33169 change from simple integers to fixed-point values somewhere in
33170 the middle) and don't really seem to be what we actually need.
33172 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33173 Signed-off-by: Eric Anholt <eric@anholt.net>
33175 commit cf74caaa91667457bba3583b216dfc157d09f39e
33176 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33177 Date: Tue Dec 1 14:32:29 2009 +0100
33179 Xv: kill an unnecessary if
33181 We always pass a non-null pointer for crtc_ret, no point to check
33184 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33185 Signed-off-by: Eric Anholt <eric@anholt.net>
33187 commit d8353c737b5b8077f499461b1e6c78f09544d226
33188 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33189 Date: Tue Dec 1 14:32:28 2009 +0100
33191 Xv: split up I830PutImage into textured and !textured case
33193 This wasn't making much sense anymore, and further cleanups will
33194 make this even more apparent. This change just makes two copies of
33195 I830PutImage and kills the not-applicable if-clauses in both
33198 There is one small functional change in here: The textured video
33199 path doesn't munch around with adaptor_priv->videoStatus anymore,
33200 which is only used by the overlay. This could prevent the overlay
33201 from being switched off if someone would use textured video at the
33204 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33205 Signed-off-by: Eric Anholt <eric@anholt.net>
33207 commit 091035146463bf1aa6674bff6947d04fc620c18f
33208 Author: Carl Worth <cworth@cworth.org>
33209 Date: Mon Jan 4 14:20:11 2010 -0800
33211 configure.ac: Bump version to 2.10.0.
33213 In preparation for the 2.10.0 release.
33215 commit 01f7d0307b3f5f085e383ff85026757de9cda359
33216 Author: Carl Worth <cworth@cworth.org>
33217 Date: Mon Jan 4 14:19:17 2010 -0800
33219 NEWS: Add final release notes for the 2.10.0 release.
33221 Mentioning that it's functionally identical to our most recent release
33224 commit e966bca3a9143cff37beb54359bea13c9f693b44
33225 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
33226 Date: Tue Dec 29 21:49:49 2009 -0800
33228 Take note for Pineview support in README and manpage
33230 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
33232 commit 6137791e9f1a4778afabbbc06de3da68648ae258
33233 Author: Carl Worth <cworth@cworth.org>
33234 Date: Mon Jan 4 14:08:24 2010 -0800
33236 man page: Remove section describing BACKLIGHT_CONTROL property.
33238 This apparently no longer exists in a KMS world, so remove it from the
33241 commit 7f36a439249cc2101d8985de5e95af652e5b984b
33242 Author: Carl Worth <cworth@cworth.org>
33243 Date: Mon Jan 4 14:07:24 2010 -0800
33245 man page: Rename PANEL_FITTING to "scaling mode".
33247 The old UMS name was PANEL_FITTING while the new KMS name is "scaling mode".
33251 commit a6fb71e6e26fb31f4cd10fad2d3a87d9114a649a
33252 Author: Carl Worth <cworth@cworth.org>
33253 Date: Mon Jan 4 14:05:27 2010 -0800
33255 man page: Add additional indentation for some output configuration options.
33257 The BACKLIGHT_CONTROL and PANEL_FITTING options appear in a list, and
33258 then each contain a sub-list of sub-options. Use indentation to make
33259 this structure more apparent to the reader.
33261 commit 25a6c8dfae240143309b14cc32ebac6008c3a378
33262 Author: Gaetan Nadon <memsize@videotron.ca>
33263 Date: Tue Dec 15 21:46:48 2009 -0500
33265 configure.ac: use backticks rather than $() for cmd subs
33267 Use "$PKG_CONFIG" rather than hard coded "pkg-config"
33269 Acked-by: Dan Nicholson <dbn.lists@gmail.com>
33270 Acked-by: Daniel Stone <daniel@fooishbar.org>
33272 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
33274 commit 2c142e421e859406b5aff16ba18624150269fc06
33275 Author: Carl Worth <cworth@cworth.org>
33276 Date: Thu Dec 10 15:25:44 2009 -0800
33278 Update version to 2.9.99.902.
33280 For the second release-candidate snapshot in preparation for 2.10.
33282 commit 8ecf70ea553083cbc26928dc3973c8f6f8b3d9d0
33283 Author: Carl Worth <cworth@cworth.org>
33284 Date: Thu Dec 10 15:17:57 2009 -0800
33286 NEWS: Add notes for 2.9.99.902 (and preliminary notes for 2.10)
33288 This comes from my cursory glance over the commit log from 2.9.1
33291 commit 88b2209a7526a7c7f3b3a2d491d48bd0d5d56e04
33292 Author: Carl Worth <cworth@cworth.org>
33293 Date: Thu Dec 10 14:52:46 2009 -0800
33295 Document the DebugFlushBatches, DebugFlushCaches, and DebugWait options.
33297 These were added in 3c0815abf28744e215bea286e71d935cd486955a . The
33298 documentation added here comes straight from that commit message.
33300 commit 8d6faf7dc546bda64e11bf6dea7f3c997f07f887
33301 Author: Tobias Doerffel <tobias.doerffel@gmail.com>
33302 Date: Wed Dec 9 09:57:20 2009 +0100
33304 Fix compiler warning in i830_copy_video_data() if XvMC is disabled
33306 The variable "intel" is unused when building i830_video.c without XvMC
33307 support which results in a compiler warning:
33309 i830_video.c: In function 'i830_copy_video_data':
33310 i830_video.c:1443: warning: unused variable `intel'
33312 Trivial fix via #ifdef.
33314 commit bd81734465912d79d6320a6fb021ce43d258b906
33315 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
33316 Date: Tue Dec 8 15:27:41 2009 +0100
33318 Xv: enable drmmode overlay
33320 Now that libdrm 2.4.16 is released (and already required) we can
33321 unconditionally enable this.
33323 Please add something like this to the release-notes/NEWS file:
33325 * Overlay support for kernel modesetting. This needs at least kernel
33326 v2.6.33 to work. A backport to 2.6.32 is available at:
33328 http://gitorious.org/daniel-s-linux-stuff/linux-kernel/commits/intel-kms-overlay-for-2.6.32
33330 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
33332 commit 37f631d669c165c4fb56ccd7a6fc0a432f453b52
33333 Author: Chris Wilson <chris@chris-wilson.co.uk>
33334 Date: Thu Dec 10 08:50:16 2009 +0000
33336 Revert "uxa-glyphs: Enable TILING_X on glyph caches."
33338 This reverts commit 3f11bbec420080151406c203af292e55177e77d1.
33340 For unknown reasons, enabling tiling for the glyph cache is causing
33341 glyph corruption both across suspend and resume and VT switching, on a
33342 wide range of chipsets (reports include both i8xx and gm45)
33344 This strongly suggests that we are handling tiling, or updates to tiled
33345 buffers, incorrectly across i915_gem_idle(). However, until we can find
33346 the root cause, we want to fix this regression before the next stable
33347 release, so simply revert this patch. :(
33350 [Bug 25406] fonts garbled after resuming from suspend since 6729b508
33351 http://bugs.freedesktop.org/show_bug.cgi?id=25406
33353 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33355 commit 093bb9ebe69760975a3fcf2322db950312e6c2d7
33356 Author: Chris Wilson <chris@chris-wilson.co.uk>
33357 Date: Tue Dec 8 23:45:28 2009 +0000
33359 i965: Only use the affine kernels if both src and mask are affine
33361 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33363 commit 0cf04ea4d736f7d7848f33b772d88a0f1b6678b1
33364 Author: Chris Wilson <chris@chris-wilson.co.uk>
33365 Date: Tue Dec 8 23:44:45 2009 +0000
33367 i965: Set src_filter before testing.
33369 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33371 commit 417ad2712edcbca635eb5eeff0d11cdb16c069ed
33372 Author: Chris Wilson <chris@chris-wilson.co.uk>
33373 Date: Tue Dec 8 23:43:57 2009 +0000
33375 Assert that we only call OUT_BATCH() inside a BATCH
33377 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33379 commit 47416b1eea09b238a997636d35998d71e0d18161
33380 Author: Chris Wilson <chris@chris-wilson.co.uk>
33381 Date: Tue Dec 8 13:47:07 2009 +0000
33383 i965: Maximum number of vertices per composite is 24, not 18
33385 Beware the potential buffer overflow.
33387 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33389 commit c1afc831c8fe4cbececee7dfa23506a6746c2425
33390 Author: Chris Wilson <chris@chris-wilson.co.uk>
33391 Date: Mon Dec 7 21:27:56 2009 +0000
33393 uxa: Cache solid fills.
33395 Maintain a small cache of pixmaps to hold SolidFill pictures. Currently
33396 we create a pixmap the size of the damaged region and fill that using
33397 pixman before downloading it to the GPU and compositing. Needless to say
33398 this is extremely expensive compared to simply emitting the solid
33399 colour. To mitigate this cost, we maintain a small cache of 1x1R
33400 pictures which is recognised by the driver as being a solid, but at the
33401 very least is maintained as a GPU ready pixmap.
33403 This gives a good boost to cairo-xcb (which uses solid fills) on a gm45:
33406 gnome-terminal-vim: 41.9s
33408 gnome-terminal-vim: 31.7s
33410 Compare with using a cache of 1x1R pixmaps in cairo-xcb:
33411 gnome-terminal-vim: 31.6s
33413 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33415 commit cd475bad23c02130d11c49882c11261c9f0d4ef1
33416 Author: Chris Wilson <chris@chris-wilson.co.uk>
33417 Date: Mon Dec 7 11:09:14 2009 +0000
33419 batch: Ensure we send a MI_FLUSH in the block handler for TFP
33421 This should restore the previous level of synchronisation between
33422 textures and pixmaps, but *does not* guarantee that a texture will be
33423 flushed before use. tfp should be fixed so that the ddx can submit the
33424 batch if required to flush the pixmap.
33426 A side-effect of this patch is to rename intel_batch_flush() to
33427 intel_batch_submit() to reduce the confusion of executing a batch buffer
33428 with that of emitting a MI_FLUSH.
33430 Should fix the remaining rendering corruption involving tfp [inc compiz]:
33432 Bug 25431 [i915 bisected] piglit/texturing_tfp regressed
33433 http://bugs.freedesktop.org/show_bug.cgi?id=25431
33435 Bug 25481 Wrong cursor format and cursor blink rate with compiz enabled
33436 http://bugs.freedesktop.org/show_bug.cgi?id=25481
33438 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33440 commit 415aab474edd1425034981306718afd8506445f1
33441 Author: Chris Wilson <chris@chris-wilson.co.uk>
33442 Date: Fri Dec 4 09:02:36 2009 +0000
33444 intel: And remember to flush the batch...
33448 Remove flush parameter from intel_batch_flush()
33450 Maxi spotted that I had broken screen updating. It appears in my haste
33451 to eliminate the extra parameter I removed a call to intel_batch_flush()
33452 when throttling, i.e. when pushing the updates to the screen before
33457 Bug 25409 [bisected] rendering corruption since a938673e
33458 https://bugs.freedesktop.org/show_bug.cgi?id=25409
33460 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33462 commit 8438372dcc9d15c7b708332cfcb3a1c16a2c5cf2
33463 Author: Kristian Høgsberg <krh@bitplanet.net>
33464 Date: Thu Dec 3 14:39:43 2009 -0500
33466 Require libdrm 2.4.16
33468 Needed for drmGetDeviceNameFromFd().
33470 commit c439207ec0cc16d7d9f523598fcdebf7cec893e2
33471 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
33472 Date: Wed Dec 2 14:43:17 2009 -0800
33474 Update man page to reflect currently available options
33476 Many have been removed or are obsolete now that UMS is gone. And some
33477 are only available on i810/i815 or i830+, so move them to the
33478 appropriate section.
33480 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
33482 commit a938673ee84d51ef655c37dfa7bbc5c35334cd28
33483 Author: Chris Wilson <chris@chris-wilson.co.uk>
33484 Date: Wed Dec 2 20:51:53 2009 +0000
33486 batch: Downgrade batch submission from a FatalError.
33488 If we wedge the GPU then we will return -EIO for the current batch and
33489 then attempt to reset the GPU. Meanwhile the X server detects the error,
33490 throws a FatalError and to all intents and purposes appears to crash to
33491 the user - whereas before it often just appeared to momentarily freeze.
33492 Of course, on older hardware the server remains frozen until we can find
33493 a way to reset those GPUs at runtime.
33495 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33497 commit 98e11210367c950e3f932419d2a4722cf971885d
33498 Author: Chris Wilson <chris@chris-wilson.co.uk>
33499 Date: Wed Dec 2 20:48:37 2009 +0000
33501 Remove flush parameter from intel_batch_flush()
33503 There is only a single caller that wishes to forcibly append a flush
33504 into the batch: intel_sync(). So move the logic there.
33506 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33508 commit 57336c26f1fb90d43851ddcf78539585b67d86d9
33509 Author: Chris Wilson <chris@chris-wilson.co.uk>
33510 Date: Wed Dec 2 20:42:41 2009 +0000
33512 Rename I830Sync() to intel_sync()
33514 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33516 commit 370157f4932cf9067ba81c8bd5a311aff610882b
33517 Author: Chris Wilson <chris@chris-wilson.co.uk>
33518 Date: Wed Dec 2 20:28:49 2009 +0000
33520 batch: Avoid flushing a NULL batch
33522 During shutdown from a FatalError during batchbuffer submission, it is
33523 possible for the batch_ptr to be NULL, so we must be careful not to
33524 append a flush on this error path.
33526 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33528 commit ad68881b670aabf8dbfd3b954e8796f91260579d
33529 Author: Chris Wilson <chris@chris-wilson.co.uk>
33530 Date: Wed Dec 2 14:14:39 2009 +0000
33532 uxa_check_composite: Minor whitespace.
33534 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33536 commit 49d2ccab2a82083110fe796636f3f91ba8c31237
33537 Author: Chris Wilson <chris@chris-wilson.co.uk>
33538 Date: Wed Dec 2 14:13:43 2009 +0000
33540 uxa_prepare_access() don't force a flush.
33542 Only the kernel knows whether the mapping requires a flush, so do not
33545 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33547 commit b68d3646f1fdfe012c16741958c7a62136a9b5aa
33548 Author: Chris Wilson <chris@chris-wilson.co.uk>
33549 Date: Wed Dec 2 14:12:19 2009 +0000
33551 Review use of errno after libdrm call
33553 Since drm may not actually set the appropriate errno after a failure, we
33554 must use the return code instead when determining the cause of failure.
33556 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33558 commit 0ff4d42a42b9e537b083343ee7dcc41cb41ae7cf
33559 Author: Chris Wilson <chris@chris-wilson.co.uk>
33560 Date: Wed Dec 2 12:12:07 2009 +0000
33562 uxa: Review uxa_prepare_access() to remove potential nesting
33564 Around a call to uxa_put_image() it is possible to mix both accelerated
33565 and fallback paths, with the fallback code making the presumed
33566 optimisation of only trying to call uxa_prepare_access() once. This
33567 fails if the accelerated path also uses prepare/finish access on the
33568 same drawable and then later fallback to the fallback path. This can
33569 happen currently if an error is reported whilst attempting to accelerate
33572 #0 memcpy () at ../sysdeps/x86_64/memcpy.S:162
33573 #1 0x00007ffff43ce4bd in fbBlt (srcLine=<value optimized out>, srcStride=40, srcX=<value optimized out>, dstLine=0xffffffffffffffff, dstStride=64, dstX=0, width=<value optimized out>, height=8, alu=3, pm=4294967295, bpp=8, reverse=0, upsidedown=0) at fbblt.c:93
33574 #2 0x00007ffff43ce740 in fbBltStip (src=0xffffffffffffffff, srcStride=156555204, srcX=34, dst=0xfffffffc, dstStride=64, dstX=40, width=304, height=8, alu=3, pm=4294967295, bpp=8) at fbblt.c:944
33575 #3 0x00007ffff4c32c53 in uxa_do_put_image (pDrawable=0x246aa410, pGC=0x2c0a4f0, depth=8, x=0, y=0, w=38, h=8, leftPad=0, format=2, bits=0x954d7c4 "") at uxa-accel.c:196 #4 uxa_do_shm_put_image (pDrawable=0x246aa410, pGC=0x2c0a4f0, depth=8, x=0, y=0, w=38, h=8, leftPad=0, format=2, bits=0x954d7c4 "") at uxa-accel.c:223
33576 #5 uxa_put_image (pDrawable=0x246aa410, pGC=0x2c0a4f0, depth=8, x=0, y=0, w=38, h=8, leftPad=0, format=2, bits=0x954d7c4 "") at uxa-accel.c:289
33577 #6 0x00000000004d574f in damagePutImage (pDrawable=0x246aa410, pGC=0x2c0a4f0, depth=8, x=0, y=0, w=38, h=8, leftPad=0, format=2, pImage=0x954d7c4 "") at damage.c:905
33578 #7 0x00000000004287db in ProcPutImage (client=0x47ca72d0) at dispatch.c:2073
33579 #8 0x000000000042bd94 in Dispatch () at dispatch.c:445
33580 #9 0x000000000042513a in main (argc=4, argv=0x7fffffffe2a8, envp=<value optimized out>) at main.c:285
33582 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33584 commit 6be26cae8379f973d9ca27e0d5371d16618e4f7b
33585 Author: Chris Wilson <chris@chris-wilson.co.uk>
33586 Date: Tue Dec 1 16:13:25 2009 +0000
33588 i830: Simplify prepare_access / finish_access
33590 Reduce the 3 conditions into the 2 distinct cases. This has the
33591 secondary benefit of also distinguishing between the reported errors.
33593 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33595 commit 637f003b047e426901cab6b1fe3a0924bcb9a38a
33596 Author: Chris Wilson <chris@chris-wilson.co.uk>
33597 Date: Tue Dec 1 13:20:20 2009 +0000
33599 uxa: Don't treat prepare_access as a flush synchronisation point.
33601 The kernel will only emit a flush iff the buffer is currently owned by
33602 the GPU. Instead of presuming that the kernel must emit a flush, it is
33603 safer to assume that it does not and so cannot mapping the buffer on to
33604 the CPU as a synchronisation point. The most obvious counter-example is
33605 when we map the same buffer twice without using it in a batch.
33607 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33609 commit cd5a9568ce0a541f030c27cdae529fe18e5f0437
33610 Author: Carl Worth <cworth@cworth.org>
33611 Date: Mon Nov 30 20:17:04 2009 -0800
33613 Add i830_bios.h and i830_display.h to EXTRA_DIST.
33615 These files have been dropped from the generated tar file since the
33616 removal of UMS support. However, the bios_reader code still includes
33617 these, so "make distcheck" fails unless these are distributed.
33619 There's probably a cleaner fix possible, but this at least fixes the
33620 build so that the snapshot can be pushed out.
33622 commit 5bdac72a5c3f68ab6d9f7f2f1dfa6e5b78f04e57
33623 Author: Carl Worth <cworth@cworth.org>
33624 Date: Mon Nov 30 20:16:19 2009 -0800
33626 NEWS: Note that the driver now has a hard dependency on KMS.
33628 Since the UMS code has all been removed.
33630 commit 813a910d7e4cc1c3deec0dcb1b536f8af337993a
33631 Author: Carl Worth <cworth@cworth.org>
33632 Date: Mon Nov 30 19:48:05 2009 -0800
33634 configure.ac: Bump version to 2.9.99.901.
33636 In preparation for a new snapshot.
33638 commit b62c72913a85895a60268a49eed1c136f63f160d
33639 Author: Carl Worth <cworth@cworth.org>
33640 Date: Mon Nov 30 19:46:59 2009 -0800
33642 NEWS: Add generic notes for the 2.9.99.901 snapshot.
33644 We plan to collect real release notes before the 2.10.0 release.
33646 commit 00aa4f7a45a318af5b651f9f3928e9da4443233a
33647 Author: Chris Wilson <chris@chris-wilson.co.uk>
33648 Date: Mon Nov 30 20:50:31 2009 +0000
33650 uxa: Limit maximum size of tiled objects
33652 On older chipsets (i.e. pre-i965) tiling is very restrictive and imposes
33653 severe size and alignment constraints. Combine that with relatively
33654 small apertures and it is very easy to create a batch buffer that
33655 cannot be mapped into the aperture (but would otherwise fit based purely
33656 on total object size). To prevent this we need to not use tiling for large
33657 buffers (the very same buffers where tiling would be of most benefit!).
33659 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33661 commit 8dd1c9eca02fb8da0c51f6fa4a38eb5e5ff41855
33662 Author: Gaetan Nadon <memsize@videotron.ca>
33663 Date: Sun Nov 22 17:13:02 2009 -0500
33665 video-intel: remove i2c_vid.h from src/Makefile.am
33667 make dist failed due to missing i2c_vid.h
33668 Commit b9b159c49854d8d9d2207946bb583537bb0d48d6 Remove UMS support.
33669 The above commit did not remove this header file from the makefile.
33671 Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
33673 commit 6729b508c44bfca0b0dbef238a8732adbed6d4c9
33674 Author: Chris Wilson <chris@chris-wilson.co.uk>
33675 Date: Mon Nov 30 16:52:10 2009 +0000
33677 uxa: Initialise lists for private pixmap structure.
33679 When updating a buffer object for the framebuffer, we may need to
33680 allocate a fresh pixmap private structure, for example if the pixmap is
33681 replaced due to resize. When doing so it is then imperative to
33682 initialise the circularly linked lists correctly.
33684 Should fix the fault:
33685 #0 i830_set_pixmap_bo (pixmap=0x24ab380, bo=0x24ab780) at i830_uxa.c:524
33686 #1 0x00007f8615c629fd in drmmode_xf86crtc_resize (scrn=0x247a320, width=1280, height=800) at drmmode_display.c:1345
33687 #2 0x000000000051246c in xf86RandR12ScreenSetSize (pScreen=0x24824f0, width=<value optimized out>, height=<value optimized
33688 out>, mmWidth=<value optimized out>, mmHeight=<value optimized out>) at xf86RandR12.c:709
33689 #3 0x0000000000512aa8 in xf86RandR12CreateScreenResources (pScreen=<value optimized out>) at xf86RandR12.c:839
33690 #4 0x0000000000514ec0 in xf86CrtcCreateScreenResources (screen=0x24824f0) at xf86Crtc.c:727
33691 #5 0x0000000000424fb3 in main (argc=<value optimized out>, argv=<value optimized out>, envp=<value optimized out>) at main.c:215
33692 as reported by 'buscher'.
33694 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33696 commit 85fe41126e83b35954dc2066eb103e0555e207a4
33697 Author: Chris Wilson <chris@chris-wilson.co.uk>
33698 Date: Mon Nov 30 15:57:42 2009 +0000
33700 uxa: Remove cache flush for copy
33702 As the copy uses the 2D blitter, it uses the render cache so the source
33703 should not require flushing if it has previously been used as a
33706 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33708 commit 2d434eed09936328a52cb942450a95aedc3eac63
33709 Author: Kristian Høgsberg <krh@bitplanet.net>
33710 Date: Mon Nov 30 10:27:42 2009 -0500
33712 Use new drmGetDeviceNameFromFd() for mapping to drm device name
33714 Should have been in libdrm to begin with.
33716 commit cfcabc45140d19bfbfa4737c0a11cdbb042d11eb
33717 Author: Chris Wilson <chris@chris-wilson.co.uk>
33718 Date: Mon Nov 30 13:58:30 2009 +0000
33720 i915: Disable centre-point sampling.
33722 I still have no idea how this is triggering failures, but it is. So
33723 revert until the problem is solved.
33725 Should fix once again:
33727 Bug 23803 [bisected i915] gnome characters disappear
33728 http://bugs.freedesktop.org/show_bug.cgi?id=23803
33730 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33732 commit 8f8b6bd03d275379918777eaf7f63c0157b7ed9d
33733 Author: Chris Wilson <chris@chris-wilson.co.uk>
33734 Date: Mon Nov 30 14:03:40 2009 +0000
33738 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33740 commit 47916ea9d351f0ce6dc34713d6c164b0032f9830
33741 Author: Chris Wilson <chris@chris-wilson.co.uk>
33742 Date: Mon Nov 30 13:58:06 2009 +0000
33744 debug: Enable dumping of batchbuffer [compile-time only]
33746 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33748 commit b118a52cd1a006321571967bd5f6c2a9e674de3a
33749 Author: Chris Wilson <chris@chris-wilson.co.uk>
33750 Date: Mon Nov 30 11:03:32 2009 +0000
33752 i915: Remove routing of alpha channel to green.
33754 This modification is redundant since the routing is done in the blend
33757 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33759 commit 5e04ded2bce4c135b57d391f5f4e24e030103e61
33760 Author: Chris Wilson <chris@chris-wilson.co.uk>
33761 Date: Mon Nov 30 10:57:04 2009 +0000
33763 i915: Fix missing texture offset for mask.
33765 In commit e581ceb, I modified the shader generation to accommodate mixed
33766 textures and solids but missed applying the new computed sampler for the
33771 Bug 23803 [bisected i915] gnome characters disappear
33772 http://bugs.freedesktop.org/show_bug.cgi?id=23803
33774 Bug 25031 rendering and color corruption since 14109abf
33775 http://bugs.freedesktop.org/show_bug.cgi?id=25031
33777 Bug 25047 [945GM bisected] rendercheck/repeat/triangles regressed
33778 http://bugs.freedesktop.org/show_bug.cgi?id=25047
33780 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33782 commit a8ea20100de5be45699e71eb6ba67bef546ad0cd
33783 Author: Chris Wilson <chris@chris-wilson.co.uk>
33784 Date: Mon Nov 30 10:13:18 2009 +0000
33786 debug: Don't always flush the batch when emitting a debugging flush
33788 I incorrectly changed the logic in 285f286 and caused the batch to
33789 always be flushed when debugging, instead of merely inserting a MI_FLUSH
33790 between operations.
33792 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33794 commit c10850c63f9f88fbf08135bc4dcef1e5a2c71ee6
33795 Author: Chris Wilson <chris@chris-wilson.co.uk>
33796 Date: Mon Nov 30 09:07:57 2009 +0000
33798 overlay: Fix build after 646b4a9483
33800 The compile cleanup was not without fault... Apparently I don't have
33801 XVMC enabled anymore and so missed that this variable is actually used.
33803 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33805 commit 3f11bbec420080151406c203af292e55177e77d1
33806 Author: Chris Wilson <chris@chris-wilson.co.uk>
33807 Date: Sun Nov 29 21:39:41 2009 +0000
33809 uxa-glyphs: Enable TILING_X on glyph caches.
33811 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33813 commit 19d8c0cf50e98909c533ebfce3a0dd3f72b755c1
33814 Author: Chris Wilson <chris@chris-wilson.co.uk>
33815 Date: Sun Nov 29 21:16:49 2009 +0000
33817 uxa: PutImage acceleration
33819 Avoid waiting on dirty buffer object by streaming the upload to a fresh,
33820 non-GPU hot buffer and blitting to the destination.
33822 This should help to redress the regression reported in bug 18075:
33824 [UXA] XPutImage performance regression
33825 https://bugs.freedesktop.org/show_bug.cgi?id=18075
33827 Using the particular synthetic benchmark in question on a g45:
33830 9542.910448 Ops/s; put composition (!); 15x15
33831 5623.271889 Ops/s; put composition (!); 75x75
33832 1685.520362 Ops/s; put composition (!); 250x250
33835 40173.865300 Ops/s; put composition (!); 15x15
33836 28670.280612 Ops/s; put composition (!); 75x75
33837 4794.368601 Ops/s; put composition (!); 250x250
33839 which while not stellar performance is at least an improvement. As
33840 anticipated this has little impact on the non-fallback RENDER paths, for
33841 instance the current cairo-xlib backend is unaffected by this change.
33843 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33845 commit f7540f06090753cba1190aa9e8cdea05a9512077
33846 Author: Chris Wilson <chris@chris-wilson.co.uk>
33847 Date: Sun Nov 29 21:12:07 2009 +0000
33849 Only flush batch during prepare access if it may modify the pixmap.
33851 As we track when a pixmap is active inside a batch buffer, we can avoid
33852 unnecessary flushes of the batch when mapping a pixmap back to the CPU.
33854 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33856 commit 9a2c18fb92659d57741bfdcacbe4f69aab361532
33857 Author: Chris Wilson <chris@chris-wilson.co.uk>
33858 Date: Sun Nov 29 21:07:45 2009 +0000
33860 batch: Emit a 'pipelined' flush when using a dirty source.
33862 Ensure that the render caches and texture caches are appropriately
33863 flushed when switching a pixmap from a target to a source.
33865 This should fix bug 24315,
33866 [855GM] Rendering corruption in text (usually)
33867 https://bugs.freedesktop.org/show_bug.cgi?id=24315
33869 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33871 commit 285f286597df5af13ac3f3d366f2fc9d0468dafa
33872 Author: Chris Wilson <chris@chris-wilson.co.uk>
33873 Date: Sun Nov 29 22:42:03 2009 +0000
33875 batch: Track pixmap domains.
33877 In order to detect when we require cache flushes we need to track which
33878 domains the pixmap currently belongs to. So to do so we create a device
33879 private structure to hold the extra information and hook it up.
33881 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33883 commit 2c3aee2b570dadd9270a08d8ff675d07ac405e33
33884 Author: Chris Wilson <chris@chris-wilson.co.uk>
33885 Date: Sun Nov 29 20:53:35 2009 +0000
33887 uxa-glyphs: Stream uploads via temporary bo
33889 Avoid mapping the glyph cache back to the cpu by allocating temporary
33890 buffer objects to store the glyph pixmap and blit to the cache.
33892 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33894 commit 646b4a9483c01509a7324cc05eaadb72bc940c6d
33895 Author: Chris Wilson <chris@chris-wilson.co.uk>
33896 Date: Sun Nov 29 10:53:36 2009 +0000
33898 Cleanup a few compiler warnings.
33900 Simple warnings for unused variables and C99-style declarations.
33902 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33904 commit 917f9bb2435ba36bb99ef4d4f7b380d7f265e862
33905 Author: Gaetan Nadon <memsize@videotron.ca>
33906 Date: Mon Nov 23 09:25:05 2009 -0500
33908 Makefile.am: add ChangeLog and INSTALL on MAINTAINERCLEANFILES
33910 Now that the INSTALL file is generated.
33911 Allows running make maintainer-clean.
33913 commit eda2bb26783f7e9ef1ae659265930308378206e1
33914 Author: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
33915 Date: Fri Nov 20 01:52:05 2009 +0100
33917 Makefile.am: Add missing .g4i to be included in the tar file.
33919 Apparently Debian packages were having to manually add these files
33920 back in. Distribute them in the first place like we meant to.
33922 commit c5e86453c3ae3709933779a9dd609bbaebe21e8e
33923 Author: Gaetan Nadon <memsize@videotron.ca>
33924 Date: Wed Oct 28 14:41:41 2009 -0400
33926 INSTALL, NEWS, README or AUTHORS files are missing/incorrect #24206
33928 Automake 'foreign' option is specified in configure.ac.
33929 Remove from Makefile.am
33931 commit 12c4a22dd0b4de92a9c3d9e0baab029292f118f0
33932 Author: Gaetan Nadon <memsize@videotron.ca>
33933 Date: Wed Oct 28 14:09:09 2009 -0400
33935 INSTALL, NEWS, README or AUTHORS files are missing/incorrect #24206
33937 Add missing INSTALL file. Use standard GNU file on building tarball
33938 README may have been updated
33939 Remove AUTHORS file as it is empty and no content available yet.
33940 Remove NEWS file as it is empty and no content available yet.
33942 commit ad4030932490f848c8ac21ba5a7a8d734994ed15
33943 Author: Gaetan Nadon <memsize@videotron.ca>
33944 Date: Mon Oct 26 12:54:21 2009 -0400
33946 Several driver modules do not have a ChangeLog target in Makefile.am #23814
33948 The git generated ChangeLog replaces the hand written one.
33949 Update configure.ac to xorg-macros level 1.3.
33950 Use XORG_DEFAULT_OPTIONS which replaces four XORG_* macros
33951 Update Makefile.am to add ChangeLog target if missing
33952 Remove ChangeLog from EXTRA_DIST or *CLEAN variables
33953 This is a pre-req for the INSTALL_CMD
33955 commit bb994ce018ae6c3e3c5548ae7439cfc5e9aa4554
33956 Author: Gaetan Nadon <memsize@videotron.ca>
33957 Date: Thu Oct 22 13:02:04 2009 -0400
33959 .gitignore: use common defaults with custom section # 24239
33961 Using common defaults will reduce errors and maintenance.
33962 Only the very small or inexistent custom section need periodic maintenance
33963 when the structure of the component changes. Do not edit defaults.
33965 commit 67bbda0bddae3e0d39bc8deb7378f467aedfd0e3
33966 Author: Gaetan Nadon <memsize@videotron.ca>
33967 Date: Thu Oct 22 12:34:17 2009 -0400
33969 .gitignore: use common defaults with custom section # 24239
33971 Using common defaults will reduce errors and maintenance.
33972 Only the very small or inexistent custom section need periodic maintenance
33973 when the structure of the component changes. Do not edit defaults.
33975 commit c180baf43b8a0e407448018f3a7e42491cf974ae
33976 Author: Chris Wilson <chris@chris-wilson.co.uk>
33977 Date: Fri Nov 13 19:35:23 2009 +0000
33979 i915: Derive the correct target color from the pixmap by checking its format
33981 Particularly noting to route alpha to the green channel when blending
33982 with a8 destinations.
33986 rendercheck/repeat/triangles regressed
33987 http://bugs.freedesktop.org/show_bug.cgi?id=25047
33989 introduced with commit 14109a.
33991 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
33993 commit e9064eacb0ad8867e320597453facbb3c376522c
33994 Author: Chris Wilson <chris@chris-wilson.co.uk>
33995 Date: Fri Nov 13 18:35:44 2009 +0000
33997 uxa: Do not remove repeat from solids for 1x1 composites.
33999 Or else we hit the buggy 1x1 source path and trigger:
34001 rendercheck/mcoords regressed
34002 http://bugs.freedesktop.org/show_bug.cgi?id=25046
34004 caused by the recent commit e581ceb.
34006 commit 14109abf285866ad4cd99d0cd16b0954a0a73a62
34007 Author: Chris Wilson <chris@chris-wilson.co.uk>
34008 Date: Tue Nov 10 11:17:23 2009 +0000
34010 i915: Fix texture sampling coordinates.
34012 RENDER specifies that texels should sampled from the pixel centre. This
34013 corrects a number of failures in the cairo test suite and a few
34014 off-by-one bug reports.
34016 Grey border around images
34017 https://bugs.freedesktop.org/show_bug.cgi?id=21523
34019 Note that the earlier attempt to fix this was subverted by the buggy use
34020 of 1x1R textures for solid sources -- which caused the majority of text
34023 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34025 commit e581ceb7381e29ecc1a172597d258824f6a1d2d3
34026 Author: Chris Wilson <chris@chris-wilson.co.uk>
34027 Date: Tue Nov 10 11:14:23 2009 +0000
34029 i915: Use the color channels to pass along solid sources and masks.
34031 Instead of allocating and utilising the texture samplers for 1x1R
34032 solid sources and masks we can simply use the default diffuse and
34033 specular colour channels and adjust the fragment shader appropriately.
34034 The big advantage is the reduction in size of batches which should give
34035 a good boost to glyph performance, irrespective of the additional boost
34036 from using simpler shaders.
34038 However, the motivating factor behind the switch is that our use of 1x1
34039 textures turns out to be buggy...
34041 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34043 commit 33cabbfca6acb5149e26f87a538a7cb79f00cad2
34044 Author: Chris Wilson <chris@chris-wilson.co.uk>
34045 Date: Tue Nov 10 11:09:52 2009 +0000
34047 i915: Check for overflow before overflowing.
34049 As the immediate victim of the overflow would be to overwrite the maximum
34050 permissible value, the test was optimistic.
34052 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34054 commit 67af5a99253b1295f8dc09b28863eb7dc8b59e1d
34055 Author: Chris Wilson <chris@chris-wilson.co.uk>
34056 Date: Tue Nov 10 11:05:20 2009 +0000
34058 Check that batch buffers are atomic.
34060 Since batch buffers are rarely emitted by themselves but as part of a
34061 sequence of state and vertices, the whole sequence is emitted atomically.
34063 Here we just enforce that batches are marked as being part of an atomic
34064 sequence as appropriate.
34066 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34068 commit 998d6b3d8c549086fbc8a9f0e309694b23398d8d
34069 Author: Chris Wilson <chris@chris-wilson.co.uk>
34070 Date: Tue Nov 10 11:00:43 2009 +0000
34072 uxa: Force alpha bits to fill remaining bits
34074 In the case of x8r8g8b8 and similar where the alpha channel is ignored,
34075 but should be interpreted as being 1, then it is convenient if those bits
34076 are set appropriately in the colour. In order to do so for these formats,
34077 where PIXMAN_FORMAT_A() returns 0 we need to compute the alpha channel
34078 width as the remaining bits instead.
34080 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34082 commit dbb68168dc909ab2ec1d935322c3fd8581e666f1
34083 Author: Eric Anholt <eric@anholt.net>
34084 Date: Thu Nov 5 15:40:20 2009 -0800
34086 Revert "configure: make --disable-dri work even if the server supports DRI"
34088 This reverts commit a851139c2141f6da370186148f2836e18b2acf83.
34089 It broke the build, and I don't see why we should be supporting this
34097 commit 4c8e783d84d2c14c8a1638b6a12307c0164d3e31
34098 Author: Eric Anholt <eric@anholt.net>
34099 Date: Wed Oct 21 13:30:38 2009 -0700
34101 Fix "Remove flow-control macros for fallbacks in the 2D driver."
34103 I guess this is the sort of failure due to rebase-happiness that makes
34104 Linus yell at us for rebasing.
34106 commit d0e08fe611681fcc840d1b9ee8d49acdf50f7e58
34107 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34108 Date: Mon Oct 26 13:15:24 2009 +0000
34110 Kill some more #defines only needed for User-Modesetting
34112 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34113 Signed-off-by: Eric Anholt <eric@anholt.net>
34115 commit 8ff2a6496413e6b12fad9352f4bd9b6736bda56c
34116 Author: Eric Anholt <eric@anholt.net>
34117 Date: Wed Oct 21 13:30:38 2009 -0700
34119 Remove flow-control macros for fallbacks in the 2D driver.
34121 It's poor style, and has confused new developers.
34123 commit c87585229b36790f883b9b8954ed061e00624df6
34124 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
34125 Date: Thu Nov 5 13:33:55 2009 -0800
34127 Increase stride limit for IGDNG
34129 It can go up to 32k. Upping this lets me use my 2560x1600 and 1920x1200
34130 monitors in an extended desktop configuration.
34132 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
34134 commit cbcfb711747e3637f20145aed00ac3d6256d0d1d
34135 Author: Rémi Cardona <remi@gentoo.org>
34136 Date: Wed Oct 28 13:12:46 2009 +0100
34138 configure: drop some more cruft
34140 DEBUGFLAGS wasn't used anywhere and the configure switch is useless.
34141 "CFLAGS=-g ./configure" does the same thing and actually works.
34143 PKG_CHECK_MODULES already calls AC_SUBST, no need to do it twice.
34145 Signed-off-by: Rémi Cardona <remi@gentoo.org>
34146 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
34148 commit 3c0a43b24cc7540cc139beca0ec2e4c202ff5593
34149 Author: Rémi Cardona <remi@gentoo.org>
34150 Date: Wed Oct 28 13:01:32 2009 +0100
34152 configure: use CWARNFLAGS from xorg-macros.m4
34154 Signed-off-by: Rémi Cardona <remi@gentoo.org>
34155 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
34157 commit a851139c2141f6da370186148f2836e18b2acf83
34158 Author: Rémi Cardona <remi@gentoo.org>
34159 Date: Wed Oct 28 12:37:38 2009 +0100
34161 configure: make --disable-dri work even if the server supports DRI
34163 XF86DRI is defined by the SDK so not defining it here just breaks the
34164 build. Define HAVE_DRI instead to avoid collisions.
34166 Note: DRI2 is still enabled/disabled entirely by SDK defines.
34168 Signed-off-by: Rémi Cardona <remi@gentoo.org>
34169 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
34171 commit 07e0b2cff6107d6c86096c7da4e31b0c976794d5
34172 Author: Rémi Cardona <remi@gentoo.org>
34173 Date: Wed Oct 28 11:45:55 2009 +0100
34175 configure: group system header checks at the top
34177 Signed-off-by: Rémi Cardona <remi@gentoo.org>
34178 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
34180 commit 0401a4c3143c5b94875cb4c66c94a0d247c32da5
34181 Author: Rémi Cardona <remi@gentoo.org>
34182 Date: Wed Oct 28 11:42:41 2009 +0100
34184 configure: check for libdrm and libpciaccess earlier
34186 Signed-off-by: Rémi Cardona <remi@gentoo.org>
34187 Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
34189 commit 10946118dd3a63f1375a1bfde0b2f0542a93c1c2
34190 Author: Albert Damen <albrt@gmx.net>
34191 Date: Sun Oct 25 18:02:07 2009 +0100
34193 Fix crash in uxa_acquire_pattern when pDst is NULL
34195 This avoids a crash when an XRenderComposite call is made with a
34196 -1 value for width/height, (which apparently compiz's gtk-window-
34197 decorator likes to do). Fixes bug:
34199 X crashes in uxa_acquire_pattern when logging in (gdm)
34200 http://bugs.freedesktop.org/show_bug.cgi?id=24724
34202 Signed-off-by: Albert Damen <albrt@gmx.net>
34203 Reviewed-by: Carl Worth <cworth@cworth.org>
34205 commit 751e0a3e4576bbf4bffa56bbd6d4de28f10db98e
34206 Author: Thomas Arnhold <thomas@arnhold.org>
34207 Date: Mon Oct 19 11:35:30 2009 -0700
34209 Fix 64-bit compiler warnings in intel_xvmc_dump_render().
34213 commit 3e8f2eae3a586aa29be4858698e666e0ec778cea
34214 Author: Eric Anholt <eric@anholt.net>
34215 Date: Thu Oct 15 13:48:56 2009 -0700
34217 XVMC: Use XCB DRI2 instead of cargo-culting our own copy of Xlib stuff. (v2)
34219 v2: Incorporate comments from Jamey on device name handling and extension
34222 commit 38ab403d7a1c461c8ac65a056bee2dd5c7f2f58e
34223 Author: Eric Anholt <eric@anholt.net>
34224 Date: Thu Oct 15 11:39:32 2009 -0700
34226 Enable XVMC by default on gen4.
34228 commit f171069608bf174d920921fa102b8619794ed272
34229 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
34230 Date: Mon Sep 28 18:35:57 2009 +0800
34234 Remove bo pin for surface buffer access, and remove access
34235 attempt for possible unmapped framebuffer. Using xv buffer
34236 pointer to pass current xvmc surface bo handler, which is
34237 assigned to src image bo and handle that the same way as in Xv.
34239 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
34240 [anholt: Fixed up for conflict against the XV rework. Not tested, because
34241 both mplayer and xine segfault with XVMC currently.]
34242 Signed-off-by: Eric Anholt <eric@anholt.net>
34244 commit aaedeffe00d9414bb03723dbc30b4938a07ce5fa
34245 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34246 Date: Wed Oct 14 18:09:08 2009 +0200
34248 Xv overlay: fix planar YUV copy for right rotated crtcs
34250 While copying and rotating the buffer, array access was out of bounds when
34251 rotated to the right (RR_Rotate_270). My buffer handling changes probably
34252 made this bug much more likely to actually result in a SIGSEGV.
34254 I've checked the logs and the bug exists since rotation has been supported,
34255 i.e. this looks like a candidate for cherry-picking for all supported
34258 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34259 Signed-off-by: Eric Anholt <eric@anholt.net>
34261 commit 703e3326bb41528cc57c0d25003707df209fc714
34262 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34263 Date: Wed Oct 14 15:56:55 2009 +0200
34265 Xv overlay: further cleanups
34267 Kill some unnecessary stuff. Small code changes, but no functional ones.
34269 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34270 Signed-off-by: Eric Anholt <eric@anholt.net>
34272 commit 1c2aedfce9e6a7561347e873d125b6889dee7941
34273 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34274 Date: Wed Oct 14 15:56:54 2009 +0200
34276 Xv: fixup the disabled drmmode overlay code
34278 This code didn't survive the global renaming of vars to saner names.
34281 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34282 Signed-off-by: Eric Anholt <eric@anholt.net>
34284 commit 909990f40a437cbd7026a10e32af1ea120f4c2a7
34285 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34286 Date: Wed Oct 14 15:56:53 2009 +0200
34288 Xv overlay: remove some more dead stuff from ums overlay support
34290 Mostly unused definitions and variables, but also some strange ums
34291 debug code. Also kill some now obsolete comments.
34293 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34294 Signed-off-by: Eric Anholt <eric@anholt.net>
34296 commit 86bc23ab5da34137c82250395c68aa92ecd88a24
34297 Author: Chris Wilson <chris@chris-wilson.co.uk>
34298 Date: Wed Oct 14 12:04:40 2009 +0100
34300 debug: Enable cache flushing after every operation
34302 If DEBUG_FLUSH_CACHES is enabled then emit a MI_FLUSH after every
34303 rendering operation. This is intended to 'fix' cases where we are
34304 missing a required flush in the middle of a sequence of operations, such
34305 as switching between 2D to 3D and render to sampler.
34307 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34309 commit 3c0815abf28744e215bea286e71d935cd486955a
34310 Author: Chris Wilson <chris@chris-wilson.co.uk>
34311 Date: Wed Oct 14 11:29:21 2009 +0100
34313 conf: Add debugging flush options
34315 Make the following options available via xorg.conf:
34317 Option "DebugFlushBatches" "1" # Flush the batch buffer after every
34318 # single operation;
34320 Option "DebugFlushCaches" "1" # Include a MI_FLUSH at the end of every
34321 # batch buffer to force data to be
34322 # flushed out of cache and into memory
34323 # before the completion of the batch.
34325 Option "DebugWait" "1" # Wait for the completion of every batch buffer
34326 # before continuing, i.e. perform synchronous
34330 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34332 commit 2f134b84445d2d0bfe3d81276bc7f6648df062ee
34333 Author: Zhao Yakui <yakui.zhao@intel.com>
34334 Date: Wed Sep 30 07:31:57 2009 +0000
34336 Return failure when the enum-type output property values are not found
34338 Failure to do so causes xrandr to report incorrect property values.
34340 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
34341 Signed-off-by: Eric Anholt <eric@anholt.net>
34343 commit fcc2ee48b866b81c79315ff10189b56fc201539d
34344 Author: Albert Damen <albrt@gmx.net>
34345 Date: Wed Oct 7 21:55:42 2009 +0200
34347 Drop frontbuffer from crtc in I830CloseScreen
34349 By dropping the frontbuffer from the crtc, the new frontbuffer
34350 can be properly added to the crtc when the xserver is reset.
34352 Signed-off-by: Albert Damen <albrt@gmx.net>
34354 commit 8a77877f9c2c6a8a1308bc1a3be9e7ad88bc7f49
34355 Author: Dave Airlie <airlied@redhat.com>
34356 Date: Fri Oct 9 14:16:06 2009 +1000
34358 drmmode: with 1.7 server, set mode major doesn't get gamma setup.
34360 Noticed this on Fedora, where 1.7 server does gamma via the randr
34361 codepaths however kms doesn't have this call which happens in the
34362 non set_mode_major path.
34364 probably should be backported to released drivers.
34366 Signed-off-by: Dave Airlie <airlied@redhat.com>
34368 commit d8c7678ddadce89ca7fc0edd1b4d9ed3d959e687
34369 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
34370 Date: Mon Oct 12 14:02:12 2009 -0700
34372 DRI2 compat build fix: it's drawable->pScreen not drawable->screen
34374 Fallout from the conversion; DRI2 compat path was broken.
34376 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
34378 commit b37ac9d317ae537d993922976f87072040b04d04
34379 Author: Chris Wilson <chris@chris-wilson.co.uk>
34380 Date: Mon Oct 12 14:31:06 2009 +0100
34382 uxa: Refactor create Picture for pixman format
34384 Pull the common methods for creating a Picture given a pixman format
34385 into its own method, and tidy the surrounding code. The benefit is that
34386 we can now composite directly to the Picture and so save an intermediate
34387 copy when creating patterns for gradients.
34389 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34391 commit 7e8f32d0a7279dce1976f87612833d9092554cfe
34392 Author: Chris Wilson <chris@chris-wilson.co.uk>
34393 Date: Mon Oct 12 14:33:08 2009 +0100
34395 uxa: Free the ScratchPixmapHeader after its associated Picture
34397 Fixes: http://bugs.freedesktop.org/show_bug.cgi?id=24459
34398 Intel Driver > 2.8: Cairo rendering bug, triggered in QtCurve GTK engine
34400 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34402 commit 1556c62e0336ea2fef866722ee44d2d188e318f3
34403 Author: Eric Anholt <eric@anholt.net>
34404 Date: Thu Oct 8 17:34:13 2009 -0700
34406 Replace dolt and shave with relying on current autotools.
34408 commit 8b2d2ff0d026eea445a071102a62d782f3bbec78
34409 Author: Eric Anholt <eric@anholt.net>
34410 Date: Wed Oct 7 16:12:25 2009 -0700
34412 Clean up more i830_memory.c madness.
34414 It was cooking up insane alignment values for buffers that new libdrm was
34415 justifiably complaining about, but it turns out we don't need the alignment
34416 values anywhere because the only case they're needed, they're computed
34417 entirely by the kernel. Also, the XVMC code was passing a completely unused
34420 commit d525a0e993a59e118ab1e8519b3d73465f8f0169
34421 Author: Eric Anholt <eric@anholt.net>
34422 Date: Tue Oct 6 19:05:24 2009 -0700
34424 Rename pScreen to screen.
34426 commit b6262dcd2754f7e87f27979dcad145133ee3b945
34427 Author: Eric Anholt <eric@anholt.net>
34428 Date: Tue Oct 6 19:00:41 2009 -0700
34432 commit fdcfeb822d15cbfe6fccceab002b75458dca3264
34433 Author: Eric Anholt <eric@anholt.net>
34434 Date: Tue Oct 6 18:57:22 2009 -0700
34436 Rename pDraw to plain old drawable.
34438 commit 7bbf4ac7137ed0e4ec0a75806a396c12f102e779
34439 Author: Eric Anholt <eric@anholt.net>
34440 Date: Tue Oct 6 18:50:54 2009 -0700
34442 Rename the xv pPriv to adaptor_priv to reflect whose private it is.
34444 commit f309d475241260cf60567100511d5f7c6c487a29
34445 Author: Eric Anholt <eric@anholt.net>
34446 Date: Tue Oct 6 18:38:53 2009 -0700
34448 Call pPixmaps plain old pixmaps.
34450 commit da0f6616ad63f1581cf91a98104e5287aa44e7ce
34451 Author: Eric Anholt <eric@anholt.net>
34452 Date: Tue Oct 6 18:37:05 2009 -0700
34454 de-pCamelHungarian the Render pictures and pixmaps.
34456 commit 050a141b7bc94b459061615124b7686a9c331e01
34457 Author: Eric Anholt <eric@anholt.net>
34458 Date: Tue Oct 6 18:30:57 2009 -0700
34460 Share several render fields between render implementations.
34462 Also, start settling on the cairo naming for things: source, mask, and dest.
34464 commit af27a3a0a5645c6f41f583611bd0f2559dc7cb2f
34465 Author: Eric Anholt <eric@anholt.net>
34466 Date: Tue Oct 6 18:24:50 2009 -0700
34468 Rename the xf86 screen private from pScrn to scrn.
34470 commit cc5d3ba3c331c3b1becf2d19277b24144bf34cfa
34471 Author: Eric Anholt <eric@anholt.net>
34472 Date: Tue Oct 6 18:02:38 2009 -0700
34474 Rename the screen private from I830Ptr pI830 to intel_screen_private *intel.
34476 This is the beginning of the campaign to remove some of the absurd use of
34477 Hungarian in the driver. Not that I don't like Hungarian, but I don't need
34478 to know that pI830 is a pPointer.
34480 commit 03e8e64f8669263e3cecb79ea57d5a26c0eaee3f
34481 Author: Chris Wilson <chris@chris-wilson.co.uk>
34482 Date: Wed Oct 7 22:15:37 2009 +0100
34484 Complete the removal of UMS build support
34486 Eric missed removing the now dead references to the deleted subdirectories
34487 in configure.ac with commit b9b159c
34489 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34491 commit 5e44a0fa4209523bee56f513a2029869f96a710a
34492 Author: Eric Anholt <eric@anholt.net>
34493 Date: Tue Oct 6 18:08:51 2009 -0700
34495 Remove more dead UMS code.
34497 commit 6a716a25abc84f1c7f766fd3bfa1c2aac91e6ae7
34498 Author: Eric Anholt <eric@anholt.net>
34499 Date: Tue Oct 6 17:58:32 2009 -0700
34501 Remove the reg_dumper code, now that it's been moved to intel_gpu_tools.
34503 commit e9aff787a55d684e3b9c6d5ea8ac2926be952280
34504 Author: Eric Anholt <eric@anholt.net>
34505 Date: Tue Oct 6 17:53:59 2009 -0700
34507 Remove the stepping debug app, replaced by version in intel_gpu_tools.
34509 commit 71b9cdaa86bb9d3deb44f32f47771c33898253c2
34510 Author: Eric Anholt <eric@anholt.net>
34511 Date: Tue Oct 6 17:53:21 2009 -0700
34513 Trim down the hotplug debug app.
34515 commit e95eb483e210def0814bc61fc36d672145da7dff
34516 Author: Eric Anholt <eric@anholt.net>
34517 Date: Tue Oct 6 17:51:09 2009 -0700
34519 Remove intel_idle, replaced by intel_gpu_top in intel_gpu_tools.
34521 commit 8ae0e44e42db645abe6d385f561260d2ae4a1960
34522 Author: Eric Anholt <eric@anholt.net>
34523 Date: Tue Oct 6 16:30:08 2009 -0700
34525 Move to kernel coding style.
34527 We've talked about doing this since the start of the project, putting it off
34528 until "some convenient time". Just after removing a third of the driver seems
34529 like a convenient time, when backporting's probably not happening much anyway.
34531 commit b9b159c49854d8d9d2207946bb583537bb0d48d6
34532 Author: Eric Anholt <eric@anholt.net>
34533 Date: Wed Sep 30 16:29:53 2009 -0700
34535 Remove UMS support.
34537 At this point, the only remaining feature regressions should be the lack of
34538 overlay support (about to land), and the need to update the XVMC code to work
34539 in the presence of KMS.
34541 Acked-by: Keith Packard <keithp@keithp.com> (in principle)
34542 Acked-by: Carl Worth <cworth@cworth.org> (in principle)
34544 commit d26f4c493721dbb2e2cbf4efcf8d37228d1b1f3b
34545 Author: Eric Anholt <eric@anholt.net>
34546 Date: Mon Oct 5 13:31:31 2009 -0700
34548 Remove UMS overlay support.
34550 The replacement code is now landed, and the rest of UMS is about to disappear.
34552 commit 33c488e83676d26e69145ea615f75ca52768f400
34553 Author: Eric Anholt <eric@anholt.net>
34554 Date: Mon Oct 5 10:38:05 2009 -0700
34556 Remove error state dumping code.
34558 This is replaced by intel_gpu_dump, and would no longer be used once UMS is
34561 commit 2370af32fe31bed8e5715639023635bdb3b83b1b
34562 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34563 Date: Tue Aug 11 16:06:47 2009 +0200
34565 Implement drmmode overlay
34567 This does not restore the overlay on EnterVT/disable it on LeaveVT.
34568 Does not look like this is necessary.
34570 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34571 [anholt: Hacked in avoiding the actual kernel calls with
34572 Signed-off-by: Eric Anholt <eric@anholt.net>
34574 commit 99230864050ea2f26c2412c4c87c50947fc3cbe1
34575 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34576 Date: Tue Aug 11 16:06:46 2009 +0200
34578 Xv: introduce an overlay hal
34580 This is the last preparatory step for overlay support with drmmode.
34581 Safe two (specially marked) function calls in the setup code, all
34582 hw accessing code goes now through these three new functions with
34583 the ums_overlay prefix.
34585 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34586 Signed-off-by: Eric Anholt <eric@anholt.net>
34588 commit bb04e184fdc55325a362ebf291e16a7c112a5f1d
34589 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34590 Date: Tue Aug 11 16:06:45 2009 +0200
34592 Xv: rework overlay buffer management
34594 The basic idea is to only pin the buffer into the gtt when
34595 the overlay hw is actually using it. This results in a few changes:
34597 - Unify data copied/buffer handling with textured video. Now offsets
34598 are always buffer relative and we just use drm_bo_map to access a
34600 - Implement double buffering using two bo's. This is necessary because
34601 we can't pin the same buffer to the gtt and map it as normal memory.
34602 - Kill XV_DOUBLE_BUFFER. With the above changes, overlay video is always
34605 There is still the XvMC passthrough case, which makes the code slightly
34606 ugly. Unfortunately we can't get at the bo behind this buffer.
34608 Changes since the last review-round:
34609 - Don't overallocate by a factor of 2.
34610 - Prevent possible use-after-free issue.
34612 Signed-off-by: Eric Anholt <eric@anholt.net>
34614 commit 60462eb5b5c1931beecc1ed2d32f91090f64174c
34615 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34616 Date: Tue Aug 11 16:06:44 2009 +0200
34618 Xv: create xvmc_passthrough helper
34620 This way all thes strange special cases make much more sense.
34622 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34623 Signed-off-by: Eric Anholt <eric@anholt.net>
34625 commit 71276dff9415322f75792a46c1a1cc0b900588b9
34626 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34627 Date: Tue Aug 11 16:06:43 2009 +0200
34629 Xv: scrap overlay offscreen pixmap support
34631 The code looks like it's been bitrotting since being copied over
34632 from the i810 driver. Furthermore painting rgb pixmaps with the overlay
34633 engine is in these days of modern compositing X an absolute no-go. And
34634 textured video doesn't support it neither, so its likely never ever
34635 used by applications.
34637 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34638 Signed-off-by: Eric Anholt <eric@anholt.net>
34640 commit 99afdf985fa9f763fda4bc49ccd8111960a4ae0f
34641 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34642 Date: Tue Aug 11 16:06:42 2009 +0200
34644 Xv: small cleanups in I830PutImage
34646 - scrap unused variable overlay
34647 - scrap an superflous if and attach the code to the preceeding else
34650 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34651 Signed-off-by: Eric Anholt <eric@anholt.net>
34653 commit d598456f1fe013c250530730c57ad720d38cde3a
34654 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34655 Date: Tue Aug 11 16:06:41 2009 +0200
34657 Xv I830PutImage splitup: extract i830_wait_for scanline
34659 Also scrap the unecessary variable sync in I830PutImage and the
34660 accompanying obfuscated logic.
34662 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34663 Signed-off-by: Eric Anholt <eric@anholt.net>
34665 commit 2ba03fa78197d1cb10d6401ad02cfe9edd2623b3
34666 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34667 Date: Tue Aug 11 16:06:40 2009 +0200
34669 Xv I830PutImage splitup: extract i830_copy_video_data
34671 Just moves the code and passes back allocation failures.
34673 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34674 Signed-off-by: Eric Anholt <eric@anholt.net>
34676 commit 81fc74cc4ce81b9be8b9ba3bcacf1284c47ce80c
34677 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34678 Date: Tue Aug 11 16:06:39 2009 +0200
34680 Xv I830PutImage splitup: extract i830_setup_video_buffer
34682 Just move the code and pass back allocation failures.
34684 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34685 Signed-off-by: Eric Anholt <eric@anholt.net>
34687 commit 0bf4cc5130482ee0f3924f34e37fa5fc988119b3
34688 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34689 Date: Tue Aug 11 16:06:38 2009 +0200
34691 Xv I830PutImage splitup: extract i830_dst_pitch_and_size
34693 Just moves the code.
34695 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34696 Signed-off-by: Eric Anholt <eric@anholt.net>
34698 commit c20ef591931b5efc5745d00c06f296c89cf32745
34699 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34700 Date: Tue Aug 11 16:06:37 2009 +0200
34702 Xv: kill destId in I830PutImage
34704 It's only used to remember that XvMC has ỲV12 as output. is_planar_fourcc
34705 already takes care of that in all necessary cases.
34707 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34708 Signed-off-by: Eric Anholt <eric@anholt.net>
34710 commit 5ce944994d57176d0ddf6f6d7c7779fdfcf39cb8
34711 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34712 Date: Tue Aug 11 16:06:36 2009 +0200
34714 Xv: kill hw double buffering logic
34716 The idea for the hw double buffering support is to program two fixed
34717 buffers and then only switch buffers in the OCMD register. But the driver
34718 as-is always programs the new buffer address (in both register sets
34719 when double buffered). Therefore we gain nothing by using this hw
34720 capability. Scrap the software support for it.
34722 When double buffered, we now allocate just a buffer of size 2*size and
34723 switch between the two parts purely in software.
34725 To make reviewing this easier, I'll shortly explain the differences of how
34726 double-buffering (i.e. tear-free video) is achieved before and after this
34729 - When double buffer, allocate a buffer twice the size (unchanged).
34730 - Depending upon the currently shown buffer-half, copy the new frame into
34731 the other buffer-half. In the old code this is done by using the right
34732 set of buffer offsets, either *Buf0Offset or *Buf1Offset. The new code
34733 simply programs the offset for the right buffer-half into the single set
34734 of offsets. The end-result is unchanged.
34736 Now the big difference in hw-programming:
34738 Old: Programm new buffer offset into both sets of _hw_ buffer offset
34739 registers. Depending upon the current _sw_ buffer, select the _hw_ buffer
34740 and program this into the OCMD register. This just complicates matters
34743 New: Just always use the hw buffer 0.
34745 And then it's again the same story in both old and new code:
34747 - Execute an overlay flip (MI_OVERLAY_FLIP) to read in the contents of the
34748 hw registers into the shadow hw registers (which are actually being used
34749 by the overlay, not the ones we write stuff into). This is synchronized
34750 with the respective crtc vblank by the hw.
34752 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34753 Signed-off-by: Eric Anholt <eric@anholt.net>
34755 commit 232418d72ee967a81e39481b9f2ee379cb685458
34756 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34757 Date: Tue Aug 11 16:06:35 2009 +0200
34759 Xv: use is_planar_fourcc helper some more
34761 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34762 Signed-off-by: Eric Anholt <eric@anholt.net>
34764 commit 219b14310efe49aff5d3d9023d2ba440c9f702c1
34765 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34766 Date: Tue Aug 11 16:06:34 2009 +0200
34768 Xv: introduce planar memcpy helper
34770 Reduced 3 copies of the same code to one.
34772 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34773 Signed-off-by: Eric Anholt <eric@anholt.net>
34775 commit b90171f16652d1145fb80a63919f76a804a94cef
34776 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34777 Date: Tue Aug 11 16:06:33 2009 +0200
34779 Xv: rename i830_display_video to i830_display_overlay
34781 This function only programs the overlay and is never called for textured
34782 video. Make this obvious.
34784 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34785 Signed-off-by: Eric Anholt <eric@anholt.net>
34787 commit 857e40a6e00db72401717da3ed79edbf65db31d4
34788 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34789 Date: Tue Aug 11 16:06:32 2009 +0200
34791 Xv i830_display_video splitup: extract i830_overlay_cmd
34793 This slightly moves around (and simplifies) the OSTRIDE reg programming,
34796 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34797 Signed-off-by: Eric Anholt <eric@anholt.net>
34799 commit 50d70e644ea0574752d0fe160ce8d873653c913b
34800 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
34801 Date: Tue Aug 11 16:06:31 2009 +0200
34803 Xv i830_display_video splitup: extract i830_calc_src_regs
34805 Also introduce an is_planar_fourcc helper. I'll use that one later.
34807 In i830_display_video this changeset moves the XVMC case (previously
34808 obscured as the default case) around. I've figured this default case
34809 does not make sense, here's why:
34811 XvMC is everywhere else handled as a planar format (e.g. in the register
34812 programming a few lines down). Furthermore the id variable gets mapped
34813 to FOURCC_YV12 if IS_I915(pI830) is true in I830PutImage. There's a
34814 second caller in the offscreen overlay support code. But I think that
34815 code is bitrotten and not reliable as an information source.
34817 So we have a different behaviour only for id=FOURCC_XVMC and i965 class
34818 hw (i830 class doesn't have xvmc). I've crawled through various
34819 sources/intel documentations. Finally in the textured video implemention
34820 for i965 class hw (src/i965_video.c) I've found a switch statement that
34821 puts XVMC into the same case as I420 and YV12. So also in i965 class hw
34822 xvmc uses a planar format.
34824 In conclusion I claim that this code was bogus and XvMC on i965 class hw
34825 over Xv overlay was most likely broken.
34827 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
34828 Signed-off-by: Eric Anholt <eric@anholt.net>
34830 commit 2841a4cd8c6c64fdf4b2203ab73b57b38a50f651
34831 Author: Carl Worth <cworth@cworth.org>
34832 Date: Mon Sep 28 20:03:11 2009 -0700
34834 Increment version to 2.9.0
34836 And add a reminder to RELEASING that the incremented version number
34837 needs to be committed, (since I forgot to do this with 2.8.99.902).
34839 commit a790aff4fef7d0c86b7b9c0da4afdc9c0aba6636
34840 Author: Carl Worth <cworth@cworth.org>
34841 Date: Mon Sep 28 20:02:28 2009 -0700
34843 NEWS: Add notes for the 2.9.0 release.
34845 Just renamed notes as from 2.8.99.902.
34847 commit a92bbcc94904684e7709b3ddaad82bc04607af26
34848 Author: Carl Worth <cworth@cworth.org>
34849 Date: Tue Sep 22 17:08:23 2009 -0700
34851 Increment version to 2.8.99.902
34853 commit ae00a0923717caec6d0dad94527da26397bd12ce
34854 Author: Carl Worth <cworth@cworth.org>
34855 Date: Tue Sep 22 15:14:38 2009 -0700
34857 NEWS: Add notes for 2.8.99.902
34859 commit 7e7db7ac530b5282b0841585959597b54fcc633b
34860 Author: Matthias Hopf <mhopf@suse.de>
34861 Date: Mon Aug 17 19:24:03 2009 +0200
34863 Add new backlight driver "samsung".
34865 commit 7ae1d0dde6cef3437b67dbc21384cb179616a6c0
34866 Author: Zhao Yakui <yakui.zhao@intel.com>
34867 Date: Mon Aug 31 13:51:01 2009 +0800
34869 Skip setting tv format property if output crtc is NULL
34871 When TV is not connected and X start, after plugging TV cable again,
34872 system will crash because output crtc is NULL. This patch will return,
34873 do not handle crtc immediately, meanwhile set value will be effective
34874 until user really enable output by xrandr command.
34876 Signed-off-by: Ma Ling <ling.ma@intel.com>
34877 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
34879 commit 02fe9be695f7e209944bd0f7b67950f93619feee
34880 Author: Zhao Yakui <yakui.zhao@intel.com>
34881 Date: Tue Sep 22 10:58:38 2009 +0800
34883 Check whether the DVI-I/D is connected or disconnected based on EDID
34885 When the monitor is digital type for SDVO-DVI D, there should exist the EDID. If
34886 there is no EDID, it should be detected as disconnected.
34888 Signe-off-by: Zhao Yakui <yakui.zhao@intel.com>
34890 commit 762e406d138ac80854b6a23b1078b52f6581f0d8
34891 Author: Chris Wilson <chris@chris-wilson.co.uk>
34892 Date: Tue Sep 22 01:34:37 2009 +0100
34894 Revert "8xx: Fallback for any non-affine transformation."
34896 This reverts commit 505025053d66d415e1c23ac858b9238fa8541d37.
34898 In theory, the non-affine paths work -- at least for the stated test case,
34899 so re-enable them and avoid the slow work-around.
34901 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34903 commit 2cc1f3cb6034dddd65b3781b0cde7dff4ac1e803
34904 Author: Keith Packard <keithp@keithp.com>
34905 Date: Sat Sep 19 17:30:57 2009 -0700
34907 i8xx: Format projective texture coordinates correctly.
34909 Projective texture coordinates must be delivered as TEXCOORDFMT_3D
34910 using TEXCOORDTYPE_HOMOGENOUS. This meant selecting the correct type
34911 in i830_texture_setup, the correct format in i830_emit_composite_state
34912 and sending only 3 coordinates in i830_emit_composite_primitive.
34914 Signed-off-by: Keith Packard <keithp@keithp.com>
34915 [ickle: tweaked to fix up a couple of use-before-initialised]
34916 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
34918 commit 00e8de212b46a243f243b437b7eb866315ab89b3
34919 Author: Chris Wilson <chris@chris-wilson.co.uk>
34920 Date: Mon Sep 21 22:14:21 2009 +0100
34922 Check the correct Picture for error during creation.
34924 commit bd817e2d733dfdb1140874b06595ccd1ef39159b
34925 Author: Keith Packard <keithp@keithp.com>
34926 Date: Mon Sep 21 17:21:17 2009 -0700
34928 Split i915/i830 composite_emit_primitive into two functions.
34930 The i915 and i830 take similar but different data when emitting the
34931 primitives, instead of trying to share code here, just split this
34932 apart and avoid potentially breaking things later on.
34934 Signed-off-by: Keith Packard <keithp@keithp.com>
34936 commit 5e80297d088e8cdbf66d765f7d252dab66c8df86
34937 Author: Keith Packard <keithp@keithp.com>
34938 Date: Fri Sep 18 21:05:23 2009 -0700
34940 Remove DGA support from the driver.
34942 The xf86DiDGA code required that the scanout buffer always be
34943 mappable, stay be at a fixed address in the aperture and have a
34944 constant size. With frame buffer resizing, the latter two are no
34945 longer true, and with KMS, we'd really prefer to not allow the former.
34947 The only option available to the driver is to completely disable DGA
34948 as the modes code has internal calls to the xf86DiDGA code when
34949 fetching new modes from the hardware.
34951 A fix for the DiDGA code will be added to the X server which will
34952 automatically initialize DGA for mode switching and input, but not
34953 frame buffer access, and not require any driver cooperation.
34955 Thus, the correct solution is for the driver to not call xf86DiDGAInit
34956 at all. For old servers, this eliminates a potential catastrophic
34957 problem where random memory is written by the X server. New servers
34958 will get the DIX-based behaviour automatically.
34960 Signed-off-by: Keith Packard <keithp@keithp.com>
34962 commit 4758311842a16600287c8f9f77ce0af1a31b9264
34963 Author: Keith Packard <keithp@keithp.com>
34964 Date: Fri Sep 18 20:59:52 2009 -0700
34966 Remove vestigial internal rotation which broke KMS DGA/VidMode modesetting.
34968 Pre-2.0, the driver supported rotation internally, rather than relying
34969 on the X server rotation support. The last piece of this dealt with
34970 rotating the mouse coordinates and also tried to preserve rotation
34971 across DGA/VidModeExtension modesetting requests.
34973 That latter bit of code broke under KMS as the rotation value was
34974 never initialized, and when set to zero would create an invalid
34975 configuration. This would confuse xrandr which would bail before
34976 making any changes, leaving the user without a way to recover.
34978 Signed-off-by: Keith Packard <keithp@keithp.com>
34980 commit 505025053d66d415e1c23ac858b9238fa8541d37
34981 Author: Carl Worth <cworth@cworth.org>
34982 Date: Mon Sep 21 13:50:09 2009 -0700
34984 8xx: Fallback for any non-affine transformation.
34986 There are definitely bugs in the 8xx code dealing with non-affine
34987 transformations. Disable that code for now to get things working.
34989 Fixes bug #22947 ([855GM, xf86-video-intel-2.8.0] "Freeze" when RENDER extension is being used)
34991 commit b4d29452b929a3ef224d3625e4bc66b787c5edb7
34992 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
34993 Date: Mon Sep 21 15:28:09 2009 +0800
34995 More dumps for Arrandale LVDS
34997 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
34999 commit 57fc09cef28bad2e3e8455b93ef2927118f8a3a3
35000 Author: Chris Wilson <chris@chris-wilson.co.uk>
35001 Date: Sun Sep 20 01:02:39 2009 +0100
35003 Avoid fallbacks for a1 src/mask
35005 Carl Worth did the hard work in identifying that the regression in
35006 cairo between X.org 1.6 and 1.7 was caused by cairo sending an a1
35007 mask to the server in 1.7 whereas in 1.6 cairo used local fallbacks
35008 (as the source was using RepeatPad, which triggers cairo's
35009 'buggy_pad_reflect' fallback for X.org 1.6). This was causing the driver
35010 to do a fallback to handle the a1 mask instead, which due to the GPU
35011 pipeline stall is much more expensive than the equivalent fallback in
35015 cairo's performance downgrades 4X with server master than server-1.6.
35016 https://bugs.freedesktop.org/show_bug.cgi?id=23184
35018 The fix is a relatively simple extension of the current
35019 uxa_picture_from_pixman_image() to use CompositePicture() instead of
35020 CopyArea() when we need to convert to a new format.
35022 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35024 commit 33f98e4056706f4c30bb4327677ac49e82058231
35025 Author: Keith Packard <keithp@keithp.com>
35026 Date: Thu Sep 17 18:16:02 2009 -0700
35028 Don't destroy bufmgr at CloseScreen time
35030 Under KMS, the bufmgr is not initialized at InitOutput time and so it
35031 won't be re-initialized during server regen. Thus we must leave the
35032 bufmgr running during regen and cannot destroy it in CloseScreen.
35034 Under UMS, each place the bufmgr is initialized, it checks to see if
35035 it has already happened. Hence, we can safely leave the bufmgr running
35036 across server regen for UMS too.
35038 Signed-off-by: Keith Packard <keithp@keithp.com>
35040 commit ee9ad853574e3ee1a210db6207b84ecd466e08ee
35041 Author: Keith Packard <keithp@keithp.com>
35042 Date: Thu Sep 17 17:57:40 2009 -0700
35044 Make sure DRM output properties are freed after we're done with them.
35046 drmmode_output_get_modes was fetching output properties but only
35047 freeing some of them.
35049 Signed-off-by: Keith Packard <keithp@keithp.com>
35051 commit b8c5c996e888485c3a16d645c8490592534a7882
35052 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
35053 Date: Tue Sep 15 19:45:47 2009 -0700
35055 Eliminate cursor flicker
35057 We can update the cursor without hiding and showing it. In fact, doing the
35058 hide/show causes noticable flicker when running in KMS mode.
35060 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
35062 commit c2abfa8e54acab61250dba1e435760e3b1499c8c
35063 Author: Chris Wilson <chris@chris-wilson.co.uk>
35064 Date: Fri Sep 4 17:10:25 2009 +0100
35066 Avoid fallbacks for compositing gradient patterns
35068 Currently when asked to composite using a gradient source or mask, we
35069 fallback to using fbComposite(). This has the side-effect of causing a
35070 readback on the destination surface, stalling the GPU pipeline. Instead,
35071 like uxa_trapezoids(), we can use pixman to fill a scratch pixmap and then
35072 copy that to an offscreen pixmap for use with uxa_composite().
35075 firefox-talos-svg: 710378.14 -> 549262.96: 1.29x speedup
35079 Thanks to Søeren Sandmann Pedersen for spotting the missing
35082 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35084 commit efbcf29dd1a1ca058b7a2a93f0685102c06c9369
35085 Author: Carl Worth <cworth@cworth.org>
35086 Date: Wed Sep 9 09:32:48 2009 -0700
35088 Update version to 2.8.99.901 for snapshot.
35090 commit 54fa27fcd97a6735d8eb6073f5ec7d0d40a7327f
35091 Author: Carl Worth <cworth@cworth.org>
35092 Date: Wed Sep 9 09:27:01 2009 -0700
35094 NEWS: Add notes for the 2.8.99.901 snapshot
35096 commit 1a77ca74bc829e019a06fe9ad559f013054c27ff
35097 Author: Chris Wilson <chris@chris-wilson.co.uk>
35098 Date: Wed Sep 9 12:09:05 2009 +0100
35100 i915: Restore nearest sampling
35102 My recent commit [94fc93] to use the pixel centre for sampling with the i830
35103 broke the i915. This restores the previous sampling coordinates for the
35104 i915 whilst preserving the correct coordinates for i830.
35106 Fixes: gnome characters disappear
35107 http://bugs.freedesktop.org/show_bug.cgi?id=23803
35109 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35111 commit ce10b5b6fca086eb4af45c1db28352e06ee4ce0b
35112 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
35113 Date: Tue Sep 8 00:56:34 2009 +0800
35115 Add B43 chipset support
35117 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
35119 commit 94fc93d4e2b88565dca17f72903d8991213c9ee8
35120 Author: Chris Wilson <chris@chris-wilson.co.uk>
35121 Date: Sat Sep 5 09:35:10 2009 +0100
35123 i830/i915: Set the sample position to the pixel center.
35125 And in particular we apply the nearest sample bias separately for
35129 device-offset-scale
35130 finer-grained-fallbacks
35131 mask-transformed-{similar,image}
35132 meta-surface-pattern
35134 surface-pattern-big-scale-down
35137 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35139 commit ced0cc8bb28106c18473777bb184872786325e99
35140 Author: Chris Wilson <chris@chris-wilson.co.uk>
35141 Date: Sat Sep 5 09:05:36 2009 +0100
35143 i830: Update comments
35145 i830_composite() is no longer shared with i915 but
35146 i830_emit_composite_primitive() is.
35148 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35150 commit 8863706e25a73f68993d946e2f1c1040bae0f9c5
35151 Author: Chris Wilson <chris@chris-wilson.co.uk>
35152 Date: Sat Sep 5 09:03:56 2009 +0100
35154 i830: Trim composite setup
35156 Remove a couple of redundant NOOPs from the setup and correct the required
35157 space checking for atomic batch operation.
35159 commit a9b12111f9787950ad6b8f295a70a872c5933c93
35160 Author: Chris Wilson <chris@chris-wilson.co.uk>
35161 Date: Sat Sep 5 00:57:39 2009 +0100
35163 i830: remove padding NOOPs from composite
35165 Bumps aa10text up from 249k to 260k!
35167 These NOOPs have existed uncommented since
35168 04d1584737fd0d14e99608a97281fd7b1549ae0e.
35170 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35172 commit 9c1bf6d01ca307b7a9b91e181ad7f341862e5e1c
35173 Author: Chris Wilson <chris@chris-wilson.co.uk>
35174 Date: Fri Sep 4 23:31:44 2009 +0100
35176 i830: do not use stale mask transform
35178 Not only were incorrectly falling back if we had non-affine
35179 transformations, but we made the decision based on a stale transformation
35183 batch_start_atomic horribly breaks performance after a while
35184 https://bugs.freedesktop.org/show_bug.cgi?id=22877
35186 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35187 Tested-by: Maximilian Grothusmann <maxi@own-hero.net>
35189 commit e903b3ebad29d943c05ff550940034c3a43e0235
35190 Author: Matthias Hopf <mhopf@suse.de>
35191 Date: Thu Sep 3 16:34:08 2009 +0200
35193 Don't set backlight level if going on->on.
35195 Otherwise changed backlight will be reset upon DPMS switch off, because first
35196 this function is called with DPMSModeOn.
35198 commit 91f26937c1a3a4f33da86aa6b62ab9b288a54e15
35199 Author: Matthias Hopf <mhopf@suse.de>
35200 Date: Tue Aug 18 18:44:38 2009 +0200
35202 Use "Backlight" as backlight brightness property.
35204 BACKLIGHT is now deprecated, though still usable.
35206 commit dbc8944ee9e6755ab11dfc7375c3394f531a49ce
35207 Author: Matthias Hopf <mhopf@suse.de>
35208 Date: Tue Aug 18 18:34:45 2009 +0200
35210 Add get_property support.
35212 So far only BACKLIGHT is changed.
35214 commit c18fc7955dc2aec878dd3bea1d51aaecf3f08858
35215 Author: Matthias Hopf <mhopf@suse.de>
35216 Date: Tue Aug 18 18:05:45 2009 +0200
35218 Don't ignore backlight level change to the same level.
35220 If set externally to a different level, this would result in a no-op.
35221 OTOH if the display is switched off (DPMS) you do not want the change to take
35222 place immediately, but rather to be saved and set later when the display is
35225 commit fe7693c94a396b9d17bb66a000178750a432ce3a
35226 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
35227 Date: Wed Sep 2 22:57:33 2009 +0800
35231 As DGA is optional in xserver, we should check this too instead
35232 of always trying to init DGA.
35234 Found when update xserver to 6fffcd5825454a7fe58ffbcfb219f007cf38e731,
35235 but not update xf86dgaproto, which caused X fails to start.
35237 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
35239 commit 5812531e08147576de776b2dd64e7f94c08eb851
35240 Author: Carl Worth <cworth@cworth.org>
35241 Date: Fri Jul 31 11:20:23 2009 -0700
35243 debug: i830_valid_command: Return invalid for subopcodes with no name
35245 Previously the code would always return the count, before ever looking
35246 into the _3d_cmds table to see if there was actually a valid command.
35248 Thanks to Alan Coopersmith who reported that the code was confusing
35251 https://bugs.freedesktop.org/show_bug.cgi?id=21666
35253 commit 7c48c21b22bf5862c5a35bda1635753cc5a7197c
35254 Author: Zou Nan hai <nanhai.zou@intel.com>
35255 Date: Thu Aug 27 09:47:12 2009 +0800
35257 set correct value for indirect access check bound
35259 commit 6361c3b9af39265df9222b1f3b6fb9c4197087c1
35260 Author: Keith Packard <keithp@keithp.com>
35261 Date: Tue Aug 25 19:23:27 2009 -0700
35263 Fix SHM functions to work with server after 1.6.0
35265 Signed-off-by: Keith Packard <keithp@keithp.com>
35267 commit 2786a66719a6dbb735eb7c551c412475c30ffa51
35268 Author: Keith Packard <keithp@keithp.com>
35269 Date: Tue Aug 25 18:35:54 2009 -0700
35271 KMS: allocate one bo per crtc for cursor
35273 The KMS API doesn't provide for sharing a single bo for multiple
35274 cursor images, so allocate one bo for each crtc to hold the cursor
35275 image. KMS also only supports ARGB cursors, so don't bother to
35276 allocate buffers for two color cursors.
35278 Signed-off-by: Keith Packard <keithp@keithp.com>
35280 commit e51126c57132492c664f86981c55b166dbb54c79
35281 Author: Keith Packard <keithp@keithp.com>
35282 Date: Tue Aug 25 18:34:25 2009 -0700
35284 Add cursor registers to debug dump output.
35286 Signed-off-by: Keith Packard <keithp@keithp.com>
35288 commit 5fa8d04d9c86f343802c05bd3e11c6e733f01b63
35289 Author: Keith Packard <keithp@keithp.com>
35290 Date: Tue Aug 25 18:33:04 2009 -0700
35292 Reload cursors as needed when setting new modes.
35294 Cursor images may need rotation, or positions updated when new modes
35295 are set. The server provides a convenience function,
35296 xf86_reload_cursors for precisely this purpose. Just call it after the
35299 Signed-off-by: Keith Packard <keithp@keithp.com>
35301 commit 1fc3f467ab3edd405adc569ac7f629077e6ffb9d
35302 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
35303 Date: Tue Aug 25 09:46:10 2009 -0700
35305 Add KMS only build flag
35307 Rather than refactoring all our init code only to have it go away when
35308 we remove UMS, this patch adds a build time flag to allow the driver to
35309 assume KMS support.
35311 With this flag active, the driver will not request that I/O or MEM be
35312 enabled at probe time, which can allow the server (if other drivers also
35313 cooperate) to run as a non-root user.
35315 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
35317 commit 5dccd1be3ab80b642ef2022446f5bdc1656ed943
35318 Author: Wu Fengguang <fengguang.wu@intel.com>
35319 Date: Fri Aug 21 14:57:11 2009 +0800
35321 Add HDMI audio registers
35323 Dump some of the audio registers at server startup time.
35325 (II) intel(0): AUD_CONFIG: 0x00000004
35326 (II) intel(0): AUD_HDMIW_STATUS: 0x00000000
35327 (II) intel(0): AUD_CONV_CHCNT: 0x00000000
35328 (II) intel(0): VIDEO_DIP_CTL: 0x20000600
35329 (II) intel(0): AUD_PINW_CNTR: 0x00000040
35330 (II) intel(0): AUD_CNTL_ST: 0x00002000
35331 (II) intel(0): AUD_PIN_CAP: 0x00000094
35332 (II) intel(0): AUD_PINW_CAP: 0x004073bd
35333 (II) intel(0): AUD_PINW_UNSOLRESP: 0x80000008
35334 (II) intel(0): AUD_OUT_DIG_CNVT: 0x00000001
35335 (II) intel(0): AUD_OUT_CWCAP: 0x00006211
35336 (II) intel(0): AUD_GRP_CAP: 0x00000004
35338 Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
35340 commit 38e97d2366738e83e76c72353b65edb13a9d2c7f
35341 Author: Wu Fengguang <fengguang.wu@intel.com>
35342 Date: Fri Aug 21 11:31:13 2009 +0800
35344 Add intel_audio reg dumping program
35346 It can dump HDMI audio registers for G45.
35348 Signed-off-by: "Wang, Zhenyu Z" <zhenyu.z.wang@intel.com>
35349 Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
35351 commit 6955fc7a74edf6034a292c31a304577c35e925e6
35352 Author: Kristian Høgsberg <krh@redhat.com>
35353 Date: Thu Aug 20 16:48:58 2009 -0400
35355 kms: Don't use fb offset when using shadow buffer
35357 commit 465a4ab416b2e5ad53b96702720331a44fffa2fe
35358 Author: Eric Anholt <eric@anholt.net>
35359 Date: Wed Aug 12 19:29:31 2009 -0700
35361 Align the height of untiled pixmaps to 2 lines as well.
35363 The 965 docs note, and it's probably the case on 915 as well, that the
35364 2x2 subspans are read as a unit, even if the bottom row isn't used. If
35365 the address in that bottom row extended beyond the end of the GTT, a
35368 Thanks to Chris Wilson for pointing out the problem.
35370 commit a3962e6f74ddd954ae1390d150a347745d7bdb24
35371 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
35372 Date: Tue Aug 18 10:56:50 2009 -0700
35374 Print block length of backlight table
35376 For debugging VBIOS dumps
35378 commit 320f21669900f99a7daf8f2294f37be9ad71d05f
35379 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
35380 Date: Mon Aug 17 15:19:41 2009 -0700
35382 Dump LVDS backlight info from bios_reader
35384 Add LVDS backlight and power VBT structures and dump from the BIOS reader.
35386 commit a50916530426a662f5ed262892080b860a445da3
35387 Author: Matthias Hopf <mhopf@suse.de>
35388 Date: Mon Aug 17 15:53:15 2009 +0200
35390 Add BACKLIGHT property support in KMS case.
35392 commit 376397c21eb9a7e4ea79d349af41da81c1af861f
35393 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
35394 Date: Tue Aug 18 10:01:12 2009 +0800
35396 Fix VGA plane disabling
35398 Only apply on G4X with SR01 bit5 workaround for VGA plane disable, and
35399 restore behavior back for other chips to make sure other modes got disabled
35402 For bug #17235, #19715, #21064, #23178
35404 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
35406 commit 926c7e7d30458078f8185d4e0b9b32b40102b3d5
35407 Author: Matthias Hopf <mhopf@suse.de>
35408 Date: Thu Aug 13 13:59:55 2009 +0200
35410 Add HP Mini 5101 to quirks list.
35412 commit 713820197755ea53003b36a920922c3c525eeeea
35413 Author: Adam Jackson <ajax@redhat.com>
35414 Date: Tue Aug 11 14:50:03 2009 -0400
35416 Fix the chip names printed in the log to be less obnoxious.
35418 Names taken from pci.ids. Pineview appears to be a platform not a GMCH,
35419 so use the G/GM convention to distinguish.
35421 commit e8f0763d405a8152c74c28792c52fe12c1d41dd5
35422 Author: Eric Anholt <eric@anholt.net>
35423 Date: Fri Aug 7 18:24:44 2009 -0700
35425 Fix math in the tiling alignment fix.
35427 commit 222b52ef16895823fbf3a0fc0be4eb23b930ed1b
35428 Author: Eric Anholt <eric@anholt.net>
35429 Date: Fri Aug 7 18:05:29 2009 -0700
35431 Align tiled pixmap height so we don't address beyond the end of our buffers.
35433 commit 62494407e529cfa68529b7267155a12d75418f21
35434 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
35435 Date: Thu Aug 6 13:52:54 2009 +0800
35437 Fix typo in bios_reader for invalid pointer cast
35439 Fixed locally for af45482a52999b52bf41468c458808e30c100e35, but pushed
35442 commit 79b6851148574419389ac8055b0c31b8bdac3ab3
35443 Author: Eric Anholt <eric@anholt.net>
35444 Date: Wed Aug 5 12:45:16 2009 -0700
35446 Fix sampler indexes on i965 planar video.
35448 We only set up one sampler, because all of our sampling is the same. By
35449 using a non-zero index for the other two samplers, we'd dereference (likely)
35450 zeroed data, resulting in using NEAREST filtering. This was a regression in
35451 40671132cb3732728703c6444f4577467fa9223f which incidentally switched from
35452 having 6 samplers to 1.
35456 commit f4e4c1a8544e264c5a1da02f4e7990a1beecf71e
35457 Author: Zou Nan hai <nanhai.zou@intel.com>
35458 Date: Wed Aug 5 15:00:37 2009 +0800
35460 It seems that indirect data upper bound check in STATE_BASE_ADDRESS
35461 is not acting like what bspec told on 965gm.
35462 G45+ follow bspec, but we have to set it to a large value for 965gm.
35464 commit af45482a52999b52bf41468c458808e30c100e35
35465 Author: Zhao Yakui <yakui.zhao@intel.com>
35466 Date: Fri Jul 24 10:44:20 2009 +0800
35468 Calculate the DVO relative offset in LVDS data entry to get the DVO timing
35470 Now the DVO timing in LVDS data entry is obtained by using the
35472 a. get the entry size for every LVDS panel data
35473 b. Get the LVDS fp entry for the preferred panel type
35474 c. get the DVO timing by using entry->dvo_timing
35476 In our driver the entry->dvo_timing is related with the size of
35477 lvds_fp_timing. For example: the size is 46.
35479 But it seems that the size of lvds_fp_timing varies on the differnt
35480 platform. In such case we will get the incorrect DVO timing because of
35481 the incorrect DVO offset in LVDS panel data entry.
35483 Calculate the DVO timing offset in LVDS data entry to get the DVO timing
35484 a. get the DVO timing offset in the LVDS fp data entry by using the
35485 pointer definition in LVDS data ptr
35486 b. get the LVDS data entry
35487 c. get the DVO timing by adding the DVO timing offset to data entry
35489 https://bugs.freedesktop.org/show_bug.cgi?id=22787
35491 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
35493 commit 50e2a6734de43a135aa91cd6e6fb5147e15ce315
35494 Author: Dave Airlie <airlied@redhat.com>
35495 Date: Tue Jul 28 18:26:25 2009 +1000
35497 intel: drop RES_SHARED_VGA not needed anymore
35499 commit f3387310f312a4a9e1d691974834c3d290c2fa32
35500 Author: Dave Airlie <airlied@redhat.com>
35501 Date: Tue Jul 28 18:27:10 2009 +1000
35503 xserver: fix up for stable build
35505 reported by Arkadiusz Miskiewicz <arekm@maven.pl>
35507 commit 9bc0096f9de4b85ca6d6a5db109e49c4364bcee7
35508 Author: Dave Airlie <airlied@redhat.com>
35509 Date: Tue Jul 28 13:55:39 2009 +1000
35511 intel: since driver depends on newer server don't need to wrap this
35513 drop resource/RAC interactions
35515 commit 9a3b568d62a0b48f4a42ea5377740b2df1af432a
35516 Author: Dave Airlie <airlied@redhat.com>
35517 Date: Tue Jul 28 13:32:30 2009 +1000
35519 intel: update for resources/RAC API removal
35521 commit 3418c6c16b108e45f67f3c868d28932266f7a0bc
35522 Author: Krzysztof Halasa <khc@pm.waw.pl>
35523 Date: Tue Jul 28 10:47:44 2009 +0800
35525 h/v bias in 3DSTATE_DEST_BUFFER_VARIABLES is 4-bits wide
35529 commit 378445738494663f2ee1615598d4c77b870745af
35530 Author: Dave Airlie <airlied@redhat.com>
35531 Date: Tue Jul 28 10:10:13 2009 +1000
35533 intel: remove unneeded includes
35535 none of these need the resource includes
35537 commit 8084f76d86f048ca5b82da089fffa9665dbbcdd5
35538 Author: Keith Packard <keithp@keithp.com>
35539 Date: Sun Jul 26 13:14:05 2009 -0700
35541 Allow DRM mode setting to include transformations
35543 This removes the explicit transform disabling code in drm_set_mode_major.
35544 Without a fixed X server, transforms will still be broken, but even a fixed
35545 X server can't work around this driver bug.
35547 Signed-off-by: Keith Packard <keithp@keithp.com>
35549 commit 9a45ace207199eb40e95dc6d2670b3096e66ecd9
35550 Author: Xiang, Haihao <haihao.xiang@intel.com>
35551 Date: Thu Jul 23 11:09:53 2009 +0800
35553 XvMC: enable XvMC/XvMC-VLD on IGDNG
35555 commit 043b4a866ab51acffc52d2d71db3475007747571
35556 Author: Xiang, Haihao <haihao.xiang@intel.com>
35557 Date: Thu Jul 23 11:07:13 2009 +0800
35559 add compiled shader programs for XvMC/XvMC-VLD on IGDNG
35561 commit 7684adaa370e8b0bd22e6e7dd1d5fa05f6142bab
35562 Author: Xiang, Haihao <haihao.xiang@intel.com>
35563 Date: Thu Jul 23 10:57:22 2009 +0800
35565 Check the version of intel-gen4asm tool in configure.ac
35567 commit 7dc95b4f1dfecbeb7e9f4a0c35fda32d020be8fe
35568 Author: Xiang, Haihao <haihao.xiang@intel.com>
35569 Date: Fri Jul 24 10:39:05 2009 +0800
35571 XvMC: pin XvMC buffers under KMS.
35573 Under KMS, the buffer allocated by i830_allocate_memory
35574 isn't pinned anymore. However currently 915 XvMC needs
35579 commit 12c5aeca7a3db92d3522d00f5daf338d522e2176
35580 Author: Eric Anholt <eric@anholt.net>
35581 Date: Thu Jul 16 13:34:09 2009 -0700
35583 8xx render: Add limited support for a8 dests.
35585 This improves aa10text performance from 74k to 569k on my 855 laptop.
35586 This also causes my 865 to hang on aa10text like it does on rgb10text,
35587 thanks to actually hitting render accel.
35589 commit 6b7728491c3b771bcba2c7ffd75330c0a0b37f44
35590 Author: Eric Anholt <eric@anholt.net>
35591 Date: Wed Jul 15 16:38:07 2009 -0700
35593 Only align DRI2 tiled pixmaps to the DRI2 tiled pixmap alignment requirement.
35595 This should save significant amounts of memory for glyph and other small
35600 commit 22f7cbc32b70a89d55c79bbea39fb10c50a310ec
35601 Author: Eric Anholt <eric@anholt.net>
35602 Date: Thu Jul 9 23:56:22 2009 -0700
35604 uxa: Tell the driver when we're just going to immediately map the pixmap.
35606 This lets the driver allocate a nice idle buffer object instead of a
35607 busy one, reducing runtime of firefox-20090601 on my G45 from 50.7 (+/- .41%)
35608 to 48.4 (+/- 1.1%).
35610 commit 5ef3db45e059df136162584d00d4b0b511456a33
35611 Author: Eric Anholt <eric@anholt.net>
35612 Date: Thu Jul 9 19:24:38 2009 -0700
35614 uxa: Skip fill of temporary alpha picture that just gets copied over.
35616 This was needed when we were doing the mask computations in this pixmap,
35617 but now they're done in a temporary and then uploaded later.
35619 This reduces runtime of firefox-20090601 from 52.6 (+/- .96%) to 50.7
35620 (+/- .41%) seconds on my G45.
35622 commit 6f3fc6b20f3daedab02e31f49678d4d2ff0fa7a3
35623 Author: Keith Packard <keithp@keithp.com>
35624 Date: Tue Jul 21 12:32:10 2009 -0700
35626 drmmode_output_get_modes: Replace existing EDID property blob with new one
35628 This synchronizes the X EDID data with the kernel EDID data each time the
35629 kernel data may have changed. Otherwise, X ends up stuck with the first EDID
35630 data it sees, failing to accomodate to different monitors.
35632 Signed-off-by: Keith Packard <keithp@keithp.com>
35634 commit 840a787a191b31ece5068e10daed04bee70bb1a2
35635 Merge: 0a4c4c5 5d50a94
35636 Author: Carl Worth <cworth@cworth.org>
35637 Date: Mon Jul 20 23:00:06 2009 -0700
35641 commit 5d50a949b3c5d0ad2bc4cf48ab25da1f707a4f6f
35642 Author: Carl Worth <cworth@cworth.org>
35643 Date: Mon Jul 20 22:59:37 2009 -0700
35645 Increment version number to 2.8.0 for release.
35647 commit b12220bd81f3a0509a3746dac3258e53f3879b23
35648 Author: Carl Worth <cworth@cworth.org>
35649 Date: Mon Jul 20 22:59:02 2009 -0700
35651 NEWS: Add notes for 2.8.0 release
35653 Many thanks to Gordon for his notes from http://intellinuxgraphics.org/2009Q2.html
35655 commit 0a4c4c5fe8ebad2dd13f5770bd90a194eebb2890
35656 Author: Peter Hutterer <peter.hutterer@who-t.net>
35657 Date: Thu Jul 16 11:40:15 2009 +1000
35659 Update to xextproto 7.1 support.
35661 DPMS header was split into dpms.h (client) and dpmsconst.h (server). Drivers
35662 need to include dpmsconst.h if xextproto 7.1 is available.
35664 SHM is now shm.h instead of shmstr. Requires definition of ShmFuncs that's
35665 not exported by the server.
35667 Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
35669 commit 57c7cbade9556e7b21867e61353f0928fd553616
35670 Author: Owain Ainsworth <zerooa@googlemail.com>
35671 Date: Thu Jul 16 20:38:43 2009 +0100
35673 accessing a pixmap if prepare_access fails is verboten.
35675 Don't do it, treat this the same as every other prepare access call in uxa.
35677 Reviewed-by: Keith Packard <keithp@keithp.com>
35678 Signed-off-by: Owain Ainsworth <zerooa@googlemail.com>
35680 commit bb3007384298cb57625ec0b3868dff9b23568f3e
35681 Author: Eric Anholt <eric@anholt.net>
35682 Date: Thu Jul 16 12:56:07 2009 -0700
35684 Really fix i915 render. Fail at commit --amend.
35686 commit 8dd7ccf37e2de6d80b556e6d18af244cefc1e417
35687 Author: Eric Anholt <eric@anholt.net>
35688 Date: Thu Jul 16 11:40:51 2009 -0700
35690 Fix 915-class Render after the 8xx-class Render fix.
35692 The two shared i830_composite.c, so giving i830 atomic batch support
35693 triggered anger about starting i830's atomic area while in i915's atomic
35694 area. Instead, split the emit-a-primitive stuff from the state emission.
35696 commit a1e6abb5ca89d699144d10fdc4309b3b78f2f7a9
35697 Author: Eric Anholt <eric@anholt.net>
35698 Date: Wed Jul 15 14:15:10 2009 -0700
35700 Use batch_start_atomic to fix batchbuffer wrapping problems with 8xx render.
35704 commit e386e7b14b139f15205e14b173e8222bf38d9e18
35705 Author: Keith Packard <keithp@keithp.com>
35706 Date: Wed Jul 15 09:43:04 2009 -0700
35708 Reset framebuffer offset when rebinding aperture (22760).
35710 scrn->fbOffset may be changed when binding objects to the aperture during
35711 server initialization or VT enter. This was accidentally removed when the
35712 NoAlloc option was eliminated.
35714 Signed-off-by: Keith Packard <keithp@keithp.com>
35716 commit b74bf3f9a65af9e72921d4e9028d9d4d023f8bc6
35717 Author: Barry Scott <barry.scott@onelan.co.uk>
35718 Date: Mon Jul 13 16:34:20 2009 -0700
35720 Fix XV scan line calculation when rotated.
35722 commit 82905c7c0b871a97ec435a765c2ca407903ba595
35723 Author: Carl Worth <cworth@cworth.org>
35724 Date: Mon Jul 13 05:29:49 2009 -0700
35726 Increment version to 2.7.99.902
35728 commit 925bc6cbd430a00928fac2ef58724dd37c3bc349
35729 Author: Carl Worth <cworth@cworth.org>
35730 Date: Mon Jul 13 05:27:40 2009 -0700
35732 RELEASING: Fix typo in instructions
35734 commit 67c0afc7b7446a7b98aa7c65043ddba4c7c72b82
35735 Author: Carl Worth <cworth@cworth.org>
35736 Date: Mon Jul 13 05:27:06 2009 -0700
35738 NEWS: Add notes for 2.7.99.902
35740 commit 34c674dd45879b8ba8395b93b16c8a9e7b848f1f
35741 Author: Keith Packard <keithp@keithp.com>
35742 Date: Sat Jul 11 22:53:42 2009 -0700
35744 Remove vestiges of NoAccel options from i830_driver.c
35746 The enum and OptionInfoRec weren't removed in the initial patch
35748 Signed-off-by: Keith Packard <keithp@keithp.com>
35750 commit 33d6e7a2355dfb8ad324c4fa28ce61c7e051b435
35751 Author: Keith Packard <keithp@keithp.com>
35752 Date: Sat Jul 11 22:53:11 2009 -0700
35754 intel.man: Mark NoAccel option as i810/i815 only
35756 The NoAccel option is not valid for other chips.
35758 Signed-off-by: Keith Packard <keithp@keithp.com>
35760 commit ed8a9a94e1a670ca35311c9ed83d0c479530d41a
35761 Author: Keith Packard <keithp@keithp.com>
35762 Date: Fri Jul 10 17:13:14 2009 -0700
35764 i830_uxa_prepare_access: Flush and wait for idle for non-bo pixmaps
35766 Without kernel support and explicit knowledge about where in the ring the
35767 last rendering operation for a specific pixmap was, we must synchronize with
35768 any outstanding rendering before accessing a pixmap which does not have a
35771 Signed-off-by: Keith Packard <keithp@keithp.com>
35773 commit cb19ac207b784d814f6f389110fd1b21a0f34e8b
35774 Author: Keith Packard <keithp@keithp.com>
35775 Date: Fri Jul 10 14:01:02 2009 -0700
35777 KMS: Keep screen pixmap devPrivate.ptr NULL during init and resize
35779 The frame buffer only has a valid address between prepare_access and
35780 finish_access calls, so remove all other attempts to compute an address from
35783 Signed-off-by: Keith Packard <keithp@keithp.com>
35785 commit 704b88dd50a7e7e3f362264b86d0401bee8603aa
35786 Author: Keith Packard <keithp@keithp.com>
35787 Date: Wed Jul 8 13:06:47 2009 -0700
35789 i830_bind_memory: Under UMS: Bind GEM bos with dri_bo_pin, else through the GART
35791 We only need to get static offsets for objects when not running KMS,
35792 otherwise the kernel will manage those as needed for us.
35794 Binding objects is done in one of two ways. For GEM buffer objects, we use
35795 dri_bo_pin. For GART allocated memory, we bind that to the GART.
35797 commit 7b273732f70e91df8b41d5c48e1379271557dd8e
35798 Author: Keith Packard <keithp@keithp.com>
35799 Date: Wed Jul 8 11:53:13 2009 -0700
35801 Allocate GTT space for GEM only under UMS
35803 GEM requires GTT space to map objects. Under KMS, the kernel driver has
35804 already provided all available GTT space to GEM, so the X server need not do
35807 Signed-off-by: Keith Packard <keithp@keithp.com>
35809 commit 56bfee8705f5d7d965227013b205dbc4c93e220c
35810 Author: Keith Packard <keithp@keithp.com>
35811 Date: Fri Jul 10 14:49:20 2009 -0700
35813 Always set screen pixmap data pointer at init and resize times
35815 For non-DRM environments, the screen pixmap will be GART allocated memory
35816 and not a libdrm buffer object and so uxa will only use devPrivate.ptr to
35817 find the associated memory. Make sure devPrivate.ptr is set each time the
35818 framebuffer is allocated so that uxa will be able to draw to it.
35820 Signed-off-by: Keith Packard <keithp@keithp.com>
35822 commit 98087a0b966d5dc69faf72719153a2c878ba3de1
35823 Author: Keith Packard <keithp@keithp.com>
35824 Date: Wed Jul 8 11:47:25 2009 -0700
35826 Make xorg.conf DRI option work under KMS. Fix name of I830AccelMethodInit
35828 KMS mode does not call I830AccelMethodInit as that does the user
35829 modesetting initialization (yes, it was misnamed), but that means that the DRI option
35830 was ignored. Create a new i830_check_dri_option function to do the option
35831 detection, then remove that from I830AccelMethodInit, which is renamed
35832 i830_user_modesetting_init to reflect what it actually does.
35834 Signed-off-by: Keith Packard <keithp@keithp.com>
35836 commit d655a3ff423e69c19a5dc07140cbf3caaa32cb86
35837 Author: Keith Packard <keithp@keithp.com>
35838 Date: Wed Jul 8 18:06:40 2009 -0700
35840 Remove NoAccel support
35842 This removes yet another 'debugging' option that hasn't seen real use in a
35843 long time, and wasn't supported under KMS in any case.
35845 Signed-off-by: Keith Packard <keithp@keithp.com>
35847 commit 4e4b947f0b13f4a62606ccfd5729d5eb26ca0e92
35848 Author: Eric Anholt <eric@anholt.net>
35849 Date: Thu Jul 9 15:52:16 2009 -0700
35851 Remove bad comment about 3DSTATE_DRAWING_RECTANGLE size.
35853 commit 9155cfca75a207bce0fad945f32f0cb33eab8c4e
35854 Author: Eric Anholt <eric@anholt.net>
35855 Date: Thu Jul 9 14:16:07 2009 -0700
35857 Fix lols in trying to figure out whether this is a 64-bit build.
35859 Noticed by: Michel Dänzer <michel@daenzer.net>
35861 commit 40e7c9505265823786cf730214db84812a5e494e
35862 Author: Eric Anholt <eric@anholt.net>
35863 Date: Mon Jul 6 11:54:50 2009 -0700
35865 Refuse to allocate giant BOs on 32-bit systems.
35867 The overcommit of address space combined with these buffers hitting SW
35868 fallbacks all the time means that we're probably better off telling the
35869 application "no" instead of likely silently failing later.
35873 commit 6337cd23e692cae789d07f429442c425c18e1d4f
35874 Author: Eric Anholt <eric@anholt.net>
35875 Date: Wed Jul 1 15:54:38 2009 -0700
35877 Initialize the other argument to drmSetInterfaceVersion.
35879 The kernel ignores it if dd_major is -1, anyway.
35881 commit 216d939858abc924f2e32c95518f937f29ea018e
35882 Author: ling.ma@intel.com <ling.ma@intel.com>
35883 Date: Tue Jul 7 14:26:02 2009 +0800
35885 enable sdvo lvds scaling function
35887 Currently we implemented basic sdvo lvds function,
35888 But except for sdvo lvds fixed mode, we can not switch
35889 to other modes, otherwise display get black. The patch
35890 intends to work for all modes whose HDisplay and VDisplay
35891 are lower than fixed mode.
35893 Signed-off-by: Ma Ling <ling.ma@intel.com>
35895 commit 0402f4f331148084552bd3963dbcb3fb900be8ea
35896 Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
35897 Date: Thu Jun 18 12:33:47 2009 +0100
35899 Raise XV limit to 2048x2048 to match hardware limits.
35901 The bigrequests limit isn't present in current X servers (tested using
35902 textured video on a 965 with both image and window at 2048x2048 on a
35903 1920x1200 display, and image at 2048x2048, window at 1024x1024).
35905 Remove the artificial limit, enabling full-screen HD video when
35908 commit a66357832388ba9db21a4b3bf8311d9d1f1ab308
35909 Author: Keith Packard <keithp@keithp.com>
35910 Date: Tue Jul 7 14:13:57 2009 -0700
35912 non-DRI FB resize failed to assign the screen pixmap devPrivate.ptr (22328)
35914 When not using DRI, the screen pixmap is not in a bo, and so the usual
35915 enable/disable access functions don't adjust the pixmap devPrivate field,
35916 leaving it to the frame buffer allocation code to assign this correctly.
35918 During mode setting and fb resizing, FB access is disabled, and the
35919 screen pixmap devPrivate is stashed away by xf86EnableDisableFBAccess,
35920 to be restored when FB access is turned back on. This means that we have to
35921 set the pixmap devPrivate.ptr (in case xf86EnableDisableFBAccess doesn't
35922 do this), along with storing the address in the scrn->pixmapPrivate field.
35924 Signed-off-by: Keith Packard <keithp@keithp.com>
35925 Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
35927 commit c889b34e432198d5410a068eff3089ff5314ac27
35928 Author: Chris Wilson <chris@chris-wilson.co.uk>
35929 Date: Tue Jul 7 18:05:01 2009 +0100
35931 Check for a valid I830Ptr before closing master.
35933 After failing to become DRM master, the X server dies attempting to close
35934 the master fd during free:
35936 (EE) intel(0): [drm] failed to set drm interface version.
35937 (EE) intel(0): Failed to become DRM master.
35938 (EE) intel(0): failed to get resources: Bad file descriptor
35939 (EE) intel(0): Kernel modesetting setup failed
35942 0: X(xorg_backtrace+0x3b) [0x8133a3b]
35943 1: X(xf86SigHandler+0x55) [0x80c7945]
35945 3: /usr/lib/xorg/modules/drivers//intel_drv.so [0xb7b4bfcc]
35946 4: X(xf86DeleteScreen+0x6b) [0x80d465b]
35947 5: X(InitOutput+0x548) [0x80b0158]
35948 6: X(main+0x1cb) [0x807220b]
35949 7: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7d107a5]
35951 Saw signal 11. Server aborting.
35952 ddxSigGiveUp: Closing log
35953 ddxSigGiveUp: re-raising 11
35956 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
35958 commit 705042f497b7b3843c2dcc5c160fb8dfeac1472a
35959 Author: Keith Packard <keithp@keithp.com>
35960 Date: Mon Jul 6 13:49:31 2009 -0700
35962 Handle DRI2INFOREC version 3
35964 This DRI2 version does not support the old CreateBuffers/DestroyBuffers
35967 Signed-off-by: Keith Packard <keithp@keithp.com>
35969 commit 2ebc7d32e47b5edd5b776c39f936ed4e053caac2
35970 Author: Keith Packard <keithp@keithp.com>
35971 Date: Thu Jul 2 13:13:14 2009 -0700
35973 Update to multi-API DRI2 interface
35975 The DRI2 interface was changed to support both old and new drivers in an
35976 API/ABI compatible fashion. This change syncs the intel driver with the new
35977 version of the DRI2 API.
35979 Signed-off-by: Keith Packard <keithp@keithp.com>
35981 commit f6f79eb629184366b1355743d601129a526da90c
35982 Author: Rémi Cardona <remi@gentoo.org>
35983 Date: Mon Jul 6 11:01:31 2009 +0200
35985 remove unused shader program
35987 This file is not even referenced by any Makefile.am
35989 Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
35991 commit 7e722ada533777c5e9ddf44bb4d770bacf8e13bf
35992 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
35993 Date: Mon Jul 6 16:25:13 2009 +0800
35995 Disable FBC on IGDNG
35997 Don't make FBC count for memory allocation.
35999 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36001 commit 74227141923a2f5049592219ab80e8733062a5d9
36002 Author: Barry Scott <barry.scott@onelan.co.uk>
36003 Date: Tue Jun 23 14:14:50 2009 +0100
36005 Fix segv for clipped movie window
36007 When playing a movie that is clipped on its left and right edges the Xorg
36008 server will SEGV sometimes. This is because the intel driver ignores the
36009 clipping info when it copies the planes out of the XV data.
36011 The check for the optimised copy was wrong to ignore the width required.
36012 Which leads to too much data being copied by the memcpy. It the source buffer
36013 happens to end exactly on a page boundary the server will SEGV.
36015 As we reviewed the code we checked the calculation of src1, src2 and src3.
36016 The patch includes additional comments to make it clear what the elements of
36017 the calculation are.
36019 This bug exists in git head and we also see it in 2.4.1.
36023 Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
36025 commit c1755599db1d9a20954b84ccc07afd892bb6ac9e
36026 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36027 Date: Thu Jul 2 13:04:51 2009 -0700
36029 Clear the bo on the rotate scratch pixmap
36031 Since the scratch pixmap header will be re-used after allocation, we
36032 need to clear its bo attachment when we stop using it, otherwise a later
36033 user will use a bogus bo.
36035 Reviewed-by: Keith Packard <keithp@keithp.com>
36036 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36038 commit 324b4686204feb3a7370eeecaff8ba44635f73ca
36039 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
36040 Date: Thu Jul 2 14:15:37 2009 +0200
36042 Xv i830_display_video splitup: extract i830_update_scaling_factors
36044 Just moved the code, no other changes.
36046 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
36048 commit c2410addbfb99fcd7069591d9f387c35ed760522
36049 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
36050 Date: Thu Jul 2 14:15:36 2009 +0200
36052 Xv i830_display_video splitup: extract i830_update_polyphase_coeffs
36054 To slightly clean up the implementation of i830_update_polyphase_coeffs,
36055 introduce the two small helper functions i830_limit_coeff and
36056 i830_store coeffs_in_overlay_regs.
36058 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
36059 Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
36061 commit 795c11c49cf10525f02127a3629d35378d802fa7
36062 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
36063 Date: Thu Jul 2 14:15:34 2009 +0200
36065 Xv i830_display_video splitup: extract i830_update_dst_box_to_crtc_coords
36067 Just moved the code ouf of line.
36069 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
36071 commit 4100abdf5d208bbcbb4ceabad0572c04221443c9
36072 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
36073 Date: Tue Jun 30 13:12:45 2009 +0200
36075 Xv: kill !textured condition
36077 This is in the overlay path and therefore always true.
36079 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
36081 commit b0df0fe91e2b800ed096f369850aa1af4be2f157
36082 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
36083 Date: Tue Jun 30 13:12:44 2009 +0200
36085 Xv overlay: implement GAMMA5 errata
36087 - also ensure that the most significant byte is zero
36088 - while I was looking at the code, add the Overlay suffix to
36089 SetPortAttribute like in the textured case.
36091 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
36093 commit 5ef4d3cde1335350d82469ebbaed1b547a59552c
36094 Author: Daniel Vetter <daniel.vetter@ffwll.ch>
36095 Date: Tue Jun 30 13:12:43 2009 +0200
36097 Xv: kill unneeded indirection
36099 overlay and textured video have the exact same QueryImageAttributes
36102 Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
36104 commit 1e4784bf26e3c154f5673f7b5add3ef7af3b1474
36105 Author: Eric Anholt <eric@anholt.net>
36106 Date: Tue Jun 30 19:52:36 2009 -0700
36108 uxa: Fix segfault on source-only picture usage with FallbackDebug.
36112 commit 7e79fc8aa93df4df37c25cf37ee0ec6c7caca1d9
36113 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36114 Date: Tue Jun 30 15:40:34 2009 +0800
36116 Fix 945GM VT switch in UMS
36118 Bug #19578. We should set private intel_crtc state according
36119 to current, as fail to do so pipe A needs active won't be taken
36120 care of. Also make sure pipe swap operation always set during
36123 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36125 commit 50392ac8234d643c8a99e1753bdb196c0062a891
36126 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
36127 Date: Mon May 18 13:53:33 2009 +0800
36129 Load fbcon too if kernel mode setting is checked on
36131 If i915 module has already been loaded and kms check is true,
36132 it would be nice to load fbcon module too.
36134 Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
36136 commit 058ad9e6ab4a00ff66046d94c2129056011ebee9
36137 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36138 Date: Tue Jun 30 16:02:19 2009 -0700
36140 Don't try to pin buffers in KMS mode
36142 The only things we try to pin in KMS mode are the cursor objects and
36143 front buffer, and those are taken care of by the kernel anyway, so we
36144 shouldn't even bother trying to pin them (well, not entirely true,
36145 XvMC tries to pin as well, but it needs work w/KMS anyway).
36147 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36149 commit cec9fc6f6cffce186606f39982d0d78ff7c63bbf
36150 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36151 Date: Tue Jun 30 12:27:59 2009 -0700
36153 Make KMS set_resource function return TRUE
36155 This is what's expected by the server, and allows the EDID for example
36156 to be exported in the KMS case.
36158 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36160 commit 53e248af4365232416d2143a791a07c6751f8319
36161 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36162 Date: Tue Jun 30 12:00:28 2009 -0700
36164 Use DVO timing block instead of fp_timing when parsing LFP data
36166 The KMS side was correct, but the UMS patch was broken. We need to use
36167 the DVO timing block of the LFP data to get the timing, not the
36170 Fixes fdo bug #22529.
36172 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36174 commit 362883c2f9a3442f7678e6e815f41b21baaa3f53
36175 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36176 Date: Mon Jun 29 16:26:15 2009 -0700
36178 Add a few error messages for DRM initialization
36180 Makes it easier to see where things go wrong.
36182 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36184 commit afac333bef4a0ac934f0e4d933dc5053d81ca88c
36185 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36186 Date: Tue Jun 30 11:10:52 2009 +0800
36188 Remove unused packed yuv sampler shader programs
36190 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36192 commit 150c9adebc0b502f19c970783f411928e5a5c3b1
36193 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36194 Date: Tue Jun 30 10:42:41 2009 +0800
36196 Xv: fix domain usage for binding table on i965+ chips
36198 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36200 commit 7770958e42d60966d95702e75c704e20ae2d90dd
36201 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36202 Date: Tue Jun 30 10:41:26 2009 +0800
36204 Add XV support on IGDNG
36206 This brings necessary change for IGDNG for texture video support
36207 from 2D render code.
36209 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36211 commit f806fe7d675b966680a63406167ce37e4f8a1ae8
36212 Author: Xiang Haihao <haihao.xiang@intel.com>
36213 Date: Thu Jun 25 15:39:32 2009 +0800
36215 Enable 2D composite on IGDNG
36217 This patch enables 2D composite on IGDNG. IGDNG requires
36218 new compiled shader programs for Gen5 and some command changes.
36219 The most notable is the layout of vertex element has changed,
36220 but we tried to keep it as origin to not change shader programs.
36221 Also vertex buffer state requires end address of vertex buffer
36222 instead of origin max index.
36224 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36226 commit 9fb34012f667e37f480085696ef9c2632d6eb7e1
36227 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36228 Date: Thu Jun 25 15:30:04 2009 +0800
36230 Add new compiled shader program for IGDNG
36232 Also check intel-gen4asm tool here for new -g option, which is
36233 required to compile new programs.
36235 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36237 commit 488acc4595bb7f40130afcb8bcb05656ff3ae82c
36238 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36239 Date: Thu Jun 25 14:05:40 2009 +0800
36241 Move shader programs under its own subdirectory
36243 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36245 commit 170cae0c8d58fc141de1d8a2f17a4328d39c1263
36246 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36247 Date: Mon Jun 29 12:55:43 2009 -0700
36249 Only get the VBIOS in non-KMS mode
36251 In KMS mode, the kernel takes care of this for us, so don't bother.
36253 commit 6511c082459789cf279e7a4528775a1c821ad8cc
36254 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36255 Date: Mon Jun 29 12:53:51 2009 -0700
36257 Output error info if we fail to get DRM resources
36259 Useful for debugging.
36261 commit 00eb73286c3512a362dce00efdeae740772d0dcd
36262 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36263 Date: Mon Jun 29 09:12:28 2009 -0700
36265 Use swapbuffers_wait control
36267 Commit 1eec83a203c48822400742a1fb184b2cb52c62f7, which added the new
36268 SwapbuffersWait option, didn't actually include the code which used it. So
36269 add a test to DRI2's CopyRegion call, only emitting the scanline wait
36270 command if the swapbuffers_wait option is set.
36272 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36274 commit f53b3239dbc0ed723774e386e07ac9d8ce96bb89
36275 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36276 Date: Thu Jun 25 10:22:23 2009 +0800
36278 Disable XvMC on 915G/GM in KMS
36280 These chips require physical address for XvMC surface, which
36281 is not available in KMS case. Instead of crashing X, disable it now.
36283 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36285 commit f0270bbb47baed78a0ff6189ae20d3ac322ec02b
36286 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36287 Date: Wed Jun 24 14:42:08 2009 -0700
36289 Fix i830_crtc_on to only check outputs associated with the given CRTC
36291 Otherwise we may end up returning a false positive if some other output & crtc
36292 are on, but not the one in question, again leading to hangs.
36294 Reported-by: Eric Anholt <eric@anholt.net>
36295 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36297 commit 5d80e24b5fc6d6028028da6ded35389c08bfce29
36298 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36299 Date: Wed Jun 24 13:12:25 2009 -0700
36301 Treat disabled CRTCs as "not covering" for scanline wait purposes
36303 Now that swapbuffers does a scanline wait to avoid tearing, it's
36304 important to take into account the CRTC status to avoid hangs. If we
36305 do a scanline wait when the CRTC is off (due to DPMS for example) we'll
36306 hang the GPU. So add some code to check the CRTC DPMS status to the
36307 i830_covering_crtc function, returning NULL if none of the covering
36308 CRTCs are actually active. KMS vs UMS logic is hidden in new i830*
36309 functions, cleaning up both DRI2 & video paths a bit.
36311 Fixes fdo bug #22383.
36313 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36315 commit 6d025e679a99778496576af9a6a6fa8c043ae811
36316 Author: Alan Coopersmith <alan.coopersmith@sun.com>
36317 Date: Tue Jun 23 09:53:14 2009 -0700
36319 Harden i830 render in case check_composite didn't throw out bad formats.
36321 Fixes a warning in a static analysis program, and the code's a little
36326 commit 1eec83a203c48822400742a1fb184b2cb52c62f7
36327 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36328 Date: Tue Jun 23 15:05:03 2009 -0700
36330 Add option to control swapbuffers behavior
36332 Until we get triple buffering, we'll want this so users can avoid taking a
36333 performance hit on apps that render slower than the refresh rate.
36335 Fixes fdo bug #22234.
36337 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36339 commit bfeeac6de096256fca82244338bb45d53ee53cbc
36340 Author: Zou Nan hai <nanhai.zou@intel.com>
36341 Date: Tue Jun 23 11:31:37 2009 +0800
36343 i915 xvmc, fix fd.o bug #22103
36345 commit 15af8ea6ab6998bbab9f4eeda227565c409da229
36346 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36347 Date: Mon Jun 22 11:11:06 2009 -0700
36349 Fix LFP data block fetch
36351 Apparently the proper way to do this is to use the LFP data pointer block to figure out the LFP data block entry size, then use that plus the panel index to calculate an offset into the LFP data block array.
36353 Fixes fdo bug #19450.
36355 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36357 commit 534e73ad4f234a04755917f2bf17ba821c27eb52
36358 Author: Zhao Yakui <yakui.zhao@intel.com>
36359 Date: Thu Jun 18 09:46:32 2009 +0800
36361 Don't change the blank/sync width when calculating scaled modes
36363 Don't the change the blank/vsync width while doing LVDS scaled modes.
36364 And use the border instead of border minus one.
36366 At the same time, make sure the horizontal border and hsync are even for
36367 the LVDS that works in dual-channel mode. So both horizontal border and hsync
36368 start are also changed to be even, even for the LVDS in single-channel
36371 https://bugs.freedesktop.org/show_bug.cgi?id=20951
36373 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36374 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36376 commit b5cd2130f97591f4a387db1b98c940c30bc6404c
36377 Author: ling.ma@intel.com <ling.ma@intel.com>
36378 Date: Tue Jun 16 16:52:25 2009 +0800
36380 TV: Set correct voltage level override values
36382 We detect TV connect status by setting DAC voltage level override
36383 values as 0.7 voltage for DAC_A/B/C. The corresponding 2-bits shold be 0x2,
36384 In order correctly to set last bit as 0, at first we must clean it.
36386 It fixed freedesktop.org bug #21204
36388 Signed-off-by: Ma Ling <ling.ma@intel.com>
36390 commit a6cdcd9fee0164f79075063fd163d142a55fcbc5
36391 Author: ling.ma@intel.com <ling.ma@intel.com>
36392 Date: Tue Jun 16 16:47:52 2009 +0800
36394 Set hot plug interrupt to detect HDMI output
36396 We detect HDMI output connection status by writing to HOT Plug Interrupt
36397 Detect Enable bit in PORT_HOTPLUG_EN. The behavior will generate an specified
36398 interrupt, which is caught by audio driver, but during one detection driver
36399 set all Detect Enable bits of HDMIB, HDMIC and HDMID, which generate wrong
36400 interrupt signals for current output, according to the signals audio driver
36401 misunderstand device status. The patch intends to handle corresponding output
36404 It fixed fredesktop bug #21371
36406 Signed-off-by: Ma Ling <ling.ma@intel.com>
36408 commit 6b93afc564a5e74b0eaaa46c95f557449951b3b9
36409 Author: Bryce Harrington <bryce@bryceharrington.org>
36410 Date: Wed May 27 03:40:52 2009 -0700
36412 add pipe a force quirk for Dell mini
36414 Add quirk to solve issue with black screen and hang occuring after closing the
36415 lid with attached external monitor, on Dell Mini.
36417 Fixes fdo bug #21960.
36419 Signed-off-by: Bryce Harrington <bryce@bryceharrington.org>
36421 commit eb09014ce10428bbcab04e155186382975545f0a
36422 Author: Li Peng <peng.li@intel.com>
36423 Date: Tue Jun 16 15:29:57 2009 -0700
36425 don't wait for vblank on rotated displays
36427 We may hang or wait for the wrong line if the display is rotated, so just skip
36428 the wait in that case.
36430 Fixes fdo bug #22196.
36432 commit 6c56521bdc0443c0656271caaa795feb13bc1d6b
36433 Author: Bryce Harrington <bryce@bryceharrington.org>
36434 Date: Wed May 27 05:18:53 2009 -0700
36436 pipe-a quirk for thinkpad x30
36438 Fixes freeze when closing lid on ThinkPad X30.
36440 Fixes FDO bug # 21976.
36442 Signed-off-by: Bryce Harrington <bryce@bryceharrington.org>
36444 commit 246cec965958e94babf5377e6f221522b05fb458
36445 Author: Zhao Yakui <yakui.zhao@intel.com>
36446 Date: Wed Jun 10 11:17:28 2009 +0800
36448 Fix EDID for LVDS output device to add the default modes
36450 Fix the EDID for the LVDS output device to add the default modes.This is
36451 similar to what we have done in UMS mode.
36452 a. When there exists the EDID, either find the DS_RANGES block or replace
36453 a DS_VENDOR block, smashing it into a DS_RANGES block with open refresh
36454 to match all the defaults modes.
36455 b. When there is no EDID, we will construct a bogus EDID and add a DS_RANGES
36456 block with the open refresh to match all the default modes.
36458 http://bugs.freedesktop.org/show_bug.cgi?id=20801
36459 http://bugs.freedesktop.org/show_bug.cgi?id=21094
36460 http://bugs.freedesktop.org/show_bug.cgi?id=21346
36461 http://bugs.freedesktop.org/show_bug.cgi?id=21417
36462 http://bugs.freedesktop.org/show_bug.cgi?id=21671
36464 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36466 commit 5d1dc7677004d445a7a781decd8c1ef9747c14fb
36467 Author: Zhao Yakui <yakui.zhao@intel.com>
36468 Date: Wed Jun 10 11:17:27 2009 +0800
36470 Get the LVDS panel limit and check whether the given modeline is valid
36472 When the connector type is LVDS, it will traverse the mode list returned by
36473 KMS kernel to get the LVDS panel limit. Then it will use the panel limit to
36474 check whether the given modeline is valid. If the given modeline exceeds
36475 the LVDS panel limit, it will be invalid.
36477 http://bugs.freedesktop.org/show_bug.cgi?id=20801
36478 http://bugs.freedesktop.org/show_bug.cgi?id=21094
36479 http://bugs.freedesktop.org/show_bug.cgi?id=21346
36480 http://bugs.freedesktop.org/show_bug.cgi?id=21417
36481 http://bugs.freedesktop.org/show_bug.cgi?id=21671
36483 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36485 commit 115e28639fbf6a1eba636dafac02fadd83036c75
36486 Author: Zhao Yakui <yakui.zhao@intel.com>
36487 Date: Wed Jun 10 11:17:26 2009 +0800
36489 Add the private data for the LVDS connector
36491 Add the private data when the connector type is LVDS.
36492 We can use the private_data to store the LVDS panel limit.
36493 For example: Hdisplay, Vdisplay.
36495 http://bugs.freedesktop.org/show_bug.cgi?id=20801
36496 http://bugs.freedesktop.org/show_bug.cgi?id=21094
36497 http://bugs.freedesktop.org/show_bug.cgi?id=21346
36498 http://bugs.freedesktop.org/show_bug.cgi?id=21417
36499 http://bugs.freedesktop.org/show_bug.cgi?id=21671
36501 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36503 commit 3da549f5b350dd42516e5cb3576f7fefe012d95e
36504 Author: Zhao Yakui <yakui.zhao@intel.com>
36505 Date: Wed Jun 10 11:37:03 2009 +0800
36507 GFX: Initialize the SDVO device based on the valid slave address
36509 When the slave address is found for the SDVO port, the SDVO device will
36511 When the slave address is not found for the SDVO port, it will return
36512 the slave address by using the following flowchart:
36513 a. If the SDVO device info is found for another SDVO port, it will return
36514 the slave address that is not used. For example: if 0x70 is used, then 0x72
36516 b. If no SDVO device info is found for another SDVO port, it will return
36517 0x70 for SDVOB and 0x72 for SDVOC.
36519 http://bugs.freedesktop.org/show_bug.cgi?id=20429
36521 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36523 commit 51b87b9913ba164d4d5de79e558233915b37a0a5
36524 Author: Zhao Yakui <yakui.zhao@intel.com>
36525 Date: Wed Jun 10 11:37:02 2009 +0800
36527 parse general definition block to get the SDVO device info
36529 The general definition block contains the child device tables, which include
36530 the child device info. For example: device slave address, device dvo port,
36532 We will get the info of SDVO device by parsing the general definition blocks.
36533 Only when a valid slave address is found, it is regarded as the SDVO device.
36534 And the info of DVO port and slave address is recorded.
36536 http://bugs.freedesktop.org/show_bug.cgi?id=20429
36538 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36540 commit 6ba148bbd78783f59eed3d898638c39b950dcd89
36541 Author: Zhao Yakui <yakui.zhao@intel.com>
36542 Date: Wed Jun 10 11:37:01 2009 +0800
36544 Dynamically get the number of child device in general definition block
36546 The size of general definition block varies on different platform/machines.
36547 In such case the number of child device is also different.
36548 And it will be better to get the number of child device in general definition
36551 The number of child device can be calculated by the following formula:
36552 (block_size - block_header_size) /
36553 sizeof( struct child_device_config)
36555 http://bugs.freedesktop.org/show_bug.cgi?id=20429
36557 Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
36559 commit cdbf84f20295c8a78624318aa6fdfff3f5c8ce27
36560 Author: Keith Packard <keithp@keithp.com>
36561 Date: Thu Jun 11 13:57:09 2009 -0700
36563 intel_batch_init: test have_gem, not directRenderingType
36565 Under KMS, directRenderingType will get set to DRI_NONE during driver
36566 initialization. When the first batch buffer is allocated, as
36567 directRenderingType is DRI_NONE, the GEM bufmgr would get trashed as
36568 intel_batch_init called a fake-bufmgr specific function.
36570 Signed-off-by: Keith Packard <keithp@keithp.com>
36572 commit 6d062e9e9fc4ac92a5dc2941b4ed0251a1dc6e11
36573 Author: Julien Cristau <jcristau@debian.org>
36574 Date: Thu Jun 11 17:15:04 2009 +0200
36578 If we don't find xext.pc, disable xvmc instead of failing configure
36579 Also add dependencies on xfixes and dri2proto (src/xvmc/dri2.h includes
36580 <X11/extensions/Xfixes.h> and <X11/extensions/dri2tokens.h>).
36582 commit 9d3c3b056c1d4da6ef7afd3f71a5994fad810632
36583 Author: Carl Worth <cworth@cworth.org>
36584 Date: Wed Jun 10 15:30:45 2009 -0700
36586 Increment version to 2.7.99.901
36588 This is the first release candidate for 2.8.0
36590 commit e2622045155f5aa79bacbec8f81c1231194f3033
36591 Author: Carl Worth <cworth@cworth.org>
36592 Date: Wed Jun 10 15:30:09 2009 -0700
36594 NEWS: Summarize new developments for 2.7.99.901
36596 commit e5bfa2702a31566fa94fa75f7289d7cbe9825420
36597 Author: Lukasz Kurylo <Lukasz.Kurylo@gmail.com>
36598 Date: Wed Jun 10 07:55:31 2009 -0700
36600 DRI2 copyregion: don't wait for scanlines that won't happen
36602 In some configurations, it's possible to wait for a scanline outside of
36603 a given CRTC range. Make sure that can't happen to fix multihead cases
36606 Fixes fdo bug #22203.
36608 Signed-off-by: Lukasz Kurylo <Lukasz.Kurylo@gmail.com>
36610 commit cda5561716891484aa6f1c4a234be8d88f17f659
36611 Author: Zou Nan hai <nanhai.zou@intel.com>
36612 Date: Wed Jun 10 16:43:30 2009 +0800
36614 add missing files for previous commit
36616 commit 6d500be3cdecd7617dc6119b58ae273eea478af7
36617 Author: Zou Nan hai <nanhai.zou@intel.com>
36618 Date: Wed Jun 10 15:53:24 2009 +0800
36620 enable MC on g45 for player without vld patch
36622 commit 88da9b48737229b305da4423e9aa43a09ccdcfac
36623 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36624 Date: Wed Jun 10 10:45:51 2009 +0800
36626 Add new chipset register dumps
36628 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36630 commit 0d56ef94be0592aec9aa83e4a5e4ff13348640c2
36631 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36632 Date: Fri Jun 5 12:59:07 2009 +0800
36634 Disable composite on new chipset for now
36636 This depends on updated intel-gen4asm be ready for new chip.
36638 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36640 commit 440ccc44f51d3a5d6f46c28cfcc576cad155fbbc
36641 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36642 Date: Fri Jun 5 12:56:04 2009 +0800
36644 Add new register definitions
36646 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36648 commit 0d8a9e2c6f58115b9b8449de52e795699ed032af
36649 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36650 Date: Fri Jun 5 12:51:51 2009 +0800
36652 Update intel_gtt utility for new chipset GTT check
36654 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36656 commit fb524caa3e2f1f516717669642bb4b2244f9e7e4
36657 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36658 Date: Mon Jun 8 10:22:14 2009 +0800
36660 Remove fixed MMIO size
36662 Use pci resource size instead, which will get the correct MMIO range.
36663 New chipset uses obviously larger MMIO range.
36665 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36667 commit 4f40b33ef4b069b18a6a18406da83a23ca6e1127
36668 Author: Zhenyu Wang <zhenyuw@linux.intel.com>
36669 Date: Fri Jun 5 11:57:57 2009 +0800
36671 Add new chipsets PCI ids
36673 Desktop and mobile version of new chipsets are added.
36674 Also do memory config like Intel 4 series chipset.
36676 Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
36678 commit accdbd23676d812d2345f86d8e3ee62f108841ff
36679 Author: Carl Worth <cworth@cworth.org>
36680 Date: Fri May 29 15:34:20 2009 -0700
36682 UXA: Rasterize trapezoids to system memory, not a pixmap
36684 Since we're only doing software rasterization right now, anyway, it
36685 makes more sense to just rasterize to system memory and then upload
36686 to a pixmap once complete. This avoids expensive read-modify-write
36689 This results in a 2.4x speedup for a real-world test case that's
36690 heavy on trapezoids, which is swfdec running on the following file:
36692 http://michalevy.com/wp-content/uploads/Giant%20Steps%202007.swf
36694 Many thanks to Chris Wilson for his cairo-traces repository and
36695 cairo-perf-trace tool which makes it so easy to measure things
36698 commit b5e32c9cf896a0b93d193d797a8e83b4aa4691fb
36699 Author: Carl Worth <cworth@cworth.org>
36700 Date: Fri May 1 13:51:39 2009 -0700
36702 Rename i830_exa.c to i830_uxa.c
36704 It was just confusing otherwise, (since the EXA code has all been
36707 commit ee539e58c3bf39766c560d625f6e4158d419e64e
36708 Author: Kristian Høgsberg <krh@redhat.com>
36709 Date: Mon Jun 8 21:14:14 2009 -0400
36711 Fall back to fb pixmaps for backing glyph pictures
36713 GEM pads pixmaps to 512 byte stride and backs them with a kernel side
36714 buffer objects. We typically don't render out of glyph pictures, so
36715 we're incurring a lot of overhead per glyph by allocating a GEM pixmap
36716 per glyph. By looking at the usage hint, we can fall back to
36717 fbCreatePixmap for pixmaps backing glyph pictures, which gives us
36718 a nice tight malloced pixmap. The fast path for text rendering is
36719 compositing from the glyph cache pixmap to the destination, which
36720 shouldn't be significantly affected.
36722 Quick bit of testing:
36724 xlib-rgba-before 384512.49: 1.01x
36725 xlib-rgba-after 389633.94: 1.00x
36726 The difference being within the margin of error for the benchmark.
36728 Signed-off-by: Eric Anholt <eric@anholt.net>
36729 Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
36731 commit 4698b3bd79452ae2066a3d195cf58dd5e30c93f3
36732 Author: Xiang, Haihao <haihao.xiang@intel.com>
36733 Date: Fri Jun 5 10:43:50 2009 +0800
36735 xvmc: don't include xf86dri.h
36737 commit 4d9f3a21ab9b2737fe7a5a088c2da1b0c2be7f04
36738 Merge: b8e360b 52054b6
36739 Author: root <root@hdmi.sh.intel.com>
36740 Date: Thu Jun 4 10:27:41 2009 +0800
36742 Merge branch 'xvmc-vld'
36744 commit b8e360bf2b77d28559d15a7c0f9c766848eb6ced
36745 Author: Eric Anholt <eric@anholt.net>
36746 Date: Wed Jun 3 10:12:25 2009 +0000
36748 Fix segfault in DRI2 vblank syncing if the region isn't onscreen.
36750 Also, fix some weirdness in the checking for whether the target was the
36753 commit 5901a67fc85ac80fabfa98b78202a388445275c3
36754 Author: Eric Anholt <eric@anholt.net>
36755 Date: Wed Jun 3 10:00:50 2009 +0000
36757 Fix some drawable abuse in i830_dri.c
36759 We were casting pixmaps to drawables even if they weren't pixmaps. They did
36760 happen to work out since we only used the drawable record out of them.
36762 commit 5e48146777921b693c828af9566a77422cc4f85f
36763 Author: Eric Anholt <eric@anholt.net>
36764 Date: Wed Jun 3 09:57:10 2009 +0000
36766 Replace some pixmap-related idioms with inline functions to do so.
36768 commit ea0b00e675281b2914450992501566122f9affe0
36769 Author: Ma Ling <ling.ma@intel.com>
36770 Date: Mon Jun 1 17:13:53 2009 +0800
36772 Provide mode line from VBT to xorg.
36774 If sdvo get modes function can't find modes from EDID,
36775 transmit sdvo lvds modes from VBT to xorg.
36777 Signed-off-by: Ma Ling <ling.ma@intel.com>
36779 commit 9e7ee50bea5f65003f1d70cc06d8c1ace1282548
36780 Author: Ma Ling <ling.ma@intel.com>
36781 Date: Mon Jun 1 17:13:22 2009 +0800
36783 Fetch mode line from VBT, then keep it.
36785 Parse SDVO LVDS option section, then according to panel type
36786 fetch fixed mode line from SDVO LVDS DTDS section .
36788 Signed-off-by: Ma Ling <ling.ma@intel.com>
36790 commit 88f766be008008d76c150e3ac16f09d4ecbb6d53
36791 Author: Ma Ling <ling.ma@intel.com>
36792 Date: Fri May 15 15:22:11 2009 +0800
36794 Wait doubled regis to be stable for load pipe detection
36796 We have two approaches for VGA detections: hot plug detection for 945G onwards
36797 and load pipe detection for Pre-945G. load pipe detection will get one free
36798 pipe ,and set border color as red and blue, then check CRT status by
36799 swf register. Because pipe registers in hires mode are double buffered,
36800 once set force border bit in pipeconf register, we have to wait for
36801 a vblank until it is effective, otherwise result is unstable.
36803 It fixed freedesktop bug #20463
36805 Signed-off-by: Ma Ling <ling.ma@intel.com>
36807 commit 6a635be5ea2b2df9d6b9a8006325244394038474
36808 Author: Ma Ling <ling.ma@intel.com>
36809 Date: Fri May 15 14:01:18 2009 +0800
36811 Set activation periods as 64 cdclk sync up with KMS
36813 We use force CRT detect trigger bit(1 << 3) to detect VGA in hot plug mode,
36814 which triggers a CRT hotplug/unplug detection cycle independent of the
36815 interrupt enable bit(1 << 9), so keep bit 9.
36816 And although spec says CRT_HOTPLUG_ACTIVATION_PERIOD_64(1 << 8) is only useful
36817 for mobile platform, it is also required to detect vga on G4x platform correctly.
36818 Tested the patch on G45/G43/Q45 platforms with no regressions
36820 It fixed freedesktop.org bug #21120 and part of bug #21210.
36822 Signed-off-by: Ma Ling <ling.ma@intel.com>
36824 commit ec2fde7c8250fdc30984f16c8a1d3587d70b0144
36825 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36826 Date: Tue Jun 2 16:42:56 2009 +0100
36828 Sync DRI2 CopyRegion to vertical retrace
36830 Akin to the Xv code, wait for the scanline to be outside the range to be
36831 copied by the DRI2 CopyRegion hook.
36833 Fixes fdo bug #20664.
36835 Reviewed-by: Eric Anholt <eric@anholt.net>
36836 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
36838 commit 704771f1c595df481d244a6f074e9f5620d4a720
36839 Author: Adam Jackson <ajax@redhat.com>
36840 Date: Thu May 28 15:03:45 2009 -0400
36842 Remove useless loader symbol lists.
36844 commit 8e942b70cb9a784b3f1311affd6fc74c4bcf68bb
36845 Author: Carl Worth <cworth@cworth.org>
36846 Date: Thu May 21 13:12:52 2009 -0700
36848 Revert "Rely on BO pixmaps being present in acceleration paths."
36850 This reverts commit 4653a7db622ad54a3182d93c81331765d930db34.
36852 Eric was getting a little too ambitious about our brave, new world.
36853 We do still want the driver to work with old, non-GEM kernels
36856 commit 1a039f4371bec455cad43f0fb7b329f2ee09a974
36857 Author: Eric Anholt <eric@anholt.net>
36858 Date: Mon Apr 27 17:45:02 2009 -0700
36860 Fold GEM detection into DRM master open.
36862 We don't have anything to do with the DRM unless it's GEM-enabled, unless
36863 we were to support GEM-but-not-DRI2, which doesn't seem useful.
36865 Compilation fixes by Carl Worth <cworth@cworth.org>
36867 commit a04a51c9bb6066454e0fda3c7897f97dab436358
36868 Author: Eric Anholt <eric@anholt.net>
36869 Date: Mon Apr 27 17:29:36 2009 -0700
36871 Open the DRM and keep the handle throughout server startup to finish.
36873 This will let us configure the server from start to finish with the
36874 most pertinent information available (KMS vs UMS, DRI2 vs non-DRI). Also,
36875 we now close the DRI2 fd at terminate, which we didn't before.
36877 This duplicates some code from DRI1 for getting a master FD like I'd done in
36878 DRI2, but given that we weren't loading DRI1 ourselves, this is also a
36879 bogosity cleanup, and avoids allocating the extra DRI1 private.
36881 commit c3bf8b980134a2761701e4bc18235695a1cb07a4
36882 Author: Eric Anholt <eric@anholt.net>
36883 Date: Thu May 21 11:05:01 2009 -0700
36885 Fix backwards logic on whether to sync to vblank or not.
36887 Thanks to Michel Dänzer for catching it.
36889 commit ad2128825ba28551cfef203da017151e2eac32ef
36890 Author: Kristian Høgsberg <krh@redhat.com>
36891 Date: Wed May 20 12:32:10 2009 -0400
36893 Only return FALSE when dri_bo_map() fails
36895 Small typo in the previous commit.
36897 commit 09beee378cecd1079e7a9fa6eee8f084d680d37e
36898 Author: Eric Anholt <eric@anholt.net>
36899 Date: Mon May 18 18:01:05 2009 -0700
36901 Don't do GTT maps on objects bigger than half the available aperture size.
36903 The basic problem is that software fallbacks will do single instructions that
36904 copy from one GTT-mapped BO into another GTT-mapped BO. If we can't get both
36905 of them bound simultanously, we fault one in, retry the instruction, fault the
36906 other in (kicking out #1), retry the instruction, fault #1 back in
36907 (kicking out #2), etc.
36909 Note that we'll still get into a nasty spot if you do a composite operation
36910 with a mask where all 3 are big-but-less-than-half-available-aperture, where
36911 you'll thrash. It at least means you'll make progress, though, since each
36912 instruction will only be operating on two BOs at at time, and the situation
36917 commit 52054b6a4c1ca5117c9750361f71aedd91220c39
36918 Author: Li Shao Hua <shaohua.li@intel.com>
36919 Date: Tue May 19 16:27:32 2009 +0800
36923 commit 34660fd2df5d61b77ed7041d32ac29053fc94f5a
36924 Author: Eric Anholt <eric@anholt.net>
36925 Date: Fri May 15 23:21:05 2009 -0700
36927 Only sync XV to vblank when drawing to the frontbuffer.
36929 This fixes emitting syncs to random pipes with boxes bigger than that
36930 pipe's vertical, leading to GPU hangs.
36934 commit b622860429e00d6ab4407980232659c283a8fe8d
36935 Author: Xiang Hai hao <haihao.xiang@intel.com>
36936 Date: Mon May 18 17:12:30 2009 +0800
36940 commit 50c10ededba15dd0c118f1b65756362061491090
36941 Merge: 13c730e 87332a7
36942 Author: Zou Nan hai <nanhai.zou@intel.com>
36943 Date: Mon May 18 17:11:28 2009 +0800
36945 Merge branch 'master' into xvmc-vld
36947 commit 87332a7cc16af82aa47e07fbf90da3635b071dbf
36948 Author: Keith Packard <keithp@keithp.com>
36949 Date: Mon May 11 13:52:00 2009 -0700
36951 Use drm_intel_bo_disable_reuse API to flag scanout and cursor buffers
36953 Buffers referenced by the kernel for scanout or cursor display should not be
36954 reused by the driver. Use the new drm API to disable reuse of these buffers.
36956 Signed-off-by: Keith Packard <keithp@keithp.com>
36958 commit ebe05200df381c0e6ee636f0f83440bfedea9bcb
36959 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
36960 Date: Fri May 15 08:50:17 2009 -0700
36962 Add --enable-debug flag to configure
36964 Defaults to enabled.
36966 commit 04772b6c09a88f0483c2a7efc48029967c77b9bc
36967 Author: Keith Packard <keithp@keithp.com>
36968 Date: Thu May 14 16:57:11 2009 -0700
36970 If DRM can't figure out which pipe to sync on, then don't sync at all.
36972 Syncing to the wrong pipe can wedge the hardware if the exclusion area is
36973 larger than the vtotal
36975 Signed-off-by: Keith Packard <keithp@keithp.com>
36977 commit 128c1c3b7d57b157604788f82bf9fd389839068f
36978 Author: Carl Worth <cworth@cworth.org>
36979 Date: Wed Apr 29 14:43:56 2009 -0700
36981 Use libdrm to lookup pipe for tear-free sync of XV
36983 Previously, the code was trying to examine a driver_private field,
36984 but those fields are only set by the userland-modesetting code so
36985 would fail in the case of KMS. This fixes bug #21076:
36987 [945GME] [KMS] XV_SYNC_TO_VBLANK does not prevent tearing of xv video
36988 https://bugs.freedesktop.org/show_bug.cgi?id=21076
36990 commit 2572fcc6196aff7a2f1095d211fd85d8668647ca
36991 Author: Wu Fengguang <fengguang.wu@intel.com>
36992 Date: Thu Apr 30 19:51:26 2009 +0800
36994 README: kill an evil dot
36996 To make copy and paste a more pleasure.
36998 Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
36999 Signed-off-by: Eric Anholt <eric@anholt.net>
37001 commit b9462516d18bc57be5f33f57adb6c3e8beede5ff
37002 Author: Eric Anholt <eric@anholt.net>
37003 Date: Tue May 12 18:53:27 2009 -0700
37005 Remove dead "avail" variable from XAA stuff.
37007 commit 1c68bc376a9cb3c0a010c8e28f69a776755c8f64
37008 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
37009 Date: Wed May 13 13:19:53 2009 -0700
37011 Add new have_gem flag
37013 Prior to this patch, code that wanted to check whether GEM was present
37014 would look at pI830->memory_manager. This turned out to be occasionally
37015 problematic in the KMS case, since memory_manager didn't always get set
37016 correctly. So add a new pI830->have_gem flag to make things clear in
37017 the various code paths, and set it after GEM initializes or when KMS is
37020 Reviewed-by: Eric Anholt <eric@anholt.net>
37021 Tested-by: Magnus Kessler <Magnus.Kessler@gmx.net>
37022 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
37024 commit 52367847087206b92f18c40d356d36ab9ee89d39
37025 Author: Keith Packard <keithp@keithp.com>
37026 Date: Thu May 7 14:58:02 2009 -0700
37028 Load i915 and fbcon when checking for KMS
37030 Signed-off-by: Keith Packard <keithp@keithp.com>
37031 Signed-off-by: Eric Anholt <eric@anholt.net>
37033 commit ff7494b4c4b1bd8bb6f169402a9edbe9780787bb
37034 Author: Keith Packard <keithp@keithp.com>
37035 Date: Thu May 7 14:58:01 2009 -0700
37037 Add DP link and GMCH M and N registers
37039 Signed-off-by: Eric Anholt <eric@anholt.net>
37041 commit e54a23bff068416ccbdb75d538dc7dcd40a6c95c
37042 Author: Keith Packard <keithp@keithp.com>
37043 Date: Thu May 7 16:35:19 2009 -0700
37045 Fallback when VT inactive
37047 While the VT is inactive, pI830->batch_bo will be NULL, so use that as a
37048 simple check for when to not use the accelerator. The alternative is to
37049 ignore VT switch and just keep drawing, which would also be fine.
37053 Signed-off-by: Keith Packard <keithp@keithp.com>
37054 [anholt: Removed extra return FALSE -- I830FALLBACK does that.]
37055 Signed-off-by: Eric Anholt <eric@anholt.net>
37057 commit 660bd745c21c56e330d844fe4c8eaae862a77483
37058 Author: Wu Fengguang <fengguang.wu@intel.com>
37059 Date: Thu Apr 30 18:42:18 2009 +0800
37061 Remove the offset parameter from i830_allocator_init()
37063 offset is redundant. i830_allocator_init() is only called
37064 in one place with offset=0.
37066 Acked-by: Magnus Kessler <Magnus.Kessler@gmx.net>
37067 Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
37068 Signed-off-by: Eric Anholt <eric@anholt.net>
37070 commit b1f5cc3349f6811ff4d8dc64cc291788d6726372
37071 Author: Carl Worth <cworth@cworth.org>
37072 Date: Tue May 12 11:45:03 2009 -0700
37074 RELEASING: Better instructions for where to send announcements
37076 The xorg-announce list doesn't need to hear about development
37077 snapshots and release candidates. The intel-gfx list is good for
37080 commit f16ee218845ec48940ea457b921d34896d80a807
37081 Author: Alan Coopersmith <alan.coopersmith@sun.com>
37082 Date: Sun May 10 16:25:24 2009 -0700
37084 Fix "Unkown" typo in two FatalError messages
37086 Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
37088 commit 8d27247829fe4f55691ce68f9f4b14810fb34b32
37089 Author: Stijn van Drongelen <tinctorius@gmail.com>
37090 Date: Tue May 5 16:37:06 2009 +0200
37092 Fix typo in (unused) INTEL_BIOS_32 macro
37094 Debian bug#527062 <http://bugs.debian.org/527062>
37096 [jcristau: fix same typo in bios_reader.c]
37097 Signed-off-by: Julien Cristau <jcristau@debian.org>
37099 commit a8a771a853478e5f45f71d0eff3c4d55bf24d0ad
37100 Author: Arkadiusz Miśkiewicz <arekm@maven.pl>
37101 Date: Wed Apr 29 00:21:18 2009 +0200
37103 Restore CFLAGS after tests.
37105 Reverts single line of broken commit e5133a94dc78411eaef324131ea8056aeb81779e.
37107 Signed-off-by: Eric Anholt <eric@anholt.net>
37109 commit 8255cca2c9092f7ecb798944aa8f03fa3efcfa6c
37110 Author: Keith Packard <keithp@keithp.com>
37111 Date: Fri May 1 18:34:43 2009 -0700
37113 Split i915 textured video commands to fit into batch buffers.
37115 i915 textured video commands are quite long, but must be contained in the
37116 same batch buffer as the 3D setup commands. When the number of clip rects
37117 for the video becomes too large for the associated commands to fit in the
37118 same batch buffer, this change breaks the sequence into pieces, ensuring
37119 that each batch contains the necessary setup sequence.
37121 Signed-off-by: Keith Packard <keithp@keithp.com>
37123 commit e5e0fb846bda3e1757b89e50d5244d28457b9fe3
37124 Author: Keith Packard <keithp@keithp.com>
37125 Date: Fri May 1 12:26:04 2009 -0700
37127 Call down to lower CloseScreen before shutting down DRM allocator
37129 Lower level functions will destroy objects that are managed by the DRM
37130 allocator, so make sure those are done before the allocator shuts down.
37132 Signed-off-by: Keith Packard <keithp@keithp.com>
37134 commit cfb98f3da936907945f48dad4c2345930f7beb60
37135 Author: Keith Packard <keithp@keithp.com>
37136 Date: Fri May 1 12:23:32 2009 -0700
37138 Revert "Leave allocator running until lower-level CloseScreens are done"
37140 Uh, oops -- the FD used to talk to the kernel is closed before the
37141 allocator_fini function is called now.
37143 This reverts commit 1872869e6ffcc8e6cab820c508fe5404d7e8ff9c.
37145 commit 11a853bd8e5d907fe7f5bd907453bcdac9032861
37146 Author: Keith Packard <keithp@keithp.com>
37147 Date: Fri May 1 11:51:13 2009 -0700
37149 Hold reference to video binding table until all rects are painted.
37151 The optimization of unreferencing the binding table when the relocation is
37152 posted causes the object to be dereferenced for each box in the clip list,
37153 causing general chaos in the buffer manager. It's easier to just hold a
37154 reference to the object until all of the boxes are painted and then drop it.
37156 Signed-off-by: Keith Packard <keithp@keithp.com>
37158 commit ed492131c13715b73c14d328d0668120acb58b40
37159 Author: Keith Packard <keithp@keithp.com>
37160 Date: Fri May 1 11:50:17 2009 -0700
37162 3D_STATE_VERTEX_BUFFERS takes four 32-bit values, not three.
37164 The spec says this command takes an extra (mbz) 32-bit value, so let's
37165 provide it with one.
37167 Signed-off-by: Keith Packard <keithp@keithp.com>
37169 commit 66d1536a2ecc7a3d44da4c0dbe6c85aaa4e05791
37170 Author: Keith Packard <keithp@keithp.com>
37171 Date: Fri May 1 11:48:51 2009 -0700
37173 Don't bother to enable VF statistics during 965 video playback
37175 This was used while bringing up the driver to debug vertext fetches.
37177 Signed-off-by: Keith Packard <keithp@keithp.com>
37179 commit 1872869e6ffcc8e6cab820c508fe5404d7e8ff9c
37180 Author: Keith Packard <keithp@keithp.com>
37181 Date: Fri May 1 11:46:51 2009 -0700
37183 Leave allocator running until lower-level CloseScreens are done
37185 The lower level close screen functions will free allocated objects, causing
37186 a crash if the allocator isn't still available.
37188 Signed-off-by: Keith Packard <keithp@keithp.com>
37190 commit 1142353b487c155a31011923fbd08ec67e60f505
37191 Author: Keith Packard <keithp@keithp.com>
37192 Date: Fri May 1 11:44:13 2009 -0700
37194 intel_batch_start_atomic: fix size passed to intel_batch_require_space (*4)
37196 intel_batch_start_atomic takes an argument in 32-bit units, and so it must
37197 multiply that by 4 before passing it to intel_batch_require_space, which
37198 takes an argument in bytes.
37200 We should figure out what units we want to use and use the same everywhere...
37202 Signed-off-by: Keith Packard <keithp@keithp.com>
37204 commit 8232177a005127a0001af4b14c3beb766958ef11
37205 Author: Keith Packard <keithp@keithp.com>
37206 Date: Thu Apr 30 15:30:05 2009 -0700
37208 Allow cursors to roam past 2048x2048 limit.
37210 There's no reason to clip cursor positions to an artificial limit; the
37211 hardware cursor limits always mirror the hardware display limits.
37213 Signed-off-by: Keith Packard <keithp@keithp.com>
37215 commit 417f3784b7fae8de3559c7607a2de60661a6a448
37216 Author: Carl Worth <cworth@cworth.org>
37217 Date: Tue Apr 28 20:42:22 2009 -0700
37219 Add doltlibtool to DISTCLEANFILES
37221 Otherwise make distcheck fails.
37223 commit bc9a688476f02233450eb05dfb40acbf337d679b
37224 Author: Carl Worth <cworth@cworth.org>
37225 Date: Tue Apr 28 20:37:13 2009 -0700
37227 NEWS: Add notes for 2.7.99.1 snapshot
37229 commit a90cab43209bba52d35b953f3d686a19c154fc43
37230 Author: Kristian Høgsberg <krh@redhat.com>
37231 Date: Tue Apr 28 13:12:55 2009 -0400
37233 Remove a handful of pointless comments
37235 CVS keywords, comments about how the source was once reformatted,
37236 and the ad-hoc changelog comment in in i830_driver.c
37238 commit c231b41f3f3eb3ba9c8436d5fd93a30704ed437c
37239 Author: Eric Anholt <eric@anholt.net>
37240 Date: Tue Apr 21 15:55:01 2009 -0700
37242 Remove checks for xf86RotateFreeShadow, introduced in server 1.4.
37244 commit e5133a94dc78411eaef324131ea8056aeb81779e
37245 Author: Eric Anholt <eric@anholt.net>
37246 Date: Tue Apr 21 15:49:01 2009 -0700
37248 unifdef LIBPCIACCESS and XSERVER_LIBPCIACCESS.
37250 Depending on new server means these are always present.
37252 commit 4653a7db622ad54a3182d93c81331765d930db34
37253 Author: Eric Anholt <eric@anholt.net>
37254 Date: Mon Apr 20 22:05:54 2009 -0700
37256 Rely on BO pixmaps being present in acceleration paths.
37258 commit 1784efc1f0c74edda8c78fb1b0cdd989f0a088ad
37259 Author: Eric Anholt <eric@anholt.net>
37260 Date: Mon Apr 20 18:25:59 2009 -0700
37262 Remove XF86DRM_MODE.
37264 We now depend on server 1.6 and current libdrm, so no need to ifdef it.
37266 commit d82ea4372fb74654eb0f37d996977003473846ed
37267 Author: Eric Anholt <eric@anholt.net>
37268 Date: Mon Apr 20 13:20:03 2009 -0700
37270 Replace I830Sync's irq emit/wait code with bufmgr use.
37272 This pre-GEM code was all sorts of broken -- see intel_bufmgr_fake.c for
37273 the hoops that must be jumped to use that kernel interface successfully.
37274 Yet we continued to use it even with KMS/DRI2/UXA, which may account for
37277 commit 612c1f1f1859ce0cef41049d79cc473657d48612
37278 Author: Eric Anholt <eric@anholt.net>
37279 Date: Mon Apr 20 16:16:48 2009 -0700
37281 Remove I830_XV ifdef.
37283 The developers of the driver don't ever run or test without XV. Don't do it.
37285 commit af6892f2b873dd508994d8bbf3d67e719b535e35
37286 Author: Eric Anholt <eric@anholt.net>
37287 Date: Mon Apr 20 16:07:05 2009 -0700
37289 Replace i830WaitSync with just I830Sync(), as nearly all callers had it wrong.
37291 People were trying to BEGIN_BATCH()/ADVANCE_BATCH() then i830WaitSync on the
37292 results, which wouldn't necessarily wait and lead to various painful bugs,
37293 since only EXA called MarkSync and only for certain rendering operations.
37295 commit 393abc15b76804580bfee0760ddbd1fb77e66c02
37296 Author: Eric Anholt <eric@anholt.net>
37297 Date: Mon Apr 20 15:58:35 2009 -0700
37299 Now that video destination pixmaps are always in BOs, no more MarkSync.
37301 commit dc71573292c215f49716f4dc1ebc416c6b172995
37302 Author: Eric Anholt <eric@anholt.net>
37303 Date: Fri Apr 17 15:13:34 2009 -0700
37305 Remove can_resize, which is now always true (except when a bug interfered)
37307 commit 96df22740b2cdc4f82a4a36c0f77663105d9337e
37308 Author: Eric Anholt <eric@anholt.net>
37309 Date: Fri Apr 17 15:11:12 2009 -0700
37311 unifdef I830_USE_UXA.
37313 We only have on acceleration architecture now, and you can always build it
37314 if you're building the driver.
37316 commit fa81ae1637bfd7b73b43355b78c88a9b84c0393c
37317 Author: Eric Anholt <eric@anholt.net>
37318 Date: Fri Apr 17 14:01:47 2009 -0700
37320 Remove EXA support.
37322 UXA has completely replaced EXA at this point. UXA is the same rendering
37323 core as EXA, but relying on kernel memory management or a fake bufmgr instead
37324 of trying to manage memory in the X Server.
37326 commit 1b10745a2528622a32271f64c35fcdb7b7154d11
37327 Author: Eric Anholt <eric@anholt.net>
37328 Date: Fri Jan 23 08:39:59 2009 -0800
37330 Remove XAA support.
37332 While EXA/UXA aren't completely good replacements (see bugzilla for
37333 performance and stability problems), we are pretty sure at this point that
37334 it's the right way to go and that having multiple acceleration architectures
37335 is getting in the way of producing a stable codebase.
37337 commit 30c226eeb35f034c3ddb881e77454051dc210298
37338 Author: Eric Anholt <eric@anholt.net>
37339 Date: Thu Jan 22 17:54:49 2009 -0800
37341 Remove DRI1 support.
37343 This was blocked on wide distribution of X Server 1.6 (now in the current or
37344 next version of major distributions) and solutions for a couple of significant
37345 architectural problems (vblank sync and frontbuffer rendering, which we now
37346 have code or good plans for).
37348 This includes disabling XVMC which is DRI1-only currently.
37350 commit 47591334a183881704a121ae06ebc2fadebe6f73
37351 Author: Eric Anholt <eric@anholt.net>
37352 Date: Fri Apr 17 13:54:47 2009 -0700
37354 Remove pre-server-1.5 support.
37356 commit 3187c7698b93f1d0e07798e9c42ca2877d55e055
37357 Author: Eric Anholt <eric@anholt.net>
37358 Date: Mon Apr 27 14:50:54 2009 -0700
37360 Fix assert at startup with DRI initialiation failure.
37362 In this path, we'd make it to allocator_init -> init_bufmgr without
37363 GEM and without FbBase being initialized, leading to assertion failure
37364 to catch this mistake.
37366 Comedy ensued when trying to move just the MapMem up, leading to the rest
37367 of the commit. Some day (like tomorrow after I rebase intel-cleanup) I'll
37368 clean this path up.
37370 Tested with 2 X Servers on 2.6.28 (one gets DRI2, one fails successfully),
37371 2 UMS X Servers on 2.6.30rc2 (each gets DRI2), and 2 KMS X Servers on
37372 2.6.30rc2 (success all around).
37374 commit fbaf13c93511547c563129527797fbef8628652e
37375 Author: Eric Anholt <eric@anholt.net>
37376 Date: Mon Apr 27 16:05:50 2009 -0700
37378 Don't clear the framebuffer if it isn't currently bound.
37380 This segfault can be triggered in non-KMS memory init before the EnterVT,
37381 and will happen anyway at EnterVT time.
37383 commit ab878118b9fb46c437a02f824d29f4a206e7fd80
37384 Author: Eric Anholt <eric@anholt.net>
37385 Date: Mon Apr 27 16:00:05 2009 -0700
37387 Don't initialize DRI2 if the fd we get is not master-capable.
37389 commit 385ac2197b64574f30620a203e675d8da0e1472f
37390 Author: Eric Anholt <eric@anholt.net>
37391 Date: Mon Apr 27 15:28:21 2009 -0700
37393 Fix drmSetMaster/DropMaster error messages.
37395 commit 5b05a589efb23b2fc09b06e4271174d922b1ab02
37396 Author: Pierre Willenbrock <pierre@pirsoft.de>
37397 Date: Sat Apr 25 22:58:20 2009 +0200
37399 format == 0 means "use default" in I830DRI2CreateBuffer
37401 Reviewed-by: Ian Romanick <idr@freedesktop.org>
37403 commit e55d943126cdd3eac7dfec5f40e794f89dbf038b
37404 Author: Eric Anholt <eric@anholt.net>
37405 Date: Mon Apr 27 11:18:07 2009 -0700
37407 Don't try to do anything for I830Sync when VT switched.
37409 Fixes a crash with fake bufmgr and UXA.
37411 commit 9748aa5594a8357754c06dd82108b131260b1a71
37412 Author: Shuang He <shuang.he@intel.com>
37413 Date: Sat Apr 18 11:48:08 2009 +0800
37415 Free front buffer bo when X exit
37417 It's needed when KMS or DRI2 is enabled, or there will be memory leak.
37418 Also fixes a segfault at startup with fake bufmgr.
37420 Signed-off-by: Eric Anholt <eric@anholt.net>
37422 commit b8ca146b060e6f76bf048f1f6cf6ee89d38d015e
37423 Author: Alan Coopersmith <alan.coopersmith@sun.com>
37424 Date: Fri Apr 24 16:03:27 2009 -0700
37426 Fix UXA to build with Sun compilers (use __func__ instead of __FUNCTION__)
37428 Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
37430 commit 5d6f4f6eb7a4dcbe1ce5a134d882e56f958ed2ba
37431 Author: Ian Romanick <ian.d.romanick@intel.com>
37432 Date: Fri Apr 24 12:28:13 2009 -0700
37434 DRI2: If the SDK supports it, use the DRI2GetBuffersWithFormat interfaces
37436 If DRI2INFOREC_VERSION is defined in the server's dri2.h and has a
37437 value greater than 1, compile to use the CreateBuffer and
37438 DestroyBuffer interfaces. The format parameter parameter to
37439 CreateBuffer is assumed to be the bits-per-pixel of the buffer.
37441 Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
37442 Reviewed-by: Kristian Høgsberg <krh@redhat.com>
37444 commit 106e4b44c5af6552cbd079c4ec34def9dcfb168a
37445 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
37446 Date: Fri Apr 24 10:55:09 2009 +0800
37448 SDVO: switch bus once for i2c transfer
37450 Before sdvo try to get edid by i2c bus, it must do switch control bus to ddc active state from sdvo only state.
37451 However if current state has been ddc active state, redundant switch operation in our driver will cause error-
37452 "Unable to write to SDVOCTRL_E for SDVOB Slave 0x70". The patch will do switch control bus only one time during
37453 whole edid transmission.
37455 It has fixed bug #19937
37457 Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
37458 Tested-by: Ma Ling <ling.ma@intel.com>
37459 Signed-off-by: Ma Ling <ling.ma@intel.com>
37461 commit 8f64837e56b2de0fb8a9100d1a844fd3f18d751c
37462 Author: Shaohua Li <shaohua.li@intel.com>
37463 Date: Mon Feb 23 15:19:25 2009 +0800
37465 Disable FBC on IGD for UMS
37467 It appears the new chip doesn't support FBC currently.
37469 Signed-off-by: Shaohua Li<shaohua.li@intel.com>
37470 Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
37472 commit 7b01aa5cc41620da5bb48f391ff98d9e82572e52
37473 Author: Eric Anholt <eric@anholt.net>
37474 Date: Wed Apr 22 09:03:21 2009 -0700
37476 Revert "fix overflow warning on videoRam"
37478 This reverts commit 08ebde4715b87867184d42b60762cd774e151f5c, which was
37479 apparently untested and broke KMS.
37481 commit ceb3a2eaf9090d67e4dfcbed188125108ceab78d
37482 Author: Zdenek Kabelac <zdenek.kabelac@gmail.com>
37483 Date: Tue Apr 21 18:58:23 2009 -0700
37485 Remember allocated offscreenImages memory for freeing it in CloseScreen()
37487 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
37488 [anholt: renamed the member to match the variable name]
37489 Signed-off-by: Eric Anholt <eric@anholt.net>
37491 commit 701c6bef14b5d464847cc42c5ccb66fdeb157f1a
37492 Author: Zdenek Kabelac <zdenek.kabelac@gmail.com>
37493 Date: Tue Apr 21 18:53:07 2009 -0700
37495 free allocated name inside sdvo_destroy
37497 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
37498 Signed-off-by: Eric Anholt <eric@anholt.net>
37500 commit fe71217ca86eb7ae8081bf3e26d30f2002d27691
37501 Author: Zdenek Kabelac <zdenek.kabelac@gmail.com>
37502 Date: Tue Apr 21 18:51:02 2009 -0700
37504 Unreference allocated bos in i965 render error paths
37506 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
37507 Signed-off-by: Eric Anholt <eric@anholt.net>
37509 commit b8b6ca830b1fd6956c76a5417462ca55174cb881
37510 Author: Zdenek Kabelac <zdenek.kabelac@gmail.com>
37511 Date: Tue Apr 21 18:47:03 2009 -0700
37513 Fix leak of some 965 render state on VT switch.
37515 Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
37516 [anholt: fixed up unneeded != NULL checks]
37517 Signed-off-by: Eric Anholt <eric@anholt.net>
37519 commit d7ca870e1ce251d42e3689a8e1e7d080ab1325fb
37520 Author: Albert Damen <albrt@gmx.net>
37521 Date: Sun Apr 5 16:36:35 2009 +0200
37523 Fix crash with XV with large virtual display
37525 If a virtual display with width > 2048 is used, the first time
37526 an XV buffer is needed will result in a BadAlloc error message,
37527 but the next time X would crash.
37529 Signed-off-by: Eric Anholt <eric@anholt.net>
37531 commit 04ed92e3db8d9fcf2a1089ed433ab67bf0969655
37532 Author: Eric Anholt <eric@anholt.net>
37533 Date: Tue Apr 21 15:17:31 2009 -0700
37535 Remove dead xoffset/yoffset from pre-randr.
37537 commit ab2c988767aec854d3c1702bee1e5a06bd00870c
37538 Author: Eric Anholt <eric@anholt.net>
37539 Date: Tue Apr 21 14:53:45 2009 -0700
37541 Remove dead monitor detect debugger.
37543 watch xrandr works fine.
37545 commit 928a37041defcca6f57f9452ba62e67524cb4510
37546 Author: Eric Anholt <eric@anholt.net>
37547 Date: Mon Apr 20 15:46:11 2009 -0700
37549 Replace a bunch of #ifdef debug flushing/syncing with a single function.
37551 This removes it from a callsite where it would have just resulted in a
37554 commit 1fc93ee184ceefaea9528bb46ae82884c44d9b36
37555 Author: Eric Anholt <eric@anholt.net>
37556 Date: Mon Apr 20 14:42:31 2009 -0700
37558 Staticize a bunch of functions and variables in the driver.
37560 This cleans up findstatic.pl output for the i830+ code, which resulted in
37561 removing some code. The only odd part of this commit is the
37562 if (0) i830_sdvo_dump() in i830_sdvo.c -- it tells the compiler that the code
37563 is used, without using it since we want the code around while debugging.
37564 It's also in a likely place to ask for the dump, so I think it's OK.
37566 commit b2d7928f195f6439ecd140e6280b958b8a1f534b
37567 Author: Eric Anholt <eric@anholt.net>
37568 Date: Mon Apr 20 14:48:59 2009 -0700
37570 Use a static inline to replace if (0) to an unused stub function.
37572 commit 093f51646b2f2301c907e14d0ad4b1f80ff60a16
37573 Author: Eric Anholt <eric@anholt.net>
37574 Date: Mon Apr 20 14:36:52 2009 -0700
37576 Remove unused i830_output_type_names
37578 commit c774ef282bfb0c594479bbab3527fe5fda2a8559
37579 Author: Eric Anholt <eric@anholt.net>
37580 Date: Mon Apr 20 14:29:17 2009 -0700
37582 Remove dead mono cursor load code.
37584 commit 75c531b9652b3919ddbf51f8eec804e6bdbb2331
37585 Author: Eric Anholt <eric@anholt.net>
37586 Date: Fri Apr 17 14:03:39 2009 -0700
37588 Rename EXA rendering functions to UXA, since we're keeping them post-EXA.
37590 commit 40dbba5cf710a6e392afca73965aa48fa3953d1f
37591 Author: Eric Anholt <eric@anholt.net>
37592 Date: Fri Apr 17 15:15:27 2009 -0700
37594 Remove some dead i830.h struct members.
37596 commit 84f69081abaeef8b05cafb64d3102eb2abdf9a8e
37597 Author: Julien Cristau <jcristau@debian.org>
37598 Date: Thu Apr 16 14:27:15 2009 +0200
37600 Require xserver 1.3
37602 I don't think anyone tests this against an old server anymore.
37604 Signed-off-by: Eric Anholt <eric@anholt.net>
37606 commit 08ebde4715b87867184d42b60762cd774e151f5c
37607 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
37608 Date: Mon Apr 20 09:43:09 2009 +0800
37610 fix overflow warning on videoRam
37612 Note that pScrn->videoRam is an 'int'.
37614 i830_driver.c: In function ‘I830ScreenInit’:
37615 i830_driver.c:3019: warning: overflow in implicit constant conversion
37617 Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
37619 commit 5a07ab502fe1e58e7e37fe554fb42d8d2c8c53ec
37620 Author: Ian Romanick <ian.d.romanick@intel.com>
37621 Date: Fri Apr 17 20:59:04 2009 -0700
37623 DRI2: Respect the src and dst parameters of CopyRegion.
37625 Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
37627 commit 2e3b95ed0197971e81ab7509245c899e96859d5b
37628 Author: Carl Worth <cworth@cworth.org>
37629 Date: Wed Apr 15 18:54:19 2009 -0700
37631 Increment version to 2.7.99.1
37633 This is the number we want for any development snapshots along
37634 what will eventually become 2.8.0. (Once we get to 'release
37635 candidates' we'll bump this up to 2.7.99.901.)
37637 commit c9e174b31ae3b9a138ae696d563420a841062f57
37638 Author: Carl Worth <cworth@cworth.org>
37639 Date: Wed Apr 15 18:52:38 2009 -0700
37641 RELEASING: Note that --with-xserver-source is needed for make distcheck
37643 I was always forgetting this without this reminder.
37645 commit 73c3be1aa033e8c5c7ee777eb2fd43c19668fa86
37646 Author: Carl Worth <cworth@cworth.org>
37647 Date: Wed Apr 15 18:18:14 2009 -0700
37649 README: Fix typos in chipset list, and point to how_to_report_bug web page
37651 Thanks to Gordon Jin for these suggestions.
37653 commit c51dddb724a79a75491369a4c3e8b7b26231e7ac
37654 Author: Carl Worth <cworth@cworth.org>
37655 Date: Wed Apr 15 18:07:17 2009 -0700
37657 AUTHORS: Add Robert Lowery to the authors file
37659 Rob got missed from my first scan since one commit lists his name as
37660 just 'Rob' and 3 commits don't attribute him as author:
37662 83d304c61ad5fdc58b0a9309dbd1e5a3f6cd9b01
37663 7552d80e367fe38bbc594fe94abd649917fe54d5
37664 6eecef4fed8a21dfdabef42eb69fd150b96167b2
37666 commit 4b5edde5da4b3e955eb2d77004de81e47bec7f69
37667 Author: Robert Lowery <rlowery@exemail.com.au>
37668 Date: Wed Apr 15 18:03:31 2009 -0700
37670 Fix typo in comment
37672 Thanks to Robert Lowery for the sharp eyes on this one.
37674 commit 3fd5a1ecd1d5140ae07ccc279298bcadd515e97f
37675 Author: Carl Worth <cworth@cworth.org>
37676 Date: Wed Apr 15 16:44:11 2009 -0700
37678 RELEASING: Update instructions to reflect some minor process improvements
37680 We've added a NEWS file now, so that needs to be updated for each release.
37682 We're also now using tag names of just <version> rather than
37683 xf86-video-intel-<version>.
37685 commit e1cace16a6130dcdd93965d2329a349d49200fa6
37686 Author: Carl Worth <cworth@cworth.org>
37687 Date: Wed Apr 15 16:33:12 2009 -0700
37689 NEWS: Add note about broken PAT code in some kernels
37691 Hoping to cut off some false bug reports here.
37693 commit 9ffd1951d1f2fd2f53273d04ea29de050f07af55
37694 Author: Carl Worth <cworth@cworth.org>
37695 Date: Wed Apr 15 16:14:44 2009 -0700
37697 Add AUTHORS and NEWS to EXTRA_DIST
37699 These new files don't do us much good if we don't distribute them in
37702 commit e4cd9de2933ada3e2a4b43552729ae3a370128bf
37703 Author: Carl Worth <cworth@cworth.org>
37704 Date: Wed Apr 15 16:14:03 2009 -0700
37706 Add a NEWS files with release-notes for 2.7.0
37708 It will be nice to have release-notes under revision control, as well
37709 being able to document issues in an obviously time-sensitive file,
37710 (as opposed to README where we were documenting some of this previously).
37712 commit 506c810f8f3db89048dda9777902f142ffeb86aa
37713 Author: Carl Worth <cworth@cworth.org>
37714 Date: Wed Apr 15 16:10:52 2009 -0700
37716 Clarify that the default acceleration is UXA if KMS is available.
37718 Stale documentation considered harmful of course.
37720 commit b9716b836cb2b4569c90b81f344932ac668dc5bf
37721 Author: Carl Worth <cworth@cworth.org>
37722 Date: Wed Apr 15 15:39:06 2009 -0700
37724 Add a new AUTHORS file
37726 This is a sorted list of everyone with more than 2 commits in the git
37727 revision history. We also list some historical authors mentioned in the
37728 man page, (with code presumably pre-dating the beginning of revision
37731 commit 8deb3a3709a9aaa549be404566715a01246354d9
37732 Author: Carl Worth <cworth@cworth.org>
37733 Date: Wed Apr 15 15:38:11 2009 -0700
37735 README: Remove almost all time-sensitive information
37737 This was all very stale, and is better convered in intel.man. We replace
37738 this with a list of pointers to where to get current information, (man
37739 page, web site, and mailing list).
37741 commit 9b615a52671aacf34666f90ecfff98651ce6afe2
37742 Author: Li Peng <peng.li@intel.com>
37743 Date: Fri Apr 10 14:39:35 2009 +0800
37745 Turn on front buffer tiling in KMS.
37747 This code disabled front buffer tiling in KMS. Turn it on since kernel
37748 handles all tiling now, this improves performance of x11perf -aa10text
37749 from 97k to 286k on my 945GME.
37751 Should help with #20761, if not totally fix it.
37753 Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
37754 Signed-off-by: Li Peng <peng.li@intel.com>
37756 commit 053432991c812146f6e7c6f13c6ace55385c825f
37757 Author: Ma Ling <ling.ma@intel.com>
37758 Date: Mon Apr 13 14:27:35 2009 +0800
37760 update manpage for BROADCAST_RGB property
37762 commit 62ba7211fe9b6aada125ebfe34cf7161e817ad6b
37763 Author: Ma Ling <ling.ma@intel.com>
37764 Date: Mon Apr 13 14:24:57 2009 +0800
37766 set broadcast RGB mode for integrated HDMI output.
37768 commit 69388953ce889080d5f014123d89bf3eb45f3d8d
37769 Author: Ma Ling <ling.ma@intel.com>
37770 Date: Mon Apr 13 14:23:06 2009 +0800
37772 set broadcast RGB mode for HDMI and TMDS from SDVOX output
37774 Almost all digital TVs accept broadcast RGB values from 16 to 235 for each channel,
37775 otherwise for those uncompensated videos, when RGB values are set from 0 to 255,
37776 they will shows black and whiter clamping, which seriously degrades picture quality.
37777 The patch will enable the broadcast RGB mode for hdtv according to user's setting.
37778 It fixed bug #14486
37780 commit 6d345c49f693cc5cffaa00b94559d2afcb3a0864
37781 Author: Carl Worth <cworth@cworth.org>
37782 Date: Fri Apr 10 14:07:14 2009 -0700
37784 Add a RELEASING file documenting the release process
37786 Thanks to Jesse Barnes for the original recipe.
37788 commit 7e516b6d24d8c0c6549a9a60fcf487e3a1615020
37789 Author: Jesse Barnes <jbarnes@jbarnes-acer.(none)>
37790 Date: Wed Apr 8 16:38:08 2009 -0700
37792 Silence warning in i830_dmi_store_field
37794 Just add a dummy ret variable to shut up gcc.
37796 commit 620e97bbd6a811ad69b8ac94df1fe2c9edf65549
37797 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
37798 Date: Wed Apr 8 15:49:00 2009 -0700
37800 Don't enable kernel execbuf fencing w/EXA
37802 If we enable kernel execbuf fence register management, it's best if the
37803 kernel manages all fence registers. This works fine if the accel
37804 method is managing pixmaps or doesn't use offscreen pixmaps. However
37805 with EXA, pixmap accesses are done relative to the framebuffer BAR
37806 mapping (pI830->FbBase) and the Screen pixmap address. So if we try to
37807 set the screen pixmap to point at a GTT mapped (and therefore properly
37808 fenced) address, later calls to intel_get_pixmap_offset() will call
37809 into EXA, which will use the pseudo-random pixmap addr and the EXA
37810 offscreen base addr (which is really just FbBase) to calculate the
37811 offset. This will fail. So disable kernel fence reg management in the
37812 EXA case (this is easier than adding proper EXA pixmap management to
37813 xf86-video-intel, and makes more sense since we'll be removing EXA soon
37818 Also happens to fix FDO #21029 (as tested by Carl Worth <cworth@cworth.org).
37820 commit 0a0731c11d10392cdc55ecc04e4e3575c8b3fe57
37821 Author: Shuang He <shuang.he@intel.com>
37822 Date: Tue Apr 7 12:31:07 2009 -0700
37824 Fix value for MI_WAIT_FOR_PIPEA_SCAN_LINE_WINDOW
37826 Since the change to scan-line based video sync, (rather than vblank-
37827 based), we've only been getting tear-free video on one of the two
37828 pipes. This fixes that bug by using the correct constant for waiting
37831 commit 940c2aad4d174b6609bdc49f8c99a4bc37926516
37832 Author: Carl Worth <cworth@cworth.org>
37833 Date: Mon Apr 6 14:36:33 2009 -0700
37835 Don't clip video to CRTC in the case of textured video
37837 Since we're not limited by a single overlay plane on a single pipe,
37838 we want to not clip at all, (so that the correct video appears on
37841 This fixes bug #20980 which shows a video spanning two pipes
37842 being rendered incorrectly.
37844 commit 63b4b5efac936c674dedad8125a8dbac4f000908
37845 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
37846 Date: Tue Apr 7 10:53:08 2009 +0800
37848 quirk LVDS on ibase MB890 855GM board
37852 commit 5d9d9a2e466474a0508a15b294a91507ccb3ccc1
37853 Author: Carl Worth <cworth@cworth.org>
37854 Date: Mon Apr 6 14:02:08 2009 -0700
37856 Fix new video sync-to-blank code for multi-head
37858 We need to account for a non-zero Y offset for the CRTC. Without
37859 this, we don't sync to the correct region, so tearing becomes
37862 commit 3d4ee3cac1d63dfdf7b54c8ba577f3b77637499f
37863 Author: Carl Worth <cworth@cworth.org>
37864 Date: Mon Apr 6 11:31:20 2009 -0700
37866 Remove support for 'auto'(-1) value of XV_SYNC_TO_VBLANK
37868 We previously had a heurstic here where we would only sync to vblank
37869 for windows that covered more than 25% of the screen. We don't need
37870 this anymore since the new approach to sync, (WAIT_FOR_SCANLINE_WINDOW),
37871 is not excessively costly for small windows.
37873 commit bc3312fd7c03d09a231dfebfe390fe668ad15d1e
37874 Author: Carl Worth <cworth@cworth.org>
37875 Date: Mon Apr 6 11:16:40 2009 -0700
37877 Use WAIT_FOR_SCAN_LINE instead of WAIT_FOR_VBLANK
37879 Either way, the goal is tear-free video playing. But waiting for
37880 a scan-line window not only has the advantage of being cheaper
37881 for small windows, but also avoids hanging the GPU in the case
37882 of the pipe getting turned off, (by screensaver, for example),
37883 while a batch is waiting for a VBLANK that will never occur.
37885 This fixes that GPU hang.
37887 commit 6cd914ef315036ce8e91c7b6492994353e8ed2d8
37888 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
37889 Date: Fri Apr 3 11:15:18 2009 -0700
37891 Fix offset in begin_gtt_access case
37893 Don't use bo->virtual in the begin_gtt_access case, use the framebuffer
37894 mapping and bo offset instead.
37896 Reviewed-by: Eric Anholt <eric@anholt.net>
37897 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
37899 commit 13c730e003e805e19deee0996b2af30f69e54c4d
37900 Author: Zou Nan hai <nanhai.zou@intel.com>
37901 Date: Wed Apr 1 15:12:55 2009 +0800
37904 1.multiple instance
37907 commit fad714c40078d22fff82dc0692a344f66ddf9680
37908 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
37909 Date: Wed Apr 1 10:11:35 2009 +0800
37911 SDVO: fix output flag dumping for unknown type
37913 Found by Hugo Jacques <hugo.jacques@verint.com>
37915 commit 00de1757dd5776962bdd4c8968181878c2ebf4c9
37916 Author: Hugo Jacques <hugo.jacques@verint.com>
37917 Date: Wed Apr 1 10:10:05 2009 +0800
37919 SDVO: add composite TV out support
37921 commit 087f72e1f5d7d11b8795ba80a842874f5a9bb01d
37922 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
37923 Date: Tue Mar 31 07:50:10 2009 -0700
37925 Match GTT unmap with map in KMS rotation case
37927 Missed this when the GTT unmap call was added. If we don't do this we
37928 trigger an assertion in libdrm, since the buffer has never been mapped
37933 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
37935 commit 4f046af760b92c07f59664359453933fb5358e3d
37936 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
37937 Date: Tue Mar 31 13:49:44 2009 +0800
37939 Disable LVDS detect methods
37941 Both methods ACPI lid and SWF bit have issues in LVDS detect from
37942 wider testing. Fallback to origin code.
37944 commit e964d4e53af3a47de6d09c884be1cc0044d03bea
37945 Author: Albert Damen <albrt@gmx.net>
37946 Date: Mon Mar 30 11:38:02 2009 -0700
37948 Non-GEM allocations incorrectly force TILE_NONE (bug 20797)
37950 With -intel 2.6.3 performance was very bad when using a non gem enabled kernel
37951 (2.6.27) and EXA. For example sauerbraten ran with 4 fps and screensaver GLBlur
37952 with 1 fps. With -intel 2.6.1 performance was good using the same kernel.
37954 Git bisecting led me to commit f1ed73c1ef3e3daa9f695194dcc813167cbcb53d (in 2.6
37955 branch) "Make i830_allocate_memory take tiling parameters" as first bad commit.
37957 Using gdb I found tiling was set exactly the same in 2.6.3 as in 2.6.1, so that
37958 was good (TILE_XMAJOR for front, back and depth buffers).
37959 Looking further I found the line mem->tiling = TILE_NONE; (line 961 in
37960 src/i830_memory.c) at the end of i830_allocate_memory suspicious, as
37961 mem->tiling now already gets set via i830_allocate_aperture and some buffers do
37962 have tiling. Removing that line indeed fixed the performance issue. Now
37963 sauerbraten runs with 30+ fps and GLBlur runs smoothly.
37965 commit 51cf8a453c2bc2e8604bfc41a649e971c1ba5026
37966 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
37967 Date: Mon Mar 30 11:27:56 2009 -0700
37969 Require libdrm 2.4.6 for GTT unmap support
37971 Need the new functions available.
37973 commit 8dabcc40747bfd478f296728741240241698f165
37974 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
37975 Date: Mon Mar 30 09:53:40 2009 -0700
37977 Tiling fixes, third set
37979 Hopefully this concludes the fixes necessary to deal with the various
37980 combinations of kernel and user level tiling. We have several cases to
37982 1) KMS (kernel handles all tiling)
37983 2) UMS w/memory management + kexec fencing (kernel handles all tiling)
37984 3) UMS w/memory mangement but no kexec fencing (userland handles tiling)
37985 4) UMS w/o memory management (userland handles tiling)
37987 For cases (1) & (2) we can use GTT mapping, which will give us good
37988 performance and take care of allocating fence registers as needed. It's
37989 important *not* to have userland set up fence regs in this case, since
37990 the kernel will be using all of them.
37992 For case (3), we use the begin/end GTT map functions provided by libdrm,
37993 in combination with pinning and fence register setup in i830_memory.c to
37994 deal with tiled surfaces. This also gives us good performance and
37997 For case (4) we use the old style virtual mapping + offset for dealing
37998 with surfaces; note that UXA doesn't seem to work in this configuration
37999 regardless of these fixes.
38003 Reviewed-by: Eric Anholt <eric@anholt.net>
38004 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
38006 commit 375b2e40fcb17e94538a75392950e2533c1bb031
38007 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38008 Date: Wed Mar 25 11:13:52 2009 +0800
38010 Disable LVDS config parsing from VBT for now
38012 As wider tests showed that this doesn't work for all VBIOS, so
38013 disable it for now and reenable it after we get reliable method.
38015 commit defa6971c6b639ab72d9d4c85c4607d5f852b2b0
38016 Author: Zou Nan hai <nanhai.zou@intel.com>
38017 Date: Thu Mar 26 15:24:29 2009 +0800
38019 enable UV half pixel
38021 commit 69c84f2c8204771b68f40ed64e64657237b54546
38022 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38023 Date: Thu Mar 26 13:52:05 2009 +0800
38025 gitignore intel_lid
38027 commit fe96e517699e37cc990a14dfb48b9611a02254ed
38028 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38029 Date: Thu Mar 26 10:54:02 2009 +0800
38031 xvmc: add missed vld g4i sources in Makefile and fix distcheck
38033 Don't miss the vld shader programs for dist.
38035 commit 91886f82a1155ca24eb895f77678ea1edfccf9fe
38036 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38037 Date: Thu Mar 26 10:51:05 2009 +0800
38039 xvmc: move 965 mc media programs to shader/mc directory
38041 Also include some missed g4i files in Makefile.am, fix shader build.
38043 commit ebcb64fdc8a45105f529d5e15c335514ecaf1a05
38044 Author: Zou Nan hai <nanhai.zou@intel.com>
38045 Date: Wed Mar 25 16:11:33 2009 +0800
38047 XvMC VLD extension support for G4X
38049 commit e6af995b24767815ab30364385768867e80605c2
38050 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38051 Date: Mon Mar 23 19:19:58 2009 +0800
38053 KMS: hook up output properties for randr
38055 This gets output properties from kernel, then hook them up
38056 for randr. So we can control output properties through randr
38059 Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
38061 commit b5b377806422c9538fa9f846819ad253c1a50846
38062 Author: Ma, Ling <ling.ma@intel.com>
38063 Date: Wed Mar 25 14:39:46 2009 +0800
38065 disable center mode for 965GM and G4X platform
38067 When disabling VGA mode, usually we don't need to touch VGA center mode.
38068 However because of hardware reason, for Cresline, Cantiga & Eaglelake platform,
38069 we have to disable center mode as well. The patch fixed bug- TV Out strobing regression,
38070 reported by Robert Lowery in intel-gfx@lists.freedesktop.org mailing list.
38072 Signed-off-by: Ma Ling <ling.ma@intel.com>
38074 commit 74d0713e002a9781ed00fdd10eb6f75907ae796c
38075 Author: Kalev Lember <kalev@smartlink.ee>
38076 Date: Fri Mar 20 02:21:41 2009 +0200
38078 Fix VT switch with XV overlay video enabled.
38080 drm_intel_bo_unpin() was called with NULL argument.
38082 Signed-off-by: Kalev Lember <kalev@smartlink.ee>
38084 commit b345b4e6adf00f8b2626e5c2563fc159e2295cc5
38085 Author: Rémi Cardona <remi@gentoo.org>
38086 Date: Fri Nov 14 13:55:52 2008 +0100
38088 clean up man page generation and remove all traces of the i810 driver
38090 The i810 compatibility symlink has been broken since libpciaccess, so just
38093 commit 48b91e066878db63a1558e4cd3e6d12ff9c49197
38094 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
38095 Date: Fri Mar 20 12:33:22 2009 -0700
38097 Don't manage fences part two
38099 Don't try to clear fences that were never installed. Missed this bit in
38100 the last fix for #20265.
38102 commit 28319d60aa2d793c209043c1ce2d38b14c66a4be
38103 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38104 Date: Fri Mar 20 13:14:15 2009 +0800
38106 Add a lid status test program "intel_lid"
38108 commit 636d252f3b1eac687f7b11952e949c383cb86ed4
38109 Author: Jesse Barnes <jbarnes@nietzche.localdomain>
38110 Date: Thu Mar 19 13:25:29 2009 -0700
38112 Don't install fences if the kernel is managing them
38114 If execbuffer is setting up fences, it also means that the kernel is
38115 managing them at pin time, so installing one in the 2D driver in that
38116 case is an error. The fence should stick around as long as the buffer
38117 is pinned (the kernel won't steal these), though it will be freed at
38118 leavevt and re-allocated at entervt.
38120 On 965+ chips, the pin ioctl will *not* install a fence reg, but that's
38121 also ok because all 965+ operations include tiling bits, and sw
38122 fallbacks will be protected by prepare/finish access hooks, which will
38123 either access the backing store or use the GTT, which will ensure proper
38124 fencing at fault time.
38128 Acked-by: Eric Anholt <eric@anholt.net>
38130 commit d92e6158c86d7da873913461ef7e828bd411d8c4
38131 Author: Carl Worth <cworth@cworth.org>
38132 Date: Thu Mar 19 10:06:03 2009 -0700
38134 Revert the rest of the EXA_VERSION_MAJOR bump
38136 This was mistakenly added in the unrelated change in revision
38137 fe08b81d0f5d6f96e0124e6286bd24aba6e140ad
38138 and wasn't completely reverted in the later revision
38139 78a60e1b66fe2e8449702dd43d9b062d279af8f1
38141 commit 85e6b528582c2397ee9eb2132cd0d05ce12eb43d
38142 Author: Ma Ling <ling.ma@intel.com>
38143 Date: Thu Mar 19 09:10:19 2009 -0700
38145 Set SSC frequency for 8xx chips correctly
38147 All 8xx class chips have the 66/48 split, not just 855.
38151 commit bedc894a565ce3c7a50990e3f78953fc2432ad40
38152 Author: Jesse Barnes <jbarnes@nietzche.localdomain>
38153 Date: Wed Mar 18 20:19:44 2009 -0700
38155 Re-enable disabled outputs after rotation
38157 The server may have made a DPMS call before doing rotation, so after we
38158 do the mode set with the rotated framebuffer, we need to re-enable the
38159 corresponding output(s).
38163 commit 6deb26ae7bd796e88a5dd90df5f6c35fbc44e798
38164 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
38165 Date: Wed Mar 18 09:36:58 2009 -0700
38167 Create known output configuration at EnterVT time
38169 Since we added the pipe A force quirk (leaving pipe A on all the time),
38170 DPMS calls to disable it have silently returned, leaving the pipe on.
38171 If another driver (like vesafb) has enabled it, we may end up with a bad
38172 configuration, leading to hangs or blank screens at VT switch time.
38176 commit 7c94227dd4fa2164bebb36234958053bf1d26c12
38177 Author: Ma Ling <ling.ma@intel.com>
38178 Date: Tue Mar 17 10:33:15 2009 +0800
38180 Use best PLL timing values for G4X platform
38182 construct function to find precise parameters from internal spreadsheet
38183 table on G4X platform.
38185 Signed-off-by: Ma Ling <ling.ma@intel.com>
38187 commit 48db5bde9298f1126dfb42f4be8a3d61166abfd8
38188 Author: Ma Ling <ling.ma@intel.com>
38189 Date: Tue Mar 17 10:41:02 2009 +0800
38191 Define documented PLL timing limits for G4X platform
38193 These timings on G4X platform were specified by internal spreadsheet from the chipset group.
38195 Signed-off-by: Ma Ling <ling.ma@intel.com>
38197 commit 78a60e1b66fe2e8449702dd43d9b062d279af8f1
38198 Author: Keith Packard <keithp@keithp.com>
38199 Date: Mon Mar 16 08:41:52 2009 -0700
38201 Revert EXA_DRIVER_KNOWN_MAJOR bump
38203 This was mistakenly added in the unrelated change in revision
38204 fe08b81d0f5d6f96e0124e6286bd24aba6e140ad
38206 commit 1a6e70cd98abc63f24f710819961c24601afb59b
38207 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38208 Date: Mon Mar 16 10:19:39 2009 +0800
38210 TV: fix contrast and saturation for 945G
38214 commit 4e95327323e3d081b565147f7738eb49c28542bc
38215 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38216 Date: Mon Mar 16 09:30:22 2009 +0800
38218 TV: force TV as connected with TV_Connector option
38220 In order to bypass failure in TV load detect, TV_Connector option
38221 will always force TV as connected with user specified connector type.
38223 commit d9dbdb325543bd747cd1bfb3e1142ea6daf2b637
38224 Author: Eric Anholt <eric@anholt.net>
38225 Date: Fri Mar 13 15:48:40 2009 -0700
38227 Add shave support, enabled by default.
38229 This cuts down build system noise so that warnings are more visible. The
38230 old style output can be reenabled for build system debugging using
38231 "make V=1", or --disable-shave at configure time.
38233 commit c3a82106a1a1a94c9e1e465c7dc0d828c1cbf50d
38234 Author: Eric Anholt <eric@anholt.net>
38235 Date: Fri Mar 13 15:29:45 2009 -0700
38237 Move contributed m4 (dolt) to a subdirectory so we can include it with others.
38239 commit e0aba04da0323efc079ce0a2b78139c4d0d49fa5
38240 Author: Eric Anholt <eric@anholt.net>
38241 Date: Fri Mar 13 13:37:35 2009 -0700
38243 Add DCC register dumping.
38245 commit a57814cc13f4287eccaf1906963b80c9205c680c
38246 Author: Eric Anholt <eric@anholt.net>
38247 Date: Fri Mar 13 13:20:35 2009 -0700
38249 Add dumping of 915 and 945 fence registers.
38251 The debug dumper functions can now return NULL to indicate no output, so
38252 we get appropriate results on 915, 945, and 965.
38254 commit 8166a7ff5c2c1e3736dd06a8453c4e4d769d8b75
38255 Author: Eric Anholt <eric@anholt.net>
38256 Date: Fri Mar 13 13:12:33 2009 -0700
38258 Add PGETBL_CTL to the debug output.
38260 It's nice to know when it gets stomped on.
38262 commit fe08b81d0f5d6f96e0124e6286bd24aba6e140ad
38263 Author: Keith Packard <keithp@keithp.com>
38264 Date: Mon Mar 2 07:39:41 2009 -0800
38266 Use CopyArea to load glyphs from per-glyph pixmap to cache pixmap
38268 With glyphs sitting in per-glyph pixmaps, there's no reason to use the CPU
38269 to move them to the cache pixmap, and lots of reasons to use the accelerator.
38271 Signed-off-by: Keith Packard <keithp@keithp.com>
38273 commit 2026c57cf0a352d9e6f9d208cfb7d4d550614477
38274 Author: Kalev Lember <kalev@smartlink.ee>
38275 Date: Fri Mar 13 21:32:08 2009 +0200
38277 Fix Xv crash with overlay video.
38281 commit 3ef9d85371a97ea5baee0c47787b3bb3cdaf5135
38282 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
38283 Date: Fri Mar 13 12:21:34 2009 -0700
38285 Use UXA when KMS is active
38287 EXA doesn't support KMS, so force UXA on if KMS is detected. And warn
38288 the user if they've specified something other than UXA in their
38293 commit dc3ff0b514b609448025680778f0e95e1980a5d8
38294 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38295 Date: Thu Mar 12 16:32:02 2009 +0800
38297 Revert "SDVO: Switch control bus only before DDC access"
38299 This reverts commit ddedf19f889da2ce6d69a3afce4665e2245682fa.
38301 After i2c STOP, control bus will return back to internal
38302 registers. So this brings back to origin code that we switch
38303 to DDC bus before START. But it's ideal to only issue DDC
38304 bus switch after STOP, not before every START, which might eliminate
38305 some complains from SDVO device, that will be another patch later.
38307 commit c6b0135d209bdad3dbc641d0e264596eaf6f99d3
38308 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38309 Date: Thu Mar 12 16:31:01 2009 +0800
38311 SDVO: reset privates before output setup
38313 For multifunction encoder, forget to reset these values
38314 will cause wrong output type in later mode setting
38317 commit 28e7f0d71fa09e15a68ab4f0de169474b6235093
38318 Author: Dan Nicholson <dbn.lists@gmail.com>
38319 Date: Tue Mar 10 20:16:03 2009 -0700
38321 Fix dist of xvmc sources
38323 The XVMC AM_CONDITIONAL is only needed around the library expression.
38324 None of the other definitons will cause anything to be built without
38325 libXvMC, but they're needed for 'make dist'.
38327 Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
38329 commit 73db44e7ac524e84e5f0fda2d60069a9e954ad1b
38330 Author: Kristian Høgsberg <krh@redhat.com>
38331 Date: Wed Mar 11 11:10:57 2009 -0400
38333 Drop Legacy3D option, only use fixed texture space with non-gem.
38335 With this change, we always expect the 3D driver to use GEM textures
38336 when the 2D driver uses GEM. When GEM is not available or disabled,
38337 we fall back to legacy fixed textures.
38339 commit 2fcf4fcccfe7cfa1425985d21a144137eca07f4e
38340 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38341 Date: Tue Mar 10 14:21:36 2009 +0800
38343 SDVO: fix pixel multiplier setting for TV
38345 We should use preferred input timing's clock for correct
38346 pixel multiplier setting, otherwise we might get inconsistent
38347 multiplier setting on pipe and SDVO device for some modes.
38349 commit fb6e00f40f713a87c760fc7603159ed11ea9b0d5
38350 Author: Eric Anholt <eric@anholt.net>
38351 Date: Mon Mar 9 20:06:30 2009 -0700
38353 Fix XV with non-GEM kernels by allocating a larger fake bufmgr.
38355 Ideally we'd not be using the EXA offscreen memory manager and just hand all
38356 that memory to the fake bufmgr for non-GEM, but the fake bufmgr's too slow for
38357 that, at least currently. So compromise and take enough memory that it will
38358 succeed at XV allocations but hopefully not anger tiny-aperture systems too
38363 commit dc12c4b3eb297b2f225409eacf1f3cd521453ab6
38364 Author: Eric Anholt <eric@anholt.net>
38365 Date: Sat Mar 7 23:22:54 2009 -0800
38367 Flip the update_dri_buffers test around to only run when DRI1 is active.
38369 Fixes segfaults at startup with DRI2 and load detection, or with DRI disabled
38372 commit 646e12a9783d1d48ef21841f0909287a876731a4
38373 Author: Xiang, Haihao <haihao.xiang@intel.com>
38374 Date: Mon Mar 9 10:59:36 2009 +0800
38378 commit 73aa23d9150121a4e4b70a78cab910acd164abf5
38379 Author: Eric Anholt <eric@anholt.net>
38380 Date: Fri Dec 5 13:06:05 2008 -0800
38382 DRI1: Update sarea (and other information) when CRTC configuration changes.
38386 Signed-off-by: Eric Anholt <eric@anholt.net>
38388 commit abb213d933ac0d808fc10d4f8d88d7b8cef76346
38389 Author: Eric Anholt <eric@anholt.net>
38390 Date: Fri Mar 6 14:33:46 2009 -0800
38392 Document the UXA AccelMethod.
38394 Signed-off-by: Eric Anholt <eric@anholt.net>
38396 commit 568297d327cc321f1186afc54b38d08db3f2914d
38397 Author: Eric Anholt <eric@anholt.net>
38398 Date: Thu Jan 22 17:08:19 2009 -0800
38400 Don't allocate the render power saving context in KMS mode.
38402 That would be the kernel's job if it chooses to do it.
38404 Signed-off-by: Eric Anholt <eric@anholt.net>
38406 commit 4e1144108424a4525bbd97c3d5a56de06760bdd9
38407 Author: Eric Anholt <eric@anholt.net>
38408 Date: Fri Mar 6 14:30:05 2009 -0800
38410 unused variable warning fix.
38412 Signed-off-by: Eric Anholt <eric@anholt.net>
38414 commit 043a76a040d4576b7a8397dca805466a99bfcdd6
38415 Author: Eric Anholt <eric@anholt.net>
38416 Date: Fri Mar 6 14:29:22 2009 -0800
38418 clean up arguments to i830_allocate_framebuffer since zaphod removal.
38420 Signed-off-by: Eric Anholt <eric@anholt.net>
38422 commit b23f57b310b693f56af273526383221a4f8b96f5
38423 Author: Eric Anholt <eric@anholt.net>
38424 Date: Thu Jan 22 17:17:09 2009 -0800
38426 Use REGION_EQUAL in place of a local implementation.
38428 Signed-off-by: Eric Anholt <eric@anholt.net>
38430 commit 6b61f9945f54df7469f2b2d702b621d4d6064c3f
38431 Author: Eric Anholt <eric@anholt.net>
38432 Date: Thu Jan 22 17:19:04 2009 -0800
38434 nuke unused define.
38436 Signed-off-by: Eric Anholt <eric@anholt.net>
38438 commit c3a747cb54acc1b037b559313e6a2113ae2ac4c7
38439 Author: Dan Nicholson <dbn.lists@gmail.com>
38440 Date: Wed Nov 26 09:57:24 2008 -0800
38442 man: Put option defaults on separate indented line
38444 The groff .IP macro is used to put the option defaults in a new indented
38445 paragraph so they are separated from the explanations.
38447 Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
38448 [anholt: hand-applied due to conflicts. mistakes are my own]
38449 Signed-off-by: Eric Anholt <eric@anholt.net>
38451 commit d2af21a66d7e1d1dd62c6aa8fb41d3fd6045bcd7
38452 Author: Eric Anholt <eric@anholt.net>
38453 Date: Fri Feb 27 23:38:00 2009 -0800
38455 Remove configurable support for disabling XV.
38457 google shows one instance of this being used a year and a half ago.
38459 commit 755757669f0cad670cfa084d33d7c3e928d27855
38460 Author: Eric Anholt <eric@anholt.net>
38461 Date: Fri Feb 27 22:57:15 2009 -0800
38463 Remove configured SW cursor support.
38465 Any time we actually need SW cursors, it gets enabled automatically.
38467 commit 917b20ead3cacf1e88314f20edde6a02b97b96d7
38468 Author: Eric Anholt <eric@anholt.net>
38469 Date: Fri Feb 27 23:01:28 2009 -0800
38471 Remove StolenOnly support.
38473 We rely on having AGPGART present to successfully allocate video memory as
38474 we configure it by default. Admit that fact, and remove support for
38475 non-AGPGART/KMS setups.
38477 commit 73b7190421132ad73179c3fb7bb0e06c427dce5c
38478 Author: Eric Anholt <eric@anholt.net>
38479 Date: Fri Feb 27 22:51:22 2009 -0800
38481 intel: Nuke shared-entity support (zaphod mode).
38483 It's been broken for years now, and KMS offers a much better chance of getting
38484 this working sensibly without making a mess of the 2D driver.
38486 commit e9d6bbbe0b41e29c58a79844decd81771da85dd4
38487 Author: Eric Anholt <eric@anholt.net>
38488 Date: Tue Mar 3 16:45:03 2009 -0800
38490 remove more page flipping leftovers.
38492 commit 1cc9b1423c5df591c615ef9588b6eefd81448f80
38493 Author: Lukáš Hejtmánek <xhejtman@ics.muni.cz>
38494 Date: Fri Mar 6 14:44:03 2009 -0500
38496 Fix another VT switch leak
38498 The batch_bo buffer object is reallocated on enter VT, so we need to
38499 unref it on leave vt.
38501 Signed-off-by: Lukas Hejtmanek <xhejtman@ics.muni.cz>
38503 commit 67fef27f4b76490be085d232aba0ca9cbb3c5e59
38504 Author: Xiang, Haihao <haihao.xiang@intel.com>
38505 Date: Fri Mar 6 09:40:07 2009 +0800
38507 Xv: free tearing on textured video
38509 Add an Xv attribute XV_SYNC_TO_VBLANK which has three values -1(auto), 0(off)
38510 and 1(on) to control whether textured adapter synchronizes the screen
38511 update to the vblank. The default value is -1(auto).
38513 commit 0d20bbbc2005a51f427a9ae6b6a66dbbb101dbab
38514 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38515 Date: Thu Feb 19 14:24:24 2009 +0800
38517 SDVO: handle multifunction encoder (try 2)
38519 For SDVO encoder that advertise multiple functions,
38520 we have to get attached display to determine current
38521 output, and update output's name according with
38524 commit d4c64f01b9429a8fb314e43f40d1f02bb8aab30f
38525 Author: Lukas Hejtmanek <xhejtman@ics.muni.cz>
38526 Date: Wed Mar 4 17:33:27 2009 -0500
38528 Fix serious memory leak at Enter/LeaveVT
38530 This fixes huge memory leak at each VT switch (about 600 BOs + 6MB
38531 of RSS of Xserver).
38533 commit 095a001f755201d3c19335b67a84c57b1d080a83
38534 Author: Kristian Høgsberg <krh@redhat.com>
38535 Date: Wed Mar 4 17:10:24 2009 -0500
38537 Use drmSetMaster() and drmDropMaster() in enter/leave VT
38539 This allows multiple X server to use DRI and makes it possible to run
38540 multiple X servers under KMS. This requires a 2.6.29 kernel to work.
38541 On older kernels it will just log a warning and DRI will fail to
38542 initialize for the second X server.
38544 commit 14bb61e0c2e28725a2f6167d3263649bc845be18
38545 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38546 Date: Mon Mar 2 16:55:45 2009 +0800
38548 SDVO: only check digital monitor when EDID exists
38550 commit 42e34e90e2e4048b38481cab61cef46f932eada7
38551 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38552 Date: Tue Mar 3 22:55:35 2009 +0800
38554 TV: add property control for TV attributes
38556 This is based on Jesse's origin patch for bug #12763.
38557 But export integer range to user instead of hardware float
38558 point format, and fix different real format on 965G and 945G
38559 for contrast and saturation.
38561 commit aa9da5e393c804019720503fe58bdd247fe1eabd
38562 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38563 Date: Tue Mar 3 20:26:19 2009 +0800
38565 TV: add option to set TV connector type
38567 This can let user override non-stable driver TV load detect,
38568 and set connector type manually, e.g for s-video to component
38569 converter, this patch seems must needed to use HD modes.
38571 commit ab1ef05cc0fd5ba9768c26cc51bc2c7b52baa45f
38572 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38573 Date: Thu Feb 19 20:23:02 2009 +0800
38575 TV: update output and crtc modes when TV format change (try 3)
38577 This is to fix bug #16566, change TV format will cause BadMatch
38578 error when crtc config apply. Everytime when we change TV format,
38579 we may generate a new list of modelines as TV clock changed. After
38580 randr get info request, new modelines will be probed and randr output's
38581 modes will be renewed too. But crtc's mode failed to be updated,
38582 as it never can find a matching mode now within new modes list.
38583 So get info will return an invalid crtc's mode, later set crtc
38584 config will pass that info, and xserver catches a bad match.
38586 This patch trys to refresh output modes and setup crtc's mode
38587 with new modelines in TV format change. So get info would be
38588 sure to turn valid crtc mode that reference in current new modelines.
38590 commit a67a911a93ff3c0b3d2e6f6fb20c0787b9d4d41d
38591 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38592 Date: Thu Feb 19 01:00:33 2009 +0800
38594 TV: sort input mode lines
38596 commit bd360e7517835626bee54bec968cdc78908c6545
38597 Author: Adam Jackson <ajax@redhat.com>
38598 Date: Mon Mar 2 10:33:35 2009 -0500
38600 KMS: Wire up output DPMS.
38602 commit 2d0aa553609a0e64fa4b2e755b9b1e244e5c3fa1
38603 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38604 Date: Mon Mar 2 15:26:44 2009 +0800
38606 Skip LVDS config parsing on pre-9xx chips.
38608 855GM laptops seems pretty broken when parsing this block.
38610 commit 55359ef9ab0a2be42a55e0279835f76a191d6c74
38611 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38612 Date: Fri Feb 27 09:02:30 2009 +0800
38614 Update driver feature block definition for missed fields
38616 commit 38a7683561cee7fffab174c2a166bfd51b51ba27
38617 Author: Eric Anholt <eric@anholt.net>
38618 Date: Fri Feb 27 22:55:44 2009 -0800
38622 commit 5bfd73cd31ba197a62f549cdbad1a1270b571027
38623 Author: Eric Anholt <eric@anholt.net>
38624 Date: Fri Feb 27 19:09:49 2009 -0800
38626 Only allocate pixmaps aligned for tiling when requested by DRI2 GetBuffers.
38628 This saves massive quantities of memory on pre-965 since the DRI2 tiling
38629 enable caused the minimum size of any pixmap to be 1MB.
38631 commit f53bdad1412f841075232455837578f00709c6ef
38632 Author: Shaohua Li <shaohua.li@intel.com>
38633 Date: Mon Feb 23 15:19:23 2009 +0800
38635 Intel video driver patch
38637 This is the intel video driver patch for a new chip, which is G33-like
38638 and has some clocking setting related register changes. This patch adds
38639 the pci id and DPLx/FPx register changes.
38641 The gtt tool should just work to me, as the chip hasn't any changes
38642 against G33 on this side.
38644 Signed-off-by: Shaohua Li <shaohua.li@intel.com>
38645 Signed-off-by: Eric Anholt <eric@anholt.net>
38647 commit 22dc9a5580d77cc4707bfb8e19474e611a06ae9a
38648 Author: Eric Anholt <eric@anholt.net>
38649 Date: Tue Feb 24 18:22:20 2009 -0800
38651 Fix UXA for server 1.4.
38653 commit cb1f7ec0876746c1b52b63cdb508544e9e4e32e3
38654 Author: Eric Anholt <eric@anholt.net>
38655 Date: Tue Feb 24 21:23:59 2009 -0800
38657 uxa: Fix composite fallback debug printing of main memory versus bo info.
38659 It was just printing whether it was a pixmap (it is), instead of whether the
38660 pixmap was offscreen.
38662 commit 635eaa511f28fb673fe306e46ed5370e78a8a534
38663 Author: Eric Anholt <eric@anholt.net>
38664 Date: Tue Feb 24 21:48:11 2009 -0800
38666 Regard the screen pixmap as suitable for acceleration.
38668 With UXA on the fake bufmgr, the screen pixmap doesn't have a BO and so
38669 no acceleration was occurring.
38671 commit 70e0261208654c6c875ad462da2734c6aa9eeb96
38672 Author: Eric Anholt <eric@anholt.net>
38673 Date: Tue Feb 24 20:54:05 2009 -0800
38675 Disable fb resizing for DRI1-only server so that DRI1 can initialize.
38677 commit d7aa330db31100b7cb54d8165f9a4b94329ece32
38678 Author: Eric Anholt <eric@anholt.net>
38679 Date: Tue Feb 24 20:34:10 2009 -0800
38681 Fix up i915 composite and common solid/copy code to use check_aperture.
38683 This could fix complaints about binding BOs and resulting failure to render.
38685 commit a625a07022ea7f6757b288fcc2ffb9e27c7f8341
38686 Author: Xiang, Haihao <haihao.xiang@intel.com>
38687 Date: Wed Feb 25 16:48:22 2009 +0800
38689 XvMC: fix broken xvmc on 965
38691 commit f6d8ae69b0f97e696c142f06c8038f336ed024f9
38692 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38693 Date: Wed Feb 25 09:57:00 2009 +0800
38695 Use LVDS config in Driver feature BDB for integrated LVDS check
38697 The LVDS config bits in VBT driver feature block is used by vendor
38698 to identify the board implement of integrated LVDS/eDP or SDVO LVDS.
38699 And video bios uses these bits for LVDS enabling or not. So check
38700 these bits for integrated LVDS might eliminate more quirks.
38702 commit 8718551f14e064b461e80a583597f0ea9fb8ca9f
38703 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38704 Date: Wed Feb 25 09:54:03 2009 +0800
38706 Update LVDS config bits definition in driver feature block
38708 ALL_LVDS is actually not defined before and include GM45.
38709 Embedded DP bit will be used for newer chips.
38711 commit 9d8e5c21a1688b915bf39261d4c3b0bf2906daef
38712 Author: Eric Anholt <eric@anholt.net>
38713 Date: Tue Feb 24 14:05:15 2009 -0800
38715 Fix distcheck from drmmode_display.h deletion.
38717 commit f4e2f522a5c5b03ea530b9eb67e9d1a9a96274ce
38718 Author: Eric Anholt <eric@anholt.net>
38719 Date: Sat Feb 21 20:36:58 2009 -0800
38721 Don't do AdjustFrame in KMS mode.
38723 This was hit by xv86vm's SwitchMode path, and for that the CRTC offsets
38724 get set at mode setting time anyway.
38726 commit 170f00e161931fdaa8c2812fc710649e1d6d977a
38727 Author: Kristian Høgsberg <krh@redhat.com>
38728 Date: Tue Feb 24 14:07:23 2009 -0500
38730 Remove a handful of unused variable warnings.
38732 commit a6b31f38ebf470c61de0e10b0ce2af0d7ee1684b
38733 Author: Kristian Høgsberg <krh@redhat.com>
38734 Date: Tue Feb 24 13:58:20 2009 -0500
38736 Update kms to work with drmModeModeInfo API update.
38738 commit 13ee9402e9822d6f57e3ebcc5ae658ce8322118e
38739 Author: Adam Jackson <ajax@redhat.com>
38740 Date: Mon Feb 23 15:44:26 2009 -0500
38742 Limit CRT DAC speed better.
38744 Verified against the public docs for i8xx parts, although not 9xx yet.
38746 commit 668b2352a47bcfba75fe0492a5805726222755eb
38747 Author: Keith Packard <keithp@keithp.com>
38748 Date: Mon Feb 23 13:31:51 2009 -0800
38750 Revert "Limit CRT DAC speed better."
38752 This reverts commit 8fd0e46571c7ba15c05f0a759113f8ca842c76a2.
38754 This doesn't even build. Please try again.
38756 commit 8fd0e46571c7ba15c05f0a759113f8ca842c76a2
38757 Author: Adam Jackson <ajax@redhat.com>
38758 Date: Mon Feb 23 15:44:26 2009 -0500
38760 Limit CRT DAC speed better.
38762 Verified against the public docs for i8xx parts, although not 9xx yet.
38764 commit 73bc7f113969834d00cd92be8374dbadc62f96a9
38765 Author: Kristian Høgsberg <krh@redhat.com>
38766 Date: Mon Feb 23 15:16:51 2009 -0500
38768 KMS: Fix bug that prevented EDID data from getting propagated.
38770 commit ef952760551ad15cb9f63025d1e087645949a227
38771 Author: Kristian Høgsberg <krh@redhat.com>
38772 Date: Sun Feb 15 22:10:42 2009 -0500
38774 Use stat() and the dev_t to find the drm device filename.
38776 Simpler and more robust. Works when we haven't set the device PCI ID,
38777 ie when not going through the legacy DRI module.
38779 commit 81c652e9a666a7459bcc5217c8a5ec518b6e00da
38780 Author: Helge Bahmann <helge.bahmann@secunet.com>
38781 Date: Sat Feb 21 10:10:04 2009 -0800
38783 Move disable_render_standby to EnterVT instead of startup.
38785 Otherwise, with a pre-2.6.28 older kernel the disable would be lost at
38786 resume time and cause hangs.
38790 commit 0621ba12a3b694720e67a49b25ca52f0e09b3802
38791 Author: Eric Anholt <eric@anholt.net>
38792 Date: Wed Feb 18 13:32:44 2009 -0800
38794 uxa: Ask for BOs ready for rendering for pixmaps.
38796 The assumption is that we're almost always accelerating our drawing to
38797 new pixmaps (fill, copy, etc.).
38799 commit 5018d0f16cb8b44c743b5b37d194fe806d955568
38800 Author: Kristian Høgsberg <krh@redhat.com>
38801 Date: Fri Feb 20 15:34:29 2009 -0500
38803 KMS: Hook up rotated shadow buffers.
38805 commit beca598bc2848093b710bd47828d622205d273df
38806 Author: Kristian Høgsberg <krh@redhat.com>
38807 Date: Fri Feb 20 11:08:51 2009 -0500
38809 Access the Xv buffer through the GTT for the non-KMS case.
38811 commit e97e2571703e3d6188bf18f211b793fc50383f9c
38812 Author: Kristian Høgsberg <krh@redhat.com>
38813 Date: Thu Feb 19 16:41:49 2009 -0500
38815 Fix i915 textured video to work with the i830_memory -> bo change.
38817 Forgot to update i915_video.c in 872aadc7102bd5131e1582ede081e22672911ba2.
38819 commit 96da26b6813a8c1da8a43036c375aa0d2bb70f16
38820 Author: Kristian Høgsberg <krh@redhat.com>
38821 Date: Thu Feb 19 16:40:19 2009 -0500
38823 Dont allocate overlay registers in KMS mode.
38825 commit 872aadc7102bd5131e1582ede081e22672911ba2
38826 Author: Kristian Høgsberg <krh@redhat.com>
38827 Date: Wed Feb 18 17:26:06 2009 -0500
38829 Make Xv used a buffer object instead of i830_memory.
38831 We still pin the buffer object in case of overlay, but for textured video
38832 we're now no longer using i830_memory for Xv anymore.
38834 commit 527e8177cde3abbabbcdccee0dbc0dcc0068a1be
38835 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38836 Date: Wed Feb 18 11:26:58 2009 +0800
38838 bios_reader: parse driver feature BDB
38840 commit 62ca1c479825ffa0e9cf444b4e25080150faa45b
38841 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38842 Date: Wed Feb 18 10:49:00 2009 +0800
38844 Fix SDVO mulitiplier setting for 945G
38846 Wrong SDVO multiplier setup has been slipped
38847 in SDVO TV patch. Thanks Michael Fu to point this out!
38849 commit 723b6065093adb56a2d7204bd990ceae41bfafc9
38850 Author: Kristian Høgsberg <krh@redhat.com>
38851 Date: Tue Feb 17 13:48:04 2009 -0500
38853 KMS: Hook up crtc::gamma_set.
38855 commit 506bbb8341c052499057834a669b82787454b426
38856 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38857 Date: Mon Feb 16 14:09:49 2009 +0800
38859 Safely init SDVO found variable
38861 Found by Michael Fu for my last SDVO detect fix.
38863 commit 9d464bd5b0d2724f5edb26e859888ceb6a248f9b
38864 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38865 Date: Fri Feb 13 11:15:48 2009 +0800
38867 Fix SDVO/HDMI detect
38869 SDVOC detect bit is only valid for HDMIC.
38870 And for SDVO devices, SDVOB detect bit should be used
38871 to probe all possible SDVO outputs.
38873 commit 48445d2e939328495b4abe0fb7e579dfcef727bb
38874 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38875 Date: Fri Feb 13 11:11:18 2009 +0800
38877 SDVO: remove ForceSDVODetect option
38879 Which is just a hack to hide our SDVO detect drawback,
38880 we will have SDVO/HDMI detect fix later.
38882 commit ddedf19f889da2ce6d69a3afce4665e2245682fa
38883 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38884 Date: Fri Feb 13 10:09:42 2009 +0800
38886 SDVO: Switch control bus only before DDC access
38888 Instead of set control bus switch before every i2c start,
38889 just set once before doing DDC. This should eliminate some
38890 encoders returning error during that.
38892 commit f1ca56e17d0ecd4f1299061a6b3272bfd289e123
38893 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38894 Date: Fri Feb 13 10:02:02 2009 +0800
38896 SDVO: Fix TV support
38898 As SDVO TV uses SDVO_TVClkIn from SDVO encoder for clock reference,
38899 it needs to generate proper PLL for current input clock. This uses
38900 fixed PLL table from vbios for this. And possible sdvo mulitiplier
38901 has to be setup correctly. This makes TV output stable on my 945GCLF2
38902 board with NTSC-M format.
38904 commit acde0ef683d6ec33d0b478923ffb11bd6785798f
38905 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38906 Date: Fri Feb 13 09:53:57 2009 +0800
38908 SDVO: fix CREATE_PREFERRED_INPUT_TIMING command
38910 commit 824b2f0c5530c3196901c961757e6677b042caf3
38911 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38912 Date: Fri Feb 13 09:50:45 2009 +0800
38914 SDVO: fix usage for SET_TV_FORMAT and GET_SDTV_RESOLUTION_SUPPORT command
38916 They both needs parameters.
38918 commit 62c0c2f5549a51c5df209f7353a19ca301f221be
38919 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38920 Date: Fri Feb 13 09:48:34 2009 +0800
38922 SDVO: fix error in modeline and DTD convert
38924 commit 37c67088a887e6380571e6eec8a8f058e3e24717
38925 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38926 Date: Wed Feb 11 15:15:27 2009 +0800
38928 SDVO: check EDID info for DVI-I
38930 For SDVO DVI-I, check EDID info for digital output,
38931 otherwise mark it to be disconnected as analog output
38932 is driven by VGA then.
38934 commit 38079bc0f1038da77048bbf6e5c10758f9fb8a55
38935 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38936 Date: Wed Feb 11 14:41:58 2009 +0800
38938 SDVO: Fix for HDMI encode and audio setup (try 5)
38940 SDVO HDMI encode and audio is not setup in detect,
38941 which fails in hotplug case for HDMI audio. Fix to
38942 check current encode type and set flag for HDMI audio
38945 Check and set HDMI encode state in get_modes.
38947 commit 3012d85cc5eb58c2447e93c05c39dc14feaae988
38948 Author: Eric Anholt <eric@anholt.net>
38949 Date: Tue Feb 10 18:47:28 2009 -0800
38951 uxa: Fix breakage from UXA_FALLBACK conversion from "do {} while (0)" construct.
38953 Thanks to keithp for post-commit review.
38955 commit 5009127de7d9527ae329d1c2fbc7283648bde2e6
38956 Author: Eric Anholt <eric@anholt.net>
38957 Date: Tue Feb 10 18:23:35 2009 -0800
38959 uxa: Fix driver against fbDoCopy -> miDoCopy change in the server.
38961 commit b53977f4c53c7c8f562f909e985b8d5a7b2526f3
38962 Author: Eric Anholt <eric@anholt.net>
38963 Date: Tue Feb 10 18:22:46 2009 -0800
38965 uxa: Fix failure to --amend in further changes in previous commit.
38967 commit 5212ec6515c6562f66b86fc16928b601bf04e49b
38968 Author: Eric Anholt <eric@anholt.net>
38969 Date: Tue Feb 10 15:35:20 2009 -0800
38971 uxa: hook up the fallback debug to the driver's fallback debug option.
38973 commit 3aa8591abfbe8db0f13912910c850fdd748808df
38974 Author: Ma Ling <ling.ma@intel.com>
38975 Date: Fri Feb 6 09:14:15 2009 +0800
38977 Don't disable vga centering bit.
38979 commit id b9f5915ce812144ffd9d2aa42e8ba856129c35e,
38980 which resolved bug #17235, but generate new regression-bug #19715.
38981 This patch intends to resolve bug #17235, and avoid regression as well.
38982 We have successfully re-tested it for bug #17235 and #19715 respectively.
38984 commit 9fe5fca3fe761a4f11857d9766138a60f471a9e6
38985 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
38986 Date: Wed Feb 4 06:00:39 2009 +0800
38988 TV quirk for HP Compaq nx6310
38990 commit 5c370091620b38447172ebeffbc6ed3256e86c9d
38991 Author: Kristian Høgsberg <krh@redhat.com>
38992 Date: Mon Feb 2 12:02:00 2009 -0500
38994 Fix front buffer memset() for non-KMS case.
38996 Missed the pI830->FbBase condition when removing the KMS hook.
38998 commit 2013799b20599a58de48cb21a5a389e898a58af1
38999 Author: Kristian Høgsberg <krh@redhat.com>
39000 Date: Mon Feb 2 11:51:47 2009 -0500
39002 Un-revert the I915_SETPARAM_NUM_USED_FENCES commit reverted by accident.
39004 Oops, my bad. Reverted 8d4bc36fae50b09a73ba2cfab920adb32141a358
39005 since my kernel doesn't yet have the new param, committed
39006 the revert by accident.
39008 commit 127330bfd53ac7571bdd12a551142528b972893f
39009 Author: Kristian Høgsberg <krh@redhat.com>
39010 Date: Mon Feb 2 11:02:59 2009 -0500
39012 Fix last-minute "cleanup" that broke the patch.
39014 commit 0cb87ccfe97b0e016e47dcf236fd5ce78dddfc4b
39015 Author: Kristian Høgsberg <krh@redhat.com>
39016 Date: Fri Jan 30 17:53:03 2009 -0500
39018 Implement front buffer resize for KMS.
39020 This adds back the resize hook so we can resize the front buffer under
39021 kernel mode setting as well.
39023 The patch also pulls the drmmode_* structs from drmmode_display.h into
39024 drmmode_display.c and eliminates the header file.
39026 commit 66bc44e8f9a0505c0b11b8042243ca74079da85f
39027 Author: Eric Anholt <eric@anholt.net>
39028 Date: Tue Jan 27 15:43:58 2009 -0800
39030 dri2: Use modesetting's master fd instead of opening our own non-master.
39032 This fixes failure to auth DRI2 clients under KMS.
39034 commit 8d4bc36fae50b09a73ba2cfab920adb32141a358
39035 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
39036 Date: Mon Jan 26 17:14:06 2009 -0800
39038 Support tiled back/depth on 915-class hardware with DRI2.
39040 Set alignments, tile settings and flags correctly in the 2D driver to support
39041 tiled rendering. UXA's create pixmap function currently assumes the worst
39042 about the alignment constraints; that should probably be fixed. Some of the
39043 1M alignment fixes could probably be done more cleanly as well.
39045 commit 6c0ca1676bf60529dd331cc739abdf68fa9e918d
39046 Author: Eric Anholt <eric@anholt.net>
39047 Date: Wed Jan 21 15:52:02 2009 -0800
39049 Don't forget the new state bos in check_aperture.
39051 They're tiny so it shouldn't have been a problem, but play it safe. This is
39052 another <5% loss on top of the previously reported value, bringing the whole
39053 series to about 8%.
39055 commit 57a02b50c60c10a25ff0f3cd93af9f37fa0d1b11
39056 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
39057 Date: Mon Jan 26 14:58:28 2009 -0800
39059 Fixup bogus VBT modes when detected
39061 Several VBT modes out in the wild have H or VSyncEnd values greater than
39062 the H or VTotal value. This clearly ends up creating a bad mode,
39063 causing some panels to either ignore the timing or display some sort of
39066 Check for these cases and fix them up by default, making things work for
39067 several Dell and Sony machines.
39069 Fixes FDO bug #17292.
39071 commit e20e1cf76fb00ba4f933a1ed6d1a4896be346c91
39072 Author: Bill Nottingham <notting@redhat.com>
39073 Date: Sat Jan 24 08:36:20 2009 +0800
39075 Quirk MSI IM-945GSE-A LVDS, TV outputs.
39077 The IM-945GSE-A claims to have a TV output, and always claims a connected
39078 LVDS output. It has neither.
39080 commit 05ff561234cc2b93fe1ea2a35041fa2e119a7e38
39081 Author: Vincent Mussard <vmussard@free.fr>
39082 Date: Sat Jan 24 08:33:16 2009 +0800
39084 quirk for AOpen MP45
39086 commit fbf003ef2767a1a9f5e4064f04a17992030d8f5c
39087 Author: Eric Anholt <eric@anholt.net>
39088 Date: Fri Dec 5 17:27:13 2008 -0800
39090 Move i965 render sampler state to BOs.
39092 This eliminates the pinned memory allocation for 965 render state.
39094 commit befd4ad8beae39377f804e06c7cbd926ec4251db
39095 Author: Eric Anholt <eric@anholt.net>
39096 Date: Tue Jan 20 14:43:08 2009 -0800
39098 Move i965 render kernels to BOs.
39100 commit 5d705de5d11297f4d6bd5237fb67619e703745bc
39101 Author: Eric Anholt <eric@anholt.net>
39102 Date: Fri Dec 5 15:30:35 2008 -0800
39104 Move 965 render unit state to BOs.
39106 This is a first step in a series of changes to avoid requiring a pinned object,
39107 which gets in the way of doing non-root KMS. This change appears to result in
39108 about a 2-6% loss in x11perf -aa10text, which better algorithms in libdrm could
39109 make up for (it hasn't really had to deal with code this bad before).
39111 commit 64b08ed5ad46ccd76964972e39bfed8721a920aa
39112 Author: Eric Anholt <eric@anholt.net>
39113 Date: Tue Jan 20 17:13:52 2009 -0800
39115 Remove 965 render wm scratch space, which was just unused.
39117 commit f126aabdf8952177bb15f392041da7a7094eb31b
39118 Author: Eric Anholt <eric@anholt.net>
39119 Date: Wed Jan 21 12:52:36 2009 -0800
39121 Fix build with server 1.4.
39123 Debian "unstable" is still stuck with this ancient version.
39125 commit 253b8db298f38676e47dc902534465054f7b58b8
39126 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
39127 Date: Wed Jan 21 09:02:13 2009 -0800
39129 Don't run in KD_TEXT mode even with KMS
39131 Leaving the VT in KD_TEXT mode keeps the kernel's blanking code active,
39132 so when a DPMS event happens, the fb console is restored rather than X's
39133 configuration. On the downside it means the kernel won't print messages
39134 in the background, which would be visible if a panic or emergency switch
39135 occurred. The proper fix here is a new kernel mode, which we can move
39138 commit 131b414feb2ecabe31b538d65725ac4427a4387a
39139 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
39140 Date: Wed Jan 21 09:01:05 2009 -0800
39142 Tear down batchbuffers unconditionally on LeaveVT
39144 Even if KMS is enabled we should do this, to avoid running batches that
39145 depend on other state we tear down in LeaveVT.
39147 commit b6f3ce32e295929f461a7bc37e61f126fb51e4aa
39148 Author: Eric Anholt <eric@anholt.net>
39149 Date: Tue Jan 20 10:45:23 2009 -0800
39151 Use drm_intel_bo_subdata to put render vb data in.
39153 This improves performance by avoiding repeated map/unmap cycles, which are
39154 a bit expensive on my machine with lock debugging on in the kernel. It could
39155 do much better if we did more than 18 or so floats at a time.
39157 commit 9a8bbb1951ad0ca0a9407a97348fc7fa03127900
39158 Author: Eric Anholt <eric@anholt.net>
39159 Date: Mon Jan 19 19:34:50 2009 -0800
39161 Move i965 render vb setup to use time, and decouple state emit from it.
39163 The require_space had failed since it only checked for the space required
39164 by the batch emits in the function itself, but not in the
39165 i965_emit_composite_state() that it called (the state we were concerned about
39166 having set up for that 12 * 4 dwords to follow!). This is replaced by
39167 intel_batch_start_atomic(), which will catch such mistakes in the future.
39169 commit 3d739597c4f5817079efd9067ad5db2f4105f765
39170 Author: Eric Anholt <eric@anholt.net>
39171 Date: Tue Jan 20 10:37:20 2009 -0800
39173 Move i965 render transform setup from emit_composite_state to prepare_composite.
39175 commit e20f7278f3abb44a3a151ac91f83c45cf1a2745a
39176 Author: Eric Anholt <eric@anholt.net>
39177 Date: Mon Jan 19 20:31:31 2009 -0800
39179 i965: Pull check_aperture out to a separate function and make it dtrt.
39181 Previously it wouldn't count the pixmaps that were about to be used, which
39182 is pretty much the only purpose of having the pain around. This also
39183 eliminates the check_twice confusion with emit_batch_header_for_composite().
39185 commit 013e2adfbf955cb21450b610091542ebd54392c2
39186 Author: Eric Anholt <eric@anholt.net>
39187 Date: Mon Jan 19 20:24:20 2009 -0800
39189 Move filter computation from emit_batch_header to prepare_composite.
39191 commit a340fe5e4227ebea5493e658eb6289624b07ab0b
39192 Author: Eric Anholt <eric@anholt.net>
39193 Date: Mon Jan 19 19:11:41 2009 -0800
39195 Use intel_emit_reloc from video to prettify 965 render bind_bo setup.
39197 commit aefe198ca427a5ad69717f49948eb3ede713bb28
39198 Author: Eric Anholt <eric@anholt.net>
39199 Date: Mon Jan 19 18:57:01 2009 -0800
39201 Move i965 render state bo setup back to prepare_composite.
39203 We want the objects to be created once per prepare/done both for efficiency and
39204 so we can handle aperture checking better.
39206 commit 946c7ef8170e74ac178c83b1465242d57fa86f2e
39207 Author: Eric Anholt <eric@anholt.net>
39208 Date: Mon Jan 19 14:43:20 2009 -0800
39210 Do check_aperture_space and batch_start_atomic for i965 video.
39212 This increases the overhead for video in the presence of cliprects, but we
39213 were already doing nasty things in that case and don't seem to care. This
39214 could fix potential bad rendering or hangs with video, particularly with
39217 commit 7be668179a12918918cad863f6936ced4ab78dbf
39218 Author: Eric Anholt <eric@anholt.net>
39219 Date: Mon Jan 19 14:29:25 2009 -0800
39221 Move 965 video setup to a separate function so we can move it around.
39223 commit c7db3201106f07f3228c989014d6db5ace5378f6
39224 Author: Owain G. Ainsworth <oga@openbsd.org>
39225 Date: Tue Jan 13 18:46:41 2009 +0000
39227 Remove the pageflipping infrastructure.
39229 It was broken on current kernels, and deprecated anyway.
39231 commit c82adfd0169317efb6c2f9de0f315651f9adbae1
39232 Author: Owain G. Ainsworth <oga@openbsd.org>
39233 Date: Tue Jan 13 17:09:00 2009 +0000
39235 Remove triple-buffering support
39237 It never worked with any upstream linux kernel, and is quite heavily
39238 deprecated. A new solution based around DRI2 will probably be
39239 forthcoming. Pageflipping itself is next.
39241 commit ada44c1c0edcd3ea9e41ed6b6fdb2bf0e87c9c67
39242 Author: Kshitij Kulshreshtha <kkhere.geo@gmail.com>
39243 Date: Tue Jan 20 11:35:36 2009 +0800
39245 Support sysfs backlight control for Sony laptops in xf86-video-intel
39247 The sony_laptop kernel module (since v2.6.23) supports backlight control
39248 via the sysfs interface:
39249 /sys/class/backlight/sony
39251 This patch will enable xf86-video-intel to use this backlight control method
39252 for Sony VAIO Laptops with Intel integrated video.
39254 commit d89de6d60a86105a198d904821853e6ed7de3305
39255 Author: Eric Anholt <eric@anholt.net>
39256 Date: Fri Jan 16 16:59:17 2009 -0800
39258 Protect i915 textured video against batchbuffer wrapping.
39260 commit 15780c53f5717936ea10ac87aff8a881c172f1dc
39261 Author: Eric Anholt <eric@anholt.net>
39262 Date: Sat Jan 17 16:23:38 2009 -0800
39264 Fix i915 batch_start_atomic counting.
39266 commit fc46cb6403387215d13aa28d720f205688ae9d67
39267 Author: Eric Anholt <eric@anholt.net>
39268 Date: Sat Jan 17 12:47:53 2009 -0800
39270 Fix libdrm version number requirement regression that got spammed in.
39272 commit 1459f794e2a4b9032807c3794b00a33fa6392837
39273 Author: Dave Airlie <airlied@redhat.com>
39274 Date: Sat Jan 17 22:14:26 2009 +1000
39276 intel: fix DRI2 should be DRI_DRI2
39278 commit db43b7870a37ea273941302096a6f00120dfae71
39279 Author: Eric Anholt <eric@anholt.net>
39280 Date: Thu Jan 15 19:05:52 2009 -0800
39282 Re-emit i915 composite setup when the batchbuffer wraps.
39284 This also introduces tests to make sure that we asked for enough reserved space
39285 and that we don't allow wrapping at the wrong time.
39287 This fixes a hang during text rendering with DRI2 and a GL client running,
39288 but could potentially affect text rendering with GEM in general with an
39289 exceptional batchbuffer setup.
39291 commit cab5b7a7b0e17414f98b2363b0961c87f32f9c05
39292 Author: Eric Anholt <eric@anholt.net>
39293 Date: Thu Jan 15 09:31:55 2009 -0800
39295 Fix invarient state emits for DRI2 (do it per batch, since there's no lock).
39297 commit 9f306193c4b128ec72c5b4db16d07302cee50ab5
39298 Author: Owain G. Ainsworth <oga@openbsd.org>
39299 Date: Tue Jan 13 18:42:44 2009 +0000
39303 I915_EMIT_IRQ is a read/write ioctl, not a write only. Found by OpenBSd's
39304 kernel code which checks these things a long more strongly than Linux.
39306 commit 1e8588ad5087c69eb77399cfaab8e4ec15eb4da9
39307 Author: Owain G. Ainsworth <oga@openbsd.org>
39308 Date: Tue Jan 13 17:02:59 2009 +0000
39310 use ifdef __linux__ where needed.
39312 since modesetting is compiled by default now, ifdef __linux__ the linux
39313 only includes and ioctls.
39315 commit db9f5915ce812144ffd9d2aa42e8ba856129c35e
39316 Author: Ma Ling <ling.ma@intel.com>
39317 Date: Wed Jan 14 14:46:52 2009 +0800
39319 Disable VGA plane reliably
39321 This fixes #17235, VGA random hang on recent G45/43 board.
39322 From spec, SR01 bit 5 should be set before VGA plane disable through
39323 control register, otherwise we might get random crash and lockups.
39325 commit a4b023c17b9c3bd65fb9466ddb8a953f60244402
39326 Author: Keith Packard <keithp@keithp.com>
39327 Date: Tue Jan 13 20:01:49 2009 -0800
39329 Assign rotation memory dri_bo to rotation pixmap.
39331 As the rotation memory and rotation pixmap are allocated separately (to make
39332 rotation at startup work), the allocate dri_bo needs to be set in the pixmap
39333 for acceleration to work. This restores the performance in rotated modes.
39335 Signed-off-by: Keith Packard <keithp@keithp.com>
39337 commit f659cc37e61364a408355b9a6a44d39f4d759935
39338 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
39339 Date: Tue Jan 13 16:49:41 2009 -0800
39341 bios_reader: make mode timing output friendlier
39343 Print out the calculated mode line values (as i830_bios.c uses) and
39344 check for validity against known problems.
39346 commit c80f1a9c5131721eaf87d12a7a67e603cdfed63b
39347 Author: Bernhard Rosenkraenzer <bero@arklinux.org>
39348 Date: Tue Jan 13 10:35:19 2009 -0800
39350 UXA: Declare glyph cache picture as component-alpha when necessary.
39352 Without this, rendering component-alpha glyphs may break without a mask.
39354 Bug #19534. Ported from fix by Michel Dänzer <daenzer@vmware.com> in
39355 xserver commit 639f289dcdbe00a516820f573c01a8339e120ed4
39357 commit 3d206f9e46b5237bda7ca3c0f92d64c45ee8bdf5
39358 Author: Ma Ling <ling.ma@intel.com>
39359 Date: Tue Jan 13 10:26:40 2009 +0800
39361 set continuous-frequency flag in get modes function
39363 http://bugs.freedesktop.org/show_bug.cgi?id=19247
39364 Because latest xorg will check whether the display is continuous frequency through one flag in monitor info structure,
39365 if not it doesn't touch default modes. When laptop failed to fetch edid, We don't set the flag. In i830_lvds.c,
39366 so currently we can not get default modes except only one mode line from bios.
39367 In order to achieve default modes from xserver successfully,I set the flag and other EDID features.
39369 commit 3354e660b0744976871683ce226f17e873f26b50
39370 Author: Keith Packard <keithp@keithp.com>
39371 Date: Wed Nov 26 16:32:12 2008 +0800
39373 Fix LVDS EDID to match all possible default modes
39375 If the EDID data from the LVDS doesn't indicate support for a wide range of
39376 continuous frequencies, it will not match any of the default modes although
39377 our LVDS scaler logic ignores refresh rates when programming LVDS modes. Fix
39378 this by smashing the compute EDID data to open up the sync rates.
39380 Signed-off-by: Keith Packard <keithp@keithp.com>
39382 commit f1e9ca4e4fb3ddb083252aea79c67c5e5e71f29c
39383 Author: Alan Coopersmith <alan.coopersmith@sun.com>
39384 Date: Fri Jan 9 16:30:15 2009 -0800
39386 Remove xorgconfig & xorgcfg from See Also list in man page
39388 commit 91c49067abab262edce54d33af8bfc03eee20f2e
39389 Author: Ian Romanick <ian.d.romanick@intel.com>
39390 Date: Wed Jan 7 21:45:54 2009 -0800
39392 Prevent redefinitions of CARD8 and friends. Fixes build.
39394 commit 934008a2cbcec0c790580750c672c0367a9e4c55
39395 Author: Eric Anholt <eric@anholt.net>
39396 Date: Tue Jan 6 10:55:59 2009 -0800
39398 Always enable KMS if server's new enough, and remove option.
39400 The API should be stable at this point, and we don't want to allow mistakes.
39402 commit 342120be0956bfc12822d1ffbfbd8aaabf3e922f
39403 Author: Eric Anholt <eric@anholt.net>
39404 Date: Mon Jan 5 23:21:07 2009 -0800
39406 Fix pin leakage with EXA GTT-mapping shortcut, and crash with UXA on KMS.
39408 commit 9a5082d2920c1a212fe935b5a093013e8035c321
39409 Author: Eric Anholt <eric@anholt.net>
39410 Date: Mon Jan 5 23:28:50 2009 -0800
39412 Disable DRI2 buffer tiling on non-965, as those need fence regs for 2D blits.
39414 This fixes glReadPixels failure on single-channel 915GM, as the software code
39415 for readpixels was actually the only code in the driver doing tiling against
39416 these buffers (everything else says "rely on fence registers", since the 2D
39417 blits don't have a "don't rely on fence registers" option).
39419 commit 7736b65be4fb4d5c59d7aedb1e64da976bb10ae9
39420 Author: Keith Packard <keithp@keithp.com>
39421 Date: Mon Dec 15 15:49:01 2008 -0800
39423 FatalError on batchbuffer map failure
39425 Yes, it would be nice to do something other than crash here.
39427 Signed-off-by: Keith Packard <keithp@keithp.com>
39429 commit 632f816c72cb4b48b690fd92d1cc1d5a9c9285c7
39430 Author: Keith Packard <keithp@keithp.com>
39431 Date: Mon Dec 15 15:43:34 2008 -0800
39433 uxa: handle uxa_prepare_access failure
39435 uxa_prepare_access may fail to map the pixmap into user space. Recover from
39436 this without crashing.
39438 Signed-off-by: Keith Packard <keithp@keithp.com>
39440 commit 90b28a56553d809374fa6d9b9529b7a8b583488c
39441 Author: Keith Packard <keithp@keithp.com>
39442 Date: Mon Dec 15 15:41:28 2008 -0800
39444 Handle drm_bo_map failure in 965 video and composite paths.
39446 These two paths allocate a number of objects directly.
39448 Signed-off-by: Keith Packard <keithp@keithp.com>
39450 commit 8237faf8f3ca73ecdf0ef009a7d361b318726f6f
39451 Author: Keith Packard <keithp@keithp.com>
39452 Date: Mon Dec 15 15:35:35 2008 -0800
39454 Resize framebuffer on screen size change (requires UXA and DRI2)
39456 Signed-off-by: Keith Packard <keithp@keithp.com>
39458 commit 21bd4e8974e4c0e83f5f95adb0fc17290444caf5
39459 Author: Keith Packard <keithp@keithp.com>
39460 Date: Mon Dec 15 15:19:00 2008 -0800
39462 Make i830_allocate_memory take tiling parameters.
39464 This eliminates the separate i830_allocate_memory_tiled function which means
39465 that all memory objects will have tiling parameters set correctly.
39467 Signed-off-by: Keith Packard <keithp@keithp.com>
39469 commit 1f61e97904dfe5f8c08bb9f284cfdfe878f7e541
39470 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39471 Date: Wed Dec 31 22:56:57 2008 +0800
39473 UXA: Fallback to dri_bo_map() if pin failed
39475 This fixes VT switch issue with UXA after Eric's
39476 aae4008096399a0e84abc7c016b35092caf9db25 on 2D side.
39478 commit 830bf916724afd21b7947f797c22a8c8aab7a0a4
39479 Author: Eric Anholt <eric@anholt.net>
39480 Date: Mon Dec 29 13:42:44 2008 -0800
39482 Don't touch the pipestat regs for detecting FIFO underrun. The kernel owns them.
39484 Since we don't perform any synchronization with the kernel on these regs, we
39485 could race with the kernel to write stale values and end up not having vblank
39486 interrupts enabled when somebody was waiting on one.
39488 commit d96f774d1bb39640486c72338fe8b19ee1ceaa23
39489 Author: Eric Anholt <eric@anholt.net>
39490 Date: Mon Dec 29 13:57:24 2008 -0800
39494 commit 59b0fbb9be880d489374b141f818948a2721a2ef
39495 Author: Henry unbongo <henryunbongo@yahoo.com>
39496 Date: Mon Dec 29 13:54:38 2008 -0800
39498 Add support for SDVO LVDS.
39500 commit 750d8e105831718d4a44a145fdb87571fa9f9d8e
39501 Author: Eric Anholt <eric@anholt.net>
39502 Date: Mon Dec 29 12:44:26 2008 -0800
39504 Fix compile failure after 45f45c73469f1bd46a1b6fb206f2e9e5e4fd66b3
39506 commit 45f45c73469f1bd46a1b6fb206f2e9e5e4fd66b3
39507 Author: Eric Anholt <eric@anholt.net>
39508 Date: Mon Dec 29 11:26:11 2008 -0800
39510 Remove logical context setup.
39512 This should be a noop. If it wasn't a noop, it means that on pre-g33 chipsets
39513 we were spamming some data into a page of system memory because we used a
39514 virtual instead of a physical address. It was also supposed to not work when
39515 we submit it from a batchbuffer, as we have been doing for some time now.
39516 This code has existed since about the beginning of the driver's existence,
39517 with no justification.
39519 commit 3544bbe22d8cf2640289e1e4febe755a47f26631
39520 Author: Eric Anholt <eric@anholt.net>
39521 Date: Mon Dec 29 10:41:02 2008 -0800
39523 Add PCI write posting to LeaveVT path when we're about to wait on write results.
39525 commit 7b67914b23b54d4d9566190440a3430e40615aa8
39526 Author: Eric Anholt <eric@anholt.net>
39527 Date: Fri Mar 28 15:51:50 2008 -0700
39529 Add SDVO LVDS register definitions.
39531 commit 8464fc285d7b22fd45b7af616fd52aa15e16638a
39532 Author: Eric Anholt <eric@anholt.net>
39533 Date: Mon Dec 29 09:51:45 2008 -0800
39535 Rely on libdrm 2.4.3 and stop checking for xf86drmMode.h.
39537 commit c1dde7ac06ce6470c74198b2560ee67d28fb0aea
39538 Author: Eric Anholt <eric@anholt.net>
39539 Date: Mon Dec 22 16:49:57 2008 -0800
39541 Remove old mergedfb includes and defines, which bother spatch.
39543 commit 75799d2834be84b016391ec95b221df32380e3e7
39544 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39545 Date: Tue Dec 30 00:55:30 2008 +0800
39547 Bug #18004: Add Aopen 915GM LVDS quirk
39549 commit cfaaf6af777ad8e56da5a077bdc01f4f2d7bc4d6
39550 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39551 Date: Mon Dec 29 21:17:44 2008 +0800
39553 Don't count vertex buffer in second aperture size check
39555 With batch flush notify vertex buffer will be unreferenced,
39556 so don't count it in later aperture check. Also adding
39557 uninitialized vertex buffer check in batch flush notify.
39559 commit b710a688a7383df320f9d4e765b48331310a4d1d
39560 Author: Dave Airlie <airlied@linux.ie>
39561 Date: Mon Dec 29 12:50:25 2008 +1000
39563 intel/kms: disable overlay when it needs physical address
39565 We can't do phy address allocations, need kernel support for this.
39567 Signed-off-by: Dave Airlie <airlied@redhat.com>
39569 commit 649374b88b330838133d78be1953ce8b18ddd2c6
39570 Author: Dave Airlie <airlied@linux.ie>
39571 Date: Mon Dec 29 12:48:11 2008 +1000
39573 intel: don't call enter/leave VT for KMS enabled systems
39575 Signed-off-by: Dave Airlie <airlied@linux.ie>
39577 commit 408f7139354a942f29334795b5480a0b6bb6142a
39578 Author: Dave Airlie <airlied@linux.ie>
39579 Date: Mon Dec 29 12:47:30 2008 +1000
39581 exa: fix map gtt call to use current API
39583 commit fcd280b730f86ea54ecd89d3e037e71f8375b5f5
39584 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39585 Date: Mon Dec 29 10:05:17 2008 +0800
39587 Bug #19239: Add a quirk for broken ACPI lid state
39589 For broken hardware/bios with incorrect ACPI LID state,
39590 there's machine that can not be fixed in ACPI way, customed
39591 DSDT that reprogram _LID method to read EC state. Although
39592 this is ACPI issue, this quirk can be used to work around that.
39594 commit 6b9f421b792c701e909d81ae2b6e6a47fb069b0b
39595 Author: Dave Airlie <airlied@linux.ie>
39596 Date: Mon Dec 29 11:51:58 2008 +1000
39598 modeset: transformPresent is a new API member
39600 commit a320541e51818833a6a445707835fbf70e9babd4
39601 Author: Dave Airlie <airlied@linux.ie>
39602 Date: Mon Dec 29 11:34:51 2008 +1000
39604 modeset: fix xf86CrtcRotate API change across server versions
39606 commit e38fd84fcccc18284b649a60b4cfd8e24eaf059d
39607 Author: Ma Ling <ling.ma@intel.com>
39608 Date: Mon Dec 22 10:35:52 2008 +0800
39610 SDVO: reset pixel repeat in avi frame
39612 For #19115, the root cause is avi_if.u.avi.PR in
39613 i830_sdvo_set_avi_infoframe() belongs to element for
39614 interlaced mode based on CEA_861B, but currently we
39615 don't support interlaced mode. So it should be set as 0.
39617 commit d8e89b26ef5ef2c15e5d34162b14d279a7f0bb1c
39618 Author: Zou Nan hai <nanhai.zou@intel.com>
39619 Date: Fri Dec 19 09:34:14 2008 +0800
39621 [965-xvmc] remove the vblank wait code, drm not support mutlple client
39622 to wait on vblank now.
39624 commit aae4008096399a0e84abc7c016b35092caf9db25
39625 Author: Eric Anholt <eric@anholt.net>
39626 Date: Wed Dec 17 14:25:22 2008 -0800
39628 uxa: Do a hack to use the aperture mapping instead of bo_map in sw fallbacks.
39630 Because of how fallbacky the uxa rendering core is, and our inability (without
39631 wfb in userland or page faulting in the kernel) to tell the kernel just where
39632 we're going to fall back, the clflush overhead can become outrageous, for
39633 example with emacs and xcompmgr. Instead of using drm_intel_bo_map, pin the
39634 buffer and do the fallback to the aperture mapping. This gets us the bad old
39635 performance that fb is designed for, instead of bad new performance.
39637 commit 74bc420940de4ed02705211ddebbc4033c032242
39638 Author: Peter Alfredsen <loki_val@gentoo.org>
39639 Date: Fri Dec 19 08:10:06 2008 +0800
39641 xvmc: fix up needed libs
39643 commit 555eea5411cf8c725df5f1b4cb80198fa6a1225b
39644 Author: Zou Nan hai <nanhai.zou@intel.com>
39645 Date: Wed Dec 17 13:55:07 2008 +0800
39647 wait vblank before render to fix tearing
39649 commit ed267072db7c58ee16a458fd3dc24ce3a2d0061d
39650 Author: Robert Noland <rnoland@2hip.net>
39651 Date: Wed Dec 17 09:27:30 2008 +0800
39653 Fix drmOpen for non-linux 965 XvMC
39655 drmOpen by name only works on linux after falling back to groping around
39656 in /proc. This doesn't work on other OS.
39658 Signed-off-by: Robert Noland <rnoland@2hip.net>
39660 commit ecdd706873c1f990f4a78fbaecf7380411edabcd
39661 Author: Eric Anholt <eric@anholt.net>
39662 Date: Sun Dec 14 16:09:25 2008 -0800
39664 uxa: Correctly prepare/finishaccess of stipple in ValidateGC (and only it)
39666 This avoids prepare/finish_access_gc overhead when we're not changing things
39667 (since GCTile is already handled) and get us the RW flag for the prepare on
39668 of the stipple pixmap so thing will be synced correctly.
39670 commit e6479f96e5d8da39fcbb5376c4a66a1f924ec4e4
39671 Author: Kristian Høgsberg <krh@redhat.com>
39672 Date: Mon Dec 15 14:54:20 2008 -0500
39674 Quiet some KMS warnings.
39676 commit 30fb0ef53e19b759715f1ee14b81b11c81d79045
39677 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39678 Date: Tue Dec 16 00:39:45 2008 +0800
39680 Revert " [965 xvmc] update dual prime g4b files"
39682 This reverts commit ea2b6b405e4c8b1bfb4bc568d0453a39a9194a8f.
39684 Duplicate with Keith's commit. No idea what's diff target
39687 commit ea2b6b405e4c8b1bfb4bc568d0453a39a9194a8f
39688 Author: Zou Nan hai <nanhai.zou@intel.com>
39689 Date: Mon Dec 15 09:19:43 2008 +0800
39691 [965 xvmc] update dual prime g4b files
39693 commit d917583c19e2eb20a559eddaa100ce71d8fbe48c
39694 Author: Pierre Willenbrock <pierre@pirsoft.de>
39695 Date: Mon Dec 15 09:08:30 2008 +0800
39697 closedir only after finishing use of any results from readdir
39699 commit 865735d8408ee6b75be8fb1a8ab32b2a53dd3adb
39700 Author: Keith Packard <keithp@keithp.com>
39701 Date: Fri Dec 12 11:07:20 2008 -0800
39703 Dump out fence registers by default, add fence end registers as well
39705 In debugging the frame buffer resize code, I needed to see what the server
39706 was doing to the fence registers, so I added this debug code. Seems useful
39707 enough to include it.
39709 Signed-off-by: Keith Packard <keithp@keithp.com>
39711 commit f3e59e59b54696667b127c82327e16b14d7bab22
39712 Author: Zou Nan hai <nanhai.zou@intel.com>
39713 Date: Fri Dec 12 11:18:25 2008 -0800
39715 Update binary versions of the dual-prime kernels
39717 Signed-off-by: Keith Packard <keithp@keithp.com>
39719 commit 80d588e9c49719ec494e705edfc14c78908bbc3f
39720 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39721 Date: Fri Dec 12 18:27:29 2008 +0800
39723 Don't warn on ring enabled in GEM
39725 commit 60c4ee9ece8ea57e61b1590dfeb69d08555e465c
39726 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39727 Date: Fri Dec 12 18:00:09 2008 +0800
39729 Fix directRenderingType check
39731 Don't miss classic texture memory allocation in DRI.
39733 commit edf765155497d89ecac328a5b268ecf60e2f377c
39734 Author: Zou Nan hai <nanhai.zou@intel.com>
39735 Date: Thu Dec 11 15:41:59 2008 +0800
39737 [965-xvmc] fix dual prime kernel, flush issue on G4x
39739 commit 80e2d90139dd99f50beb4f9353599608624b777d
39740 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39741 Date: Thu Dec 11 13:45:17 2008 +0800
39743 Let lid status be unknown if no acpi lid object found
39745 commit 83377b543defdca7226d7c1a7794e4ff3d8b4c84
39746 Author: Bryce Harrington <bryce@canonical.com>
39747 Date: Thu Dec 11 09:38:27 2008 +0800
39749 Pipe-A quirk for HP 2730p (bug #18852)
39751 commit 6c4e134a1a30785c2e5c6d57b21fde54a2dd3413
39752 Author: Bryce Harrington <bryce@canonical.com>
39753 Date: Thu Dec 11 09:34:15 2008 +0800
39755 PipeA quirk for Quanta/W251U (launchpad bug #244242)
39757 commit 1e974ff6b8446ecd64677b3c9aba60ca850923cc
39758 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39759 Date: Wed Dec 10 10:54:31 2008 +0800
39761 Try to always probe SDVOC on 965G/965GM
39763 Detect bit of SDVOC is reserved on 965G/965GM, instead of ignore SDVOC
39764 this trys to always probe it on these chipsets.
39766 commit f5f67e1b54e67b4bfc3db3482b2693211be81d63
39767 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39768 Date: Wed Dec 10 10:36:07 2008 +0800
39770 bug #17395: Quirk CRT for Sony VGC-LT71DB
39772 commit d8b764fbd27dc9c8b28386093931b8d38855bd19
39773 Author: Keith Packard <keithp@keithp.com>
39774 Date: Tue Dec 9 21:51:14 2008 -0800
39776 Add RandR 1.3 panning support by supporting the crtc set_origin function
39778 RandR 1.3 panning support can use the regular mode setting interface, but
39779 that's really slow. Providing set_origin makes it nice and snappy.
39781 Signed-off-by: Keith Packard <keithp@keithp.com>
39783 commit 4d7a95959d8223aec41550eb19f60b3edd7210a1
39784 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39785 Date: Tue Dec 9 19:48:11 2008 +0800
39787 Try to add LVDS detect support
39789 This one trys to use lid status for LVDS detect,
39790 which works when internal panel is not used as primary
39791 display alone, or there's no internal panel at all.
39792 ACPI button driver's lid state interface is preferred,
39793 and SWF state is also checked if ACPI method failed.
39795 commit 0fe61b0b7e3bbe8ced1b0ad2be72c438d200c64b
39796 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39797 Date: Tue Dec 9 19:19:58 2008 +0800
39799 Remove Cappuccino SlimPRO SP625F 855GM LVDS quirk
39801 It breaks bug #18462 on IBM 855GM with same subdevice ids.
39803 commit bea98cdfd93fc1181a06c51e57fcab227ff4827e
39804 Author: Carl Worth <cworth@cworth.org>
39805 Date: Fri Dec 5 15:42:53 2008 -0800
39807 Set vertex_buffer_bo to NULL after unreference.
39809 Which is just being tidy. We initially were looking at this code
39810 path due to a report of a crash on server shutdown which started
39811 after this unreference call was added. Setting this to NULL
39812 apparently didn't avoid the crash, but it's a good thing to do
39815 commit ce7efc2e3676c8f80206415480dda91e5e021396
39816 Author: Carl Worth <cworth@cworth.org>
39817 Date: Thu Dec 4 11:41:02 2008 -0800
39819 Don't smash fixed_mode if skip_panel_detect is set.
39821 Without this change, setting LVDSFixedMode to false is not effective
39822 as i830_bios_init calls i830_parse_panel_data which in turns sets
39823 a fixed_mode. To fix this we still call parse_panel_data to set
39824 the various lvds_options but we return before setting fixed_mode.
39826 commit e8b95efbf5d9c3a5b75b2bb8b5b51844b5fcdfbc
39827 Author: Carl Worth <cworth@cworth.org>
39828 Date: Wed Dec 3 13:49:52 2008 -0800
39830 i965: Add batch_flush_notify hook to create new vertex-buffer bo
39832 This avoids mapping a buffer object which is being referenced
39833 by a batch that has already been flushed, (which is a terribly
39834 expensive operation).
39836 On my machine this brings the performance of x11perf -aa10text
39837 from 85k back to 150k, (where it was before a recent kernel
39838 upgrade). Also, before this patch, when I used my X server
39839 actively performance would drop as low as 15k---hopefully that
39840 bug is gone with this change.
39842 commit 768f317cf0da4cd6682af2e71e71c3e130e05182
39843 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39844 Date: Mon Dec 8 10:33:13 2008 +0800
39846 Fix DRI2 compiling warning
39848 commit 95596f51503bb468364719aec9083d59999e34b7
39849 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39850 Date: Mon Dec 8 10:30:12 2008 +0800
39852 Fix TV compiling warning
39854 commit c47b6d1def917fad5ad2b5a3e4167edfd354f2c8
39855 Author: Kristian Høgsberg <krh@redhat.com>
39856 Date: Sat Dec 6 21:20:52 2008 -0500
39858 Make sure DRI/DRI2 can initialize properly with KMS.
39860 commit 70af658d4e94cc372f9e9c831611f70b3c1cecab
39861 Author: Kristian Høgsberg <krh@redhat.com>
39862 Date: Sat Dec 6 19:19:21 2008 -0500
39864 Simplify crtc preinit a bit.
39866 commit 3ad9c9a82d7b359b9b711070628e6ff07a2aa9f7
39867 Author: Kristian Høgsberg <krh@redhat.com>
39868 Date: Sat Dec 6 19:17:39 2008 -0500
39870 Fix KMS compilation.
39872 commit 2e3c098c5ed9a8451713dc754a5f086992249336
39873 Author: Eric Anholt <eric@anholt.net>
39874 Date: Fri Dec 5 12:21:53 2008 -0800
39876 uxa: Reject solid/copy to under-8bpp destinations.
39878 EXA wouldn't create pixmaps for under-8bpp, but UXA does. Fixes
39879 mis-rendering in xfwm, evolution message compose, firefox link
39880 drag'n'drop, and I'm sure more. Big thanks to Pierre Willenbrock for
39881 debugging the issue!
39885 commit 261c20a479f6ec1e94c2ba801323072227cc3ade
39886 Author: Eric Anholt <eric@anholt.net>
39887 Date: Fri Dec 5 12:13:26 2008 -0800
39889 uxa: Add in EnableDisableFBAccess handling like examodule.c did.
39891 This fixes assertion failures when rendering text while VT switched.
39893 commit caecd6031e416705b1f0a7051535211feaebdedd
39894 Author: Eric Anholt <eric@anholt.net>
39895 Date: Thu Dec 4 15:18:08 2008 -0800
39897 Remove the extra memory allocation for 965 video state now that it's all in BOs.
39899 commit 48c113ade26d5ad64999cd5cc288495c10e02fc4
39900 Author: Eric Anholt <eric@anholt.net>
39901 Date: Thu Dec 4 15:10:49 2008 -0800
39903 Emit proper relocations to pixmaps in BOs in i965 video.
39905 commit ae2cd8b75e41393e072e27064ecefecf1aa1bc6c
39906 Author: Eric Anholt <eric@anholt.net>
39907 Date: Thu Dec 4 15:02:12 2008 -0800
39909 Move i965 video surface state and binding table to BOs.
39911 commit 1b3c3c9d79305abe1785fdaef26a950dafa8890f
39912 Author: Eric Anholt <eric@anholt.net>
39913 Date: Thu Dec 4 14:30:56 2008 -0800
39915 Move i965 video vertex data to BOs.
39917 This eliminates extra syncing when clipping is involved.
39919 commit 84825972597042f0aa6784594dace96be96a0234
39920 Author: Eric Anholt <eric@anholt.net>
39921 Date: Thu Dec 4 14:26:09 2008 -0800
39923 Move remaining i965 video programs to BOs.
39925 commit 40671132cb3732728703c6444f4577467fa9223f
39926 Author: Eric Anholt <eric@anholt.net>
39927 Date: Thu Dec 4 14:20:21 2008 -0800
39929 Move i965 video wm and sampler state to BOs.
39931 commit 48803eb7463ad14f3109f67fcf4ccff4362baaa2
39932 Author: Eric Anholt <eric@anholt.net>
39933 Date: Thu Dec 4 13:22:12 2008 -0800
39935 Stop allocating unused scratch space for i965 video.
39937 commit f3fe46b557a4dd7e212c2790fb47142d375c1c75
39938 Author: Eric Anholt <eric@anholt.net>
39939 Date: Thu Dec 4 13:18:06 2008 -0800
39941 Move i965 video vs/sf state to BOs.
39943 commit 4b9d3eac57a972c055c4acd7a10dfe8aa918131c
39944 Author: Eric Anholt <eric@anholt.net>
39945 Date: Thu Dec 4 13:11:13 2008 -0800
39947 Move i965 video cc state to BOs.
39949 commit 176e92d89fdfb199780014722feab6ac25836dcc
39950 Author: Eric Anholt <eric@anholt.net>
39951 Date: Thu Dec 4 12:53:37 2008 -0800
39953 Move I965DisplayVideoTextured unit state setup to separate functions.
39955 commit 61929f4c641e2ecb145ad2b22f7092d40e31ae6d
39956 Author: Eric Anholt <eric@anholt.net>
39957 Date: Thu Dec 4 11:44:35 2008 -0800
39959 Move I965DisplayVideoTextured surface/sampler setup to separate functions.
39961 commit 73d03cb93ca761ee555b87558882e26bd2f10d91
39962 Author: Eric Anholt <eric@anholt.net>
39963 Date: Thu Dec 4 11:37:38 2008 -0800
39965 Move debug code for I965DisplayVideoTextured to separate functions.
39967 commit df01d8f0791bd825c96d9d5e20313d5ffcfeb532
39968 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39969 Date: Thu Dec 4 16:11:33 2008 +0800
39971 Change master version to 2.6.99.1
39973 commit b156b3165e1aae5df0353737d0335ac2e653f5fd
39974 Author: Julien Cristau <jcristau@debian.org>
39975 Date: Wed Dec 3 04:35:38 2008 +0100
39977 Don't unconditionally define DRI2
39979 xorg-server.h will define it if the server has dri2 support.
39981 commit 8d7cbab267e8fbcb2fcf90b18346b60607277266
39982 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39983 Date: Thu Dec 4 18:43:03 2008 +0800
39985 Fix maxium backlight level restored issue
39987 The origin check for bring back max value for '0'
39988 backlight level is ok for legacy or combo control method
39989 as '0' mostly doesn't act in ideal lowest level. But it
39990 breaks in using kernel control method which should provide
39991 a reasonable backlight range.
39993 This is tested fine on T61 with thinkpad_acpi module.
39995 commit 32e7d47925b445a85847d77651b258e117aaaffe
39996 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
39997 Date: Thu Dec 4 09:24:31 2008 +0800
39999 Disable XvMC on DRI2 for now
40001 commit 293f6232c6c0ce12802a2123edfab4551d25e266
40002 Author: Dave Airlie <airlied@linux.ie>
40003 Date: Wed Dec 3 16:51:25 2008 -0800
40005 uxa: don't call composite routines with no buffer.
40007 We can get a case with gnome-terminal + links, where we get two arrays
40008 of glyphs all with 0 width and 0 heights in them. If this happens
40009 we manage to get to this case without any buffer setup and segfault.
40011 (cherry picked from commit 717c7492a0f6ba3fb3eabda33515881eef314155)
40013 commit 457a680afd0d8f835131ea72be0c3c618c2892c7
40014 Author: Eric Anholt <eric@anholt.net>
40015 Date: Wed Dec 3 13:17:26 2008 -0800
40017 Enable tiling for DRI2 back/depth buffers.
40019 This results in allocation overhead for small (8x8-128x128 or so) pixmaps with
40020 DRI2, but we're interested in looking at tiling them in general in the near
40023 commit b662ecccb5c036fcc4aa19026642bde0a1ca2ac8
40024 Author: Eric Anholt <eric@anholt.net>
40025 Date: Wed Dec 3 10:50:33 2008 -0800
40027 Re-enable composite accel on 965 with UXA.
40029 This was blocked on getting 965 render aware of BO pixmaps, which is happily
40032 commit d426f799f541cb32fab134f2af068c6cd31923d9
40033 Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
40034 Date: Wed Dec 3 14:59:21 2008 -0200
40036 Include <X11/Xfuncproto.h> prior to including "edid.h" from the sdk.
40038 commit 07f5a8223187c1abc79c104d2fa5859a54cecd30
40039 Author: Robert Lowery <rglowery@exemail.com.au>
40040 Date: Wed Dec 3 09:48:23 2008 +0800
40042 TV: add support to set TV margins in xorg.conf
40044 commit 00ae7a571b6413aa2530e4f1310f8f4646631946
40045 Author: Eric Anholt <eric@anholt.net>
40046 Date: Tue Dec 2 13:08:23 2008 -0800
40048 UXA: Add support for tiled front/back/depth by cutting over to the GTT map.
40050 commit 34d54db945c67a2fee0a38cc9eafc463b8413669
40051 Author: Eric Anholt <eric@anholt.net>
40052 Date: Tue Dec 2 13:01:56 2008 -0800
40054 Remove DRI_MM defines which are always true now.
40056 commit f7540df184a75a749cabcfa2da8974d3d0714ea6
40057 Merge: 5f34702 ee87f7b
40058 Author: Keith Packard <keithp@keithp.com>
40059 Date: Mon Dec 1 19:06:09 2008 -0800
40061 Merge commit 'origin/dri2'
40063 commit 5f347020a52b583e19d333361ea7bf2591255899
40064 Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
40065 Date: Sat Nov 29 02:35:59 2008 -0200
40067 Export libIntelXvMC and libI80XvMC symbols.
40069 This is a side effect of adding the visibility flags to XORG_CFLAGS.
40071 commit ec7fd3d1b477928f9bffde1a72d7f3fdea89e629
40072 Merge: d978cd4 baec75a
40073 Author: Zou Nan hai <nanhai.zou@intel.com>
40074 Date: Thu Nov 27 10:48:09 2008 +0800
40076 Merge branch '965-xvmc'
40084 commit d978cd4b453ea588ed2fc2f2cb4ec26856fe00d4
40085 Author: Wu Fengguang <wfg@linux.intel.com>
40086 Date: Thu Nov 20 00:47:25 2008 -0800
40088 refresh batch_bo reference after intel_batch_flush()
40090 The call into intel_batch_flush() will invalidate the pI830->batch_bo
40091 stored in bo_table[0]. Fix it by re-read the refreshed value.
40093 Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
40094 Signed-off-by: Eric Anholt <eric@anholt.net>
40096 commit ee87f7becc2138348ec0d948829db6e86cee86da
40097 Merge: 3dd7f0f a5b1e62
40098 Author: Kristian Høgsberg <krh@redhat.com>
40099 Date: Wed Nov 12 16:40:37 2008 -0500
40101 Merge commit 'origin/master' into dri2
40108 commit 30c29627c859f1f6780aa652d6de4fa985b63dd0
40109 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40110 Date: Thu Nov 13 12:35:02 2008 -0800
40112 Make sure DRM library paths are included
40114 We need @DRM_LIBS@ in order to find drm_intel if installed in a
40117 commit 42ec6938bdea3a5b065a4248fdca5f031415191e
40118 Author: Ma Ling <ling.ma@intel.com>
40119 Date: Fri Nov 7 14:23:41 2008 +0800
40121 enable Intel G35 SDVO HDMI audio output
40123 Set the SDVO_AUDIO_ENABLE bit to enable SDVO HDMI audio output of the
40126 Signed-off-by: Ma Ling <ling.ma@intel.com>
40127 Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
40129 commit 42b3903c3fda95a7c536cb957d8163d8dc54a9a1
40130 Author: Wu Fengguang <wfg@linux.intel.com>
40131 Date: Fri Nov 7 14:23:40 2008 +0800
40133 enable Intel G45 integrated HDMI audio output
40135 Enable audio output for the integrated HDMI of Intel G45 chipset
40136 by introducing the SDVO_AUDIO_ENABLE bit.
40138 Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
40140 commit 293f4a8b6d1ca8a47d14c4d4b5904b85a8167e12
40141 Author: Wu Fengguang <wfg@linux.intel.com>
40142 Date: Thu Nov 13 09:50:37 2008 +0800
40144 introduce i830_hdmi_priv.has_hdmi_sink
40146 HDMI is compatible with DVI, and we've seen many boards that
40147 use HDMI port for DVI output.
40149 So Zhenyu proposed this flag: i830_hdmi_priv.has_hdmi_sink
40150 to indicate the presence of HDMI capable monitors.
40152 Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
40154 commit 667923559219429b0c5fec12a0164f7eba1f8f2d
40155 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40156 Date: Tue Nov 11 12:55:06 2008 -0800
40158 Default to FULL_ASPECT panel fitting
40160 Now that 8xx is fixed, we should be able to preserve aspect ratio by
40163 Fixes fdo bz #18033.
40165 commit 65cd0fbb018b2c18f1571dc0924c7d92eaf794ad
40166 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40167 Date: Tue Nov 11 09:36:50 2008 +0800
40169 TV: fix contrast and saturation for 915/945G
40171 915/945G uses exponent-mantissa format instead of
40172 fixed-point format on 965G.
40174 commit a5b1e62337d4e8840347bb186db48697f0690a19
40175 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40176 Date: Fri Nov 7 16:06:26 2008 +0800
40178 quirk LVDS on Asus Eee box
40180 Reported by Yang Sheng <sheng.yang@intel.com>
40182 commit 9942cfa6dcc70a09ea38f738b1e73e3f005080b9
40183 Author: Keith Packard <keithp@keithp.com>
40184 Date: Thu Nov 6 15:04:27 2008 -0800
40186 Use long crt hotplug activation time on GM45.
40188 The GM45 b-spec requires the use of the longer hotplug activation period,
40189 but does not require looping twice over the detection logic. With this
40190 patch, CRT detection appears solid on my GM45.
40192 Signed-off-by: Keith Packard <keithp@keithp.com>
40194 commit 226efa7c74c0ed8a7b974a3727c5d7322c8ef815
40195 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40196 Date: Wed Oct 29 22:41:20 2008 +0800
40198 TV: subcarrier fix for NTSC and PAL
40200 Using subcarrier DDA values from spec give more close result to
40201 standard frequency.
40203 commit 3651341292d90b7ded4c3f013bcb0f46537a113a
40204 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40205 Date: Wed Oct 29 20:41:46 2008 +0800
40207 TV: fix timing parameters for PAL, 480p, 1080i
40209 commit b404afb755b608b02bcf0be1f8fe8a38d3d7bc1e
40210 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40211 Date: Wed Oct 29 20:32:18 2008 +0800
40213 TV: save serveral TV_CTL register fields in mode set
40215 Driver should keep those fields according with spec.
40217 commit 2ae91f0ffdadfb393d526b94e21914a31aa14232
40218 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40219 Date: Wed Oct 29 20:26:44 2008 +0800
40221 TV: fix default contrast and saturation modifier
40223 Color knobs was set with higher modifier which caused strong color
40224 on TV screen. Setting fixed point modifier to default 1.0 makes picture
40227 commit c4cab00ef7f57fc27776f53263aacec2edf6f959
40228 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40229 Date: Wed Oct 29 20:26:03 2008 +0800
40231 TV: white space cleanup
40233 commit b130f94f26111e1a12ce869292915504588d0369
40234 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40235 Date: Thu Nov 6 10:14:00 2008 +0800
40237 SDVO: fix more command definition errors
40239 commit 3ab979da5c867fb862e33a80b973a35d56d56a55
40240 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40241 Date: Mon Nov 3 16:05:28 2008 +0800
40243 SDVO: add command for set monitor power state
40245 SDVO encoder power state must be higher than monitor's state.
40246 This only adds command and doesn't enable monitor power state
40249 commit f558a4972d1cca03785da025170259c61541ee63
40250 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40251 Date: Mon Nov 3 16:01:40 2008 +0800
40253 SDVO: add GetScaledHDTVResolutionSupport command
40255 commit cd8cbff5d7921d2f4fdac106c3d54414310af94d
40256 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40257 Date: Mon Nov 3 15:20:18 2008 +0800
40259 SDVO: fix sdvo tv format and sdtv resolution request/reply definition
40261 commit 61672b8c458ecf449138c51aeb12cf24dbbd59d9
40262 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40263 Date: Mon Nov 3 14:59:53 2008 +0800
40265 SDVO: add HDMI audio encrypt change bit for GetInterruptEventSource command
40267 commit 45fb04474d0e67643a633dcc3aeb4efe7fd71ea1
40268 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40269 Date: Mon Nov 3 14:51:14 2008 +0800
40271 SDVO: fix wrong order of sdvo version's major/minor
40273 commit 5cbf1e72711e906b5309694045b56933e14dca06
40274 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40275 Date: Thu Nov 6 09:46:54 2008 +0800
40277 Make IS_GM45 into IS_G4X define
40279 Those are identical that using one define is much clear.
40280 And it can also apply fixes for GM45 too, which is missing with
40283 commit 78b6cff3b6ec9513ede3e71d81a828967e6c0068
40284 Author: Eric Anholt <eric@anholt.net>
40285 Date: Wed Nov 5 14:39:19 2008 -0800
40287 i965: Support render acceleration with pixmaps in BOs.
40289 This required reordering the relocation emits for surface/binding table
40290 so that we didn't add new relocations to things that had already been
40291 relocated at (the check_aperture requirement).
40293 commit 801f0eac4f7b0071ed21eb9c68c5a951e1cc409a
40294 Author: Eric Anholt <eric@anholt.net>
40295 Date: Wed Nov 5 13:29:39 2008 -0800
40297 Make I830FALLBACK debugging a runtime instead of compile-time option.
40299 commit d828b0802c939cb12664db4dd157a4944a4db01b
40300 Author: Maxim Levitsky <maximlevitsky at gmail.com>
40301 Date: Wed Nov 5 13:47:09 2008 -0800
40303 Add an option to make the overlay be the first XV adaptor.
40305 commit 040d9bf9d8748d1ed8f977a6356d198def978b51
40306 Author: Eric Anholt <eric@anholt.net>
40307 Date: Thu Oct 30 13:55:07 2008 -0700
40309 Remove the CheckDevices timer.
40311 This was an early attempt to support display switch hotkey. It never really
40312 worked, as the events didn't show up when they should. It also cost a wakeup
40313 every second to check just in case, instead of taking an interrupt like we can
40314 now using ACPI. Additionally, some machines apparently get stuck with the
40315 event set and end up DDCing checking for new devices every second and
40316 smashing your video modes.
40318 commit 08914cceda6c57530023cdcdb5ad7e4024f36a6e
40319 Author: Carl Worth <cworth@cworth.org>
40320 Date: Thu Oct 30 16:46:06 2008 -0700
40322 Use buffer objects for binding table and surface-state objects.
40324 Instead of having a static array for these and doing an ugly sync
40325 everytime we recycle the array, we now simply allocate short-lived
40326 buffer objects for this dynamic state. The dri layer, in turn, can
40327 take care of efficiently reusing objects as necessary.
40329 On a GM965 this change was tested to improve the performance of
40330 x11perf -aa10text from roughly 120000 to 154000 glyphs/sec.
40332 commit 47cc3d79da8174ba30ca130b0fb6c7d9c871caed
40333 Author: Carl Worth <cworth@cworth.org>
40334 Date: Thu Oct 30 16:53:57 2008 -0700
40336 Unreference the vertex_buffer_bo in gen4_render_state_cleanup
40338 This avoids leaking one buffer object.
40340 commit 88700acf30f9eab8f96c197c7d113ce38c0af6e7
40341 Author: Carl Worth <cworth@cworth.org>
40342 Date: Mon Oct 27 14:23:02 2008 -0700
40344 Rename gen4_dynamic_state to gen4_vertex_buffer
40346 We don't actually plan to put any other data in this structure, so it
40347 doesn't make sense to have a generic name, (since we'll only be using
40348 it for our vertex buffer).
40350 commit 9e95722763e2379d14a6b46c3750a44713da5135
40351 Author: Carl Worth <cworth@cworth.org>
40352 Date: Thu Oct 23 15:43:19 2008 -0700
40354 965: Move composite setup to new _emit_batch_header_for_composite
40356 This function is the new name for _allocate_dynamic_state now that
40357 it also emits everything to the batch necessary for setting up a
40358 composite operation. This happens in prepare_composite() every
40359 time and in composite() whenever our vertex buffer fills up.
40361 It's not yet strictly necessary to be redoing this setup in
40362 composite() but it will be soon when the setup starts referring
40363 to buffer objects for surface state and binding table. This
40364 move prepares for that.
40366 commit fcb2a5a1253c505913e66b08107c0a9f57b07bad
40367 Author: Carl Worth <cworth@cworth.org>
40368 Date: Wed Oct 22 17:12:47 2008 -0700
40370 Use buffer object for vertex buffer (in new gen4_dynamic_state)
40372 This begins the process of separating the dynamic data from the
40373 static data, (still to move are the surface state and binding
40374 table objects). The new dynamic_state is stored in a buffer
40375 object, so this patch restores the buffer-object-for-vertex-buffer
40376 functionality originally in commit 1abf4d3a7a and later reverted
40379 A notable difference is that this time we actually do use
40380 check_aperture_space to ensure things will fit, (assuming
40381 there's a non-empty implementation under that).
40383 commit 7e68786cf746b6b984b184080578db1947b268ed
40384 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40385 Date: Thu Oct 30 13:40:23 2008 -0700
40387 Don't modify render standby if kernel mode setting is active
40389 Since we can't bang on registers directly in that case.
40391 commit 87ea531c5dc5b39809395b277c330854aaaaf019
40392 Author: Adam Jackson <ajax@redhat.com>
40393 Date: Fri Oct 24 14:47:27 2008 -0400
40395 Quirk: No LVDS on Dell Studio Hybrid
40397 commit 4d5360b1f3666bca4ed44c7f1b9217b8157c74b9
40398 Author: Carl Worth <cworth@cworth.org>
40399 Date: Wed Oct 22 16:17:39 2008 -0700
40401 Use consistent idiom for obtaining static_state
40403 This doesn't make any difference, but it's cleaner to have
40404 each function follow the same idiom for obtaining these pointers.
40406 commit 36fc53d068673497e77796ba69e010d08c359447
40407 Author: Carl Worth <cworth@cworth.org>
40408 Date: Wed Oct 22 15:45:51 2008 -0700
40410 Rename state_base_offset to static_state_offset in gen4_static_state_init
40412 More cleanup here, and again no functional change.
40414 commit 8fda7df0071046d057eeaf74328668b227af52a9
40415 Author: Carl Worth <cworth@cworth.org>
40416 Date: Wed Oct 22 15:28:24 2008 -0700
40418 Rename gen4_state_init to gen4_static_state_init
40420 This follows naturally from the structure rename.
40422 Also we make things less muddled by having this function
40423 actually accept a pointer to a gen4_static_state_t rather
40424 than a gen4_state_t, (and then fetching the desired pointer
40427 Again, no intended change in functionality here.
40429 commit acba4238116f2ead7be88d49c4c56c7cd6d7f3b8
40430 Author: Carl Worth <cworth@cworth.org>
40431 Date: Wed Oct 22 15:12:57 2008 -0700
40433 Rename gen4_state_t to gen4_static_state_t
40435 It doesn't contain only static data yet, but it will soon, so
40436 this renaming prepares for that. Also, this helps make things
40437 more clear between gen4_render_state_t and gen4_state_t which
40438 were muddled before, (particularly because the corresponding
40439 identifiers were render_state and card_state). The card_state
40440 identifier is now known as static_state which should be less
40443 This change is strictly search-and-replace with no functional
40446 commit c6550fa53d92c21dc677b86ebcd528a4ac640764
40447 Author: Carl Worth <cworth@cworth.org>
40448 Date: Wed Oct 22 14:51:42 2008 -0700
40450 Ignore intel_gtt binary
40452 We simply forgot to add this to .gitignore when it was created.
40454 commit ee3df629eaa388f09bec33ad04d443359d63ea7f
40455 Author: Dave Airlie <airlied@linux.ie>
40456 Date: Mon Oct 20 18:46:49 2008 -0700
40458 Default kernel mode setting to off, add configure flag to enable
40460 Should help avoid unpleasantness.
40462 commit 55cb65c6fdcf932d09e97abfa5374ee574676798
40463 Author: Eric Anholt <eric@anholt.net>
40464 Date: Sun Oct 19 21:22:37 2008 -0700
40466 Enable Option "Legacy3D" for 965 as well, and clarify both the docs and code.
40468 Signed-off-by: Eric Anholt <eric@anholt.net>
40469 Acked-by: Keith Packard <keithp@keithp.com>
40471 commit 28bb056f8cd326770590157412835318863b4134
40472 Author: Keith Packard <keithp@keithp.com>
40473 Date: Sat Oct 18 19:37:09 2008 -0700
40475 Use pipes, not planes for sarea geometry data
40477 Vblank in the kernel is far simpler if it deals with pipes instead of
40478 planes, so we're changing both user and kernel side. This is the user mode
40479 side. This fixes both i830_crtc_dpms and i830PipeSetBase, the two functions
40480 which copy geometry from the crtc to the sarea.
40482 Signed-off-by: Keith Packard <keithp@keithp.com>
40484 commit 546e2aca5b8b5d9f486d119ea2872ccd2b056054
40485 Author: Keith Packard <keithp@keithp.com>
40486 Date: Tue Oct 14 14:38:05 2008 -0700
40488 Reduce incidence of MI_FLUSH usage.
40490 This tracks whether the last command in each batch is an MI_FLUSH command
40491 and avoids appending another MI_FLUSH in the non-GEM cases.
40493 Signed-off-by: Keith Packard <keithp@keithp.com>
40495 commit 9b91cdf7ac88982880c887ffb0811c92157a7945
40496 Author: Keith Packard <keithp@keithp.com>
40497 Date: Sun Oct 19 14:24:46 2008 -0700
40499 No need to check for drmCommandWriteRead; we link against libdrm
40501 My checks for DRM enabled were a bit too extensive; drmCommandWriteRead is
40502 part of libdrm which the driver is always linked against. Only the symbols
40503 in the DRI module need to be checked here.
40505 Signed-off-by: Keith Packard <keithp@keithp.com>
40507 commit c25437a69490b09acff335a81a38aa540f0822f5
40508 Author: Keith Packard <keithp@keithp.com>
40509 Date: Sun Oct 19 13:39:04 2008 -0700
40511 Remove accidental 'return;' left at the top of I830DRIInitBuffers
40513 I was testing the behaviour of the XAA-based DRI buffer drawing code for
40514 tiled buffers and accidentally left I830DRIInitBuffers disabled.
40516 Signed-off-by: Keith Packard <keithp@keithp.com>
40518 commit ae224be27bb4432f25dae851debea6966bf31261
40519 Author: Keith Packard <keithp@keithp.com>
40520 Date: Tue Oct 14 14:35:36 2008 -0700
40522 Check for DRI module before initializing GEM mm
40524 GEM requires the DRI extension module currently, so make sure that is loaded
40525 (by checking for the DRIQueryVersion symbol) before trying to call it. This
40526 allows the server to start with the DRI extension disabled.
40528 Signed-off-by: Keith Packard <keithp@keithp.com>
40530 commit 7ddea0447c8972104d43cd7966f5ce89b4cca20c
40531 Author: Keith Packard <keithp@keithp.com>
40532 Date: Fri Oct 17 00:07:09 2008 -0700
40534 Handle differently tiled front/back/depth/third in DRI window management
40536 When moving or clearing the extra buffer contents associated with DRI
40537 windows, the XAA code needs to see which buffer is being manipulated in the
40538 Setup functions to program the tiling values correctly. Calling
40539 I830SelectBuffer and not then calling I830Setup... would result in mis-tiled
40542 Signed-off-by: Keith Packard <keithp@keithp.com>
40543 Signed-off-by: Eric Anholt <eric@anholt.net>
40545 commit c946383afc644ae7740e3c3146424fdd86c05285
40546 Author: Keith Packard <keithp@keithp.com>
40547 Date: Fri Oct 17 00:07:08 2008 -0700
40549 XAA tiling support was mis-computing adjusted pitch (>>4 instead of >>2)
40551 This may well explain why XAA never worked well on tiled front buffers;
40552 tiled buffers require a different pitch programming on 965 than non-tiled
40553 buffers, in dwords rather than bytes.
40555 Signed-off-by: Keith Packard <keithp@keithp.com>
40556 Signed-off-by: Eric Anholt <eric@anholt.net>
40558 commit 871728a0cbba12579c830e67020cc0c69c1611bb
40559 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40560 Date: Thu Oct 16 10:35:17 2008 +0800
40562 Don't handle irq in GEM mode
40564 commit 2419bce9efbff63529074e64af5ec5c2e62e368b
40565 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40566 Date: Thu Oct 16 10:28:16 2008 +0800
40568 Make GTT dumper work on other 9XX chips
40570 commit 986c8df79e83c369655ad786a6bf6342b53c535d
40571 Author: Adam Jackson <ajax@nwnk.net>
40572 Date: Thu Oct 16 10:24:04 2008 +0800
40574 Fix Mac mini crash in DDC mode probe
40576 SDVO on Mac mini trys to get EDID from CRT port, which
40577 failed with recent DVI-I change.
40579 commit 6707371176147340fabc9ab6f1e3d6d5ac980662
40580 Author: Eric Anholt <eric@anholt.net>
40581 Date: Wed Oct 15 08:12:11 2008 -0700
40583 Remove gratuitous flushing in EXA after solid operations.
40585 This snuck in with the UXA rename commit.
40587 commit 3dd7f0f9423bb891bc99cd3b77dc3d57e057a7ef
40588 Author: Kristian Høgsberg <krh@redhat.com>
40589 Date: Tue Oct 14 23:57:10 2008 -0400
40591 Update to DRI2 changes.
40593 Implement DRI2CopyRegion and provide drm device filename to DRI2 module.
40595 commit 282f51c3f0e5bc2cedd2f60f458ca2662290d471
40596 Merge: 8a54e3b 4dd0068
40597 Author: Kristian Høgsberg <krh@redhat.com>
40598 Date: Tue Oct 14 23:13:40 2008 -0400
40600 Merge commit 'origin/master' into HEAD
40602 commit 4dd00681dd0f9fce8dfd4592b46418edbbd2eeb4
40603 Author: Eric Anholt <eric@anholt.net>
40604 Date: Tue Oct 14 11:33:33 2008 -0700
40606 Fix broken stolen memory counting on G4X.
40608 On the GM45 we were assuming too little stolen memory (mostly harmless,
40609 except when it wasn't, until the AGP fix), and on the G45 we were assuming too
40610 much stolen memory, which was quite harmful when we touched the page that
40613 Future stolen memory accounting should use src/reg_dumper/intel_gtt before and
40614 after enabling AGP on the chipset to confirm that only the GTT entries not
40615 mapped to stolen are replaced, and that all of the unmapped GTT entries are
40616 replaced with the constant scratch page.
40618 commit e7e49bed7e254256f8cc0d4afcdfadc6dadf19e6
40619 Author: Eric Anholt <eric@anholt.net>
40620 Date: Tue Oct 14 10:38:05 2008 -0700
40622 Add a GTT dumper for G4x debugging.
40624 commit 1c6ea4ab4772453da915306045add8d147d567f2
40625 Author: Julien Cristau <jcristau@debian.org>
40626 Date: Sun Oct 12 15:10:22 2008 +0200
40630 Reported by Tomas Carnecky on IRC
40632 commit 6cb4150160bb1e1365773561fb53294ad9248a0e
40633 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40634 Date: Sun Oct 12 08:07:44 2008 +0800
40636 Bug #17892: Fix possible crash in CRT probe
40638 Introduced in patch for DVI-I CRT probe, if no EDID
40639 CRT monitor is connected, origin code destroys default
40640 DDC bus which causes crash in later get_modes call.
40641 Change it to setup and destroy DDC bus as needed in
40642 get_modes, so we always reprobe and get current state.
40644 commit 140451abb41fbd68d5aac4736c1cf1ec2805dd9e
40645 Author: Keith Packard <keithp@keithp.com>
40646 Date: Thu Oct 9 18:18:34 2008 -0700
40648 For non-DRM, add NOOPs after BATCH_BUFFER_START to verify completion
40650 In non-DRM mode, the driver waits for the hardware by checking the ring
40651 pointers; when the ring is empty, it assumes the hardware is idle. However,
40652 the hardware updates the ring pointers before executing a command, so if the
40653 MI_BATCH_BUFFER_START is the last command in the ring, the driver will think
40654 the hardware is idle while it may still be processing the contents of the
40655 batch buffer. Placing NOOPs after the BATCH_BUFFER_START allows the driver
40656 to know that the hardware has completed the batch buffer.
40658 Signed-off-by: Keith Packard <keithp@keithp.com>
40660 commit d09d938bf26ea0c44352ff6e7a57ba43f4047fdb
40661 Author: Keith Packard <keithp@keithp.com>
40662 Date: Thu Oct 9 18:17:13 2008 -0700
40664 i830 nondrm batch buffer insertion was missing ADVANCE_LP_RING() call
40666 The ring commands to insert a batch buffer to the ring in i830 form were not
40667 terminated by a call to ADVANCE_LP_RING(). This surely would have caused
40670 Signed-off-by: Keith Packard <keithp@keithp.com>
40672 commit d24010b7b3f2419beb40dc5ae1e8aeb3e04b5a93
40673 Author: Carl Worth <cworth@cworth.org>
40674 Date: Thu Oct 9 14:29:04 2008 -0700
40676 Disable frame buffer compression by default for GM965.
40678 We haven't found a way to make FBC work reliably with GM965 yet,
40679 (it often fails to notice CPU writes). This appears to be a
40680 specific problem with this device, (as we haven't gotten similar
40681 bug reports for subsequent devices such as GM45). So FBC is
40682 now disabled by default for GM965 but can still be enabled with
40683 the FrameBufferCompression option for experimenting/debugging.
40685 This resolves bug #16257:
40687 [GM965 EXA] Frame-buffer compression broken for CPU writes (XPutImage)
40688 https://bugs.freedesktop.org/show_bug.cgi?id=16257
40690 commit 4128b01ec84ffa0f03c335c36738f383376d9381
40691 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40692 Date: Thu Oct 9 09:19:41 2008 +0800
40694 Remove Lenovo T61 TV quirk
40696 BIOS already says no integrated TV, and this quirk blocks TV
40697 on R61i which has the same subdevice id as T61.
40699 commit 74571363539426abeb0a1af11f3bb545d91ed6c2
40700 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40701 Date: Wed Oct 8 15:36:10 2008 +0800
40703 Put forware VBIOS data parsing
40705 i830_bios_init() is called too late after output init, which
40706 makes bios data mostly useless, and caused all TV init fail as
40707 tv_present flag is not set.
40709 commit 111dd151dedf535464d638461f9485ec2939780e
40710 Author: Lukas Hejtmanek <xhejtman@ics.muni.cz>
40711 Date: Tue Oct 7 15:07:08 2008 +0200
40713 Fix driver build against server master.
40715 Broken by 3a4151b69daa478ac6edf042d604ee41e8429c0d
40717 commit 3a4151b69daa478ac6edf042d604ee41e8429c0d
40718 Author: Eric Anholt <eric@anholt.net>
40719 Date: Mon Oct 6 17:00:08 2008 -0700
40721 Fix driver build against server 1.4.2.
40723 This disables UXA and DRM modesetting pre-1.5, due to privates handling
40726 commit 3621183cf4acef23414e8d69c34b1e587f52ec67
40727 Author: Eric Anholt <eric@anholt.net>
40728 Date: Mon Oct 6 16:59:16 2008 -0700
40730 Fix bios_reader build against old servers.
40732 The server's pci_device structure ends up conflicting with libpciaccess's.
40733 Just don't include the server structure for this non-server tool.
40735 commit 080d36648f92d3b9a559e5458687fa82afab411f
40736 Author: Eric Anholt <eric@anholt.net>
40737 Date: Wed Oct 1 16:13:09 2008 -0700
40739 Add support for RepeatPad and RepeatReflect to 915 and 830-class Render accel.
40741 commit 5e04b5ad1b618655e201cbf1f5b2585aa6be0c17
40742 Merge: 979bb10 260cbcf
40743 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40744 Date: Mon Oct 6 16:08:12 2008 -0700
40746 Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
40748 commit 979bb109c6d76fe18b9e29f55afb2548f890bbfe
40749 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40750 Date: Mon Oct 6 16:05:46 2008 -0700
40752 Document more VBIOS functionality
40754 Document a few more VBT structures and comment on VBIOS communication a
40755 bit. There should be enough there now for a sufficiently motivated
40756 developer to start implementing support for hotkeys and other features
40757 on pre-IGD OpRegion machines.
40759 commit 260cbcfe61868175ba3e649ce07d43f57601f9be
40760 Author: Carl Worth <cworth@cworth.org>
40761 Date: Mon Oct 6 13:17:59 2008 -0700
40763 Document and use 'legacy' border color mode
40765 It's very convenient that the hardware supports this non-default
40766 mode since it's exactly what is specified by the Render extension.
40767 This provides a more efficient means of fixing bug #16820:
40769 [EXA] Composition result in black for areas outside of source-surface bo
40770 https://bugs.freedesktop.org/show_bug.cgi?id=16820
40772 without the software fallback we had in the earlier fix,
40773 (commit 76c9ece36e6400fd10f364ee330faea470e2da64 ).
40775 commit 98ca9e2af1b0a67ed91e10cf661af70dfba44d43
40776 Author: Carl Worth <cworth@cworth.org>
40777 Date: Mon Oct 6 13:16:37 2008 -0700
40779 Rename default_color to border_color
40781 This is consistent with the documentation, (and just plain makes
40784 commit 175b30e3828266a89021c2439119545099d72715
40785 Author: Carl Worth <cworth@cworth.org>
40786 Date: Mon Oct 6 12:48:07 2008 -0700
40788 Revert "Fallback to software for RepeatNone with transformed RGB-only pictures."
40790 This reverts commit 76c9ece36e6400fd10f364ee330faea470e2da64.
40792 We've learned a new technique that should let us avoid this fallback
40793 to software. See following commit.
40795 commit 21cce1998a10db520c390eb5061814867564c583
40796 Author: Shuang He <shuang.he@intel.com>
40797 Date: Mon Oct 6 09:48:21 2008 +0800
40799 Fix a typo in G965 texture video code
40801 commit 76c9ece36e6400fd10f364ee330faea470e2da64
40802 Author: Carl Worth <cworth@cworth.org>
40803 Date: Thu Oct 2 20:40:14 2008 -0700
40805 Fallback to software for RepeatNone with transformed RGB-only pictures.
40807 We wish it wouldn't, but the hardware ignores the alpha in the
40808 BorderColor we set when the source picture format has no alpha
40809 in it, (and it uses alpha of 1.0 where we want 0.0). For now,
40810 fallback for these cases. This gives a correct result, but
40811 obviously is not as fast as we would like.
40813 This fixes bug #16820:
40815 [EXA] Composition result in black for areas outside of source-surface bounds
40816 https://bugs.freedesktop.org/show_bug.cgi?id=16820
40818 commit f1dbc266ccfe26c6b9a272e40a5bbe9afaa4f2e0
40819 Author: Eric Anholt <eric@anholt.net>
40820 Date: Thu Oct 2 14:45:12 2008 -0700
40822 Work around libpciaccess reporting a 0 rom size by guessing.
40824 I required the following patch on top of this to work around libpciaccess
40825 brokenness. libpciaccess reports 0 rom size if there's no rom resource,
40826 even if the rom file exists in sysfs.
40828 commit 8304b405e0dc2f31fd2d2fd82e150ba502ab74e2
40829 Author: Keith Packard <keithp@keithp.com>
40830 Date: Mon Sep 29 17:37:28 2008 -0700
40832 Eliminate INT10 call to get BIOS contents
40834 libpciaccess (and the old X server PCI code as well) provides a function to
40835 get the ROM contents. Code to use that was already present in the driver and
40836 used if the INT10 function failed. Skip the INT10 and just use libpciaccess
40837 as that eliminates several module loads and scary use of vm86.
40839 Signed-off-by: Keith Packard <keithp@keithp.com>
40841 commit ab2b70f831314f81a581bfb1e48d059a3a2b0b06
40842 Author: Carl Worth <cworth@cworth.org>
40843 Date: Wed Oct 1 16:05:16 2008 -0700
40845 Prefer repeatType field over using both repeat and repeatType.
40847 Eric informed me that the repeat field exists only for backwards
40848 compatibility with old drivers that weren't prepared for values
40849 other than 0 or 1 here. Since we are, we can just ignore that
40850 field and examine only repeatType. So the code's a (tiny) bit
40853 commit 128223ee9b7880e640056475462eca9a88415492
40854 Author: Carl Worth <cworth@cworth.org>
40855 Date: Wed Oct 1 15:29:04 2008 -0700
40857 Add support for RepeatPad and RepeatReflect.
40859 It's quite simple to support these modes---we simply need to
40860 turn on the support for them in the hardware.
40862 These changes have been verified with the extend-pad and
40863 extend-reflect tests in cairo's test suite. However, this
40864 currently required using a custom-modified version of cairo.
40865 The issue is that released versions of cairo, (and even
40866 cairo master so far), don't pass RepeatPad and RepeatReflect
40867 to Render, (due to various bugs and workarounds in cairo
40868 and pixman). I do plan to fix those issues in cairo, so that
40869 in a future release of cairo, (1.8.2 perhaps?), the cairo
40870 test suite will usefully test these new repeat modes in our
40873 commit b7279f1be1b913c1c6ee8ebfb95c97800217a821
40874 Author: Carl Worth <cworth@cworth.org>
40875 Date: Wed Oct 1 15:26:39 2008 -0700
40877 Examine picture repeatType as well as repeat field.
40879 The existing switch statement was switching on the Boolean
40880 repeat field rather than the correct repeatType field. This
40881 had not caused any problem before as only two possible repeat
40882 values were supported (RepeatNone = 0 and RepeatNormal = 1)
40883 so they were always the same as the repeat field.
40885 Soon, however, we'll be supporting more repeat types, so we'll
40886 need to switch on the correct value.
40888 commit ffcbbb071f1cde90fe0dc4887a05dd66c0e66985
40889 Merge: 1eb46d0 bf8a3ff
40890 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40891 Date: Tue Sep 30 16:59:35 2008 -0700
40893 Merge branch 'master' into xf86-video-intel-2.5-branch
40895 commit bf8a3ffeaa4df9ecbad7abff9a44d714de2fa104
40896 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40897 Date: Tue Sep 30 16:59:02 2008 -0700
40899 Revert "Use -Werror by default"
40901 This reverts commit f315e9d1ad92562195ce42b956d4be6b31e8a13e. The world
40902 isn't ready for a warning free build.
40904 commit 1eb46d055b147941fa529ac582c336a114f47892
40905 Merge: 8408995 497b742
40906 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40907 Date: Tue Sep 30 12:54:27 2008 -0700
40909 Merge branch 'master' into xf86-video-intel-2.5-branch
40911 commit 497b7420c4a76df553ed53322a390a1c754ce30c
40912 Author: Olivier Fourdan <fourdan@xfce.org>
40913 Date: Tue Sep 30 12:53:11 2008 -0700
40915 Fix ordering of VGA vs. plane disable
40917 From the Intel 965 Programmer’s Reference Manual, volume 3,
40918 chapter 2.2.2 "Mode Switch Programming Sequence".
40920 The disable sequence should be:
40922 - Disable planes (VGA or hires)
40924 - Disable VGA display in 0x71400 bit 31
40926 This patch implements that order plus waits for a vblank at the end.
40930 commit a4568740ee296bc392843fd324d2a047e7862187
40931 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40932 Date: Tue Sep 30 12:48:32 2008 -0700
40934 Be more verbose about panel data in VBIOS dumper
40936 Dump more panel data, including number of expected entries. Had to
40937 refactor things a bit, but now each function should get size information
40938 so further checking can be added more easily.
40940 commit fa2586a40f20e73ec7420466638e8f595e0da987
40941 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40942 Date: Tue Sep 30 12:46:20 2008 -0700
40944 Use VBT LFP info pointers by default
40946 On some machines it appears that the LFP info pointers give us more
40947 accurate panel info than if we index into the LFP data table using the
40948 panel type index. Early reports indicate that using the pointers
40949 doesn't cause regressions, so switch to them by default to help 8xx
40952 Fixes bug 17310 (and hopefully 17658 too).
40954 commit f315e9d1ad92562195ce42b956d4be6b31e8a13e
40955 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40956 Date: Tue Sep 30 12:07:52 2008 -0700
40958 Use -Werror by default
40960 commit f082e877d54dbafa437fd05e9c07e870cd513be9
40961 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
40962 Date: Tue Sep 30 12:06:46 2008 -0700
40964 Work around gcc uninitialized variable warnings
40966 GCC isn't smart enough to analyze the control flow and figure out that
40967 these are false positives, but initializing them shouldn't hurt, so work
40970 commit 836d24f2cc3836fbd1695cf1f88c2af975b4862c
40971 Author: Keith Packard <keithp@keithp.com>
40972 Date: Mon Sep 29 19:02:51 2008 -0700
40974 Use uintptr_t instead of uint64_t to hold pointer value
40976 commit baec75a56d91546443a2e04812face4997bc764b
40977 Author: Zou Nan hai <nanhai.zou@intel.com>
40978 Date: Sun Sep 28 11:18:40 2008 +0800
40980 [965-xvmc] 965q indirect media data workaround
40982 commit 11d304e99c0e11c28901ec28115d9c8b81a2b9cc
40983 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40984 Date: Sun Sep 28 10:08:26 2008 +0800
40986 Bug #16631: add option for SDVO force detect
40988 Some ADD2 card doesn't get SDVO detect status setup right,
40989 which disabled outputs on those cards. This adds a new
40990 option "ForceSDVODetect" to probe all SDVO ports anyway.
40992 commit 1cc15ba454fdf54a7dea9da066e0a023a4742fab
40993 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
40994 Date: Fri Sep 26 10:01:52 2008 +0800
40996 Render register clock gating disable fix on 4 series chipset
40998 commit d6b2696f9ac14a81598e0147698209ad428fd45d
40999 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41000 Date: Fri Sep 26 09:53:03 2008 +0800
41002 Do force CRT detect sequence twice on 4 series chipset
41004 commit d8a007b056e3d6c3e132ecac06368b21d2ac4123
41005 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41006 Date: Fri Sep 26 09:44:59 2008 +0800
41008 Bug #16515: Fix VT switch with DVI on G45
41010 On DVI, centered VGA mode is used instead of native mode VGA, and
41011 PLL on pipe is used instead of VGA PLL setting. So make sure PLL
41012 settled down in restore time.
41014 commit a428892b76dba4c55e2ae28933d30e98d9fd646d
41015 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41016 Date: Tue Sep 23 12:36:11 2008 -0700
41018 Update supported hardware list
41020 commit 3c2f7d84785954eb91341f85abb175f299b2797d
41021 Author: Zou Nan hai <nanhai.zou@intel.com>
41022 Date: Tue Sep 23 17:15:42 2008 +0800
41024 [965-xvmc] IGD support
41026 commit 08ab23bb87563517d5653bddf82763b5ef641bed
41027 Author: Zou Nan hai <nanhai.zou@intel.com>
41028 Date: Tue Sep 23 14:57:18 2008 +0800
41030 [965-xvmc] optimize, pack some payload
41032 commit 8408995ffbf705aa0bc09ab72c58c2e31a4b70c3
41033 Merge: d18b86a afa630b
41034 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41035 Date: Mon Sep 22 18:38:54 2008 -0700
41037 Merge branch 'master' into xf86-video-intel-2.5-branch
41039 commit afa630b448e5993850433c9f0b129758ec4d37b5
41040 Author: Bryce Harrington <bryce@canonical.com>
41041 Date: Mon Sep 22 18:38:35 2008 -0700
41043 Add TV out quirk for HP Compaq nx6110
41045 Adds a TV out quirk for HP Compaq nx6110. Fixes bz #17683.
41047 commit 204a6281aa59e29027793ba5c84a5d9af7dc9e90
41048 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41049 Date: Mon Sep 22 18:38:03 2008 -0700
41051 Revert "Add no TV out quirk for HP Compaq nx6110"
41053 This reverts commit 6a46022df57d09109fc119fedce99c156715a4d4. It should have
41054 been attributed to Bryce.
41056 commit d18b86aa7fc9270e24ed9ea861ebbbbb54558fee
41057 Merge: 9a7da65 6a46022
41058 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41059 Date: Mon Sep 22 18:03:39 2008 -0700
41061 Merge branch 'master' into xf86-video-intel-2.5-branch
41063 commit 6a46022df57d09109fc119fedce99c156715a4d4
41064 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41065 Date: Mon Sep 22 13:26:58 2008 -0700
41067 Add no TV out quirk for HP Compaq nx6110
41069 Hopefully we can fix the VBT parsing code so we don't need this, but for the
41070 time being it seems necessary. Fixes bz #17683.
41072 commit 10909d9b665864bda2b1654de009d556cd068726
41073 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41074 Date: Mon Sep 22 13:11:55 2008 -0700
41076 Add Cappuccino SlimPRO SP625F to no LVDS quirks list
41078 Looks like this platform might contain VBTs that indicate an LFP is present
41079 even though it isn't. Intended to fix bz #11368.
41081 commit 9a7da65e624060a37eef989963a9c79f13df30d3
41082 Merge: 043b6e7 2f93cfb
41083 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41084 Date: Mon Sep 22 09:59:52 2008 -0700
41086 Merge branch 'master' into xf86-video-intel-2.5-branch
41088 commit 2f93cfbc7e96acc32efb5e1ca49b817a81cba6e3
41089 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41090 Date: Fri Sep 19 15:20:55 2008 +0800
41092 Fix output detection for DVI-I
41094 For CRT this trys to probe all possible port for EDID and
41095 detects got confirmed by EDID's d/a type bit.
41096 For HDMI/DVI, also using EDID d/a type bit to ensure it should
41097 handle the connect or not.
41099 commit 1fbe4d602816c9dfc5fba917b9fdc257d8d025b0
41100 Author: David Schleef <ds@schleef.org>
41101 Date: Thu Sep 18 15:37:00 2008 +0800
41103 Bug #17277: fix upscaling limit
41105 Oh duh (i830_video.c):
41107 /* Clamp dst width & height to 7x of src (overlay limit) */
41108 if(drw_w > (src_w * 7))
41111 if(drw_h > (src_h * 7))
41114 The condition I see in the documentation appears to be src_h/drw_h < 8, that
41115 is, src_h < 8*drw_h. It appears this was "fixed" incorrectly in e784e152.
41116 It seems difficult to believe that this limitation would exist at all for the
41119 commit bc36608e321e01a2be792688b4b734bb7c0667f7
41120 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41121 Date: Thu Sep 18 10:42:33 2008 +0800
41123 Check display stride limit when allocate front buffer
41125 commit 043b6e71b83eb05339a6f8c4814e6941f8b9695a
41126 Author: Xiang, Haihao <haihao.xiang@intel.com>
41127 Date: Tue Sep 16 13:21:43 2008 -0700
41129 Move bufmgr init earlier so it's available at I830DRIDoMappings time.
41131 Fixes a crash with non-GEM mode. Bug #17540.
41132 (cherry picked from commit 62b75df84c893bf28e20014cf88ce45064611dc9)
41134 commit 0a2d17f7c2ea9b695df1855aab4ccc519546d8e5
41135 Author: Xiang, Haihao <haihao.xiang@intel.com>
41136 Date: Tue Sep 16 11:49:39 2008 -0700
41138 Put back check for pI830->hw_status in setting hws in non-GEM mode.
41140 Fixes crashes on non-GEM systems with physical hardware status page.
41143 (cherry picked from commit 188d58dac9a87b56dbc34ec219cd196928bbcf64)
41145 commit 62b75df84c893bf28e20014cf88ce45064611dc9
41146 Author: Xiang, Haihao <haihao.xiang@intel.com>
41147 Date: Tue Sep 16 13:21:43 2008 -0700
41149 Move bufmgr init earlier so it's available at I830DRIDoMappings time.
41151 Fixes a crash with non-GEM mode. Bug #17540.
41153 commit 188d58dac9a87b56dbc34ec219cd196928bbcf64
41154 Author: Xiang, Haihao <haihao.xiang@intel.com>
41155 Date: Tue Sep 16 11:49:39 2008 -0700
41157 Put back check for pI830->hw_status in setting hws in non-GEM mode.
41159 Fixes crashes on non-GEM systems with physical hardware status page.
41163 commit e2743a409a02978ca8e953dbbeeb96f61bbd5ce6
41164 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41165 Date: Fri Sep 12 14:22:48 2008 -0700
41167 Only BO map render state if kernel mode setting is active
41169 We'll probably end up doing this differently, but avoid this path for now.
41171 commit 95b3e3a65a60aa264fbbfd141df980b534005264
41172 Merge: 94a86fa ec17c88
41173 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41174 Date: Fri Sep 12 13:42:06 2008 -0700
41176 Merge branch 'master' into xf86-video-intel-2.5-branch
41180 configure.ac - resolved version number conflict
41182 commit 8a54e3be5c5057fe8e3c52c03401fdada7978c45
41183 Author: Keith Packard <keithp@keithp.com>
41184 Date: Fri Sep 12 12:47:13 2008 -0700
41186 Fix mis-merge of DRI2 changes related to pI830->directRenderingType
41188 The old code had a directRendering boolean, the new one has an enum. A bunch
41189 of the merges failed to remove the '!' in front of the old name, breaking a
41190 bunch of 'we don't have DRI' tests.
41192 commit d8d95d8c71f2cd4bab277f44132ece7963714a5b
41193 Author: Eric Anholt <eric@anholt.net>
41194 Date: Thu Sep 11 16:11:46 2008 -0700
41196 Fix build failures that should have been in the previous merge commit.
41198 commit ec17c88a0ed7c9cf4ad68aa52a7a891946a1c0f4
41199 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41200 Date: Thu Sep 11 16:14:47 2008 +0800
41202 Add support for G41 chipset
41204 G41 is another 4 series chipset like G45/43.
41206 commit f9c625e1e5ddfff06b38fdd4e596fd8eae5934d5
41207 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41208 Date: Thu Sep 11 15:35:27 2008 +0800
41210 Disable render standby
41212 Render standby is known to cause possible hang issue on some
41213 mobile chips, so always disable it.
41215 commit 7e8eb6e20b6b1a2a3c1ef28f694fa23c68a15c48
41216 Merge: fba9b5a 58a3817
41217 Author: Eric Anholt <eric@anholt.net>
41218 Date: Wed Sep 10 14:10:26 2008 -0700
41220 Merge branch 'master' into dri2
41224 src/i830_batchbuffer.c
41227 commit 58a3817305ef1455a11be6ead8a8521dfc5875c7
41228 Author: Eric Anholt <eric@anholt.net>
41229 Date: Tue Sep 9 11:02:49 2008 -0700
41231 Track move of exec to bufmgr, and restoration of emit/wait funcs for non-drm.
41233 commit f367334c6392a717f6cd2f4ed02200be1c6d356a
41234 Author: Eric Anholt <eric@anholt.net>
41235 Date: Tue Sep 9 11:01:33 2008 -0700
41237 Track the move of irq emit/wait to fake bufmgr.
41239 commit 0b4f7b630312b148ce4e172cb7cd9f673751b2a3
41240 Author: Eric Anholt <eric@anholt.net>
41241 Date: Tue Sep 9 10:57:08 2008 -0700
41243 Track move of bufmgr functions to libdrm_intel.
41245 commit 0f804bfa1e1e972e9b4e3b7c8db61e9877c50f50
41246 Author: Eric Anholt <eric@anholt.net>
41247 Date: Tue Sep 9 19:17:05 2008 -0700
41249 Bug #17446: Don't try to manage IRQs in GEM mode.
41251 The kernel told us that it was already doing so, resulting in failure.
41253 commit da63b5adec8555cc7b3e71f33933f4c9dd6f714e
41254 Author: Eric Anholt <eric@anholt.net>
41255 Date: Sun Aug 31 14:27:29 2008 -0700
41257 Add some MCHBAR registers for debugging tile swizzling issues.
41259 commit c7aaf0118baa34e583df5f1c29c9dab9a6af6eb7
41260 Author: Dave Airlie <airlied@linux.ie>
41261 Date: Tue Sep 9 18:13:56 2008 +1000
41263 mode: fix missing comma
41265 commit b9ef0ed7d7b96eca6394cd0d367369ec511d1bcd
41266 Author: Daniel Stone <daniel@fooishbar.org>
41267 Date: Fri Sep 5 05:02:08 2008 +0300
41269 i830: Fix timer leak
41271 TimerCancel just cancels the timer: it still leaves the TimerRec intact and
41274 commit fba9b5aff450d874ef98cafd3ecc0fd85f9597b5
41275 Author: Eric Anholt <eric@anholt.net>
41276 Date: Wed Sep 3 16:49:07 2008 +0100
41278 DRI2: Move pixmap pitch alignment for use with depth to pixmap create.
41280 The previous location for pitch fixup would have only worked when depth was
41281 used with the backbuffer, and no page flipping or other adventures occurred.
41283 commit 71ec627c3a65cfc7bca7353af43c60b18e73230d
41284 Author: Eric Anholt <eric@anholt.net>
41285 Date: Wed Sep 3 16:37:25 2008 +0100
41287 DRI2: Emit the MI_FLUSH before flushing batch in swapbuffers.
41289 Should fix issues with swapbuffers flushing to front buffer on 965.
41291 commit daaefd173b4c98d0ebabd43352bfa3a030a62e4b
41292 Author: Eric Anholt <eric@anholt.net>
41293 Date: Wed Sep 3 16:26:27 2008 +0100
41295 UXA: Re-enable non-965 render.
41297 commit 32bb5fc7db14911d0438aea1504e1551b2c9fc25
41298 Author: Eric Anholt <eric@anholt.net>
41299 Date: Wed Sep 3 16:19:03 2008 +0100
41301 Don't set up sarea or drm mappings in DRI2 mode.
41303 commit 4a5e31138115f435be7f8f5999b1b2329a18d669
41304 Author: Kristian Høgsberg <krh@redhat.com>
41305 Date: Wed Sep 3 16:18:23 2008 +0100
41307 Fix broken test for DRI1 in DRI2 conversion.
41309 commit fca7a4e9a5c73e1c129e94eeccf8fc795f729931
41310 Author: Stefan Dirsch <sndirsch@suse.de>
41311 Date: Wed Sep 3 15:32:11 2008 +0200
41313 Pipe A force quirk for Toshiba Satellite A30.
41315 commit 57ad9cc689724a2f3583eda862250eab0a6798d5
41316 Author: Fabio <fabio.ped@libero.it>
41317 Date: Mon Sep 1 13:33:50 2008 +0800
41319 Man page patch to clarify meaning of VideoRam option with i810/i815
41321 commit fd6a9ca5b59a868260eb91272091ea6169091737
41322 Author: Kristian Høgsberg <krh@redhat.com>
41323 Date: Fri Aug 8 15:36:35 2008 -0400
41327 commit a1c802e4a0324386cf7370594a46354c89e3b646
41328 Merge: 3733a1b 808b72f
41329 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41330 Date: Fri Aug 29 09:23:35 2008 -0700
41332 Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
41334 commit 3733a1b54c95aa378f32577f9e996946e8e8e48c
41335 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41336 Date: Fri Aug 29 09:12:05 2008 -0700
41338 Fix build when using kernel DRM headers
41340 Unfortunate mismatch between kernel and DRM master headers.
41342 typedef struct _drm_i915_batchbuffer { ... } drm_i915_batchbuffer_t;
41344 typedef struct drm_i915_batchbuffer { ... } drm_i915_batchbuffer_t;
41346 so use the typedef rather than the morphing structure name.
41348 commit bd7ee7fcf51138d21da92a3d8cbaff49871d7188
41349 Author: Zou Nan hai <nanhai.zou@intel.com>
41350 Date: Fri Aug 29 11:57:46 2008 +0800
41352 [965-xvmc] There is some issue with uv half pixel, remove it till fix
41354 commit 973a5fb25652a97e28af03295d430a79557be756
41355 Author: Zou Nan hai <nanhai.zou@intel.com>
41356 Date: Fri Aug 29 10:59:17 2008 +0800
41358 [965-xvmc] some shader optimize
41360 commit e7379d1c48a522f137cf1a06a334748c88471946
41361 Author: Zou Nan hai <nanhai.zou@intel.com>
41362 Date: Thu Aug 28 09:49:57 2008 +0800
41364 [965-xvmc] half-pixel bug fix
41366 commit be49ae6bd683ee9eb778f7ea937aaee4d72c51f0
41367 Author: Zou Nan hai <nanhai.zou@intel.com>
41368 Date: Wed Aug 27 11:29:15 2008 +0800
41370 [965-xvmc] add missing g4a file
41372 commit e813b139b7c353a0930c3e00408700619d3949e4
41373 Author: Zou Nan hai <nanhai.zou@intel.com>
41374 Date: Wed Aug 27 11:11:41 2008 +0800
41376 [965-xvmc] support field prediction and dct type, optimize in GPU
41378 commit 808b72f81454061c815321e51a9b9f925c4bf786
41379 Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
41380 Date: Tue Aug 26 22:24:36 2008 -0400
41382 Change uxa private keys to integer variables.
41384 Prepares for a devPrivates system that will store an index.
41386 commit 62ce9e8f9c8cc2014645d58f4249c496aebc36e8
41387 Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
41388 Date: Tue Aug 26 22:23:06 2008 -0400
41390 Remove unused exa_pixmap_key.
41392 commit 087ade8e66cf7a34b8a96e1efe438099376cd896
41393 Author: Robert Noland <rnoland@2hip.net>
41394 Date: Tue Aug 26 16:35:07 2008 -0400
41396 Fix typo in last commit
41398 commit 6404ac178bf2c32a715a66e7fd6ab0184a05af36
41399 Author: Robert Noland <rnoland@2hip.net>
41400 Date: Tue Aug 26 15:09:47 2008 -0400
41402 Check for drm before calling modeset ioctl.
41404 commit c889198ab54336248a03050ad410254912d79bb7
41405 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41406 Date: Tue Aug 26 10:12:21 2008 +0800
41408 Fix X exit crash in NoAccel
41410 NoAccel should still be needed for performance evaluation,
41411 so don't crash me when exit.
41413 commit 0929a191ae36473acb0244a818098e4acf26d675
41414 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41415 Date: Tue Aug 26 10:01:23 2008 +0800
41417 Destroy bufmgr after allocation finish
41419 Fixed X exit crash for me, allocation reset will
41420 unreference bo which still requires bufmgr to be live.
41422 commit 94a86fa51a931f0059ea17728f857ec92cec5b16
41423 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41424 Date: Mon Aug 25 16:59:03 2008 -0700
41426 Fix UXA build for distcheck
41428 Was missing proper path info in CFLAGS.
41430 commit 1c224f72d5f8fa8b6c4d9a6e06698f2462bb43b6
41431 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41432 Date: Mon Aug 25 16:59:03 2008 -0700
41434 Fix UXA build for distcheck
41436 Was missing proper path info in CFLAGS.
41438 commit 09a235af0f216052d3101b153c18be1fa75abf15
41439 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41440 Date: Mon Aug 25 16:38:17 2008 -0700
41442 Update version to 2.4.97 for first 2.5 test release
41444 commit 8e3243b8e5549033aa2f901a3bdd882418f6a69b
41445 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41446 Date: Mon Aug 25 16:37:34 2008 -0700
41448 Update version to post-2.5
41450 commit 7a5b090abc39c380955489bad5e46d1278f38fee
41451 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41452 Date: Thu Aug 21 09:40:57 2008 -0700
41454 Hide kernel mode setting EXA code behind XF86DRM_MODE
41456 Prevents compiler warnings in the non-kms case.
41458 commit 4937b98981923c599786f5007ab0580ad08e1909
41459 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41460 Date: Thu Aug 21 09:39:33 2008 -0700
41462 Fix compiler warnings in VBIOS utils
41464 commit 08326827fdb8abecbff20c7b051537dbad433c40
41465 Merge: 5af5041 0de8ca3
41466 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41467 Date: Thu Aug 21 09:30:38 2008 -0700
41469 Merge branch 'modesetting-gem'
41471 commit 5af504166f14f4401818f3b7c27ac2ccba7b9ff5
41472 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41473 Date: Wed Aug 20 15:06:36 2008 -0700
41475 Don't disable planes in i830_update_dsparb
41477 We need to program DSPARB with only one plane enabled at most, and that's
41478 guaranteed to be the case when we're called during mode set, so just assert
41479 that case but otherwise leave the planes alone to avoid flicker on active
41480 displays during output detection for example.
41484 commit 7b6f4d22211d71480caf6335a3eacaacff369371
41485 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41486 Date: Wed Aug 20 14:40:29 2008 -0700
41488 Don't allocate a pipe for hotplug detection
41490 It shouldn't be needed...
41492 commit ba4a4b7887d5c2da21cadfeeed601eaec8515550
41493 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41494 Date: Wed Aug 20 14:39:34 2008 -0700
41496 Add more panel debugging info to register dump & vbios reader
41498 commit 0de8ca36306c506356e82d500134b4eeb96a6080
41499 Merge: f26bcb9 5eccb5e
41500 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41501 Date: Wed Aug 20 12:11:00 2008 -0700
41503 Merge branch 'master' into modesetting-gem
41509 commit 5eccb5ec7ea78b5800f98decd4216d39b9ebe5ec
41510 Author: Julien Cristau <jcristau@debian.org>
41511 Date: Wed Aug 20 15:33:59 2008 +0800
41513 Don't skip the checks for DRI headers with --enable-dri
41515 commit 1baebabc3af9f21a5b8084788ecd88dc66ce78de
41516 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41517 Date: Tue Aug 19 16:02:55 2008 -0700
41519 Add swf_dumper to .gitignore
41521 commit f26bcb9cb879e2e9ee2b26c95438cf46286c0208
41522 Merge: 67ab546 7a1cc48
41523 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41524 Date: Tue Aug 19 15:53:18 2008 -0700
41526 Merge branch 'master' into modesetting-gem
41532 commit 67ab5462527c5ed94ed073421b60e85a213ea267
41533 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41534 Date: Tue Aug 19 15:46:52 2008 -0700
41536 Use GTT mapping for EXA PrepareAccess function
41538 Makes software fallbacks *much* faster.
41540 commit 547cf774801bea2177d1224204f7de0dbf6baf3c
41541 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41542 Date: Tue Aug 19 15:45:26 2008 -0700
41544 Don't allocate EXA offscreen space if kernel mode setting is active
41546 We'll use GEM alloc & free routines in that case.
41548 commit 2ddd85f8aef5da1aebbfe9fab510858e7a41860d
41549 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41550 Date: Tue Aug 19 15:41:41 2008 -0700
41552 Bail out if kernel mode setting is active but DRI fails
41554 commit 646bbdc4b6a7b887ee8f0e51fdff5d07c4586329
41555 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41556 Date: Tue Aug 19 15:05:00 2008 -0700
41558 Add quirk for pre-915s with working PFIT regs
41560 The regs are undocumented, but on some machines they work fine, so add this
41561 quirk to indicate it.
41563 commit 7a1cc48276f974d04e1c5ef7c92d98fe5ae9d4fa
41564 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41565 Date: Tue Aug 19 10:44:04 2008 -0700
41567 Add VBIOS software flags dumper program
41569 Still need to actually print all the flags & test on various platforms, but at
41570 least the bits are documented now.
41572 commit 385dc85881c56e218e7d17d7e1d022568aa29cfb
41573 Author: Zou Nan hai <nanhai.zou@intel.com>
41574 Date: Tue Aug 19 16:27:12 2008 +0800
41576 Fix from He Shuang(shuang.he@intel.com),
41577 XvMCDestroyContext to check pointer valid
41579 commit 3147fa775c7e2a933dc74af1372a3318e77d3218
41580 Author: Zou Nan hai <nanhai.zou@intel.com>
41581 Date: Tue Aug 19 16:20:16 2008 +0800
41583 Fix from He Shuang(shuang.he@intel.com),
41584 XvMCDestroyContext to check pointer valid
41586 commit 078a59dd48ed00a6512d155883956c18752da9ce
41587 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41588 Date: Mon Aug 18 12:53:33 2008 -0700
41590 Remove unused VBIOS flag defines
41592 We just use the structure definitions now.
41594 commit 7048141254b7c1948fe77b1742b463f8723610d8
41595 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41596 Date: Mon Aug 18 12:51:31 2008 -0700
41598 Pack bdb_general_definitions block
41600 More correct & consistent with other definitions.
41602 commit 0e6759bb6ad7293b6447bd20e3d7d8fdee46aba7
41603 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41604 Date: Tue Aug 19 07:30:01 2008 +0800
41606 Fix reverted LVDS bios capability dword definition
41608 Issue found by Peter Clifton.
41610 commit 76b5a7ff5c943cfaa254d2fd94b6397c91d1f8e6
41611 Author: Shaohua Li <shaohua.li@intel.com>
41612 Date: Mon Aug 18 11:13:20 2008 +0800
41614 [PATCH] avoid duplicate mode set in lvds
41616 xf86SetDesiredModes() already sets lvds to full mode. later when
41617 xf86CrtcScreenInit() initialized randr12, i830_lvds_set_property will
41618 recall xf86CrtcSetMode and set mode to full. This patch is to remove the
41619 duplication. In my test, this can save about 0.2 - 0.4s x startup time.
41621 commit b8ee31cd80a945d1fbbfc45234ffe34a109b5717
41622 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41623 Date: Sat Aug 16 12:18:01 2008 -0700
41625 Remove last TTM bits
41627 Makes it build again with drm#modesetting-gem.
41629 commit 6eb3e0f2f4e43e436029fc82e458ac8de1f94745
41630 Author: Adam Jackson <ajax@redhat.com>
41631 Date: Fri Aug 15 15:47:35 2008 -0400
41633 Don't touch pScrn->monitor->DDC directly.
41635 commit ed7269e0a377b3135b06cacecc59119f06958feb
41636 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41637 Date: Fri Aug 15 10:10:43 2008 +0800
41639 Always check and free driver private in LVDS destroy
41641 commit f4cb9a135dca9279af8186a9b18bf62ef7351019
41642 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41643 Date: Fri Aug 15 10:02:44 2008 +0800
41645 Make skip_panel_detect clear for its meaning
41647 commit 22918f62c89a4314fb5d01c58f22fee5b9a15a27
41648 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41649 Date: Thu Aug 14 16:02:37 2008 -0700
41651 Fix pipe A force quirk
41653 Last commit introduced a logic buglet, we went from (foo & BLAH) -> (!foo &
41654 BLAH) rather than !(foo & BLAH), so fix it up to make my laptop work again.
41656 commit c1687f9ca9ce7b52043272e877e070c810e6599f
41657 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41658 Date: Thu Aug 14 15:48:02 2008 -0700
41660 Map/unmap render state only when bo is available
41662 Otherwise just use the GTT address.
41664 commit 380c80712f78b3673b64ea07746a8e25e15fba8e
41665 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41666 Date: Thu Aug 14 15:39:57 2008 -0700
41668 Fixup AccelMethod kernel mode setting code
41670 Allow UXA or EXA in the kernel mode setting case, defaulting to EXA.
41672 commit 4475dfb541c988ad19b177e60f31f333e2fb3355
41673 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41674 Date: Thu Aug 14 15:38:07 2008 -0700
41676 Use pwrite for cursor updates
41678 Don't open code map/memcpy/unmap, let libdrm do that for us if necessary.
41680 commit cb217d4bfd941d0fa9ceae3e483dd1ca1d768e86
41681 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41682 Date: Wed Aug 13 16:55:39 2008 -0700
41684 Make EXA & UXA share bo getting function
41686 Needed for proper acceleration & batch buffer handling.
41688 commit 1ab7239b73a5e298cadaf44e5605cf05855f08e9
41689 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41690 Date: Tue Aug 12 18:16:37 2008 -0700
41692 Map gen4 render state buffer before initializing
41694 commit 603f48e31b021c4dc0bbf7b5efbb2e68aeb421d5
41695 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41696 Date: Tue Aug 12 18:04:08 2008 -0700
41698 Don't set tiling (yet) if kernel mode setting is active
41700 commit f744aa8d4b22374e1de6dda4facb673c3c428d4d
41701 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41702 Date: Tue Aug 12 18:02:21 2008 -0700
41704 Don't run old accel init code
41706 Had the wrong logic. Throw in a couple of cleanups while we're there.
41708 commit 2321c865f3da9fd321910236a92b32e0af1aa458
41709 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41710 Date: Tue Aug 12 17:59:16 2008 -0700
41712 Don't wait for ring if kernel mode setting is active
41714 We won't have the ring mapped.
41716 commit a1b334a4c8ae090744e802f8d41a36ea631d6c84
41717 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41718 Date: Tue Aug 12 17:58:29 2008 -0700
41720 Update DRM based modesetting support
41722 Just a checkpoint, still needs a lot of work to properly handle resize, rotate
41725 commit 5d78cf7cf0d23f60e6d61a92cbc2065b440027d3
41726 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41727 Date: Tue Aug 12 17:56:59 2008 -0700
41729 Add EXA pixmap management functions for kernel mode setting
41731 Mostly pulled over from the old kernel modesetting code; a few updates were
41734 commit 9ec36e0c8bd8a4bd7c40569412fc1a21219b5af9
41735 Author: Olivier Fourdan <ofourdan@redhat.com>
41736 Date: Mon Aug 11 12:07:32 2008 -0400
41738 Fix DPMS off in the presence of the pipe A quirk.
41740 Still turn off the VGA plane, and also handle the DRI path at the end.
41742 commit 445c2e9ef8fc1d49bbb03e07bf36e0339be16a80
41743 Author: Olivier Fourdan <ofourdan@redhat.com>
41744 Date: Mon Aug 11 11:30:51 2008 -0400
41746 Apply pipe A quirk to 845 as well.
41748 commit 915c370964d8dbab48ff24ede2c5e78e7b39132d
41749 Author: Zou Nan hai <nanhai.zou@intel.com>
41750 Date: Mon Aug 11 15:45:35 2008 +0800
41752 [965-xvmc] some fixes for earlier 965 chips
41754 commit 19ebdf1b24184c0a293a698d44c4e0571109d153
41755 Author: root <root@cr-znh.(none)>
41756 Date: Mon Aug 11 15:31:13 2008 +0800
41758 [i965-xvmc] bypass copy when put image with xvmc
41760 commit d592eabc806b752053ade3c18e9dd8e0f39b45a3
41761 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41762 Date: Mon Aug 11 15:16:44 2008 +0800
41764 Fix possible spurious interrupts in hotplug detect on 4 series chip
41766 commit d0018a96064ee0adfe87c2d50c341bf7d2e45eb0
41767 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41768 Date: Mon Aug 11 15:16:09 2008 +0800
41770 Disable display clock gating for 4 series chips
41772 commit 173b909a71955997ed6366e70faebf63fe922a8e
41773 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41774 Date: Fri Aug 8 18:11:13 2008 -0700
41776 Make it actually build the kernel stuff if possible
41778 Fixup the kernel stuff to have a slightly better chance of working. Still need
41779 to fixup the EXA pixmap functions.
41781 commit fe90c0522f65f60c4c431787c889b7fb639ab61b
41782 Author: Eric Anholt <eric@anholt.net>
41783 Date: Fri Aug 8 16:35:25 2008 -0700
41785 Set lvds_ddc_mode before use to avoid a segfault on mac mini.
41787 commit b1aef6f63c151dcb202fce869e4b80598b4b2052
41788 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41789 Date: Fri Aug 8 11:54:18 2008 -0700
41791 Initial port of kernel modesetting from old intel-kernelmode branch
41795 commit 91e7b93c12f612e41b04a2d6106ff29b53b22a43
41796 Author: Eric Anholt <eric@anholt.net>
41797 Date: Thu Aug 7 11:02:19 2008 -0700
41799 intel_idle: Instead if #if 0, add an ignore flag for unreliable INSTDONE bits.
41801 This lets us produce a decent total value, and still show the unreliable bits
41802 in case they aren't on your platform.
41804 commit f9504eff31eb3c9c6c6b33dced9875866ff8307e
41805 Merge: f911347 5c9a62a
41806 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41807 Date: Wed Aug 6 12:40:47 2008 -0700
41809 Merge branch 'drm-gem'
41811 commit 5c9a62a29f62a9ecce37fae98cb01f8217eaba15
41812 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
41813 Date: Wed Aug 6 12:39:03 2008 -0700
41815 Revert "Switch to using a buffer object for the vertex buffer"
41817 This reverts commit 1abf4d3a7a203ff5d6e5ceda29573e7fd69ddf8e.
41821 src/i965_render.c - flushing was removed, keep it that way
41823 commit f91134795b545c8baebf218975b261c76a0e5873
41824 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
41825 Date: Wed Aug 6 16:19:29 2008 +0800
41827 Fix SDVO HDMI encoding detect (#16920)
41829 Check return value of get supported encode command and current
41830 encoding mode, which could be DVI or HDMI.
41832 commit fc3e287e6b6db21b113aa40ec4d397802c067f8b
41833 Author: Keith Packard <keithp@keithp.com>
41834 Date: Tue Aug 5 22:50:01 2008 -0700
41836 [uxa] Remove unused pixmap size limits.
41838 All size-related rendering limits should be managed by the driver in the
41839 pixmap_is_offscreen call. There's no need for uxa to even know these values.
41841 commit 68f0872db6ec4d5dc3b524ee08ecad0aa125acd9
41842 Author: Keith Packard <keithp@keithp.com>
41843 Date: Tue Aug 5 22:36:03 2008 -0700
41845 [uxa] Check xalloc returns and deal with failure
41847 Failing xalloc in a rendering function means just dropping the drawing on
41848 the floor (that's what we've always done).
41850 commit 8f10bfb127bfe73d83d58f1f306fb9a4dfd825d6
41851 Author: Keith Packard <keithp@keithp.com>
41852 Date: Tue Aug 5 22:34:24 2008 -0700
41854 Use EXA by default instead of UXA until we have GTT mapping
41856 commit b2d058d80ccd08d9e02ef866ee7a95b58686f6a3
41857 Author: Keith Packard <keithp@keithp.com>
41858 Date: Mon Aug 4 23:43:13 2008 -0700
41860 Rename uxa using _ instead of caps
41862 commit fc4d9c55a7fa8001786c1e4da10f005406c57ece
41863 Author: Keith Packard <keithp@keithp.com>
41864 Date: Mon Aug 4 18:17:20 2008 -0700
41866 Change PrepareAccess to take access mode rather than index
41868 commit c155bb3cb17a3bd3b2e90be52cd1fc90147c4e17
41869 Author: Keith Packard <keithp@keithp.com>
41870 Date: Mon Aug 4 04:11:13 2008 -0700
41872 Add batch flush in i830_uxa_prepare_access
41874 commit 66706718553cd272eab6f817b5a059df3e0a4347
41875 Author: Keith Packard <keithp@keithp.com>
41876 Date: Mon Aug 4 04:10:21 2008 -0700
41880 commit 12df8f40d2fb41f5446db1b49beeb442da18bee2
41881 Author: Keith Packard <keithp@keithp.com>
41882 Date: Mon Aug 4 01:43:44 2008 -0700
41884 Use dri_bo for all object allocations, including pixmaps under uxa
41886 commit 4cc20b7f6e25f4be4598f8edbe0077117126b4ee
41887 Author: Keith Packard <keithp@keithp.com>
41888 Date: Mon Aug 4 01:42:48 2008 -0700
41890 Don't call sync on prepare_access -- just let the driver deal with it.
41892 Let the driver do whatever sync is necessary from the prepare_access hook
41893 rather than forcing a full sync.
41895 commit b0b0998b5d52d105eb1e631f688aa8f1bd55ef39
41896 Author: Keith Packard <keithp@keithp.com>
41897 Date: Wed Jul 30 16:15:37 2008 -0700
41899 Make EXA functions work for UXA as well
41901 EXA and UXA have the same acceleration interface, but UXA doesn't provide
41902 pixmap stride information as it doesn't manage pixmaps. Move all of that
41903 into the driver structure so that the acceleration functions needn't
41904 reference the EXA structure.
41906 commit 59774e9aca2d743e82d616bb644d20ff6d60d492
41907 Author: Keith Packard <keithp@keithp.com>
41908 Date: Tue Jul 29 22:57:09 2008 -0700
41910 Add UXA - the unified memory acceleration architecture.
41912 This eliminates the cost of EXA migration management while providing full
41913 pixmap allocation control to the driver. The goal is to make something
41914 useful for UMA drivers.
41916 commit a21d4794b6812ce05d08f06dc47b26c4fb1c1fef
41917 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
41918 Date: Tue Aug 5 13:37:38 2008 -0700
41920 Reorganize VBIOS code
41922 Make VBT parsing happen at driver init time rather than in each output init
41923 function, to save time and better separate VBIOS code into i830_bios.[ch]. The
41924 changes end up touching the output files due to field name changes, and allow
41925 us to reorder & simplify our LFP mode detection code.
41927 commit c2f0df4dc97c87539b66525a277c7d1e2c421f61
41928 Merge: 4f1f308 750bd0b
41929 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
41930 Date: Tue Aug 5 13:03:46 2008 -0700
41932 Merge branch 'drm-gem' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel into drm-gem
41934 commit 4f1f30854260e18b8b9c4103eebbad5dcf729ce6
41935 Merge: e9e6b47 2aaa207
41936 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
41937 Date: Tue Aug 5 13:02:47 2008 -0700
41939 Merge branch 'master' into drm-gem
41945 commit 1a59cc6b9acf312de1755d67757bf7f1967342e4
41946 Author: Julien Cristau <jcristau@debian.org>
41947 Date: Mon Aug 4 12:18:12 2008 +0200
41949 Fix up the HP Pavilion ze4944ea quirk
41951 The chip is 855GM, not GM45.
41953 commit 27e9506fda86e90fd67de1715fa32d23aaa8a683
41954 Author: Alan Coopersmith <alan.coopersmith@sun.com>
41955 Date: Thu Jul 31 20:02:21 2008 -0700
41957 Man page typo fixes
41959 commit da2eb83fb9a52291ea98f3285aee7bee2d55e0ca
41960 Author: Jesse Barnes <jbarnes@eee.(none)>
41961 Date: Thu Jul 31 19:21:36 2008 -0700
41965 Add example dual head config, add info on bug reporting.
41967 commit 2aaa207db2cea03ba05c439b5074807689f2c5be
41968 Merge: 2049ba2 76eb8e6
41969 Author: Jesse Barnes <jbarnes@hobbes.lan>
41970 Date: Thu Jul 31 13:10:19 2008 -0700
41972 Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
41978 commit 2049ba211e7cdc383976c09f52c2b43acdd59481
41979 Author: Jesse Barnes <jbarnes@hobbes.lan>
41980 Date: Thu Jul 31 13:07:20 2008 -0700
41982 Update DSPARB while planes are still off
41984 This avoids the flickering people reported in the 2.4.0 release.
41986 commit 750bd0bde09adf956c17bbb49c5a6020f12e60a4
41987 Author: Carl Worth <cworth@cworth.org>
41988 Date: Tue Jul 29 15:22:39 2008 -0700
41990 Call DRM_I915_GEM_THROTTLE from I830BlockHandler
41992 This prevents the CPU from ridiculously outrunning the GPU.
41994 commit a893f176dda0b64f7dadfda6bf0331240037851e
41995 Author: Carl Worth <cworth@cworth.org>
41996 Date: Fri Jul 25 15:56:35 2008 -0700
41998 Add call to intel_bufmgr_gem_enable_reuse
42000 This instructs GEM to reuse buffer objects and improves the
42001 performance of my favorite 'x11perf -aa10text' from about
42002 169k to about 188k glyphs/sec.
42004 commit b3c1a148679a4d943e556f996ef6b9004f549a41
42005 Author: Carl Worth <cworth@cworth.org>
42006 Date: Fri Jul 25 15:18:28 2008 -0700
42008 Eliminate unnecessary flush from i965_composite
42010 This improves 'x11perf -aa10text' performance from ~144k to ~169k
42012 commit 1abf4d3a7a203ff5d6e5ceda29573e7fd69ddf8e
42013 Author: Carl Worth <cworth@cworth.org>
42014 Date: Fri Jul 25 14:48:45 2008 -0700
42016 Switch to using a buffer object for the vertex buffer
42018 commit e5fab0b7681be06a5a3be4bbd769ba5c435e2128
42019 Author: Carl Worth <cworth@cworth.org>
42020 Date: Fri Jul 25 13:44:29 2008 -0700
42022 Add OUT_RELOC macro and backing intel_batch_emit_reloc function
42024 commit 76eb8e6f1f0c6962b23550564f4273f392567857
42025 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42026 Date: Thu Jul 31 16:59:43 2008 +0800
42028 Fix SDVO reg definition
42030 Remove wrong set tv resolution command, adding HDMI regs in dump.
42032 commit 77ed3d7600c1d92bf4a3ef4f54405cde8c232986
42033 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42034 Date: Thu Jul 31 13:13:45 2008 +0800
42036 Don't program dsparb on new Intel chip
42038 On new chip, DSPARB is controlled by hardware only.
42040 commit e9e6b4738573ffa3822ae68088bd53a645eedbce
42041 Author: Eric Anholt <eric@anholt.net>
42042 Date: Wed Jul 30 14:15:57 2008 -0700
42044 intel-gem: Use new getparam to detect kernel GEM support.
42046 commit 42fb06f3f14fbec070350cf48361be4a0be0af04
42047 Author: Tomas Carnecky <tom@dbservice.com>
42048 Date: Sun Jun 15 14:27:16 2008 +0200
42050 Reorder visuals reported by the intel driver
42052 The root window visual can not be changed. Neither at runtime nor
42053 through the configuration file. The xserver simply selects the first one
42054 that matches the class (usually TrueColor). I need a root window visual
42055 with stencil buffer because my compiz plugin uses the it for some
42056 operations. This patch reorders the visuals that the 3D driver reports
42057 and puts the one with stencil (and depth) bits as first.
42059 commit 1e90c82161789d1fd50b79db50edb7860b934639
42060 Author: Julien Cristau <jcristau@debian.org>
42061 Date: Sun Jul 27 12:07:21 2008 +0200
42063 Link with -lm for sin/cos
42065 commit d77d42e543ee82e801a8563a911826521c0ba557
42066 Author: Eric Anholt <eric@anholt.net>
42067 Date: Sat Jul 26 16:10:05 2008 -0700
42069 intel-gem: Give a better error message if the kernel rejects the tiling mode.
42071 commit 6ac352f146bafe3b4bb5832340eb0ccb730565d2
42072 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42073 Date: Wed Jul 23 15:10:07 2008 +0800
42075 shut up gcc warning
42077 commit 075cf7fce57da4e51acafb3bfd3d61ba1a8096d5
42078 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42079 Date: Wed Jul 23 15:07:48 2008 +0800
42081 Fix distcheck for clean doltcompile
42083 commit 29d982c6bf873c6685bd0e395a8b0e5e3b1063e0
42084 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42085 Date: Wed Jul 23 11:16:14 2008 +0800
42087 Fix HDMI output number
42089 Output 33 is confusing.
42091 commit fd060ce89d86f6e8ff742d5b287abe8ecea32927
42092 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42093 Date: Tue Jul 22 15:29:34 2008 +0800
42095 Bug #16801: fix X crash when NoAccel on 965
42097 commit 577c4d1a0c753fb46d43be05a9781d2e204fac45
42098 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42099 Date: Tue Jul 22 15:14:14 2008 +0800
42101 only check pipe underrun if vtSema is true.
42103 commit 04bbed2248e3e8f86efd64225abf2db8eb552653
42104 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42105 Date: Tue Jul 22 15:08:57 2008 +0800
42109 Fix regression on my 855GM for VGA output.
42111 commit 7defa4f1f0eba82f39e74f96d2ad7ed2481b537e
42112 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42113 Date: Tue Jul 22 10:51:01 2008 +0800
42115 Check underrun on enabled pipe
42117 commit b7765b0837af4cc80f1257ba04495140ef5d536e
42118 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42119 Date: Tue Jul 22 09:51:54 2008 +0800
42121 Thinkpad R60e TV quirk via DMI info
42123 commit 286ff63f83db8ea16514512385e6f8c6875871ef
42124 Author: Eric Anholt <eric@anholt.net>
42125 Date: Thu Jul 17 17:06:28 2008 -0700
42127 Don't set up the HWS page in GEM mode now that the kernel manages it.
42129 commit ed0fbd016b8fbcf01eddfd17fd25e745e7af2ba4
42130 Author: Jesse Barnes <jbarnes@hobbes.lan>
42131 Date: Mon Jul 21 11:48:38 2008 -0700
42133 Add no LVDS quirk for Transtec Senyo 610 mini PC
42137 commit 62a037b0585d6ecb43daac9d4eb0927a4618a367
42138 Author: Eric Anholt <eric@anholt.net>
42139 Date: Fri Jul 18 14:50:08 2008 -0700
42143 commit 3cbfc6c76bcd2bc9194a944092f6ce0881ff9da2
42144 Author: Eric Anholt <eric@anholt.net>
42145 Date: Fri Jul 18 14:56:18 2008 -0700
42147 Fix uninitialized-use warning in i830_debug.c ring dumping.
42149 commit 1b1c4975768f63ed0f3933ec384d24c324a383b3
42150 Author: Eric Anholt <eric@anholt.net>
42151 Date: Fri Jul 18 14:53:26 2008 -0700
42153 Get prototype for i830_bios_get_tv().
42155 commit 978abf84e8cf3ba3e915774981fe03610e15efeb
42156 Author: Eric Anholt <eric@anholt.net>
42157 Date: Fri Jul 18 14:50:42 2008 -0700
42159 Bump version number past the 2.4 stable branch.
42161 commit f4ab1f6ad47b8f9a33d45b35c1cb2ba9610b96a6
42162 Author: Carl Worth <cworth@cworth.org>
42163 Date: Fri Jul 18 14:31:45 2008 -0700
42165 Add intel_statuspage to .gitignore
42167 commit c0dce28892d7a31ebcca615f87ba8d4fa6872204
42168 Merge: a6ad504 b8ca1c7
42169 Author: Keith Packard <keithp@keithp.com>
42170 Date: Fri Jul 18 10:49:50 2008 -0700
42172 Merge commit 'origin/master' into drm-gem
42174 commit b8ca1c747a679c931267363639fc0bc690cae2d6
42175 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
42176 Date: Thu Jul 17 14:41:51 2008 -0700
42178 Choose a split for DSPARB based on the configured modes for both planes.
42180 Previously, we were attempting to give both planes equal space in the FIFO
42181 to be fair. However, larger modes require more fifo space, so split it based
42182 on the relative HDisplay of the modes. This should resolve some fifo underrun
42183 issues with differently-sized displays, or single large ones.
42187 commit b37a2a8ca82279468e3806dcf77d5fa7bdd0e874
42188 Author: Jesse Barnes <jbarnes@hobbes.lan>
42189 Date: Thu Jul 17 12:30:57 2008 -0700
42191 Don't disable pipe A on 855 chips
42193 It needs to stay enabled or we may see hangs when trying to re-enable it (say
42194 at VT switch time).
42196 Fixes FDO bug #15168.
42198 commit a6ad50402cb879b08de95f94959bb1f59701c1ff
42199 Merge: e9916b9 c4565a9
42200 Author: Carl Worth <cworth@cworth.org>
42201 Date: Thu Jul 17 00:38:54 2008 -0700
42203 Merge branch 'master' into drm-gem
42208 src/reg_dumper/Makefile.am
42210 commit c4565a9811487402d899d0933cc63e27ffe1ff08
42211 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42212 Date: Wed Jul 16 10:45:28 2008 +0800
42214 Only initialize integrated TV encoder for mobile chips
42216 commit bca316863b79148b2f51a97d58b94e52ba3eae9f
42217 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42218 Date: Tue Jul 15 13:35:21 2008 +0800
42220 xvmc: use vector based structure
42222 This is more clear to read MV components.
42224 commit 0c67219d0e6cfc858af2eb375c2f9473d0f6db60
42225 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42226 Date: Tue Jul 15 13:29:33 2008 +0800
42228 move FBC register dump out of display registers
42230 This makes it easier to read.
42232 commit a76ae7cab3001ce6639ae23fb0769c4e46b57e71
42233 Author: lipeng <peng.li@linux.intel.com>
42234 Date: Tue Jul 15 10:26:33 2008 -0700
42236 Fix GPIO pin usage for DDC on second HDMI port.
42238 commit 5054a86b1b362f3f63310afb9ef2d37c9f1cfca5
42239 Author: Eric Anholt <eric@anholt.net>
42240 Date: Mon Jul 14 13:43:31 2008 -0700
42242 Fix DSPARB setting on 845/865, which have only the AEND field and 96 entries.
42244 commit 8ac00ca97995e36514ff593fec3c0f0d316ed138
42245 Author: Julien Cristau <jcristau@debian.org>
42246 Date: Fri Jun 6 13:40:20 2008 +0200
42248 Link the driver with -lpciaccess and -ldrm if needed
42250 This makes sure the driver ends up with a DT_NEEDED reference to
42251 the libraries it's using.
42253 commit d61182511b1520b04133d28480e1ffaacbaeb753
42254 Author: Julien Cristau <jcristau@debian.org>
42255 Date: Sun Jul 6 12:00:52 2008 +0200
42257 Fix gen4asm rule to work with a build dir
42259 commit 757c00927a6f5760135136450b8d02d0f999ac1c
42260 Author: Carl Worth <cworth@cworth.org>
42261 Date: Mon Jul 7 15:58:27 2008 -0700
42263 Use up to 256 separate vertex buffers
42265 This allows us to only call i830WaitSync once every 128 calls to composite
42266 rather than on every call. However, we do need to also call MI_FLUSH to
42267 avoid the vertex cache getting in our way, (since our "separate" buffers
42268 are all allocated as one contiguous chunk).
42270 commit 0c548cd040d0c5e1812470ccdf6de86b6a2926d7
42271 Author: Carl Worth <cworth@cworth.org>
42272 Date: Mon Jul 7 14:01:15 2008 -0700
42274 Allow for multiple vertex buffers (though only use one for now)
42276 Using more than one (in the future) will allow for doing less frequent calls
42279 commit cc2249333cd462b4d99d110a12c454ca141b2be8
42280 Author: Carl Worth <cworth@cworth.org>
42281 Date: Mon Jul 7 13:25:42 2008 -0700
42283 Move VERTEX_BUFFERS setup from prepare_composite to composite
42285 This is in preparation for having larger (or multiple) vertex buffers
42288 commit e9916b9d496fd0b4df717892dda33f81a2e2990a
42289 Author: Eric Anholt <eric@anholt.net>
42290 Date: Fri Jul 11 18:59:05 2008 -0700
42292 Set tiling state for buffers allocated using GEM.
42294 commit 3c0f2bcc9965155fecab606edae68874427d3423
42295 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42296 Date: Thu Jul 10 12:57:25 2008 -0700
42298 Add VBIOS based TV connector detection
42300 Now that the VBIOS code supports it, we can use the general features block to
42301 detect whether a TV connector is present on a given platform.
42303 Reviewed by Nanhai Zou.
42305 commit e00d9435609bcff1afb71aa6638a6b42a64f5178
42306 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42307 Date: Thu Jul 10 12:55:12 2008 -0700
42309 Improve VBIOS feature detection, add SSC support
42311 Improve the VBIOS feature detection and use it to find whether the platform
42312 supports spread spectrum clocking. Use the specified reference clock, but
42313 disable SSC if multiple heads are active, since it can cause problems in cloned
42316 Reviewed by Nanhai Zou.
42318 commit 7332132a79e5b5c208d43e93dfe0c8b12eb1728d
42319 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42320 Date: Mon Jul 7 10:48:56 2008 -0700
42322 Improve FBC size checking
42324 In hindsight, this is obvious, since nowhere do we tell the FBC unit how much
42325 memory it has available. We need to make sure the compressed buffer is big
42326 enough to handle the uncompresed buffer, both in terms of vertical size and
42327 total framebuffer size, or the compressor could overwrite the memory
42328 immediately following the compressed buffer.
42330 commit f2ec3fd8f2c63b1bc0745802dcf737eda4bb42b6
42331 Author: Stefan Dirsch <sndirsch@suse.de>
42332 Date: Thu Jul 3 23:12:13 2008 +0200
42334 Add pipea force enable quirk for another 855GM machine.
42336 Fixes Novell Bug #406123, thanks Christoph.
42338 commit a34a4e3f6420e2b06bbdaa124fe0ccb1bc6a0bd9
42339 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42340 Date: Thu Jul 3 01:03:14 2008 +0800
42342 Fix official name for GM45 chipset
42344 commit 6506ac7aa008436158db4f4196802cab462f3446
42345 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42346 Date: Wed Jul 2 11:12:54 2008 +0800
42348 Give asus and eeepc backlight method higher priority
42350 On #16418, Evgeniy Manachkin <sfstudio@mail.ru> reported that
42351 last asus and eeepc backlight patch is wrong, as acpi_video0 method
42352 will take priority and doesn't work.
42354 commit 55678c64bc6e3ed613ea6db14c105c18a0cf28ce
42355 Author: Dave Airlie <airlied@linux.ie>
42356 Date: Tue Jul 1 18:51:18 2008 +1000
42358 intel: fix drm check.
42360 The -mm check is broken now that 2.3.1 is out.. make it 2.4.0, to be fixed
42363 commit 0597d94001afe0a9a34104285ce4dffc75b68071
42364 Author: Roland Test-tools Bär <roland@verifysoft.de>
42365 Date: Mon Jun 30 10:32:42 2008 -0700
42367 i810: Remove an effectively unused variable (only used in an incorrect free())
42371 commit 37661d916e8384f537d5ed8cbec66a93795c6a05
42372 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42373 Date: Fri Jun 27 11:21:59 2008 +0800
42377 commit 7834a3b118ae4e034f064257762d5c25ada5fe52
42378 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42379 Date: Fri Jun 27 09:54:35 2008 +0800
42381 xvmc: fix motion_type dump for frame/field picture
42383 commit 989ec9e8a69f909cb64f17e4465982613b4b054d
42384 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42385 Date: Fri Jun 27 09:39:02 2008 +0800
42387 xvmc: Don't copy on xvmc surface in PutImage
42389 As xvmc rendering result has already been in fb, we shouldn't
42390 do extra copy on it. Although special care is required for i915
42391 xvmc surface pitch alignment, which must be at least 1KB aligned.
42392 So video display function should take it into acount instead of
42393 always setting Y pitch to be double of U/V pitch.
42395 commit 550082070a3fdb951e3cf08974dc56276c0a739c
42396 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42397 Date: Thu Jun 26 14:22:34 2008 +0800
42399 xvmc: enlarge batch buffer size
42401 commit d50cec6ef9e2178ea663e58d548390f0f3da7692
42402 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42403 Date: Thu Jun 26 14:20:05 2008 +0800
42405 xvmc: only allocate memory requiring physical address on 915G
42407 Later 945-ish chipset can use graphics address instead.
42409 commit d1c34d8c0e7ae2a1d952791343e131837fbfda99
42410 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42411 Date: Thu Jun 26 10:05:21 2008 +0800
42413 xvmc: init mc render load indirect command once
42415 commit 75e38fa8348198ba151afa37e10be3b0b0b468f8
42416 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42417 Date: Thu Jun 26 09:47:28 2008 +0800
42419 xvmc: init dest and reference buffer once
42421 Init them right after context create, and only update buffer address
42424 commit b4d8ca8b38e495b56bb3b4143e5dfe91ee651f15
42425 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42426 Date: Wed Jun 25 14:27:16 2008 +0800
42428 xvmc: init one-time mc context once
42430 Don't setup one-time mc context everytime, as the content is always
42431 unchanged. And several structs got packed layout inside to ease static
42432 state initialization.
42434 commit 54f3f528e4804b958936ac773e7a507b2676f815
42435 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42436 Date: Mon Jun 23 15:51:49 2008 +0800
42438 xvmc: environ debug option support
42440 export INTEL_XVMC_DEBUG=1
42442 commit 7a705081788bf957409e17baa5f189241589057c
42443 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42444 Date: Mon Jun 23 15:49:29 2008 +0800
42446 xvmc: formalize surface type definition
42448 We should also carefully select surface type id.
42450 commit 0283d22f510ba47dfdac2806a58e46649c26580b
42451 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42452 Date: Fri Jun 27 13:39:31 2008 -0700
42454 Add pipea force enable quirk for HP Pavilion ze4944ea
42456 Fixes bug #16540, thanks Bryce.
42458 commit 08903abe4dc0295c7ed7d1ff1a22e0e579540c15
42459 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42460 Date: Fri Jun 27 13:37:10 2008 -0700
42462 Add pipe a force enable quirk for Lenovo T60
42466 commit 05df8c0b31721a9ccc7215fb1cda1115758367c7
42467 Author: Hong Liu <hong.liu@intel.com>
42468 Date: Fri Jun 20 10:57:14 2008 +0800
42470 Fix SDVO HDMI output.
42472 While some cards had enough initialized at startup to work already, others
42473 required that the driver actually initialize the required AVI info frame.
42475 commit 59f39e0ea0ad48c52d9b8f7afb6e04ce95db717e
42476 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42477 Date: Tue Jun 24 11:01:46 2008 -0700
42479 Remove ErrorF debugging from modeset ioctl
42483 commit 1056d79809a3765f81113f3988bd948c1f9c673a
42484 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42485 Date: Tue Jun 24 10:44:23 2008 -0700
42487 Fix back buffer damage handler for 965+ chips
42489 When page flipping is enabled, we need to make sure any changes to the front
42490 buffer are reflected in the back buffer(s) or corruption might occur at page
42491 flip time. So make the damage handler work on 965 by adding appropriate tiling
42492 flags and pitch adjustments.
42494 commit 65eee25d7d2ca979205f3776d620dbb36bf68a13
42495 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
42496 Date: Tue Jun 24 10:41:46 2008 -0700
42498 Add support for keeping vblank counters sane across mode setting
42500 The DRM supports disabling of vblank interrupts when not in use, but in order
42501 to function properly it must also be aware of mode setting, which will reset
42502 the frame counter to 0. Add code to call into the DRM before and after mode
42503 setting, so that it can account for any lost vblank events.
42505 commit 277632d8c37cf43c7c1178ae7d5a2e6d5705968c
42506 Author: Zou Nan hai <nanhai.zou@intel.com>
42507 Date: Tue Jun 24 16:43:40 2008 +0800
42509 965 xvmc, current only frame based DCT_type support
42511 commit 9d7929436488969ca4fd1fef4d2e92be5e1a93aa
42512 Author: Keith Packard <keithp@keithp.com>
42513 Date: Fri Jun 20 00:23:45 2008 -0700
42515 Flush pending batch in block handler
42517 Make sure any pending rendering commands are delivered to the hardware
42518 before the server goes to sleep.
42520 commit 2f8a0aa8cd24ede824aa52102b45a295f1b6b5be
42521 Author: Keith Packard <keithp@keithp.com>
42522 Date: Tue May 20 20:29:25 2008 -0700
42524 Remove some unreliable regs for i915
42526 commit 6b6be2b2a7179c1c9fa12989e16076aae105f4e2
42527 Author: Eric Anholt <eric@anholt.net>
42528 Date: Thu Jun 19 15:01:00 2008 -0700
42530 Fix a crash in i830_sdvo_init error paths by setting up dev_priv earlier.
42532 commit 65ad29d78793c7804f133a58de80ffaa0404ca28
42533 Author: Eric Anholt <eric@anholt.net>
42534 Date: Thu Jun 19 14:27:30 2008 -0700
42536 Automatically detect the presence of HDMI.
42538 Now, SDVO is only probed if the SDVO detected bit is set. If the SDVO probe
42539 fails, but the detect bit is set, assume that it's an HDMI output.
42541 commit ea0d21006ec71bc56acde7291e5f8d28e54b83fe
42542 Author: Eric Anholt <eric@anholt.net>
42543 Date: Thu Jun 19 14:10:29 2008 -0700
42545 Fix hdmi POSTING_READ to use the register number instead of the register value.
42547 We won't dwell too long on how this ever worked.
42549 commit 2fa9240a3b685a0c3d0e2644fc144c927b67b6b3
42550 Author: Eric Anholt <eric@anholt.net>
42551 Date: Thu Mar 20 15:14:46 2008 -0700
42553 Get HDMI output working.
42555 HDMI, despite claiming to have pixel multiply support, actually doesn't
42556 want it, at least in the way that SDVO did. Disable it.
42558 Also disable the NULL_PACKETS_DURING_VSYNC bit. Despite the docs telling us
42559 to set it, the output doesn't like it when you do that, and the BIOS doesn't
42562 Also add a posting read to SDVOx setting. Without it, half the X starts
42563 wouldn't get anything on the screen. (interestingly, it was every other
42564 startup, not just a chance).
42566 commit dc8ab9d35b261b6105a62949cfb47a3554531d0b
42567 Author: Eric Anholt <eric@anholt.net>
42568 Date: Thu Mar 20 13:49:47 2008 -0700
42570 Set the sync active bits like we're supposed to, matching the BIOS.
42572 commit af13826ed063b60a1e4179101e8889b4ab7a9d0b
42573 Author: Eric Anholt <eric@anholt.net>
42574 Date: Mon Mar 10 11:51:05 2008 -0700
42576 Add pixel multiplier support for HDMI
42578 commit 9a05662918223477eb1cf6f80ffac08712721c70
42579 Author: Eric Anholt <eric@anholt.net>
42580 Date: Mon Mar 10 11:47:15 2008 -0700
42582 The phase shift its are now reserved, and add HDMI clock limits.
42584 commit beba1dd3561e38573ed9f507328caf7f8fb9f84a
42585 Author: Eric Anholt <eric@anholt.net>
42586 Date: Thu Mar 6 16:05:17 2008 -0800
42588 Initial HDMI work. Not currently hooked up at startup.
42590 commit da58dc3b02999f3244d0eaf77180b828d85bd609
42591 Author: Eric Anholt <eric@anholt.net>
42592 Date: Thu Mar 6 14:57:57 2008 -0800
42594 Add DisplayPort registers.
42596 commit 52a414d1158aa7512fb79560dc220e2a78456a2d
42597 Author: Eric Anholt <eric@anholt.net>
42598 Date: Thu Mar 6 14:00:18 2008 -0800
42600 Remove VGA regs from debug output.
42602 commit b60bbbfd7b7dcf31726ad615a493f34f17cc6309
42603 Author: Eric Anholt <eric@anholt.net>
42604 Date: Thu Mar 6 13:59:06 2008 -0800
42606 Add little hotplug detector app.
42608 commit 64fa32e2eba21cb94aa0f8d87a6ee6b7c178a7a5
42609 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42610 Date: Thu Jun 19 11:02:29 2008 +0800
42612 Add backlight kernel method support on Asus and Eeepc
42614 Noted and tested by Evgeniy Manachkin <sfstudio@mail.ru>
42615 for asus-laptop support, also add eeepc support.
42617 commit 1cfe769c74d1a3a392bf1aaaf5c2dcc8273daf66
42618 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42619 Date: Tue Jun 17 12:39:28 2008 -0700
42621 Add support for Intel 4 series chipsets.
42623 commit 289790c0467d27e96b537598a6589fc6a36da8b8
42624 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42625 Date: Tue Jun 17 09:46:12 2008 +0800
42627 Fix compiling with server master in LVDS backlight patch
42629 commit 66418cc99cfd360e303efcdf74a9ee31d435ff15
42630 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42631 Date: Thu Jun 12 14:04:41 2008 +0800
42633 Fix compiler warning when disable xvmc config
42634 (cherry picked from commit df0bbdc7cbb6ff357a81ed28d12e56c9c7d643f7)
42636 commit 078d05865884ca5784b532c0a79c8440632002ac
42637 Author: Shuang He <shuang.he@intel.com>
42638 Date: Mon May 26 09:20:53 2008 +0800
42640 Fix typo in xvmc block destroy
42641 (cherry picked from commit 87ace420a34df7425641d089f71830e44fced098)
42643 commit 4c4ef27779aebf4df90b6233de05be2bb972de4c
42644 Author: Lukas Hejtmanek <xhejtman@ics.muni.cz>
42645 Date: Mon Jun 16 02:49:41 2008 +0800
42647 Fix maximum backlight issue
42649 commit d775ddc64dc8349b8ef9e85b0be9e93cb1997fea
42650 Author: Eric Anholt <eric@anholt.net>
42651 Date: Fri Jun 13 10:03:24 2008 -0700
42653 [gem] Catch -EINTR from blocking GEM ioctl and restart.
42655 commit 02d7141ac395446613edba0f2e2a951e750e4e9f
42656 Author: Dave Airlie <airlied@redhat.com>
42657 Date: Fri Jun 13 11:30:58 2008 +1000
42659 [gem] remove one more unused bit
42661 commit f07acbdaac95931d184ea2c557edb632e577eb47
42662 Author: Eric Anholt <eric@anholt.net>
42663 Date: Thu Jun 12 13:47:35 2008 -0700
42665 Require libdrm 2.4.0 always since we need the bufmgr code.
42667 This lets us remove a lot of conditional compile stuff in the DRI case, as
42668 if we're doing DRI and have 2.4.0, we can rely on GEM ioctls existing.
42670 commit 754df0aa55c4b5f0c3c415ad97d7612504410229
42671 Author: Eric Anholt <eric@anholt.net>
42672 Date: Wed Jun 11 14:43:38 2008 -0700
42674 [gem] Chase move of create ioctl from generic to device-specific.
42676 commit 339f00e340c3a9c8ac8f17147e37906494ef370b
42677 Author: Eric Anholt <eric@anholt.net>
42678 Date: Tue Jun 10 16:14:06 2008 -0700
42680 [gem] Don't set up the ring in GEM mode, as that'll be handled by the kernel.
42682 commit e930c0dc5b5f98d256a4167ed70ba813c88924d2
42683 Author: Eric Anholt <eric@anholt.net>
42684 Date: Sat Jun 7 13:57:59 2008 -0700
42686 Use bufmgr_gem when available instead of the fake bufmgr.
42688 This is a bit unstable still sometimes.
42690 commit ecf19e1cda60a938d41413075ae6e00f24e0ec1a
42691 Author: Eric Anholt <eric@anholt.net>
42692 Date: Fri Jun 6 15:27:44 2008 -0700
42694 Change most usage of pixmap offsets to using a reloc macro.
42696 This is based on airlied's RING->BATCH commit. The 965 code still needs to
42697 be fixed up for relocations.
42699 commit bade7d7d2505a10a8a7d24b084aff9742e2d6d64
42700 Author: Eric Anholt <eric@anholt.net>
42701 Date: Fri Jun 6 14:03:25 2008 -0700
42703 Use the DRM for submitting batchbuffers when available.
42705 There are some concerns with this, as the DRM will be setting the nonsecure
42706 flag on the batchbuffer, and the server may be submitting some secure-only
42707 commands. It appears to work on the 915GM test system currently.
42709 commit 6e94affcc2240e668bcf1aa41f3c8b19929d144b
42710 Author: Eric Anholt <eric@anholt.net>
42711 Date: Fri Jun 6 14:01:05 2008 -0700
42713 Avoid needless flush emits in the blockhandler.
42715 The EmitFlush in i830_dri.c was added as a pageflipping workaround, and was
42716 noted to not even be sufficient then. There's no reason for it to be there,
42717 so it's removed. After that, we just have to not emit an MI_FLUSH if we
42718 already had, and cursor movement no longer bashes memory manager.
42720 commit b2216e7bc2f1a35f9fc1794bad83208cd5c583d1
42721 Author: Eric Anholt <eric@anholt.net>
42722 Date: Wed Jun 4 16:31:16 2008 -0700
42724 Use batchbuffers instead of ring emits for general commands.
42726 The batchbuffers are managed using libdrm and bufmgr_fake, and dispatched from
42727 the ring from userland.
42729 commit d0fda9d24c5fe4870cdbd2ba0213091165666a8e
42730 Merge: cc98d41 79f94ee
42731 Author: Eric Anholt <eric@anholt.net>
42732 Date: Tue Jun 10 11:32:49 2008 -0700
42734 Merge branch 'master' into drm-gem
42736 commit 79f94ee3c4652019395b9c329d94f7779c558571
42737 Author: Eric Anholt <eric@anholt.net>
42738 Date: Tue Jun 10 11:31:22 2008 -0700
42740 Set up/restore PWRCTXA from enter/leavevt not server start/exit.
42742 This should improve behavior in the presence of VT switching, but also avoids
42743 a crash on X exit from writing the register after unmapping mmio.
42745 commit 231a302013981cc597ba09ee89b367c8ab56e8ba
42746 Author: Bryce Harrington <bryce@bryceharrington.org>
42747 Date: Tue Jun 10 14:10:39 2008 +0800
42749 Two more Dell quirks
42753 commit cc98d41df63f1ee2561440fab3aa471a80a12920
42754 Author: Eric Anholt <eric@anholt.net>
42755 Date: Mon Jun 9 12:00:14 2008 -0700
42757 Add a little program to dump out the first 64 dwords of the status page.
42759 commit 1142be53eb8d2ee8a9b60ace5d49f0ba27332275
42760 Author: Jesse Barnes <jbarnes@hobbes.lan>
42761 Date: Mon Jun 9 08:52:59 2008 -0700
42763 Fix TV programming: add vblank wait after TV_CTL writes
42765 Fxies FDO bug #14000; we need to wait for vblank after writing TV_CTL or followi
42766 ng "DPMS on" calls may not actually enable the output.
42768 commit 64a8f2433d7774d06119793b57cec6d3be6389c1
42769 Author: Robert Lowery <rglowery@exemail.com.au>
42770 Date: Tue May 20 21:09:23 2008 +1000
42772 Fix TV out connection type detection
42774 Make sure we wait for vblank when using the TV DAC to detect the connection
42777 Fixes FDO bug #14000.
42779 commit a993a8f3ce3a2079c5b5ebaca2ec2564047d7d4e
42780 Author: Tomas Carnecky <tom@dbservice.com>
42781 Date: Fri Jun 6 10:03:23 2008 +0800
42783 Fix compile error in LeaveVT
42785 commit 432d07fb175b385715eb3ccf97a303d2ad7bfe11
42786 Merge: 3a967b8 bff180e
42787 Author: Eric Anholt <eric@anholt.net>
42788 Date: Thu Jun 5 14:57:18 2008 -0700
42790 Merge branch 'master' into drm-gem
42792 commit bff180e6cac4452ef491c81855eb12bfa03d0bf3
42793 Author: Eric Anholt <eric@anholt.net>
42794 Date: Thu Jun 5 14:55:23 2008 -0700
42796 Move debug clock printout from ErrorF to X_INFO.
42798 commit ad459b21b7de4a79552ac155803d5930432fb84b
42799 Author: Eric Anholt <eric@anholt.net>
42800 Date: Thu Jun 5 14:37:59 2008 -0700
42802 Remove SVG_WORK_CONTROL init.
42804 The bit set is now reserved -- used to be a workaround for early revisions.
42806 commit e41dc98a0aeec389b221038753056c725ecee1c5
42807 Author: Eric Anholt <eric@anholt.net>
42808 Date: Thu Jun 5 14:37:26 2008 -0700
42810 Initialize clock gating from EnterVT and save/restore it appropriately.
42812 commit 8061e5ac27a5f61f940bccc940be922999cc1d3f
42813 Author: Eric Anholt <eric@anholt.net>
42814 Date: Thu Jun 5 14:09:45 2008 -0700
42816 Move BIOS scratch register setup to EnterVT instead of PreInit.
42818 We want these to always be set when our driver's in control. They are
42819 already appropriately save/restored at leave/entervt.
42821 commit b61cb9283185eb5211e84eb7d8e68beea607c2eb
42822 Author: Eric Anholt <eric@anholt.net>
42823 Date: Thu Jun 5 13:57:54 2008 -0700
42825 Remove gratuitous wait_ring_idle after I830Sync. Syncing implies that.
42827 commit b65f67b74ff99aea1490d52f170ef4fe80cef6c7
42828 Author: Eric Anholt <eric@anholt.net>
42829 Date: Thu Jun 5 13:40:55 2008 -0700
42831 Move the remaining two lines of SetHWOperatingState to the caller.
42833 commit adb4f5a5e826e584ab212d23fc8d474c3e7bb8e8
42834 Author: Eric Anholt <eric@anholt.net>
42835 Date: Thu Jun 5 13:39:08 2008 -0700
42837 Remove duplicated i830_stop_ring()/SetHWOperatingState() in EnterVT().
42839 commit c2e2fe48113667c683c6e1e9b1237635c41c61c3
42840 Author: Eric Anholt <eric@anholt.net>
42841 Date: Thu Jun 5 13:21:55 2008 -0700
42843 Move DSPARB setup into a separate function, and save/restore it at VT switch.
42845 commit 7e51384c973a96366b02ea646392c43574674111
42846 Author: Eric Anholt <eric@anholt.net>
42847 Date: Thu Jun 5 11:56:42 2008 -0700
42849 Replace a couple of wait-for-ring-idles with a single function to do so.
42851 commit 65306cdd71dad71e4ca7229764f81a0880dd70bf
42852 Author: Eric Anholt <eric@anholt.net>
42853 Date: Thu Jun 5 11:26:02 2008 -0700
42855 Just remove the mprotect kludge.
42857 Besides not being #ifdef __linux__ed as requested, some linux kernels break
42858 in exciting new ways when you try to mprotect from PROT_NONE back to
42859 PROT_READ|PROT_WRITE. Yes, there are bugs in the code we're calling in a
42860 bug-exploiting bug workaround.
42862 If you want this workaround for the original bug exposed when moving to
42863 libpciaccess, it's already in libpciaccess.
42865 commit beb72ae5aa053303f5cc419e9c9d7c6db964f160
42866 Author: Hong Liu <hong.liu@intel.com>
42867 Date: Wed Jun 4 16:58:05 2008 +0800
42869 Set SDVO sync polarity to default on 965
42873 commit c7fee208fd46e143965ea173984d284e1eec2a9b
42874 Author: Hong Liu <hong.liu@intel.com>
42875 Date: Wed Jun 4 16:56:50 2008 +0800
42877 Fix DSPARB change on 965G
42879 From the spec, only 965GM and IGD_GM have 128 FIFO entries.
42880 With DSPARB change introduced by commit bd137a, I've got PIPE B
42881 underrun when dual-headed on G35 platform.
42883 commit 64ccc8a036a1d80fc918e9b124d24ce0d26bdbbb
42884 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42885 Date: Wed Jun 4 22:30:04 2008 +0800
42887 xvmc: add render dump function
42889 commit 1729a4f29dd52346a9fa997f818d57884047657f
42890 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42891 Date: Wed Jun 4 22:25:39 2008 +0800
42893 xvmc: move block size lookup helper to generic
42895 commit 3cc8876db7ba82fa18e8fab4f267ac6e87964c47
42896 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42897 Date: Tue Jun 3 16:15:36 2008 +0800
42899 xvmc: a little cleanup
42901 Also safe check context size to not exceed surface max.
42903 commit 96955003557686469c7ae9f7d42620d2851e6fad
42904 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42905 Date: Tue Jun 3 16:13:52 2008 +0800
42907 xvmc: remove unused dri drawable
42909 commit bd137a19dc29dd466eac030e040f729ed0807e3f
42910 Author: Jesse Barnes <jbarnes@hobbes.lan>
42911 Date: Wed May 28 15:47:47 2008 -0700
42913 Fixup DSPARB for 855 & 945
42915 It turns out 855 has a different DSPARB layout than 915+. And 945+ have more
42916 FIFO entries, so we have to allocate things differently. So on 855 split the
42917 FIFO evenly again between A & B planes, and do the same on 945, where we have a
42918 larger FIFO. Fixes an issue reported by Daniel Stone with the previous default
42921 commit 3a967b8359bd6c05a10111076cdb02df15cbf2e8
42922 Author: Eric Anholt <eric@anholt.net>
42923 Date: Mon May 12 13:51:11 2008 -0700
42925 [gem] Note if pinning a buffer fails.
42927 commit 363ccc3dfbf8ab91608d502e0376a43653581856
42928 Author: Jesse Barnes <jbarnes@hobbes.lan>
42929 Date: Mon May 26 09:43:40 2008 -0700
42933 What I originally checked in was a bit misleading.
42935 commit 2e1425246ccc75216247b0c2fa6fce2635db472b
42936 Author: Jesse Barnes <jbarnes@hobbes.lan>
42937 Date: Mon May 26 09:40:10 2008 -0700
42939 Handle display FIFOs better
42941 Add some debug code to catch FIFO underruns, which are normally bugs (unless
42942 they occur during mode setting) and remove any plane C FIFO allocations, since
42943 we don't use that plane at all. We may eventually need to be a little smarter
42944 about this on platforms that use plane C for the popup.
42946 commit 89bb53cc7a853d88fc34a0ca65ae2b6227a8dd24
42947 Author: Jesse Barnes <jbarnes@hobbes.lan>
42948 Date: Mon May 26 09:34:34 2008 -0700
42950 Fixup power saving registers
42952 Update clock gating disable bits to match docs and allocate a power context
42953 memory area so that newer chips can save state and power down the render unit.
42955 commit ced4b4bb0c44b55feb5e32ddb860f1c1dc3bbde5
42956 Author: Eric Anholt <eric@anholt.net>
42957 Date: Thu May 22 22:10:25 2008 -0700
42959 [gem] Reduce console spam from debugging.
42961 commit 165c0865d849b7d280a3a119fe9ae0ad34637df0
42962 Author: Alan Coopersmith <alan.coopersmith@sun.com>
42963 Date: Wed May 21 14:56:27 2008 -0700
42965 Define DEFFILEMODE for OS'es that don't have it
42967 commit 71180653825a1b141a08590e4b767d33d9b5d8c1
42968 Author: Jesse Barnes <jbarnes@hobbes.(none)>
42969 Date: Wed May 21 11:51:55 2008 -0700
42971 Revert "Disable FBC by default on 965GM"
42973 This reverts commit 53e3693ef13f31f3fc33bcff7286ab2b03b2d430.
42977 src/i830_driver.c - default FBC on for 965+
42979 commit 740a73dad1d9a9b0ebf1cc7d6f2a056abad57aaf
42980 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
42981 Date: Tue May 20 08:54:01 2008 +0800
42983 Revert "Add FIFO watermark regs to register dumper"
42985 This reverts commit 0c00a638ef57aa9d6a3047176b0bfad733f781f0.
42987 Those FIFO watermark regs are 945-ish, and cause problem
42990 commit c81050c0058e32098259b5078515807038beb7d6
42991 Merge: 9c9a5d0 e9532f3
42992 Author: Keith Packard <keithp@keithp.com>
42993 Date: Sat May 17 23:26:14 2008 -0700
42995 Merge commit 'origin/master' into drm-gem
42997 commit 9c9a5d0e48c6e911574695fc417d2dc1a0fd1c20
42998 Author: Keith Packard <keithp@keithp.com>
42999 Date: Tue May 13 12:24:50 2008 -0700
43001 Add check for GEM, use that to enable driver GEM support
43003 commit e9532f3abf1e0b62b2569d4aae60a5a63add571f
43004 Author: Rémi Cardona <remi@gentoo.org>
43005 Date: Wed May 14 14:52:37 2008 +0800
43007 Add glproto to DRI dependencies
43009 commit fe455a38cbac8d47a5c82b3eeb3cc6a59a6e8443
43010 Author: Eric Anholt <eric@anholt.net>
43011 Date: Tue May 13 11:56:10 2008 -0700
43013 Formatting fixes for intel_idle.
43015 commit 5c1720ecbc5528c40bb48084dc68ef233e9bb43b
43016 Author: Eric Anholt <eric@anholt.net>
43017 Date: Mon May 12 16:48:05 2008 -0700
43019 Add i915 support to intel_idle.
43021 commit 089011daf3da8db3bd16d50e1d6a6457da82300e
43022 Author: Keith Packard <keithp@keithp.com>
43023 Date: Mon May 12 12:18:19 2008 -0700
43025 Track name changes in GEM ioctls.
43028 unreference → close
43031 commit 10f1d835b8ac7bf3153ac39d295ad2a72c4aa076
43032 Author: Keith Packard <keithp@keithp.com>
43033 Date: Mon May 12 12:15:09 2008 -0700
43035 Record alignment requirements in mem structure for use by GEM.
43037 GEM needs memory alignment requirements sent at pin time, which is a bit
43038 after the allocation itself. Store the required alignment in the memory
43039 object for later use by pin.
43041 commit 5103e5a39a4869a714b8a59da8bf51ecc8c65e4f
43042 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
43043 Date: Mon May 12 10:58:24 2008 -0700
43045 Panel fitting: fix letterbox modes
43047 In full_aspect mode, we try to preserve the aspect ratio by adding
43048 either top & bottom or left & right borders. In the letterbox case (top
43049 & bottom borders) we were miscalculating the top border which led to
43050 programming a bad mode. Fix the calculation and bug #15559.
43052 commit f280c50dd5d29c5717adc6da2368363cb3d0adbb
43053 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43054 Date: Mon May 12 08:57:56 2008 +0800
43056 Revert last HP 965GM pipe A quirk
43058 Seems not resolve the issue (fdo bug #15885).
43060 commit d91d9e6a2f2ba18b35cb6fd7bc3fe8bc617eb44f
43061 Author: Bryce Harrington <bryce@bryceharrington.org>
43062 Date: Sat May 10 07:58:59 2008 +0800
43064 More Pipe A force quirks
43066 commit 548d4c0f979a72eac5d0c9c76da12f0bd05e64e1
43067 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43068 Date: Thu May 8 12:43:30 2008 +0800
43070 Fix i965 render getting dest format error
43072 Depend on value returned by function within assert is wrong.
43073 Fixed weird render corrupt on i965.
43075 commit 9f324860431ff8199a78d19bbaa74046e1476b89
43076 Author: Mike Isely <isely@pobox.com>
43077 Date: Thu May 8 10:00:17 2008 +0800
43079 Implement option to ignore external fixed mode settings
43081 The Intel xorg driver tries mightily to determine the native fixed
43082 panel mode settings for the LVDS output. It does this through various
43083 means, including scanning video BIOS tables, and noticing if the pipe
43084 in question has already been set up by somebody else (and adopting
43085 those timings). This strategy works well for say a laptop where the
43086 LCD panel is an integral part of the machine. But for other
43087 applications where the display is unrelated to the system's BIOS or
43088 other software, then the BIOS will likely have no clue how to
43089 configure the LVDS output. Worse still, the BIOS can simply "get it
43090 wrong", leaving the pipe misconfigured. Unfortunately the Intel
43091 driver can potentially notice this, adopt the same settings, leaving a
43094 All of this complexity normally happens independently, behind the
43095 scenes, from the mode timings that might otherwise be specified by the
43096 user. This driver has a concept of fixed, i.e. "native" mode, and
43097 then user-specified mode. If the corresponding resolutions between
43098 those concepts don't match, then the driver in theory will arrange for
43099 scaling to take place while adhering to the actual native mode of the
43100 panel. Said another way, if the user says 800x600 but the driver
43101 mistakenly (see above) thinks the native mode is 640x480, then 640x480
43102 is the mode set with scaling to an 800x600 frame buffer. If the
43103 driver gets the wrong native mode, then the result is a miserable mess
43104 with no way for the user to override what the driver thinks is right.
43106 This patch provides a means to override the driver. This implements a
43107 new driver option, "LVDSFixedMode" which defaults to true (the normal,
43108 probe-what-I-need behavior). However when set to false, then all the
43109 guessing is skipped and the driver will assume no fixed, i.e. "native"
43110 mode for the display device. Instead with this option set to false,
43111 the driver will directly set the timings specified by the user,
43112 providing an escape hatch for situations where the driver can't
43113 correctly figure out the right mode.
43115 Under most scenarios of course, this option should not be needed. But
43116 in situations where the Intel video BIOS is hopelessly fouled up
43117 related to the LVDS output, this option provides the escape hatch for
43118 the user to get a working display in spite of the BIOS situation.
43120 Signed-off-by: Mike Isely <isely@pobox.com>
43122 commit 33f033cbf346c13a687e469e8879579fcd5bb2fb
43123 Author: Jesse Barnes <jbarnes@hobbes.lan>
43124 Date: Wed May 7 14:46:52 2008 -0700
43126 Save/restore cursor registers
43128 There are lots of good reasons for doing this, one of them is fdo bug #11305.
43130 commit a7188b1f2dd9a69fa7daefe478d283735226f9f3
43131 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43132 Date: Wed May 7 14:21:59 2008 +0800
43134 Use I830FALLBACK when check pitch for debug
43136 commit 79fde3ad7a949fb7f0dc9907cb6eea6ca9a910ae
43137 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43138 Date: Sat Apr 19 08:54:38 2008 +0800
43140 Check pitch for EXA operation
43142 2D pitch limit applys to all chips. Pre-965 chip has
43143 8KB pitch limit for 3D. 965 supports max pitch by current
43145 (cherry picked from commit 8187a5a16f8bd8f0ba5e7f5357f355928b3b8f07)
43147 commit 36ec93300926084fb2951d69b001e4c67bc6ff79
43148 Author: Eric Anholt <eric@anholt.net>
43149 Date: Tue May 6 18:48:20 2008 -0700
43151 Bug #15807: Fix use of the ring while VT-switched, hit by fast user switching.
43153 The fix for flushing at blockhandler with no DRI on 965 was broken and would
43154 try to flush the chip even when the driver wasn't in control of the VT.
43157 commit 0c00a638ef57aa9d6a3047176b0bfad733f781f0
43158 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
43159 Date: Tue May 6 14:38:48 2008 -0700
43161 Add FIFO watermark regs to register dumper
43163 commit c81a4687fca80bf7367d7f0e9a00a6a09737c5bb
43164 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43165 Date: Tue May 6 13:52:22 2008 +0800
43167 Only check xvmc lib when xvmc is enabled.
43169 Don't check xvmc lib if user has already wanted to disable it.
43170 Fix fdo bug #15762.
43172 commit be746a90a87d7a9807fa4243493e7e4d48f7f1c0
43173 Author: Bryce Harrington <bryce@bryceharrington.org>
43174 Date: Tue May 6 13:51:46 2008 +0800
43176 More quirks from ubuntu/dell
43178 commit 417f86d80525d2f5038628afba8f70f8c0fecca3
43179 Author: Keith Packard <keithp@keithp.com>
43180 Date: Mon May 5 13:36:19 2008 -0700
43182 Move a declaration under #ifndef HAVE_FREE_SHADOW
43184 commit 587dc5bfc4ef32942bd7511f932a3a2e2fcdccb0
43185 Author: Keith Packard <keithp@keithp.com>
43186 Date: Mon May 5 13:35:22 2008 -0700
43188 Fix up ring dumping code for non-i965
43190 commit 65877d32bb2cd71ccc8d6a66244f029fc3577f69
43191 Merge: b9d12da fff17b9
43192 Author: Keith Packard <keithp@keithp.com>
43193 Date: Mon May 5 12:05:00 2008 -0700
43195 Merge branch 'master' into drm-gem
43197 commit b9d12da91eb1afb05bd78a40e15f4333e697093d
43198 Author: Keith Packard <keithp@keithp.com>
43199 Date: Mon May 5 10:44:09 2008 -0700
43201 Use GEM for buffer naming now.
43203 commit 0741020f3725c32c48ad15535b9c0549e4092c23
43204 Author: Keith Packard <keithp@keithp.com>
43205 Date: Mon May 5 10:17:01 2008 -0700
43207 Set pin alignment for gem on non-965
43209 non-965 tiled frame buffers have fairly strict alignment requirements, 512K
43210 on 8xx and 1MB on 9xx, plus they must be aligned to the size of the
43213 commit 271abfd0ec29851273d8ca5241d963be5c64e43c
43214 Author: Eric Anholt <eric@anholt.net>
43215 Date: Fri May 2 10:06:13 2008 -0700
43217 Add initial GEM hacks to bring the server up.
43219 commit a0ced923bb793aa22e6bfbeeec0888d3b42ce176
43220 Author: Jesse Barnes <jbarnes@hobbes.lan>
43221 Date: Tue Apr 29 13:19:02 2008 -0700
43223 Add a new quirk for BIOSes that reprogram regs at lid close/open time
43225 Dell Latitude D500s seem to have this problem. At lid close/open, the DSPABASE
43226 reg gets reset to 0, so we either need to keep the framebuffer at offset 0 or
43227 make sure we reprogram the CRTCs after the lid opens again. Since we can't
43228 make sure the former is always true (buffer resize, etc.), this patch adds a
43229 quirk to reset the modes at lid open time.
43231 Fixes FDO bug #14890.
43233 commit fff17b9d1b58cb53032d153094826dd306836d59
43234 Author: Keith Packard <keithp@keithp.com>
43235 Date: Tue Apr 29 10:32:14 2008 -0700
43237 Use new xf86RotateFreeShadow function to clean up shadow buffers.
43239 This simply moves code from the driver up into the X server; use it where
43242 commit 1d467a8038946a37844795e8860be113d43219ac
43243 Author: Keith Packard <keithp@keithp.com>
43244 Date: Wed Apr 23 11:08:38 2008 -0700
43246 Overlay video doesn't require that the target pixmap be in video memory.
43248 I830PutImage was checking to make sure the target pixmap resided in video
43249 memory, but this isn't necessary when using the overlay. Test
43251 commit 40e0a03af57d7b3180d7066bcb15e03dcc9ca295
43252 Author: Keith Packard <keithp@keithp.com>
43253 Date: Mon Apr 21 01:03:22 2008 -0700
43255 Dump batch buffers found from main ring
43257 commit c8ae3b781f0d8e325876a74c91cd0a685d34454b
43258 Author: Keith Packard <keithp@keithp.com>
43259 Date: Sun Apr 20 02:11:15 2008 -0700
43261 Add a bunch of 965 ring stuff to the debug dump
43263 commit 2c135ef8ac40f8e7cd071de7414adfae019f9198
43264 Author: Eric Anholt <eric@anholt.net>
43265 Date: Wed Apr 16 21:39:58 2008 -0700
43267 Make the binding table and surface state be arrays to reduce syncing.
43269 commit c3fb62df4e60b63295f94c99b3c5de70dbf94e1c
43270 Author: Keith Packard <keithp@keithp.com>
43271 Date: Thu Apr 17 10:04:55 2008 -0700
43273 Add a kludge-around to fix cd/wt bits in fb ptes on linux.
43275 Mmap from /sys/devices/pci* on linux forces the cache-disable and
43276 write-through bits, which turns our write-combining map into an
43277 uncached-map, seriously impacting performance. It turns out that a bug in
43278 mprotect allows us to fix this by disabling access to those pages and then
43279 immediately re-enabling them.
43281 commit 0ae283582d21776d3317d5fc1c25751d50d562c7
43282 Author: Hong Liu <hong.liu@intel.com>
43283 Date: Thu Apr 17 10:57:34 2008 +0800
43285 fix possible segfault in I830FreeScreen
43287 pI830 may point to NULL if I830PreInit fails
43289 commit 7008eac855302fd467e001495483c2dccfddf68f
43290 Author: Eric Anholt <eric@anholt.net>
43291 Date: Wed Apr 16 12:09:41 2008 -0700
43293 Update dolt from git, fixing fallback to libtool.
43295 commit 23d1df22d177e54bfc46304053d8115047ff85d4
43296 Author: Eric Anholt <eric@anholt.net>
43297 Date: Mon Apr 14 16:46:17 2008 -0700
43299 Abstract surface setup into a separate function.
43301 commit e27ce4238f748d024635835824c120eb298b84b2
43302 Author: Eric Anholt <eric@anholt.net>
43303 Date: Mon Apr 14 23:30:21 2008 -0700
43305 Move the vertex buffer to the state structure as well.
43307 commit 69709fe36cfdb59c140f5cba233dcb606f8dbc85
43308 Author: Eric Anholt <eric@anholt.net>
43309 Date: Mon Apr 14 18:34:13 2008 -0700
43311 Move the binding table for 965 to the state structure.
43313 commit 049d057dfd73b6c191a8ba5d73eddb77cf7748e7
43314 Author: Eric Anholt <eric@anholt.net>
43315 Date: Mon Apr 14 17:09:22 2008 -0700
43317 Move the 965 static variables local to the function they're used in.
43319 commit 092962c962fdf3e33a8d387221448fe7c8c36376
43320 Author: Eric Anholt <eric@anholt.net>
43321 Date: Mon Apr 14 14:02:58 2008 -0700
43323 Make the VS unit state static for 965 render.
43325 commit 4bbdd7096d4927fe48673006fff7df922972f116
43326 Author: Eric Anholt <eric@anholt.net>
43327 Date: Mon Apr 14 13:53:36 2008 -0700
43329 Rename some variables in i965_render.c for clarity.
43331 The gen4_render_state is now always called "render_state" (i965_render.c
43332 bookkeeping) and gen4_state_t is now always called "card_state" (the buffer
43333 for state used by the chip).
43335 commit 456bb529335c28cf60f3b6e5900b804efa5c185a
43336 Author: Carl Worth <cworth@cworth.org>
43337 Date: Fri Nov 9 17:24:21 2007 -0800
43339 Associate one sf_state object with each sf_kernel
43341 (cherry picked from a2b5c23184d19b386fdfd04f578a55566df60132 commit)
43343 commit 90d6b178473ba32cf66e6e654e608cb4374e4a19
43344 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43345 Date: Mon Apr 14 16:31:03 2008 +0800
43347 Change default panel fitting mode to origin
43349 i8xx currently only works in FULL mode.
43350 (cherry picked from commit 33ffd781bbca3d0dee8c1b47e7b90be5824b9a4f)
43352 commit c1ad0a1a344a87a537aa5f93f6c774997a3189e4
43353 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43354 Date: Mon Apr 14 10:02:00 2008 +0800
43356 Remove old sf/wm program from Makefile
43359 (cherry picked from commit 1d8383923bebf38ab8ada6a262cb4fc8c30e074a)
43361 commit 18ef4158e5574bfc6621b268821532f13e261d13
43362 Author: Keith Packard <keithp@keithp.com>
43363 Date: Sun Apr 13 21:37:49 2008 -0700
43365 Only use FOURCC_XVMC when INTEL_XVMC is defined
43367 The XVMC code uses a magic FOURCC code to signal frame updates, but that
43368 code is only defined when the XVMC code is used.
43370 commit 269809030ee9d08b56ad0178a7a9c0194e25785e
43371 Author: Keith Packard <keithp@keithp.com>
43372 Date: Sun Apr 13 12:47:02 2008 -0700
43374 Move wm_scratch first to meet alignment requirements
43376 commit 08d3206f3b891e611e3e16c9eb79cfca21ab8a51
43377 Author: Keith Packard <keithp@keithp.com>
43378 Date: Sun Apr 13 12:35:14 2008 -0700
43380 Kludge to make planar video work - duplicate sampler state
43382 This patch duplicates all of the surface sampler state needed to read from
43383 the separate YUV surfaces and then has the planar sampler read from samplers
43384 0, 2, 4 instead of 0, 1, 2. This appears to make things work, instead of
43385 having the samplers break at random. I do not understand why this works.
43387 commit 6366e4de540c4713cbc8402c89516464bbdc35b9
43388 Author: Keith Packard <keithp@keithp.com>
43389 Date: Sat Apr 12 16:39:00 2008 -0700
43391 Skip copying on FOURCC_XVMC surfaces
43393 commit 0fec42b64ee529eb7ec15febdaa5e4986ec1ab17
43394 Author: Keith Packard <keithp@keithp.com>
43395 Date: Sat Apr 12 11:01:14 2008 -0700
43397 Set some mask bits to try and get yuv conversion working
43399 commit 91d3e19786e22061f4b617cf39a3cd713139a2d9
43400 Author: Keith Packard <keithp@keithp.com>
43401 Date: Sat Apr 12 11:00:51 2008 -0700
43403 Use compressed instructions for yuv conversion
43405 commit 03836067b77606c134c71b30c7078d09d77c95fa
43406 Author: Eric Anholt <eric@anholt.net>
43407 Date: Fri Apr 11 12:17:28 2008 -0700
43409 Fix the offset to sampler default color, and a compiler warning.
43411 commit d3138d9ff0f821fb3adbd27684b0c22a8a910c3e
43412 Author: Carl Worth <cworth@cworth.org>
43413 Date: Tue Nov 13 12:37:13 2007 -0800
43415 Enumerate all possible wm_state objects
43417 We have a collection of wm_state objects for each ps kernel,
43418 (one for each combination of src and mask extend and repeat
43421 Thanks to Dave Airlie for noticing an errant write through a
43422 wild wm_state pointer in an early version of this commit.
43424 (cherry picked from 7763706a93d3021907273f9b330750ba110e2fc3 commit)
43426 This cherry-pick required more reformatting than most, due to the
43427 projective texturing merge.
43429 commit 2293a3677d1dcf294de6a8712bf0d2f65b50dcc3
43430 Author: Carl Worth <cworth@cworth.org>
43431 Date: Mon Nov 12 16:09:00 2007 -0800
43433 Enumerate all possible src,mask sampler state pairs
43435 This will eventually allow for the elimination of sampler state
43436 updates while compositing---and initializing everything in the
43437 initialization function.
43439 (cherry picked from commit d0874697be8086cd64740c24698df8cd4d31c76f)
43441 commit da85d1b822dcb31698e9c5ab85a7bb27ad745eee
43442 Author: Keith Packard <keithp@keithp.com>
43443 Date: Fri Apr 11 11:22:29 2008 -0700
43445 Revert "remove '#line NUM ...' in macro process"
43447 This reverts commit f47486fab3dffcbb03e7ad89f777abba1e887299.
43449 Sorry for failing to push the #line processing changes to intel-gen4asm;
43450 those are now pushed, so this change should no longer be necessary. Having
43451 correct file name and line numbers in the error messages is awfully nice
43454 commit f47486fab3dffcbb03e7ad89f777abba1e887299
43455 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43456 Date: Fri Apr 11 10:12:40 2008 +0800
43458 remove '#line NUM ...' in macro process
43460 intel-gen4asm doesn't allow '#' line
43462 commit 5c9cde37e769287fb7bf4e08c3600a33c2e92dce
43463 Author: Hong Liu <hong.liu@intel.com>
43464 Date: Fri Apr 11 09:54:34 2008 +0800
43466 Bug #14935: Fix i9xx reference clock for spread spectrum.
43468 commit 152a50703aa5e9ebaa9abbe448518742734a5eb7
43469 Author: Eric Anholt <eric@anholt.net>
43470 Date: Thu Apr 10 16:06:41 2008 -0700
43472 Fix compiler warning from 24-bit lvds change.
43474 commit baa55ffe3b67ca952e4d1ef7087335d76eb2f24b
43475 Author: Carl Worth <cworth@cworth.org>
43476 Date: Tue Nov 13 12:56:13 2007 -0800
43478 Enumerate all possible cc_state objects
43480 We need one for each possible combination of src and dst
43481 blend_factors. Again, as with recent changes, this eliminates
43482 state updates from prepare_composite and allows that function
43483 to instead simply reference an existing object initialized
43484 within gen4_state_init.
43486 Thanks to Dave Airlie (and git-bisect) for pointing out that with
43487 gnome-terminal all text was appearing as solid black with an early
43488 version of this commit. As expected the bug was an alignment issue.
43490 (cherry picked from 0c0ab52c2d100c47f38c7ef826ef585c8b9815e9 commit)
43492 Performance is approximately equivalent on text tests, but may be
43495 commit d5a80e1e3ab5724d34b20f9ee6f830efd0f5b076
43496 Author: Keith Packard <keithp@keithp.com>
43497 Date: Wed Apr 9 16:27:40 2008 -0500
43499 Single memcpy when pitches align on planar image transfer
43501 commit ac97f2b1487df5574875350a9cded958dae33afa
43502 Author: Keith Packard <keithp@keithp.com>
43503 Date: Wed Apr 9 16:27:23 2008 -0500
43505 Use available symbolic register name
43507 commit f270456e5612cb88933e6aabcd9a816c5c292229
43508 Author: Keith Packard <keithp@keithp.com>
43509 Date: Wed Apr 9 10:00:08 2008 -0500
43511 Remove .g4b files on clean
43513 commit fb6ed8b8b59e9dd18801afef5d7c62042ad176d7
43514 Author: Keith Packard <keithp@keithp.com>
43515 Date: Wed Apr 9 09:58:44 2008 -0500
43517 Update yuv->rgb conversion programs to write to src_sample regs
43519 The YUV->RGB code was written to write directly to the dataport registers,
43520 but that didn't work for the compositing functions (cause still unknown).
43521 This change makes that code write RGB values to the src_sample registers as
43522 with the other sample computation fragments.
43524 commit b01d582e23fc99e32bc47a395e9caa366731372a
43525 Author: Keith Packard <keithp@keithp.com>
43526 Date: Wed Apr 9 09:30:35 2008 -0500
43528 Revert "Compute pixel values directly into data port"
43530 This reverts commit 346cf57deabb4c336612df4c13650a87b5ef6775.
43532 Mixing randr transforms and video caused screen corruption for Render
43533 operations. No, I don't understand why.
43535 commit 3fc3d1a701bae257b70aa7b7654c722f30e71399
43536 Author: Keith Packard <keithp@keithp.com>
43537 Date: Wed Apr 9 02:02:56 2008 -0500
43539 Remove sync after 965 video put.
43541 The hardware has been marked as needing a sync, so the next video put will
43542 block waiting for the previous one to complete. Adding a sync here just
43543 stalls the video playback for no good reason.
43545 commit 825d9e50c59450f07178a54fed2616e551dc0455
43546 Author: Keith Packard <keithp@keithp.com>
43547 Date: Wed Apr 9 00:09:34 2008 -0500
43549 Add planer video decode kernel
43551 Support for planar video reduces bus bandwidth by 25% and also reduces CPU
43552 usage during planar->packed conversion.
43554 commit a03eaaa67b33c57530e92c53d28917e2563b4427
43555 Author: Keith Packard <keithp@keithp.com>
43556 Date: Tue Apr 8 16:24:55 2008 -0500
43558 Use shared exa_wm code for packed yuv decode
43560 Eliminate special video sf and ps programs.
43562 commit 32ef98518394d29cb87405005c660278489396bb
43563 Author: Keith Packard <keithp@keithp.com>
43564 Date: Tue Apr 8 16:21:55 2008 -0500
43566 Compute pixel values directly into data port
43568 Instead of leaving pixel values in src_sample registers, compute the pixel
43569 values directl to the data port to save 8 moves. This cannot work when no
43570 computation is done as there is both no way to wait for the sampler to
43571 finish and because the sampler returns data in a different order from that
43572 required by the data port (sigh).
43574 commit b68d9f4245d0ebe3371c179401ff145f1a4d101b
43575 Author: Keith Packard <keithp@keithp.com>
43576 Date: Tue Apr 8 16:09:00 2008 -0500
43578 Use symbolic names for channels in YUV code
43580 commit 781be9d47289713b0a8fcd95c769a9c6241d62e9
43581 Author: Keith Packard <keithp@keithp.com>
43582 Date: Tue Apr 8 16:08:20 2008 -0500
43584 Rename src/mask/data registers to indicate channel
43586 commit 2871ac8eefd0192080bb0569140c3f5d0e1d9b44
43587 Author: Eric Anholt <eric@anholt.net>
43588 Date: Thu Apr 10 13:34:13 2008 -0700
43590 Statically allocate the sampler default color, which we never change.
43592 Performance change is in the noise. Also from Carl Worth.
43594 commit 80dd784e33847e431403d4659a7b8d8425b2676f
43595 Author: Eric Anholt <eric@anholt.net>
43596 Date: Thu Apr 10 13:24:51 2008 -0700
43598 Add copyright information for recent editors of this file.
43600 commit b606278db83ec84b1db562a2d65697c50561b169
43601 Author: Eric Anholt <eric@anholt.net>
43602 Date: Thu Apr 10 13:17:58 2008 -0700
43604 Keep static copies of the 965 render programs in video memory.
43606 This reduces the CPU overhead of memcpying them in every time, for a speedup
43607 in aa24text of around 30%. This is based on work by Carl Worth which is
43608 in the intel-batchbuffer branch.
43610 commit bd41d2bb858770c109d5013274a4354e8e6ae32d
43611 Author: Eric Anholt <eric@anholt.net>
43612 Date: Thu Apr 10 10:24:48 2008 -0700
43614 Add dolt to speed up compile time and reduce insanity.
43616 Hopefully some day we can be rid of libtool entirely.
43618 commit 96ae9a9f0abc582e45ff4e17dd213c8739c24cb3
43619 Author: Eric Anholt <eric@anholt.net>
43620 Date: Wed Apr 9 14:17:29 2008 -0700
43622 SDVO: Rename output_dtd variable to input_dtd, which is its new job.
43624 commit 72c3995d5158286d20e97296f4552697f32f935f
43625 Author: Eric Anholt <eric@anholt.net>
43626 Date: Tue Apr 8 15:00:18 2008 -0700
43628 SDVO: Set adjusted_mode in mode_fixup from the preferred input timings for TV.
43630 commit b61178524bece756dc0fd4f798c6c7743b5725db
43631 Author: Eric Anholt <eric@anholt.net>
43632 Date: Tue Apr 8 14:11:42 2008 -0700
43634 SDVO: Fix lols TV mode list handling.
43636 commit 5db73f18d2badec37523850024b9aa824e4e6bd8
43637 Author: Eric Anholt <eric@anholt.net>
43638 Date: Tue Apr 8 13:53:51 2008 -0700
43640 SDVO: Use the TV reference clock for SDVO TV outputs.
43642 Based on a patch by Will Wiseman.
43644 commit e031cc02e65acfbafb48136dad414751e04425c5
43645 Author: Mike Isely <isely@pobox.com>
43646 Date: Wed Apr 9 14:15:03 2008 +0800
43648 Implement support for 24 bit pixel format
43650 The Intel driver appears to be coded to only work with displays
43651 expecting 18 bit pixels. However I have an application using a LCD
43652 display that expects pixel data in 24 bit format. The difference is
43653 only 2 bits in a single GPU register. This patch implements that
43654 change, controlled by a new driver option, "LVDS24Bit". The default
43655 value is false, which is the previous behavior. When set to true,
43656 then 24 bit panels should work (at least the one I'm testing here
43661 Signed-off-by: Mike Isely <isely@pobox.com>
43663 commit 79b18980ac9f60b1978abe421352df965aed1681
43664 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43665 Date: Wed Apr 9 14:03:45 2008 +0800
43667 Include <stdint.h> in i810.h for integer types definition
43669 commit 2b720262e1235f1c9da860ba3e9181f0c377aa5e
43670 Author: Jesse Barnes <jbarnes@virtuousgeek.org>
43671 Date: Wed Apr 9 17:34:17 2008 +0800
43673 Fix LVDS regression
43675 Disable panel fitting on 855GM, and fix dither setting.
43677 commit af92f4f885c8eae2211d09080b2289aa5bab9ec1
43678 Author: David Schleef <ds@schleef.org>
43679 Date: Tue Apr 8 12:43:20 2008 -0700
43681 Bug #15219: Fix default overlay CSC factors, and use BT709 if available.
43683 Using the updated factors even when BT709 conversion isn't available
43684 (non-965) should still give us better color reproduction. Tested on a
43685 945GM, examining the +/-5% of black bars of videotestsrc.
43687 commit 37bc23660a8c346f1eaa6c93ed2c7a840828f0b0
43688 Author: Bryce Harrington <bryce@bryceharrington.org>
43689 Date: Mon Apr 7 17:26:49 2008 +0800
43691 Quirks from Ubuntu/Dell
43693 FD bug #15353. Launchpad bug ID is available for reference.
43694 (cherry picked from commit 24248097988775d62f6c416f2988e74d31c91cd0)
43696 commit dcbdc901d64d2fab489292ef1c9ca1490bdd1e94
43697 Merge: 0147c1c 13563bd
43698 Author: Keith Packard <keithp@keithp.com>
43699 Date: Fri Apr 4 19:19:35 2008 -0700
43701 Merge commit 'origin/master' into projective-965
43703 commit 0147c1c84872f7a109721a53d88a539932d9be81
43704 Author: Keith Packard <keithp@keithp.com>
43705 Date: Fri Apr 4 19:17:55 2008 -0700
43707 Forgot to build exa_wm_src_sample_argb.g4b
43709 commit 13563bd65cbbbb3e697dacd814efa3517e1493dc
43710 Author: Hong Liu <hong.liu@intel.com>
43711 Date: Wed Apr 2 16:05:43 2008 -0700
43713 SDVO: Fix DDC bus selection based on output priority list.
43715 commit 6ad2d6ba86689674876f5f4c473f11e39243ac38
43716 Author: Eric Anholt <eric@anholt.net>
43717 Date: Wed Apr 2 15:16:17 2008 -0700
43719 SDVO: Fix mixups with input and output channels.
43721 The 2-bit input_mask was actually an input count -- in0 is always there, and
43724 The output flags weren't being reported in the log, so I mistakenly took
43725 controlled_output == RGB0 to mean that the device only reported an RGB0,
43726 while it actually reported RGB0|SVID0|YPRPB0|misc|other. Move SVID0 up
43727 in priority and remove the RGB-is-it-really-TV hack I had just come up with.
43729 Finally, set the input/output mapping at mode set time. We're always
43730 supposed to do this, but haven't had to so far as we've never handled
43731 devices with more than one output.
43733 commit 3a17400dc67534f0eb474ece080f01061469569c
43734 Author: Eric Anholt <eric@anholt.net>
43735 Date: Wed Apr 2 14:21:23 2008 -0700
43737 SDVO: Handle RGB outputs that are really TV outputs, and select a TV format.
43739 Still doesn't light anything up.
43741 commit 3c1701797b61afaae826a78455079e2115483053
43742 Author: Eric Anholt <eric@anholt.net>
43743 Date: Mon Mar 31 10:04:18 2008 -0700
43745 Add WIP SDVO TV-out support.
43747 Doesn't include properties for selecting TV formats or picture enhancements,
43748 and totally untested.
43750 commit 9d5ba26fb5c337388920b45eadda85e43bc564fa
43751 Author: Eric Anholt <eric@anholt.net>
43752 Date: Fri Mar 28 14:28:28 2008 -0700
43754 Add SDVO TV-out registers.
43756 commit 90886f9a602d58b754e9a8d0f1a9c40803d34fa2
43757 Author: Keith Packard <keithp@keithp.com>
43758 Date: Tue Apr 1 00:54:13 2008 -0700
43760 Remove extra flushing
43762 commit 0836373dc6e2f8612f120074980561f7ac11f6f7
43763 Author: Keith Packard <keithp@keithp.com>
43764 Date: Tue Apr 1 00:16:05 2008 -0700
43766 Add projective versions of the PS kernels
43768 commit bfd803e085e938866efb45c67a79facef78ec399
43769 Author: Keith Packard <keithp@keithp.com>
43770 Date: Tue Apr 1 00:06:08 2008 -0700
43772 Shrink WM thread to 32 registers and 1024 scratch space.
43774 Saving registers means we can run more in parallel.
43776 commit a6492661ae07310128eb73c3ef037c42ce7ab184
43777 Author: Keith Packard <keithp@keithp.com>
43778 Date: Mon Mar 31 23:50:20 2008 -0700
43780 Fix composite with mask using new compositing thread code
43782 Clean up register allocation to never overlap
43783 Always write 4 values for each texture vertex.
43785 commit f8081178eb6fda0e405967cbacad532561619262
43786 Author: Keith Packard <keithp@keithp.com>
43787 Date: Mon Mar 31 12:06:37 2008 -0700
43789 remove old shader source files
43791 commit 879f8717b09f79156b218ee9cc2107700190d586
43792 Author: Keith Packard <keithp@keithp.com>
43793 Date: Mon Mar 31 12:05:56 2008 -0700
43795 remove old monolithic shaders
43797 commit 6db8faeb754897b21af045d00f50db9640b080bb
43798 Author: Keith Packard <keithp@keithp.com>
43799 Date: Mon Mar 31 12:04:00 2008 -0700
43801 remove old monolithic shaders
43803 commit 6bb92213374f278387c539bbe05b773e87e11b90
43804 Author: Keith Packard <keithp@keithp.com>
43805 Date: Mon Mar 31 11:59:14 2008 -0700
43807 Back to new shaders, fix urb read length
43809 commit 05710145b6fc4ed2c528128b2e6022591a53d050
43810 Author: Keith Packard <keithp@keithp.com>
43811 Date: Mon Mar 31 11:31:31 2008 -0700
43813 Dont set the compr bit on 8-unit sends
43815 commit 08500507284f13ad7084eb231b43e117e9728129
43816 Author: Keith Packard <keithp@keithp.com>
43817 Date: Mon Mar 31 02:20:43 2008 -0700
43819 Use m4 to clean up gen4 asm progs. Start adding projective transform support.
43821 Use macros for register names, modularize functions into separate files.
43823 commit 949d73271d7100c1f028fd60f185f4929461304e
43824 Author: Keith Packard <keithp@keithp.com>
43825 Date: Sun Mar 30 19:19:46 2008 -0700
43827 Remove rotation sf and wm progs
43829 commit 492ff1494f782240e6ca68919b2d0b9aa400fc53
43830 Author: Keith Packard <keithp@keithp.com>
43831 Date: Sun Mar 30 19:14:18 2008 -0700
43833 Fix the sf_mask program to compute and pass corrected uvw cooefficients
43835 sf_mask is the same as sf except that it must compute both src and mask uvw
43836 cooefficients, which are conveniently adjacent in the same registers, and so
43837 need only an extended execution width
43839 commit 6304b38423f99190a5e54f1a7dcaa75adfad4f2a
43840 Author: Keith Packard <keithp@keithp.com>
43841 Date: Sun Mar 30 00:54:51 2008 -0700
43843 Reimplement wm program for nomask case to handle affine transforms
43845 This involves correctly computing u/v locations based on x/y vectors and
43846 line constants computed in new sf program.
43848 Also, use fewer instructions to make this go a bit faster (2X for 500x500
43851 commit 771a56b1ed0df69345c723cb62a73b6842cd8227
43852 Author: Keith Packard <keithp@keithp.com>
43853 Date: Sun Mar 30 00:54:07 2008 -0700
43855 Clarify comment in exa_sf program
43857 commit 4f469189fed541549e5d470b2529275a29cc2f20
43858 Author: Keith Packard <keithp@keithp.com>
43859 Date: Sat Mar 29 14:28:05 2008 -0700
43861 Compute du/dv/dw in no-mask SF prog
43863 commit 7dcb6e627449c80cea9812462ce6a3e125bd1240
43864 Author: Keith Packard <keithp@keithp.com>
43865 Date: Sat Mar 29 14:27:35 2008 -0700
43867 Disable cursors while switching modes
43869 commit d50264fc7c4e0f80d9cc68dfdf322f68520a0f1e
43870 Author: Eric Anholt <eric@anholt.net>
43871 Date: Fri Mar 28 15:15:26 2008 -0700
43873 Add a "total" field to intel_idle output (set when any unit is busy).
43875 commit 1450acd046d47e1739a3ffbd146c73ad2974a935
43876 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
43877 Date: Fri Mar 28 13:23:02 2008 -0700
43879 Use combo backlight control rather than pure legacy
43881 If the legacy bit is set, use both the BLC_PWM_CTL and LBB regs to control the
43882 backlight, rather than just LBB. Looks like more platforms want that than what
43883 the current code does. Note that kernel provided interfaces will always be
43884 used if available, so this shouldn't affect users with /sys/class/backlight
43889 commit 49f409fef5e8b04a8595309ba8959ca36b2e41b1
43890 Author: Matthias Hopf <mhopf@suse.de>
43891 Date: Wed Jan 9 16:31:41 2008 +0100
43893 Include missing stdlib.h for free()
43895 commit 3414313ac9d0faa95797ee18cc33afa231ec7581
43896 Author: Matthias Hopf <mhopf@suse.de>
43897 Date: Wed Jan 9 16:08:57 2008 +0100
43899 Fix obviously wrong verification of DRI_SUPPORTS_CLIP_NOTIFY.
43901 commit c40f195f2dc3467259b4588e087aac9741003ebf
43902 Author: Matthias Hopf <mhopf@suse.de>
43903 Date: Wed Jan 9 14:40:49 2008 +0100
43905 Fix obviously stray '(' in non-I830_USE_EXA case.
43907 commit 7bba2c13310ed5ac22a355a3cc0ec8b7afaa79cf
43908 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43909 Date: Fri Mar 28 11:41:24 2008 +0800
43911 Disable DRI earlier if fb width > 2048
43913 commit b1f358ba97473b792ec2b7ed5170152faebe7262
43914 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
43915 Date: Fri Mar 28 11:29:40 2008 +0800
43917 Disable overlay on new mobile chipset
43921 commit 588371ba15afffa9343629f0ea7ce530586eef0b
43922 Merge: 52d6ced c13efdd
43923 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
43924 Date: Wed Mar 26 14:53:15 2008 -0700
43926 Merge branch 'master' into pfit
43928 commit c13efdd3d6d3846078766814b57306a7519bce18
43929 Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
43930 Date: Wed Mar 26 10:11:36 2008 +0800
43932 More compile warning fixes.
43934 o Check for RANDR_GET_CRTC_INTERFACE before defining functions that
43935 are used only if it is defined.
43936 o Declare a variable before code, and rename it from ret to xvmc_status
43937 to better describe it.
43938 o if 0 some static functions not used.
43939 o Don't declare some unused variables.
43940 o Declare as static some functions that are used only in the file defining it.
43941 o Add a default/fallback return True to the Bool function
43942 src/xvmc/intel_batchbuffer.c:intelInitBatchBuffer().
43943 o Ansify src/xvmc/xf86dri.c.
43944 o Add missing prototype to src/xvmc/xf86dri.h and follow pattern of other
43945 headers by adding "extern" before function prototype.
43947 commit 8cdbd55f8075cd18b563badde35815665d7d053e
43948 Author: Eric Anholt <eric@anholt.net>
43949 Date: Tue Mar 25 10:06:14 2008 -0700
43951 Fix 965+ rendering issues with DRI disabled.
43953 The new chips no longer automatically flush the rendering cache, so if we
43954 don't flush the RC at blockhandler, the last rendering done may not
43955 appear on the screen. This was particularly noticable with a bare Xorg with
43956 some missing root weave, and terminals where the last character wouldn't
43957 appear until the cursor blinked. A flush in the DRI blockhandler path had
43958 hidden this issue for most people.
43960 commit c02ab432dd7058c700c35eecf6215daf5f262c51
43961 Author: Eric Anholt <eric@anholt.net>
43962 Date: Mon Mar 24 13:48:20 2008 -0700
43964 Revert "Use mprotect on unbound AGP memory to attempt to catch use while unbound."
43966 While I still like the idea, the mprotect calls themselves are failing on
43967 Linux and causing more trouble than they're worth.
43969 This reverts commit a1612b7728d4153499fe86b6713a13c8702cc7d9.
43976 commit 552a1b824db31a234d7c5cb71057ed0e0ce64477
43977 Author: Eric Anholt <eric@anholt.net>
43978 Date: Mon Mar 24 13:25:37 2008 -0700
43980 Disable a bunch of clock gating disables on IGD_GM, which doesn't need them.
43982 Besides our driver having fallen through to the GM965 path for
43983 RENCLK_GATE_D1, the BIOS was turning some of these on. It may be relevant
43984 for previous platforms as well to zero out the fields that should be zero
43985 in the other registers.
43987 commit c1a2bc357fc122d3718eca771530367ba2915815
43988 Author: Julien Cristau <jcristau@debian.org>
43989 Date: Mon Mar 24 13:29:24 2008 +0800
43991 Don't try to build XvMC if DRI is disabled
43993 Instead of failing the build, don't even try unless --enable-xvmc was used.
43995 commit 5fcc002f77ebb66c1dc0f901a377ccf9bbc1d0bd
43996 Author: Maxim Levitsky <maximlevitsky@gmail.com>
43997 Date: Mon Mar 24 13:24:12 2008 +0800
43999 Fix crash triggered by dpms low power mode with hardware overlay running
44001 When overlay switch crtc, it ignores current crtc dpms mode which
44002 might lead to hang.
44004 commit 645980596450ed21c3b8927410a6bfe38a0c55d1
44005 Author: Keith Packard <keithp@keithp.com>
44006 Date: Fri Mar 21 03:08:24 2008 -0700
44008 Transformed coord computed using floats. Don't move bilinear composite dst.
44010 The homogeneous coordinate computation in the core server cannot be used for
44011 many legal matrices as it overflows. Just use floats in the driver; faster
44012 and avoids troubles.
44014 When compositing with bilinear filter, don't push the dst coordinates around
44015 as that makes the output blurry when pixels are aligned.
44017 commit bf725b92cf93d74a6f414f81f319dddbc2c78525
44018 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44019 Date: Fri Mar 21 11:50:27 2008 +0800
44021 xvmc: fix distcheck
44023 Add i915 headers in Makefile.am
44025 commit e8da1a5418ab240ae7b1ef95bfea6bee22f1e0c1
44026 Author: Joerg Sonnenberger <joerg@britannica.bec.de>
44027 Date: Fri Mar 21 11:41:49 2008 +0800
44029 Fix DRI config for NetBSD
44031 Reported by Joerg. This does a little more clearn up.
44033 commit 224c36cc14a7552bc5df793ad582f0ea5349eb1e
44034 Merge: 4b9b7b0 03ccffd
44035 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44036 Date: Thu Mar 20 10:31:57 2008 +0800
44038 Merge branch 'xvmc'
44040 commit 52d6ced652059989e6d9780a149488ccd16e3a22
44041 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44042 Date: Wed Mar 19 11:59:15 2008 -0700
44044 Fixup backlight control and panel fitting property names
44046 They should be listed as lower case, since that's what you'd pass to xrandr.
44048 commit 4b9b7b007d729f94b01b0031d8ae478134b501da
44049 Author: Keith Packard <keithp@keithp.com>
44050 Date: Tue Mar 18 14:08:19 2008 -0700
44052 Handle projective transforms on 9xx for Composite.
44054 Projective transforms require un-normalized texture coordinates and the use
44055 of the texldp instruction. The coordinates are passed as x/y/z/w (the z is
44056 unused, but there isn't a vertext format for just x/y/w).
44058 commit f699389818f1f11f3edddcdddcd0a43be21ba4c0
44059 Author: Keith Packard <keithp@keithp.com>
44060 Date: Tue Mar 18 14:06:47 2008 -0700
44062 Add i830_transform_is_affine and i830_get_transformed_coordinates_3d.
44064 These are needed to deal with projective transforms in the composite
44067 commit a55974b435224fe5726fa44298d051c71e4056a5
44068 Author: Keith Packard <keithp@keithp.com>
44069 Date: Tue Mar 18 13:47:20 2008 -0700
44071 Add defines for 9xx S3 word in 3D_STATE_LOAD_STATE_IMMEDIATE
44073 These aren't used at this point, but what the heck.
44075 commit 4f5500abe209b92b39ae1f2d7a1118362ac95034
44076 Author: Keith Packard <keithp@keithp.com>
44077 Date: Tue Mar 18 13:45:46 2008 -0700
44079 8xx/9xx can handle textures to 2kx2k. 965 can do 8kx8k
44081 commit b1b173d03b3acd300c3b0f0ceffeddf1a8137839
44082 Author: Julien Cristau <jcristau@debian.org>
44083 Date: Tue Mar 18 12:02:25 2008 +0100
44085 Add i8[13]0_ring.h to _SOURCES to fix distcheck
44087 Thanks to Bruno <bonbons67@internet.lu>.
44089 commit 03ccffd5b9b8108166ea143884a1a0a6417f3236
44090 Merge: ac76363 d72e18c
44091 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44092 Date: Tue Mar 18 16:54:15 2008 +0800
44094 Merge branch 'master' into xvmc
44096 commit ac763634069fe070b3afc073ce437959612d39fe
44097 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44098 Date: Tue Mar 18 16:52:29 2008 +0800
44100 Disable XvMC if DRI is not enabled.
44102 Default XvMC to disabled.
44104 commit d72e18c10995079121eded758a60ed99b5b8ce1a
44105 Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
44106 Date: Mon Mar 17 16:59:12 2008 +0800
44108 Compile warning fixes.
44110 Move some declarations and don't declare an extra variable with the
44111 same name, to fix warnings about mixed declarations and code.
44113 commit 81df48d8d151907c2e9dd33a729860c39aa78175
44114 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44115 Date: Mon Mar 17 10:05:57 2008 +0800
44117 bug 15060: Dell Latitude X300 needs pipe A quirk
44119 commit ffb5a115cc62ed63110f77815b7f04976f720065
44120 Merge: 05cf070 24e6eb2
44121 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
44122 Date: Mon Mar 17 15:19:05 2008 -0700
44124 Merge branch 'master' into pfit
44126 commit 24e6eb2ae683ed6614ac79c0d7811ac9fe94209f
44127 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
44128 Date: Mon Mar 17 14:17:27 2008 -0700
44130 Add support for new ->get_crtc output callbacks
44132 Using the new interface allows the server to avoid some flicker at startup.
44134 commit 69fbc17441d0f894d17b058e65ae22300cd2a54c
44135 Author: Eric Anholt <eric@anholt.net>
44136 Date: Fri Mar 14 10:04:18 2008 -0700
44138 Change OUT_RING and similar calls to OUT_BATCH for batchbuffer merge
44140 commit 65f92cfb7a05c9c028cf73ce1221cc0a651b50b3
44141 Author: Eric Anholt <eric@anholt.net>
44142 Date: Fri Mar 14 10:03:11 2008 -0700
44144 Remove h0h0 UploadToScreen implementation that used pattern blits.
44146 It never worked and was never that hot of an idea. Most of the motivation is
44147 gone with glyph pixmaps anyway.
44149 commit 05cf07071e25f84dec3476a9bed7235ed50cd249
44150 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
44151 Date: Thu Mar 13 18:12:00 2008 -0700
44153 Initial panel fitting changes
44155 Basic support for panel fitting.
44157 commit 402fbd36221d2ed27913850f6bfcce768b970fe8
44158 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
44159 Date: Thu Mar 13 16:12:20 2008 -0700
44161 Remove fbc enable/disable messages
44163 Tired of them filling up my logs.
44165 commit c20d78a7bc512219fa36dfc9707fe65f66ab80f1
44166 Author: Eric Anholt <eric@anholt.net>
44167 Date: Mon Mar 10 17:27:45 2008 -0700
44169 Add workarounds for 830/845's lack of a8/x8r8g8b8/x8b8g8r8 support.
44171 When we have unset channels (color for a8, or alpha for x8*), force them
44172 to the appropriate value in the texture combiner rather than relying on getting
44173 the desired results from teture mapping. Performance is the same on 865, so
44174 multiple paths aren't necessary and we can use the less friendly texture
44175 formats everywhere.
44177 commit f425181695b1b3864373e1919cf651229f19fa69
44178 Merge: f4cfc98 1c0e495
44179 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44180 Date: Thu Mar 13 23:58:56 2008 +0800
44182 Merge branch 'master' into xvmc
44184 commit f4cfc98ab038af4d5d52af12e4d60ded62cf4a6e
44185 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44186 Date: Thu Mar 13 23:56:48 2008 +0800
44188 nuke unneed error_base and event_base
44190 commit f269c69b4e7be076dc9838d0e0801c940ccdba08
44191 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44192 Date: Thu Mar 13 23:41:22 2008 +0800
44194 remove unused init/fini hooks for xvmc driver
44196 commit 1c0e4958564588162bc22a980567436dfc9d1e56
44197 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44198 Date: Thu Mar 13 10:15:55 2008 +0800
44200 Always bind memory after agp aperture allocation
44202 This fixed the crash appearing for Xv buffer memory, which appears
44203 allocation and free happening in stolen memory. We should always
44204 try to bind for stolen memory for correct protect setup, otherwise
44205 we might get fault that stolen memory protect reverted to non-r/w
44206 after previous free and unbind.
44208 Credit goes to Lukas Hejtmanek to identify the real problem of Xv
44211 commit 2e2372912ed9bc0d86e8960653ef0bfce5cf99ab
44212 Author: Eric Anholt <eric@anholt.net>
44213 Date: Wed Mar 12 11:38:56 2008 -0700
44215 Eliminate local variables defined in i830 BEGIN_LP_RING().
44217 This lets us get better sanity asserts, and avoid mysterious braces when you
44218 do two BEGIN_LP_RING()s in a single function. Potential minor performance
44219 loss isn't too exciting, as ring access is about to become a compat path
44220 anyway. This change also removes the requirement for ring emits to be aligned
44223 commit abf11a274e14535630742fe4c41cc0ae92555293
44224 Author: Eric Anholt <eric@anholt.net>
44225 Date: Wed Mar 12 11:03:26 2008 -0700
44227 Separate i810 and i830 ring macros out into separate files.
44229 I want to hack on i830 for changing it into a compat path for batchbuffer
44230 without having to worry about the i810 stuff getting broken.
44232 commit e04908853db74dc76718f6bc0dfb23b5ab3c5837
44233 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44234 Date: Wed Mar 12 15:44:58 2008 -0700
44236 Add sysfs backlight support for Fujitsu laptops
44238 Tested by Dan Williams.
44240 commit 907399660b83b3939aab3dc7d0697197441e4ec8
44241 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
44242 Date: Wed Mar 12 12:33:36 2008 -0700
44244 Add stdint.h to DVO drivers
44246 Fixes the build for me (just following the git development guide on a
44249 commit 8563effc96195116c96f89a801c1de6ab982630a
44250 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44251 Date: Wed Mar 12 11:14:17 2008 -0700
44253 Add stdint.h to i830_bios.h
44255 Needed for new uint* usage.
44257 commit c82d8e51c6ff9522ad0deb33399d13f3d6fef19d
44258 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44259 Date: Wed Mar 12 11:13:50 2008 -0700
44261 Add pipe A force quirk for Thinkpad R31
44265 commit a8f4108bea0966406b0fcae8d062ae1cfb4afbb3
44266 Author: Hong Liu <hong.liu@intel.com>
44267 Date: Wed Mar 12 21:55:56 2008 +0800
44269 Fix rotation mem with kernel memory manager in VT switch
44271 Unbind and bind a DRM BO may change the buffer offset, thus
44272 crtc may reference a wrong rotated memory after a VT switch cycle.
44273 Destroying it here will cause its reallocation when entering VT.
44275 commit cd73c6a5a78eac726b14ef242b8452c559c4c003
44276 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44277 Date: Wed Mar 12 20:37:41 2008 +0800
44279 Pass default visual in dri context create
44281 commit 21580add63307ca68790bf911243b8675f5f7c64
44282 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44283 Date: Wed Mar 12 20:37:17 2008 +0800
44285 disable memory alloc debug
44287 commit 5f272471878c896038b2c3f2bf704655d908828d
44288 Author: Eric Anholt <eric@anholt.net>
44289 Date: Tue Mar 11 12:14:51 2008 -0700
44291 Nuke more CARD* types from i810 driver. Untested.
44293 commit 9a62d3b598cca3c28fa8b6313bba82983cc29ecf
44294 Author: Eric Anholt <eric@anholt.net>
44295 Date: Tue Mar 11 12:07:52 2008 -0700
44297 Remove i830+ driver's use of CARD*/INT* types for great justice.
44299 Several uses are actually left, which are determined by the X Server
44300 interfaces we're implementing.
44302 commit 15f17aa682129c57d3864a355834edff06e91540
44303 Author: Eric Anholt <eric@anholt.net>
44304 Date: Tue Mar 11 11:27:10 2008 -0700
44306 Cleanse DVO drivers of CARD* types.
44308 commit 642ba6278876ec945ad687c2b9624e532cd3f629
44309 Author: Eric Anholt <eric@anholt.net>
44310 Date: Tue Mar 11 09:59:16 2008 -0700
44312 Add support for brightness control to i915 textured video output.
44314 commit ecdb5963ef6873ab82998dca6ca5186644666ec8
44315 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44316 Date: Tue Mar 11 11:08:25 2008 -0700
44318 Add pipe A force enable quirk for Toshiba Portege R205-S209
44322 commit 4ea67e9b853e9ed61557780b87bd184fc81fd731
44323 Author: Kristian Høgsberg <krh@redhat.com>
44324 Date: Tue Mar 11 13:54:33 2008 -0400
44326 Crack down on warnings.
44328 Fix printf formatting warnings, wrap a couple of long lines, nuke
44329 unused variables, add missing #include <unistd.h>.
44331 commit 9fd13e6773371c82b9799a5bda7c96ffa5cafe8c
44332 Author: Kristian Høgsberg <krh@redhat.com>
44333 Date: Tue Mar 11 13:48:07 2008 -0400
44335 Silence warnings about possible uninitialize use of dst_format.
44337 Whoa, gcc got a lot smarter about warnings. If iXXX_get_dest_format()
44338 doesn't support the picture format passed in it won't initialize the
44339 uint32_t pointed to by dst_format and return FALSE.
44341 What gcc now can detect is if dst_format is used without checking the
44342 return value, it might be used uninitialized. This patch makes sure
44343 we always check the return value before using dst_format.
44345 commit 31654f52b084ebb32f69854dfb1c342c9da07928
44346 Merge: a03aca1 02c0ffb
44347 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44348 Date: Tue Mar 11 14:37:41 2008 +0800
44350 Merge branch 'master' into xvmc
44357 commit 02c0ffb1d8112a6fb03f9c8eae9ec1a1fca123ef
44358 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44359 Date: Mon Mar 10 11:51:37 2008 -0700
44361 Add pipe A force enable quirk for Dell Latitude D500
44365 commit a1cd37f94cb8efbde84ba0a46839fb8735dd3595
44366 Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com>
44367 Date: Mon Mar 10 14:43:24 2008 +1000
44369 quirk: add DVO ivch quirk
44371 commit f2b9f387af7f618ad083b1ec8bee513ff2157d06
44372 Author: Matthieu Herrb <matthieu.herrb@laas.fr>
44373 Date: Sat Mar 8 23:36:56 2008 +0100
44375 Makefile.am: nuke RCS Id
44377 commit bf6d59ee98127ac6447848c4202c54d861a510d4
44378 Author: Eric Anholt <eric@anholt.net>
44379 Date: Fri Mar 7 14:26:22 2008 -0800
44381 Avoid setting reserved bits in SDVOx on everything 965-class.
44383 This will likely fix SDVO outputs on G35 hardware.
44385 commit 5915c75422c5277d530e7f8ecbdfe94654706efd
44386 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44387 Date: Fri Mar 7 13:18:09 2008 -0800
44389 Enable hardware overlay on 965
44391 Based on Maxim Levitsky's patch for the same. Many thanks to Maxim for
44392 discovering that this features still works in recent hardware!
44394 commit 13ae79b415d944891cfdc2cd2d5397fec36a4e8e
44395 Author: Dave Airlie <airlied@redhat.com>
44396 Date: Fri Mar 7 08:04:37 2008 +1000
44398 fixup pciaccess version detect
44400 commit be08b61540e1edbb7673f0b6add0f1e32b136b3c
44401 Author: Jesse Barnes <jbarnes@jbarnes-t61.(none)>
44402 Date: Thu Mar 6 13:55:13 2008 -0800
44404 Implement new get_property feature for backlight control
44406 This should keep the backlight value reported by xrandr --prop & xbacklight
44407 consistent with changes by other software in the system (like the hotkey driver
44408 or kernel backlight driver).
44410 commit 502e41f6dff91393cfeaf08c9ba312781d1b7267
44411 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44412 Date: Thu Mar 6 19:32:29 2008 +0800
44414 remove extra dri checking
44416 And directRenderingDisabled already has config check result.
44418 commit ea58d66cf778c0477c30e5bfe3142b64e8b971d5
44419 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44420 Date: Thu Mar 6 19:31:33 2008 +0800
44424 commit c823fb00a20f523f50a2779c36aa1c014ba01535
44425 Author: Jesse Barnes <jesse.barnes@intel.com>
44426 Date: Thu Mar 6 19:29:22 2008 +0800
44428 Pull chipset detection out of PreInit into its own function
44430 commit 6b16cee0786b636dc849826193b26e898832a73c
44431 Author: Jesse Barnes <jesse.barnes@intel.com>
44432 Date: Thu Mar 6 19:27:42 2008 +0800
44434 Account for FBC LL cache in memory reservation
44436 When we reserve memory (memory that won't be managed by the kernel) make
44437 sure to reserve space for the FBC LL cache as well.
44439 commit 6f04d02471f1cebc76eab637a9a78c4004fd3cc9
44440 Author: Jesse Barnes <jesse.barnes@intel.com>
44441 Date: Thu Mar 6 19:27:07 2008 +0800
44443 Move MTRR fixup to its own routine
44445 commit 750beb9232b51223d8e650878ce8dad071f0d0b3
44446 Author: Jesse Barnes <jesse.barnes@intel.com>
44447 Date: Thu Mar 6 19:24:47 2008 +0800
44449 Refactor memory allocation into a separate function
44451 This simplifies the memory allocation code and fixes a number of bugs.
44452 Prior to this change, some flags may have been set after memory
44453 allocation occurred, meaning they had no effect. It should also make
44454 the allocation logic clearer.
44456 commit d466b25d09bbe85abe84d9d7714ffb65fafe7593
44457 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44458 Date: Wed Mar 5 13:38:58 2008 +0800
44460 Really print adjusted_mode for mode debug
44462 commit d767eabe534be7f7bdcbf00614a8901124ac911f
44463 Author: Hong Liu <hong.liu@intel.com>
44464 Date: Thu Feb 28 13:51:18 2008 +0800
44466 Fix DPLL programming in CRTC mode set
44468 The following patch fixes the display problem on internal development machines.
44469 The code in commit 3c22ed633be2ac96eea7bc533839e956f1f31b84 (Jesse's force pipe
44470 A enable patch) broke DPLL programming. Moving the DPLL set back up in the
44471 function solves the problem.
44473 Fix for internal bug #309.
44475 commit 4936e097028b91f4bdc2d9101dc49f6fe586e718
44476 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44477 Date: Tue Mar 4 08:52:40 2008 -0800
44479 Change FBC idle mode back to default
44481 Use the "immediate idle" mode for FBC. Rather than waiting for the current
44482 compression pass to finish before signalling to the CPU that it's idle, this
44483 mode should stop any current compression pass and tell the CPU that the GPU is
44486 Seems to fix #13326.
44488 commit 02a3b80ee8341968dbce5b7cb871b187db5ff805
44489 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44490 Date: Fri Feb 29 11:13:34 2008 +0800
44492 Bug 14591: comment out MI_DISPLAY_POWER_DOWN as for 965GM only
44494 commit ea7c7d088e0a3c3b7ab2bdd0475aa13379764bf4
44495 Author: Peng Li <peng.li@intel.com>
44496 Date: Wed Feb 27 11:56:55 2008 +0800
44498 HP Compaq 6730s has no TV output
44500 commit fdc25cb0ca7d479051cee66a032f9ee1b68db98a
44501 Author: Mark Kettenis <mark.kettenis@xs4all.nl>
44502 Date: Tue Feb 26 17:26:50 2008 -0800
44504 Bug #14649: Fix build if pciaccess headers are not with other xorg headers.
44506 We failed to include the pciaccess header flags in our build. Even if the
44507 server was to include those flags for us, it would leave us broken for the
44508 tools-only case, and it's easier to just put the flags everywhere so we avoid
44509 future copy'n'paste mistakes.
44511 commit e55d86f654aa9dd4951b441636c8ca4b37786daf
44512 Author: Eric Anholt <eric@anholt.net>
44513 Date: Tue Feb 26 12:26:23 2008 -0800
44515 Fix SDVO I2C access on Mac Mini in EFI mode.
44517 The GMBUS was being left in SDVO pin access mode, which blocked our bit-banging
44518 access to those pins. Thanks to Peter Jones for quick debugging turnaround
44519 in getting this fixed.
44521 commit 6935c732c351585f31e2094c4201a00e3d5529b3
44522 Author: Julien Cristau <jcristau@debian.org>
44523 Date: Mon Feb 25 14:47:02 2008 +0100
44525 Look for src/{modes,parser} in $srcdir, not in objdir
44527 This makes the detection work in case configure is run from another directory.
44529 commit 66cdccb021a4748b2af41e415c36ed58ca808df6
44530 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44531 Date: Mon Feb 25 14:29:09 2008 +0800
44533 Fix overlay rotation
44535 Track current covered crtc rotation in port's private,
44536 and fix overlay rotation transform to be against crtc mode
44539 commit 975d7833bc844485c7043538336bddcba7f7f90e
44540 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44541 Date: Fri Feb 22 11:41:25 2008 -0800
44543 Add quirk for DVO channel selection
44545 Some machines want DVOA, some DVOB. We can use this quirk to differentiate
44546 them until we come up with a better solution. Patch from Hong Liu.
44550 commit 33802b6ee3c298bae4d2b6f464ee5cc730a58506
44551 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44552 Date: Thu Feb 21 08:43:34 2008 -0800
44554 Add pipe A force enable quirk for ThinkPad X40
44556 ThinkPad X40s need this quirk or they'll crash frequently.
44560 commit 444984a578aae92ff55c06da897ea1d23679e706
44561 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44562 Date: Thu Feb 21 08:37:42 2008 -0800
44564 Remove side effects from VGA debug code
44566 The VGA register dumping code was leaving ARX in data mode rather than index
44567 mode, which could cause problems for later software accessing AR* registers.
44568 Fix it to make sure it's in index mode when we're done.
44572 commit 28049540d8a9f79401fcfeb90784f5a528e7b34f
44573 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44574 Date: Wed Feb 20 13:08:27 2008 -0800
44576 Fix DSP*CNTR restoration
44578 When checking which pipe a given plane was associated with, we weren't properly
44579 masking the pipe selection bits. Fixes #14481 and should allow the driver to
44580 work with vesafb again.
44582 commit 293120bfc40a5b828567551954d8312639e73578
44583 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44584 Date: Tue Feb 19 21:13:45 2008 +0800
44586 Don't leak memory if no DMI info is provided by kernel
44588 commit cdc5cb7c3d24f36a30fb931b43ec6df8294709e2
44589 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44590 Date: Tue Feb 19 08:51:53 2008 +0800
44592 Fix last 8XX clock's p2 value commit
44594 My fault to change Hong's origin patch reversely.
44596 commit be0591880f20bdcbae159d5ab47113b6cc6fbffe
44597 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44598 Date: Mon Feb 18 14:35:06 2008 +0800
44600 Fix Lenovo X60 TV quirk
44602 Z61 has same subsys ids with X60, but does have one S-video
44603 TV out. Use DMI info instead to quirk TV on X60, X60s.
44605 commit d5e564c5f24b0d9320bc4908693d37e7e6d68bc6
44606 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44607 Date: Mon Feb 18 14:32:32 2008 +0800
44609 Add DMI info for i830 quirks
44611 Linux kernel has CONFIG_DMIID to export dmi info
44614 commit 4a42b01f5ee5a673716d6959dfe0e693b037eb48
44615 Author: Keith Packard <keithp@keithp.com>
44616 Date: Sat Feb 16 18:16:12 2008 -0800
44618 Decode DSPCLK_GATE, dump PIPE*STAT, MI_MODE, MI_DISPLAY_POWER_DOWN, MI_ARB_STATE, MI_RDRET_STATE, ECOSKPD
44620 commit 663313559873208c00c3d3748b99d837949ecdf2
44621 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44622 Date: Fri Feb 15 13:20:22 2008 +0800
44624 Bug #14440: fix stolen mem size mask on i830M
44626 For i830M stolen mem size mask should always be 0x70.
44627 Use 0xF0 for later chipsets should be ok, so behavior is
44628 identical to kernel agp.
44630 commit 8c8ca16120927fc2d3cf8bcd5212a4227a280708
44631 Author: Eric Anholt <eric@anholt.net>
44632 Date: Fri Feb 8 11:34:04 2008 -0800
44634 Add a little program to output the stepping information for several chipsets.
44636 commit f5896dcd1f5a964dce024ab8887ac8a79f7093a1
44637 Author: Jesse Barnes <root@nietzche.virtuousgeek.org>
44638 Date: Thu Feb 7 11:56:28 2008 -0800
44640 Add CACHE_MODE_0 register to dump output
44642 commit d59eaa8b1e6eeb9775c9d21c7a5fd28f25b2bc3a
44643 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44644 Date: Wed Feb 6 09:35:43 2008 -0800
44646 Only disable FBC if registers are available
44648 The call to disable FBC should only occur if the FBC feature is actually
44649 present or we may end up hanging on a read from a non-existent register.
44651 commit 83a071e2c01eb71462273f25d86d77159ac51114
44652 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44653 Date: Tue Feb 5 17:09:19 2008 -0800
44655 Fix build warnings on 64 bit
44657 Use PRIx32 for printing CARD32 types, and PRIx64 for portably printing uint64_t
44658 types. Requires the addition of a new include, inttypes.h, to work. Hope C99
44659 is ok with everybody...
44661 commit abf4b0cafa038e97bd7a6e5ba107b822117b9ccc
44662 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44663 Date: Tue Feb 5 13:30:51 2008 -0800
44665 Only enable FBC if one pipe is active
44667 Some chips can't support FBC if multiple pipes are active. So if more than one
44668 pipe is on or we're going from one->two pipes enabled, make sure FBC is
44671 Intended to fix 13418, 13326, 13152.
44673 commit f79a585d57c61fbe9af5684a336918e1fe3e8416
44674 Author: liuhong <liuhong@devlinux-hong.sh.intel.com>
44675 Date: Tue Feb 5 11:04:34 2008 +0800
44677 xf86-video-intel: fix ivch using wrong dvo register
44679 We use the wrong DVO output register for ivch which is used for DVO LVDS.
44680 Vbios enables DVOB and driver never touches it, so everything seems
44681 working correctly now.
44683 commit cabed67b6a0e4dd408eeb8d2083269100a65933c
44684 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44685 Date: Tue Feb 5 11:36:24 2008 -0800
44687 Program FBC fence offset register
44689 Just a partial fix for some of the FBC issues people have been seeing. The
44690 other half is to disable FBC if both pipes are running.
44692 commit 332587081fb05d284864aabe608dd9988eee3d44
44693 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44694 Date: Tue Feb 5 10:34:55 2008 +0800
44696 Fix last commit on i8xx debug p2 value
44698 commit 2d18a299cf39f2ecf9bd3ba8407f892566b2edcc
44699 Author: Hong Liu <hong.liu@intel.com>
44700 Date: Tue Feb 5 09:10:48 2008 +0800
44702 Fix PLL reference clk debug dump
44704 commit 9536515d7717969795edc1b80d6e6a36820dd575
44705 Author: Hong Liu <hong.liu@intel.com>
44706 Date: Tue Feb 5 09:06:14 2008 +0800
44708 Allow non-strict free order for bo_list
44710 commit 11de831cb706025c003e98fef6d666d273192bcd
44711 Author: Hong Liu <hong.liu@intel.com>
44712 Date: Tue Feb 5 08:58:44 2008 +0800
44714 Bug 10584: Mac Mini EDID data assigned to TMDS output
44716 EDID data for TMDS output got from crt should be applied
44717 to TMDS output on mac mini.
44719 commit 95327571134d8ebb9d936d80b70c2f642393aa9f
44720 Author: Hong Liu <hong.liu@intel.com>
44721 Date: Mon Feb 4 17:14:23 2008 +0800
44723 Bug 10773: fix i8xx pll p2 value in i830_crtc_clock_get()
44725 Also fix debug dump, slightly modified to use macro instead.
44727 commit be1b568bc2379856c3eaaed365002512bebc218c
44728 Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com>
44729 Date: Thu Jan 31 22:47:18 2008 -0200
44731 Make sure symbols used by other modules are public.
44733 These symbols must be explicitly exported, otherwise if compiled with
44734 hidden symbols, LoaderSymbol (and dlopen) will fail to find them.
44736 commit cc1031a834b81aa6f835c282ad86deddabe229a0
44737 Author: Mark Kettenis <mark.kettenis@xs4all.nl>
44738 Date: Thu Jan 31 21:02:48 2008 -0800
44740 Bug #14246: Fix biuld on OpenBSD.
44742 commit b1e1c7cd993ab4f1935d750eb8852e74be757e53
44743 Author: Eric Anholt <eric@anholt.net>
44744 Date: Wed Jan 30 23:01:04 2008 -0800
44746 Add detail on different units to intel_idle.
44748 commit aa3ac79759581b5eb05293a8cbcf89eb5b76712c
44749 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44750 Date: Thu Jan 31 18:26:46 2008 +0800
44752 Don't crash if SW cursor
44754 In case of device option or hw cursor allocation fails.
44756 commit 2e43bec8731ba1b172f7a0bf867bbb5c1adbda2d
44757 Author: Jesse Barnes <jesse.barnes@intel.com>
44758 Date: Wed Jan 30 18:59:12 2008 +0800
44760 Frame buffer compression support on new chipset
44762 commit bf629466a46c4037ec7b7cc5ee16be947618bd68
44763 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44764 Date: Wed Jan 30 18:55:20 2008 +0800
44766 hardware status page initialization rework
44768 Order hardware status page setup more reasonable after
44769 all memory bound, in case new chipset requires non-stolen
44770 page and that could be bound then.
44772 Also clean up drm irq handler install function, and put
44773 first install in starting stage later than status page setup,
44774 so we won't make device cry for uninitialized status page.
44776 commit 04032dad28baab80131edbe8fe58aade8149bb71
44777 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44778 Date: Wed Jan 30 18:52:32 2008 +0800
44780 Wrap up chipsets which needs graphics address for status page
44782 Also add support on new chipset.
44784 commit b9c3fa79d9fd48c34536f92e6c1c70fe3a619410
44785 Author: Keith Packard <keithp@keithp.com>
44786 Date: Fri Jan 11 15:51:00 2008 -0800
44788 Add intel_idle to measure GPU load ratio.
44790 Samples GPU idle register to detect when the GPU is busy; prints out the
44791 number of busy samples per 100 total samples once per second.
44793 commit 5d9e67aec3ce458d22b7febd3819542abb360534
44794 Author: Erik Andren <erik.andren@gmail.com>
44795 Date: Fri Jan 25 11:06:01 2008 +0800
44797 Clevo M720R has no TV-out
44799 commit 6bf53eb48f40ad0c8ea9679ee634447410821b4f
44800 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44801 Date: Fri Jan 25 16:59:13 2008 +0800
44803 Set vtSema before EnterVT
44805 Which was missing in our ScreenInit and initial EnterVT.
44806 This not only causes failure in initial rotation with TTM,
44807 as we won't bind in rotate_mem alloc in this case, and hide
44808 another bug that we call randr12 function in I830LoadPalete
44809 before we call xf86RandR12Init.
44811 commit 94a18fa1f8141837bdab32e545da7a7aed1cc396
44812 Author: Julien Cristau <jcristau@debian.org>
44813 Date: Thu Jan 24 15:24:40 2008 +0100
44815 Don't build reg_dumper if we don't have pciaccess 0.10.0
44817 The pci_device_map_range() function was added in libpciaccess 0.10.0, and
44818 is used by the reg_dumper tool. Don't try to build it if we have an older
44820 Also make sure that util-macros >= 1.1.3 is available when running autoconf,
44821 because it's required for the PACKAGE_VERSION_* macros.
44823 commit d341e41c863f1212bf2c6b84782a7e472b6612a1
44824 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44825 Date: Thu Jan 24 10:40:18 2008 +0800
44827 Clear shadow memory after allocation
44829 commit 88a1041c5361964d37107c22d77feaa1b2160656
44830 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44831 Date: Thu Jan 24 10:36:06 2008 +0800
44833 Fix i830 block handler wrap
44835 which was observed in rotation crash with stack overflow.
44837 commit db0a7c569e383436a2725e1e74f35fb426da1196
44838 Author: Andreas Stawinoga <a.stawinoga@gmx.de>
44839 Date: Thu Jan 24 08:51:09 2008 +0800
44841 Samsung Q45 has no TV output
44843 commit f54bba7600647c76fe3bd2d0133edc06dbea6a05
44844 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44845 Date: Tue Jan 22 00:48:23 2008 +0800
44847 Add missing i830M and 845G pci ids info
44849 commit 633dbd28218fd8a746c1735fd44a14f35bd07faa
44850 Author: Julien Cristau <jcristau@debian.org>
44851 Date: Thu Jan 17 22:34:21 2008 +0800
44853 Bug 14032: i810, set default depth to 16
44855 X fails to start without an xorg.conf on i810, because the default depth set by
44856 the server (32) is not supported by the driver. I think the driver should
44857 default to 16 bits (it supports 24, but when I tested it that didn't seem to
44858 work very well, and dri is disabled at 24 bits depth anyway) instead of letting
44859 the server use 32 and failing later on.
44861 commit 4fec18e07f2b23306e885d63c74661ae0e4b04d9
44862 Author: Jesse Barnes <jesse.barnes@intel.com>
44863 Date: Fri Jan 11 10:28:59 2008 -0800
44865 Turn on backlight when LVDS panel is powered up
44867 In some configurations, the LVDS may be off at startup along with the
44868 backlight. So when turning the LVDS on for the first time, we may also need to
44869 set the backlight to a non-zero value. So try to use the saved value if
44870 possible, but if it's zero, make the backlight full brightness when turning on
44873 Note that this is slightly sub-optimal for configurations where zero is a valid backlight brightness.
44875 Fixes fdo bz #13958.
44877 commit b5cf8faf8102e6313c23de87a680cc327ac95478
44878 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44879 Date: Thu Jan 10 14:25:18 2008 -0800
44881 Remove unnecessary quirk code in CRT probing
44883 Since we return early in the CRTC DPMS code, keeping the pipes & clocks
44884 enabled, this code is no longer necessary. As an added bonus, VGA connections
44885 should be reported correctly even when the pipe A quirk is active.
44887 Updates fix for fdo bz #11432.
44889 commit 3c22ed633be2ac96eea7bc533839e956f1f31b84
44890 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44891 Date: Wed Jan 9 09:47:38 2008 -0800
44893 Add pipe A force enable quirk
44895 On some platforms, the firmware may read & write GPU registers on lid close,
44896 suspend/resume time or during various SMM events. If one of the graphics pipes
44897 is disabled at that time, the GPU may hang due to the programming dependencies
44898 of the various registers.
44900 This patch adds a quirk to force the driver to keep pipe A enabled if
44901 necessary, through user configuration in xorg.conf or via a platform specific
44902 quirk. Leaving the pipe enabled comes at a power cost however, so the quirk
44903 should only be enabled when strictly necessary.
44905 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=11432.
44907 commit d3b41bcf231a6ffd289c128c06aa619f554a776c
44908 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44909 Date: Wed Jan 9 21:28:25 2008 +0800
44911 Disable frame buffer compression on new chipset now.
44913 commit 2f0c0427dc4113bac64cda07c6365df1f0637697
44914 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44915 Date: Wed Jan 9 21:27:42 2008 +0800
44917 Update PIPELINE_SELECT instruction and surface state format for new chipset
44919 commit 139e9affdd49dcbe08e260a0a2bd001de16566e6
44920 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44921 Date: Wed Jan 9 21:26:57 2008 +0800
44923 GTT access change for new integrated graphics device
44925 commit 4f288ff2d1e7d28bf49167f073fbce21492a7c32
44926 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44927 Date: Wed Jan 9 21:25:47 2008 +0800
44929 Add new integrated graphics chipset ids
44931 commit f089942689df622b1af00569eade4883a84eb7a3
44932 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44933 Date: Thu Jan 3 11:28:18 2008 +0800
44935 Change origin i965G_1 to chipset market name G35.
44937 commit 71143200ed66cb9dcc1f9edf98128be80b05619c
44938 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44939 Date: Fri Jan 4 15:01:25 2008 -0800
44941 Don't modify low bit of BLC_PWM_CTL when using combo backlight control
44943 On some platforms, the low bit of BLC_PWM_CTL is wired as a 'max brightness'
44944 flag, rather than a regular part of the backlight duty cycle. So when in the
44945 combo mode, divide the total number of backlight levels available by two
44946 (tossing one bit) and adjust the programming in the set_brightness routine.
44948 Note that platforms with this behavior may need quirks added so that they work
44951 commit b434bf4029d1a405216454629bcea717c779ae5a
44952 Author: hulifox008 <hulifox008@163.com>
44953 Date: Fri Jan 4 02:46:54 2008 -0800
44955 Fix copy'n'paste-o in unused airlied i2c code.
44957 commit f2ffc0f6e038357dda268363e52c11ada1d0b810
44958 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
44959 Date: Thu Jan 3 11:16:15 2008 -0800
44961 Fix compilation error when not using DRI
44963 This patch complements 88f8b688e2316ae4a1f7485f0010ce90de54783a which
44964 added uint64_t typed variables to avoid unsigned long overflows in
44965 32-bit architectures but didn't include <stdint.h> with the required
44968 When XF86DRI and _XF86DRI_SERVER_ are defined this header gets
44969 indirectly included through "i830_dri.h", thanks to "i830_common.h"
44970 which masquerades this problem as released in 2.1.0 and that manifests
44973 In file included from i810_driver.c:88:
44974 i830.h:137: error: expected specifier-qualifier-list before 'uint64_t'
44975 i830.h:240: error: expected specifier-qualifier-list before 'uint64_t'
44977 Patch from Carlo Marcelo Arenas Belon <carenas@sajinet.com.pe>
44979 commit a03aca101d4d51ed2279a99d1137ce3b7ef364e2
44980 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44981 Date: Wed Jan 2 14:11:31 2008 +0800
44983 xvmc: setup driver info for XvMCGetDRInfo request
44985 User don't need to have XvMCConfig file then, if ld.so can
44986 find libIntelXvMC.so in ld path.
44988 commit 0fd769b5882058e5ebf17ee1c1e35b4a3d46e6b0
44989 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
44990 Date: Wed Jan 2 13:57:23 2008 +0800
44992 xvmc: add context id info in debug/trace
44994 commit 83d304c61ad5fdc58b0a9309dbd1e5a3f6cd9b01
44995 Author: Nanhai Zou <nanhai.zou@intel.com>
44996 Date: Wed Jan 2 10:50:44 2008 +0800
44998 TV: fix 576p refresh rate
45000 commit a2bb2afd24c95e056b801c85dc12daeac54b37f0
45001 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45002 Date: Fri Dec 28 15:52:12 2007 +0800
45004 xvmc: move drmOpen earlier
45006 remove wrong comments
45008 commit 96246c27cb836bae8ee02c46c68a1f2102efd14b
45009 Author: Joakim <elupus@ecce.se>
45010 Date: Thu Dec 27 17:09:02 2007 +0800
45012 Aopen Minipc 965GM LVDS quirk
45014 commit 666e01bd426a553f23eeb5040ecbc28b4abb87da
45015 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45016 Date: Wed Dec 26 09:51:09 2007 +0800
45018 xvmc: add info in intel.man
45020 commit e75d0c20e3b02f649b62ac9da7f41bb43d0f3d80
45021 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45022 Date: Wed Dec 26 09:37:34 2007 +0800
45024 xvmc: make XvMC default on
45026 commit ed9ff6f30fff33fe2d60715f1b503047f8e1ef01
45027 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45028 Date: Tue Dec 25 08:32:47 2007 +0800
45030 xvmc: consolidate put surface interfaces
45032 Make surface list track like context list, and move
45033 generic PutSurface code out of i915. So MC driver just
45034 needs to fill specific the command params, and we can
45035 get one single type of intel xvmc command to issue,
45036 ddx driver's put_image_size got removed.
45038 Fix last commit with missing XvMCContext assignment.
45040 Fix priv_data free with apropriate X11 XFree.
45042 commit f0ced5edaeec3e7741828eab728417843326d56f
45043 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45044 Date: Mon Dec 24 10:53:57 2007 +0800
45046 xvmc: store XvMCContext pointer instead of context_id in internal ctx list
45048 commit 216ad7398742f2f80b5477e89f827acb351b4d72
45049 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45050 Date: Thu Dec 20 13:39:32 2007 +0800
45052 xvmc: add disable xvmc build support
45054 Move xvmc source files under XVMC config
45056 commit a811ca30659e8c477027ea9c91fb9fc516fdabc7
45057 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45058 Date: Thu Dec 20 13:08:07 2007 +0800
45060 xvmc: handle surface create/destory in generic code
45062 commit 5f961f28310a1cd19967df2034b0452bf4446394
45063 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45064 Date: Thu Dec 20 12:51:06 2007 +0800
45066 xvmc: move dri context handling to generic code
45068 Use XvMCContext's context_id for dri context handling instead
45069 of driver private id. Remove unnecessary field for i915 private
45072 commit 73827e78357ab3f30a21f159ecdfb26dd0ea889b
45073 Merge: 3cf8627 f69b48f
45074 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45075 Date: Wed Dec 19 22:47:02 2007 +0800
45077 Merge branch 'master' into xvmc
45079 commit f69b48fe24ef94dac44b8123884ca71df675be4b
45080 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45081 Date: Tue Dec 18 18:10:33 2007 -0800
45083 Unconditionally restore pipe configuration
45085 This is a partial fix for #13196, which covers both leaving pipes disabled at
45086 server exit time and problems with restoring the pipe configuration on certain
45087 chipsets. It restores the pipe configuration unconditionally (previously we
45088 made sure the PLL was running and we weren't in VGA mode) but also adds some
45089 additional PLL settle time to the PLL register write paths.
45091 commit 4757a218d733af12d04674455fc6e1fad48a1cd0
45092 Author: Michel Dänzer <michel@tungstengraphics.com>
45093 Date: Tue Dec 18 19:37:46 2007 +0100
45095 Always set pPriv->buf to NULL after freeing the memory it pointed to.
45097 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=13108 .
45099 commit 5cbb4110116f6d4187fd27660a78203caa1da46b
45100 Author: Adam Jackson <ajax@redhat.com>
45101 Date: Thu Dec 13 15:10:43 2007 -0500
45103 i830_sdvo_mode_valid: Fix return values to match what we actually check.
45105 commit e03bc385e924222d3d5f3c9a8d65e1ef63b858dc
45106 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45107 Date: Tue Dec 11 13:28:01 2007 -0800
45109 Add cscope files to .gitignore
45111 commit e720ae4476c3f986f623ce0f0ab9775b8b9b7e05
45112 Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com>
45113 Date: Mon Dec 10 13:00:14 2007 -0800
45115 CRT hotplug detection improvements
45117 Patch from Hong Liu.
45119 Fixup CRT detection by making sure the pipe is enabled before CRT
45120 detection actually occurs. Fixes bugs Hong was seeing on G35 and other
45123 commit d9df93578b74785c08ba860b4c9aa23b0c89c91c
45124 Author: Dave Airlie <airlied@linux.ie>
45125 Date: Mon Dec 10 16:41:24 2007 +1000
45127 ivch: fails on address mismatch as I seem to get this on my 865 system
45129 commit cd6f83519c69f72f146fea59afbd6a3542da783a
45130 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45131 Date: Mon Dec 10 05:49:58 2007 +0800
45133 Fix tv quirk for Dell Latitude X1
45135 which is actually 915GM, reported by
45136 Andreas Schildbach <andreas@schildbach.de>
45138 commit 13ec9c8141a9f794258869a04a6bab59dac5eefa
45139 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45140 Date: Sun Dec 9 00:52:13 2007 +0800
45142 exa: fix rendering issue on some 855GM laptops
45144 This trys to initialize more 3d states on i830 like mesa
45145 code, which fixes exa rendering failure for some 855GM users.
45147 commit e2d9fd6425584119f28768d85c4b7d106ecf8742
45148 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45149 Date: Fri Dec 7 16:30:46 2007 -0800
45151 Describe output properties in more detail
45153 Add descriptions for LVDS and TV output properties and also mention the EDID
45154 property a new output configuration section.
45156 commit 6f92200db39a790c95c1fac64baf0bd41a70e481
45157 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45158 Date: Wed Dec 5 08:10:17 2007 -0800
45160 Fix backlight setting save/restore
45162 We need to save the current backlight value at LVDS init time, as well as when
45163 we change the DPMS setting. Also, since 0 is a valid backlight value, don't
45164 set the backlight value to maximum at startup if the value happens to be zero.
45166 These fixes should make the backlight user experience much more consistent and
45167 hopefully less frustrating.
45169 commit 4c1fa5235555c9ef3de1eebd006c4e91283362bf
45170 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45171 Date: Tue Dec 4 18:27:24 2007 -0800
45173 Add BCM_ to backlight control method enums
45175 Avoids polluting the global namespace with such generic terms.
45177 commit 30d74f1b1f30e751d2c47aa63a4a2043e71467dc
45178 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45179 Date: Tue Dec 4 18:23:44 2007 -0800
45181 Use "LEGACY" backlight method if backlight control is such
45183 This may regress the user experience a bit (fewer backlight levels) on machines
45184 where both the LBB and native registers work, but it's better that it work for
45185 everyone than work extra well for some and not at all for others.
45187 commit 0e554e6dd4da56be24d2f971ff4cd4d585936724
45188 Author: Eric Anholt <eric@anholt.net>
45189 Date: Tue Dec 4 10:32:45 2007 -0800
45191 Remove extra have_libpciaccess=no that broke tools build with old servers.
45193 commit 3cf862717688aa679b5e667cd041ddd6f5db0b58
45194 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45195 Date: Thu Nov 29 23:12:16 2007 +0800
45197 xvmc: add "XvMC" option, default is false.
45199 commit 6da4b7a80cdf730018435fe84485ecdad62bc860
45200 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45201 Date: Thu Nov 29 22:56:13 2007 +0800
45203 xvmc: move batch buffer init/fini code into generic
45205 commit 75ef3e669dac1259d282dcc8f54b197fc19f22b3
45206 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45207 Date: Wed Nov 28 22:13:55 2007 +0800
45209 Replace ALLOCATE_LOCAL/DEALLOCATE_LOCAL with xalloc/xfree
45211 commit df47b7ba57f3e7dc8bac1898679186d9d48639ac
45212 Merge: df718a7 7f9ceff
45213 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45214 Date: Tue Nov 27 01:13:43 2007 +0800
45216 Merge branch 'master' into xvmc
45218 commit 7f9cefffbf44a2d07b51cc13aaf2d54fcd8f2f22
45219 Author: Alan Hourihane <alanh@tungstengraphics.com>
45220 Date: Fri Nov 16 15:56:08 2007 +0000
45224 commit df718a77c854c9110a81c2bc2a32a9aac85850f2
45225 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45226 Date: Fri Nov 16 16:08:06 2007 +0800
45228 xvmc: add xvmc decoder type string
45230 commit a03a758db0df615fbb44dae1ab6910bfeaeb54f9
45231 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45232 Date: Fri Nov 16 15:49:10 2007 +0800
45234 xvmc: move some i915 context destroy code to generic function
45236 commit 7bfbcb4dc7806ae7543f230373e67f35ba30668d
45237 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45238 Date: Fri Nov 16 15:43:04 2007 +0800
45240 xvmc: clean batch buffer finish
45242 commit a6e8f4ef0481c66fda4f22fcc4bb0ef9c000f0cd
45243 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45244 Date: Fri Nov 16 13:51:44 2007 +0800
45246 xvmc: remove i915 drm map type
45248 commit 4a2b0f340357c4ca58dc9586fad1337b83966362
45249 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45250 Date: Thu Nov 15 18:36:54 2007 -0800
45252 Fix typo in 1920x1080 resolution entry
45254 commit 6c68fcc93917565712b6a9386df54d0204e4ea8f
45255 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45256 Date: Fri Nov 16 10:04:46 2007 +0800
45258 xvmc: more i915 xvmc cleanups
45260 commit 04b60b6d3436fb57163c23acf335364f0ea2f79c
45261 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45262 Date: Thu Nov 15 11:49:53 2007 -0800
45264 Bump version to 2.2.0
45266 commit 6eecef4fed8a21dfdabef42eb69fd150b96167b2
45267 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45268 Date: Thu Nov 15 10:10:59 2007 -0800
45270 Adjust default TV out paramaters
45272 According to several users, a default brightness of 0 results in much better TV
45273 output. Improved control of these parameters will be provided by Randr1.3,
45274 which will standardize several output properties across various chips.
45276 commit 7552d80e367fe38bbc594fe94abd649917fe54d5
45277 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45278 Date: Thu Nov 15 10:07:34 2007 -0800
45280 Add 1920x1080 mode to TV out
45282 Allows users to use 1080p modes on TV out, see bug #13034.
45284 commit e56c166ceb8c1ceddd5020e8de2d0b9d51f535a4
45285 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45286 Date: Thu Nov 15 18:34:18 2007 +0800
45288 remove unnecessary i830_reg.h includes
45290 i830_reg.h only contains 3d engine cmds for 8XX chips.
45292 commit 6f3de19159cd7a2d73ed212add909edb3aedce9c
45293 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45294 Date: Thu Nov 15 18:33:27 2007 +0800
45296 Move fb compression reg definition into i810_reg.h
45298 where we put MMIO control reg in, and shared with intel_reg_dump
45301 commit 1ff97991a4954d508cc260dda82b7678010d13c9
45302 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45303 Date: Thu Nov 15 17:01:31 2007 +0800
45305 xvmc: intel_xvmc.c cleanups
45307 commit 1cfdf192882b856bfc93b735cfb9cb813cac3fdb
45308 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45309 Date: Thu Nov 15 16:38:21 2007 +0800
45311 xvmc: remove unused SAREAPTR macro
45313 commit d9ea2fbdcce3c7c91428846eff43b65ac340f38d
45314 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45315 Date: Thu Nov 15 15:25:35 2007 +0800
45317 xvmc: i915 rename files
45319 commit 06c0579df8d45f7f04b96956353d5f86cf59148f
45320 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45321 Date: Thu Nov 15 15:01:56 2007 +0800
45323 xvmc: some code clean up
45325 commit 4983e3401390654098dbf99dcfb45bab060b068c
45326 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45327 Date: Thu Nov 15 13:56:24 2007 +0800
45329 xvmc: move locking code into generic place
45331 commit bcd656c8dfe7309a5cf02ce1da8bc747ccd6e21c
45332 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45333 Date: Thu Nov 15 11:18:38 2007 +0800
45335 xvmc: oops, should first map batch buffer before init
45337 commit 0464af581b0fcfc4d1c7fec8160d812262086177
45338 Merge: 2256519 e56c166
45339 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45340 Date: Thu Nov 15 11:05:44 2007 +0800
45342 Merge branch 'master' into xvmc
45344 commit 22565198b360826ba10e7ca05d272c419054dc08
45345 Merge: 619044b e64f492
45346 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45347 Date: Thu Nov 15 18:53:22 2007 +0800
45349 Merge branch 'master' into xvmc
45351 commit e64f4929690a57701241334fbca17e95b39ba3d4
45352 Author: Jesse Barnes <jesse.barnes@intel.com>
45353 Date: Wed Nov 14 16:44:34 2007 -0800
45357 Open the "actual_brightness" file as read only, since we only read from it.
45358 Also set an initial backlight_duty_cycle at init time so we don't set the
45359 brightness to 0 at startup.
45361 commit 53e3693ef13f31f3fc33bcff7286ab2b03b2d430
45362 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45363 Date: Wed Nov 14 16:24:56 2007 -0800
45365 Disable FBC by default on 965GM
45367 Several people have reported that they see frequent FBC related display
45368 corruption on 965GM, so disable it for now. Users wanting to enable it can use
45369 the driver option "Framebuffercompression" to override the default.
45371 commit b2726899bc6ebd108aa4a5dd66fe1d881bb778b3
45372 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45373 Date: Wed Nov 14 15:49:41 2007 -0800
45377 Needed for the new debug code
45379 commit b7bc694c6b98dac17763426d905a22d3ae17a018
45380 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45381 Date: Wed Nov 14 15:46:57 2007 -0800
45383 Fix kernel get_brightness function
45385 We need to look at "actual_brightness" rather than "brightness". The former
45386 contains the brightness value the kernel driver has actually set, while the
45387 latter is merely what the user requested.
45389 commit 9ee96ed02aacd953ab58288fddf0fa60f60b668e
45390 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45391 Date: Wed Nov 14 15:44:56 2007 -0800
45393 Correct FBC debug message
45395 To be consistent, it should say 'plane' rather than 'pipe'.
45397 commit 675b9471d84059496ca41d427a4f95d419f82be9
45398 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45399 Date: Wed Nov 14 15:17:05 2007 -0800
45401 Add FBC registers to register dump output
45403 Just for completeness.
45405 commit a2a0821e74a61f53cc7f0c41ce629644ad712114
45406 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45407 Date: Wed Nov 14 11:30:20 2007 -0800
45409 Don't stop ring before restoring hw state
45411 Some of the hw state restoration callbacks may want to use the ring for various
45412 things like stopping video playback, so leave the ring enabled until we get
45413 back from RestoreHWState. Also rename the functions so that their purpose is
45414 clearer and remove a couple of redundant lines.
45416 commit 619044b242d4abf605b42cfff538b63f445fe709
45417 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45418 Date: Thu Nov 15 02:23:07 2007 +0800
45420 xvmc: move batch buffer init/fini to intel_batchbuffer.c
45422 commit ae805ac379ccfdae9ee272896665e73f5bc7d2cf
45423 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45424 Date: Thu Nov 15 02:13:07 2007 +0800
45426 xvmc: take out batch buffer handling code
45428 commit ad3bc0158d37e98fcbbe6a8e31413c142a260424
45429 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45430 Date: Wed Nov 14 22:46:46 2007 +0800
45432 Don't enable fbc with XAA or tiling is off.
45434 This slightly reworks my last fbc patch. We don't
45435 support tiled front buffer with XAA now, so also
45436 disable fbc on it. If tiled alloc failed, disable
45439 commit 26194e19e1c80615697016e25640d4c8c244353f
45440 Author: Keith Packard <keithp@koto.keithp.com>
45441 Date: Tue Nov 13 22:26:40 2007 -0800
45443 Review PLL spreadsheet and update register ranges.
45445 The PLL spreadsheet makes the precise register ranges allowed for each mode
45446 quite clear, and shows a few inaccuracies in the b-spec. In particular, the
45447 N register value may range from 1 to 6 instead of 3 to 8. This should close
45448 the gap we've seen in the reachable frequencies.
45450 commit 581509321e51becb8ee6ddcfaa15a2a713c9ef8e
45451 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45452 Date: Wed Nov 14 22:16:04 2007 +0800
45454 Remove one redundant line.
45456 commit 8cea3c8cfa86600b5f2c3e1f21c19d4b2a320fd4
45457 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45458 Date: Wed Nov 14 22:13:30 2007 +0800
45460 Don't setup fb compression if fb is not tiling
45462 commit 539b669d6e23833a917a9b0d3a355d4283e0b76a
45463 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45464 Date: Tue Nov 13 17:59:54 2007 -0800
45466 Remove harmless error message
45468 This really isn't an error in general. If vblank pipe setup really fails for
45469 some reason, it'll be obvious enough when the user tries to use vblank events
45472 Patch from Hong Liu.
45474 commit e8db8584ae4860aa991c916126335b287a972d5b
45475 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45476 Date: Wed Nov 14 00:11:20 2007 +0800
45478 xvmc: attempt to move batch buffer code out of i915 private
45480 commit 11c0e0dbef42f1ae77a77d534870d3e9821fb017
45481 Merge: c838f72 43fbc54
45482 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45483 Date: Tue Nov 13 22:44:06 2007 +0800
45485 Merge branch 'master' into xvmc
45487 commit 43fbc547786cf26514f95bce79fab58a66c291bf
45488 Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com>
45489 Date: Mon Nov 12 12:23:41 2007 -0800
45491 Improve backlight control
45493 This commit fixes backlight support for several platforms.
45495 Except on recent machines supporting the IGD OpRegion specification,
45496 backlight control is rather platform specific. In some cases, we can
45497 program the native backlight control regsiters directly without any
45498 trouble. On others, we need to use the legacy backlight control
45499 register. On still others, we need a combination of the two. And on
45500 some platforms, none of the above will work, so we go through the
45501 kernel backlight interface, which provides a platform specific driver
45502 for backlight control.
45504 commit 10988c5e6ec0f3c40d56bbf209b7976627cca706
45505 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45506 Date: Mon Nov 12 09:19:22 2007 -0800
45508 Fix typo in PLL enable check
45510 Check against DPLL_A instead of DPLL_B before writing PIPEACONF.
45512 Thanks to Colin Guthrie for his sharp eyes.
45514 commit ab2055ebb20aa6de121fa377e488ce91913035ae
45515 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
45516 Date: Fri Nov 9 13:55:46 2007 -0800
45518 Add 'ret' variable to i830_allocator_init
45520 Needed in the XF86DRI_MM case.
45522 commit 24280edbdf54e9559b26ea27c1b005eea73f51bc
45523 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
45524 Date: Fri Nov 9 13:53:58 2007 -0800
45528 Remove unused 'ret' variable, move pI830 under #ifdef XF86DRI_MM in
45529 i830_allocate_memory.
45531 commit 4fe507957bf826d81a71cd63af17c5547d1023a1
45532 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
45533 Date: Fri Nov 9 13:51:11 2007 -0800
45535 Remove unused 'palette_enable' variable
45537 commit 62404a9c9c0d98efb69ac96209dea17331e0489c
45538 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45539 Date: Fri Nov 9 07:55:01 2007 -0800
45541 Update man page to reflect EXA by default
45543 commit c838f7297f675ea9d82678ed37715955cb0a8a78
45544 Merge: a9449ff 6f53568
45545 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45546 Date: Fri Nov 9 19:37:07 2007 +0800
45548 Merge branch 'xvmc' of git+ssh://zhen@people.freedesktop.org/~zhen/xf86-video-intel into xvmc
45550 commit a9449fff4be8e6b3cc82914376420e3a66b960f5
45551 Merge: a470f6b 177924e
45552 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45553 Date: Fri Nov 9 19:36:25 2007 +0800
45555 Merge branch 'master' into xvmc
45557 commit 177924e879564b7e9e70fd607141978bfd053fff
45558 Author: Jesse Barnes <jesse.barnes@intel.com>
45559 Date: Thu Nov 8 18:16:48 2007 -0800
45561 Bump driver version to 2.1.99 in preparation for 2.2 release
45563 commit cb4e5796f0537ea5e0e646d473930c7b826c85d8
45564 Author: Jesse Barnes <jesse.barnes@intel.com>
45565 Date: Thu Nov 8 18:13:47 2007 -0800
45569 If EXA is compiled into the driver, default to using it for acceleration.
45570 Hopefully we can remove XAA entirely one day.
45572 commit 43a59ab26b09fcc24de1ed7bd770bb622f899ceb
45573 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45574 Date: Thu Nov 8 16:52:41 2007 -0800
45576 Clear current video crtc on DPMS off
45578 When calling the video DPMS off function, make sure we zero out the current
45579 crtc so that it will be properly re-set up next time video is turned on.
45581 Fix from Peter Clifton with changes by Keith Packard.
45583 commit e784e152a8e84b6e447b55a5c7019e7b47e17621
45584 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45585 Date: Thu Nov 8 16:37:53 2007 -0800
45587 Fix overlay destination clamping
45589 The overlay width & height scaling clamp check was reversed. Fix that and
45590 update the comment.
45592 commit b8770f710729d616b3ac72544aa522161a78f819
45593 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45594 Date: Thu Nov 8 16:19:01 2007 -0800
45596 Setup 3D state at EnterVT time
45598 In the absence of full suspend/resume support in the kernel, we have to
45599 save/restore state in Enter/LeaveVT. For 8xx chips, 3D state may be lost
45600 during suspend/resume, so re-emit the basic setup at EnterVT time.
45602 Patch from Peter Clifton.
45604 commit 5f92b4c2db9712496b829ee239468e3d14e27d2f
45605 Author: Eric Anholt <eric@anholt.net>
45606 Date: Thu Nov 8 13:31:51 2007 -0800
45608 Bug #12059: Add ch7019 to the list of supported devices for ch7017.
45610 While I'm here, fix the chip description to be LVDS instead of TMDS in i2c
45613 commit eecd3ccedee6c4acf101591f7e60673660379e62
45614 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
45615 Date: Thu Nov 8 09:31:08 2007 -0800
45617 Check DPLL status before writing PIPEnCONF regs
45619 If the DPLL isn't enabled or is in VGA mode, writing the PIPEnCONF registers
45620 may cause a hang or crash. So ensure the DPLL is in the proper state before
45623 Another excellent fix from Peter Clifton.
45625 commit b434c1a437c407de88396b219560649c2dae82b2
45626 Author: Jesse Barnes <jesse.barnes@intel.com>
45627 Date: Thu Nov 1 12:06:07 2007 -0700
45629 Framebuffer compression fix: front buffer may not be at fence 0
45631 Fix a long standing bug in the framebuffer compression code (thanks to
45632 Pierre Willenbrock!) that prevented FBC from working correctly if the front
45633 buffer was anywhere but fence register 0.
45635 commit 50d3693a6862028e50bc5ba8c788e7ea573c3eb8
45636 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45637 Date: Thu Nov 1 22:22:27 2007 +0800
45639 Fix allocation reset for really not free BO allocator
45641 commit ecd995d533d28b622afc71a20504d47c33ff5b8d
45642 Author: Alan Coopersmith <alan.coopersmith@sun.com>
45643 Date: Tue Oct 30 18:20:49 2007 -0700
45645 Fix builds without DRI
45647 commit 7c88b58a93fce9fda59b6344acb87af16336e287
45648 Author: Alan Coopersmith <alan.coopersmith@sun.com>
45649 Date: Tue Oct 30 18:20:15 2007 -0700
45651 Clear compiler error: "void functions cannot return values"
45653 commit a470f6b2ca291e305979dd9223885a50dc0b334b
45654 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45655 Date: Fri Oct 26 22:40:47 2007 +0800
45657 try to move some fields to xvmc driver structure
45659 Still a lot of things to move include batch buffer, lock, context,
45660 surface resource handling, etc.
45662 commit 6f53568cd2aeeec6bd8c1f1d9f9c5ecadc5c62f9
45663 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45664 Date: Fri Oct 26 14:58:43 2007 +0800
45666 Fix DRI context creation for XvMC
45668 commit d2c78f82c20f33fc9c22cab8a7ca161e57a34bf8
45669 Author: Thomas Hellstrom <thomas-at-tungstengraphics-dot-com>
45670 Date: Wed Oct 24 19:34:12 2007 +0200
45672 Adapt to DRM Lockfree and setStatus changes.
45674 commit 78aaec0ffc711742bf8ad77757ed8c15cc3f7a9f
45675 Author: Eric Anholt <eric@anholt.net>
45676 Date: Wed Oct 24 11:50:54 2007 -0700
45678 Fix typo in my hand-application of rglowery's patch.
45680 commit 021265fbec9fd4ee31bdc9767c0c0453479ea22c
45681 Author: Rob <rglowery@exemail.com.au>
45682 Date: Wed Oct 24 11:06:53 2007 -0700
45684 Fix a crash in TV mode handling by initializing the prev field of modes.
45686 commit ce9802edd25255f6c9c11e512b2ef905384b914d
45687 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45688 Date: Wed Oct 24 00:31:53 2007 +0800
45690 fix xvmc return values
45692 Use values defined by xvmc lib spec
45694 commit 9c84543eb0fde4643eb7529eb6b08eacf57c12d0
45695 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45696 Date: Tue Oct 23 22:16:55 2007 +0800
45698 add missing license header for new files
45700 commit 87345e820d20f1a98216c52e156c2e18c15ffa44
45701 Author: Jesse Barnes <jesse.barnes@intel.com>
45702 Date: Mon Oct 22 14:22:37 2007 -0700
45704 intel_reg_dumper - dump VGA AR registers too
45706 Add a VGA AR dumping function so we can debug text mode problems too.
45708 commit 90e4eafdf59edd966ed626107c3909b24737da8d
45709 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45710 Date: Mon Oct 22 17:09:15 2007 +0800
45712 revert xvmc memory allocation hacks
45714 commit 167c56bcf82f0a6d72151b202e6d56f6e0d627f6
45715 Merge: fc9368c 78e251d
45716 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45717 Date: Tue Oct 23 00:12:22 2007 +0800
45719 Merge branch 'master' into xvmc
45721 commit fc9368ce7dacd52ef222559d20188ddcf2a85df5
45722 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45723 Date: Tue Oct 23 00:09:06 2007 +0800
45725 split xvmc user lib chipset drivers
45727 commit 78e251db671e21bc859c9b505d391b70babee2dc
45728 Author: Eric Anholt <eric@anholt.net>
45729 Date: Fri Oct 19 15:04:10 2007 -0700
45731 In the clock graph, draw the VCO as erasures in the lines representing clocks.
45733 This shows one of the reasons for the gaps: with the other settings, the VCO
45734 is too low inside the gap. However, it also points out another issue: we
45735 aren't using the high end of the VCO range due to some other limits being hit.
45737 commit 9f9b888525b274036d301d6e06351583d0415f9e
45738 Author: Eric Anholt <eric@anholt.net>
45739 Date: Thu Oct 18 11:25:24 2007 -0700
45741 Warn in the log if we choose a PLL clock that's way out of line.
45743 commit 1f8bf110394cc1df66aae9acf5c818145ae19b52
45744 Author: Eric Anholt <eric@anholt.net>
45745 Date: Thu Oct 18 11:17:38 2007 -0700
45747 Add some nickle scripts for looking at PLL issues.
45749 While here, move similar nickle scripts under src/scripts/
45751 commit 79636b8c776ae024518103c9fa137c8498c21d48
45752 Author: Eric Anholt <eric@anholt.net>
45753 Date: Wed Oct 17 17:10:24 2007 -0700
45755 Bump pciaccess version requirement for new API usage.
45757 commit c824c45e72908b5b3fa832fec91c7b47f305180c
45758 Author: Dave Airlie <airlied@optimus.(none)>
45759 Date: Wed Oct 17 16:35:29 2007 +1000
45761 intel: make block handler hook happen no matter what video is in use
45763 This shouldn't make a difference now, but for upcoming EXA/batch changes it
45764 will be more useful
45766 commit aa88d23e615e7f154cf850927a0a7bb9fbb0d2f9
45767 Author: Rob <rglowery@exemail.com.au>
45768 Date: Wed Oct 17 09:19:19 2007 +0800
45770 AOpen 965GM mini pc has no LVDS
45772 commit 38900f55e2b50ea383f78adb016e3a02c3bd17ad
45773 Author: Kyle McMartin <kyle@freedesktop.org>
45774 Date: Tue Oct 16 16:38:28 2007 -0400
45776 toshiba satellite u300 has no tv out
45778 commit 74ac5de14ebb77aeb39d698e9e8d188c9d9abd76
45779 Author: Michel Dänzer <michel@tungstengraphics.com>
45780 Date: Tue Oct 16 15:08:26 2007 +0200
45782 Adapt to libdrm buffer object API changes.
45784 commit d66e924680a3504aa1feb9c228145a95d1a16b1e
45785 Author: Eric Anholt <eric@anholt.net>
45786 Date: Mon Oct 15 11:31:51 2007 -0700
45788 In update_front_offset(), set pScrn->fbOffset even at EnterVT during init.
45790 Failure to do so led to the memset() at EnterVT being done to an invalid
45793 commit cc892d1ba1d87d08bdd04f857efc6b8f64620018
45794 Merge: b52ce0b fd97edb
45795 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45796 Date: Mon Oct 15 22:09:14 2007 +0800
45798 Merge branch 'xvmc' of git+ssh://zhen@people.freedesktop.org/~zhen/xf86-video-intel into xvmc
45800 commit c94cdfd6ddbc580523737f596e97b96a7ce100b9
45801 Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com>
45802 Date: Sat Oct 13 15:57:45 2007 -0700
45804 Fix palette save/restore
45806 When restoring the palette, we weren't checking to make sure the
45807 associated pipe was enabled before writing the registers. In some
45808 configurations, this led to a driver crash. Add new routines to handle
45809 palette save/restore and verify that the pipes are on before we touch
45810 the registers (could easily be changed to enable/disable the pipes
45811 around the save/restore as well).
45813 commit e04333a6352040bc883655d606923c912d005981
45814 Author: Eric Anholt <eric@anholt.net>
45815 Date: Fri Oct 12 12:42:45 2007 -0700
45817 Fix failure in tiling setup on non-power-of-two allocations on pre-965.
45819 commit a1612b7728d4153499fe86b6713a13c8702cc7d9
45820 Author: Eric Anholt <eric@anholt.net>
45821 Date: Fri Oct 12 12:42:06 2007 -0700
45823 Use mprotect on unbound AGP memory to attempt to catch use while unbound.
45825 This doesn't help with the most common use-while-unbound cases, which are
45826 from the hardware side.
45828 commit 3af442ba52550a9d183e215d49cc12dac0cb9e4b
45829 Author: Eric Anholt <eric@anholt.net>
45830 Date: Thu Oct 11 11:56:06 2007 -0700
45832 Don't double-free the memory manager allocation.
45834 commit 6c485ff5bd25e9aa6a3c1eb669843e6b969d94df
45835 Author: Eric Anholt <eric@anholt.net>
45836 Date: Fri Oct 5 15:52:56 2007 -0700
45838 Update memory manager sizing for the current set of LIFETIME_FIXED bufffers.
45840 commit d47cf148776d74f9035863b23eefdc2b5893af08
45841 Author: Eric Anholt <eric@anholt.net>
45842 Date: Fri Oct 5 15:45:51 2007 -0700
45844 Move tiling fence register setup to bind time instead of allocate time.
45846 This allows us to allocate tiled buffers in buffer objects. In the process
45847 I removed the fence division that we had for tiled buffers on pre-965. If we
45848 resurrect that code, it should probably be managed by just dividing all the
45849 objects in roughly half and fencing those halves (to reduce the alignment
45850 requirement), instead of using giant fences until we run out of space and then
45851 trying to deal with scarce space on the last (or not) buffer. Halving
45852 our tiled objects would use 6/8 of our fence registers on that hardware.
45854 commit a6dc81a0864f9ab2f6fc1aa31002c0191f674ceb
45855 Author: Eric Anholt <eric@anholt.net>
45856 Date: Fri Oct 5 12:37:19 2007 -0700
45858 Allow front/back/depth to move over the lifetime of the server.
45860 commit 132dce7565feeea1055899f8c1627766fe84c88c
45861 Author: Eric Anholt <eric@anholt.net>
45862 Date: Fri Oct 5 12:12:06 2007 -0700
45864 Delay SAREA and mapping setup until EnterVT when using the memory manager.
45866 Otherwise, we would use uninitialized offsets in the early setup.
45868 commit 4ca3550fb7d488741f8dc1ba3c8722393277c3b8
45869 Author: Eric Anholt <eric@anholt.net>
45870 Date: Thu Oct 4 17:02:15 2007 -0700
45872 Rework DRI buffer mappings and sarea setup to allow for moving buffers.
45874 While this has been a desired feature for some time, to allow for reallocation
45875 of the front buffer, it was made more necessary by the desire to avoid
45876 requiring a NO_MOVE buffer type in TTM because buffer objects may not be left
45877 pinned over VT switch. This is a step towards making those buffers
45878 movable and resizable.
45880 commit f393a12d21eed668cf4771f022beded9a4c547c7
45881 Author: Eric Anholt <eric@anholt.net>
45882 Date: Thu Oct 4 13:03:14 2007 -0700
45884 Move drmMMLock to after we have unbound our (pinned) buffers.
45886 There are still issues due to the fact that we're allocating
45887 NEED_LIFETIME_FIXED memory as buffer objects, which we refuse to unpin because
45888 we have no way of pinning it back in the same location.
45890 commit 8fc4e3078210f726c7c375faa2f2fd8d05017c09
45891 Author: Eric Anholt <eric@anholt.net>
45892 Date: Fri Oct 5 11:51:55 2007 -0700
45894 Refuse to allocate LIFETIME_FIXED objects in buffer objects.
45896 We can't guarantee the offset will stay the same using the current DRM
45897 interface, but the correct solution is fixing our code to allow these objects
45898 to move. Breaks TTM mode of the DRI driver for now.
45900 commit 0de747f7d219a56434dad49a8a6d1d9d4c251ca7
45901 Author: Eric Anholt <eric@anholt.net>
45902 Date: Thu Oct 11 10:36:01 2007 -0700
45904 In i830_allocate_memory_bo, bind if we control the VT, not on lifetime-fixed.
45906 commit 0040bb95445e800af80ca9fa1b92f5db33f1b4ac
45907 Author: Eric Anholt <eric@anholt.net>
45908 Date: Wed Oct 10 16:48:12 2007 -0700
45910 Fix potential use-after-free in XV overlay code on video stop.
45912 commit 56ae767180da4a8bd4b7a4c35b92e019ab67d753
45913 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45914 Date: Thu Oct 11 10:09:29 2007 +0800
45916 Revert "Fix G33 GTT stolen mem range"
45918 This reverts commit 2a8592f2ebcba86b1127aa889155d58a3dc186ca.
45920 This causes compat issue between ddx and kernel apggart version.
45921 Revert it as it shouldn't hurt normal people's default GTT size,
45922 but if you change it to 1MB in bios, it's supposed to be broken now.
45924 commit 5faee02e97b7543b41bb1f8e3751558c48802833
45925 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45926 Date: Wed Oct 10 13:55:30 2007 +0800
45928 EXA: fix tiled dest rendering on i8XX chips
45930 commit 5bbe9f09fda1238da04e6b584813d40fc646f49d
45931 Author: Keith Packard <keithp@koto.keithp.com>
45932 Date: Wed Oct 3 00:18:25 2007 -0700
45934 Include xf86mm.h if using XF86DRI_MM
45936 This header file doesn't appear to be getting automatically included for
45939 commit 41dacf97ec1d4ba6f54057f4fcc7bb76033a454e
45940 Author: Keith Packard <keithp@koto.keithp.com>
45941 Date: Wed Oct 3 00:17:41 2007 -0700
45943 Quirk for Samsung Q35 which has no TV output.
45945 commit b52ce0bd178c268d20334e2be8fad46037e47499
45946 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45947 Date: Mon Oct 8 22:56:44 2007 +0800
45949 use field name to initialize i915 xvmc driver
45951 commit e0fb228401b10ea4281eee48497bf1d584999705
45952 Merge: 84d7b2b 9e770bf
45953 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45954 Date: Mon Oct 8 22:45:11 2007 +0800
45956 Merge branch 'master' into xvmc
45962 commit fd97edb62463e50a1860290d54684a68d5490bdc
45963 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45964 Date: Mon Oct 8 16:13:26 2007 +0800
45966 move some definition out of i915 specific header
45968 commit 3a2b5222341b37322b788e329c2b9006dc14ecfd
45969 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
45970 Date: Mon Oct 8 15:46:11 2007 +0800
45972 Temp work around ttm allocation for xvmc, which should be fixed later.
45974 commit 9e770bf69edaa8944048049e11266dbe1ef145e5
45975 Author: Eric Anholt <eric@anholt.net>
45976 Date: Thu Oct 4 14:31:37 2007 -0700
45978 Replace setting of LIFETIME_FIXED on cursors with just updating the offsets.
45980 commit c9d6e90c2523096c45d330552f471e6bf1752704
45981 Author: Eric Anholt <eric@anholt.net>
45982 Date: Thu Oct 4 14:27:07 2007 -0700
45984 Don't set overlay registers LIFETIME_FIXED. It always uses the current offset.
45986 commit 1cc6bec4628a4ca04a68919a84148a1e6312072f
45987 Author: Eric Anholt <eric@anholt.net>
45988 Date: Thu Oct 4 14:23:48 2007 -0700
45990 Revert setting of exa_965_state as LIFETIME_FIXED.
45992 The current offset is used every time in render setup.
45994 commit dfa22df8ce7713b491c4766ce6ce652db91b5f90
45995 Author: Eric Anholt <eric@anholt.net>
45996 Date: Fri Oct 5 12:53:07 2007 -0700
45998 Don't leak buffer object allocations in i830_reset_allocations().
46000 commit 7db9d706a1bc9da7afc4823cee92b23e96ae55b8
46001 Author: Eric Anholt <eric@anholt.net>
46002 Date: Fri Oct 5 12:40:48 2007 -0700
46006 commit bf33c4e2d91172efd08690fadda7b3743dd3de62
46007 Author: Eric Anholt <eric@anholt.net>
46008 Date: Fri Oct 5 12:40:37 2007 -0700
46010 Fix formatting of error message.
46012 commit 04e936935f0b0045600241424f1d04a6721a2432
46013 Author: Eric Anholt <eric@anholt.net>
46014 Date: Mon Oct 1 17:29:35 2007 -0700
46016 Bring the CH7017 driver closer to spec.
46018 This is also closer to what my hardware is programmed with, except for some
46019 very confusing off-by-one bugs in an unexpected direction.
46021 commit 263d48ad93a0fb7100729d0c6e0443797c20c80a
46022 Author: Eric Anholt <eric@anholt.net>
46023 Date: Mon Oct 1 16:23:30 2007 -0700
46025 Fix probing of the sample CH7017 device I found by allowing GPIO overrides.
46027 It may be that the LVDS chips need to be DVOA and GPIOB only on mobility
46028 devices with them, and DVOC/GPIOE on non-mobility like this 845. But until
46029 more examples are found, just make this one device probe.
46031 commit 219354af212c7b68c20df689692c55331e36a705
46032 Author: Eric Anholt <eric@anholt.net>
46033 Date: Mon Oct 1 16:09:13 2007 -0700
46035 Restore building of pciaccess-based tools even if the server doesn't use it.
46037 commit 1b83e5e589105d05c1b78032fa47eecee478cb99
46038 Author: Eric Anholt <eric@anholt.net>
46039 Date: Mon Oct 1 14:32:28 2007 -0700
46041 Remove more DVO probing noise.
46043 commit 29d3188ee9676f697f6fcb859c7f680f5df30318
46044 Author: Eric Anholt <eric@anholt.net>
46045 Date: Mon Oct 1 13:45:09 2007 -0700
46047 Suppress I2C failure error messages during DVO device detection.
46049 Otherwise, errors would be printed for every chip probed which wasn't present
46052 commit 6dd8228a386334d05836353672aae9f8c456e435
46053 Author: Eric Anholt <eric@anholt.net>
46054 Date: Mon Oct 1 12:07:22 2007 -0700
46056 Enable CH7017/7018 DVO driver by default.
46058 On hardware containing one of these chips (generally single pipe with an LVDS
46059 connected), it's probably better to try to enable it and fail in some cases
46060 than to not probe the output at all.
46062 commit 994ee6721d9f1650ad37e21198b86848352b73af
46063 Author: Eric Anholt <eric@anholt.net>
46064 Date: Mon Oct 1 12:01:16 2007 -0700
46066 Disable TV/VGA output on the CH7017/7018 devices.
46068 We don't support those outputs on this chip yet, so turn them off if the BIOS
46071 commit 0e5574e842eeaac75271ccd50027d3e54c57722d
46072 Author: Regis Prevot <regis@myrix.net>
46073 Date: Sat Sep 29 16:21:08 2007 +0200
46075 Ignore detection of TV output on Panasonic CF-Y7
46077 commit 84d7b2b955107264b67526a04028a8b79d8804e3
46078 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46079 Date: Sat Sep 29 19:41:15 2007 +0800
46081 rename xvmc init function names
46083 commit 1845e7c6e9caf7a9ad7f46a8fbfa321ddec69a1e
46084 Author: Jesse Barnes <jesse.barnes@intel.com>
46085 Date: Fri Sep 28 10:12:11 2007 -0700
46087 Go back to using old drm_i915_flip_t field name
46089 This field shouldn't have been renamed in the first place. Go back to using
46090 the old name so that the tree is backward and forward compatible again.
46092 commit f3d92ab474de11babe507b0e3c15aca146b6cb66
46093 Author: Kyle McMartin <kyle@freedesktop.org>
46094 Date: Fri Sep 28 10:17:55 2007 -0400
46096 Dell XPS 1330 has no TV out
46098 commit bbf69c7446aa9748dacecccfe0b63d803f865b00
46099 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46100 Date: Thu Sep 27 17:37:19 2007 +0800
46102 Pin cursor, overlay(no physical) and exa state buffers
46104 And cursor mem counting for dri mem manager is not relate
46107 commit 64188246bb25e95a916ebb8a50a09298e56d91d9
46108 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46109 Date: Wed Sep 26 15:48:19 2007 +0800
46111 clean up old i915 xvmc function declare
46113 commit 320b7f08f457455f1d16b4c36d149bdc3f6ceead
46114 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46115 Date: Wed Sep 26 15:02:33 2007 +0800
46117 Revert to origin xvmc setup steps
46119 We have to grab xvadaptor's PutImage earlier than
46120 xv init, otherwise no chance to do that..
46122 commit 1e0ff38f44e3b5a8b4a5d509c7faa266270d522c
46123 Author: Hong Liu <hong.liu@intel.com>
46124 Date: Thu Sep 20 11:28:52 2007 +0800
46128 Use consistent interface for counting pixmap offset.
46130 commit 85e4bd109464a4d90856e17bc960872eeaebc2ad
46131 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46132 Date: Wed Sep 26 14:32:46 2007 +0800
46134 Log xvmc driver name
46136 commit e58b76a24a29e0a502650807de81e650e5dc184d
46137 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46138 Date: Wed Sep 26 14:40:04 2007 +0800
46140 clean up err report
46142 Still need to clean up return value for XvMC API
46144 commit b10a9668dc50d10489b2ae531d89c8ab7c408170
46145 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46146 Date: Wed Sep 26 14:01:57 2007 +0800
46148 fix i915 xvmc devPrivate reference
46150 commit f12eaaf385e7a0b85524588a5d865d1cc3f1a0cf
46151 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46152 Date: Wed Sep 26 13:49:55 2007 +0800
46154 comment out i965_xvmc_driver for now
46156 commit 513e8a2f8abde1b11b8058e94a650796fd51779b
46157 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46158 Date: Wed Sep 26 13:45:42 2007 +0800
46160 Create xvmc driver interface
46162 convert i915 to new xvmc driver interface
46164 commit ca67fa767dc762dac369e84b27a7ef15673d527c
46165 Author: Eric Anholt <eric@anholt.net>
46166 Date: Mon Sep 24 11:12:22 2007 -0700
46168 Remove logic for supporting i915tex_dri.so vs. i915_dri.so.
46170 There can be only one.
46172 commit 130f79613bfe6a8cfa7f431c8cce06cbb93cc91a
46173 Merge: b73235f 1260316
46174 Author: Eric Anholt <eric@anholt.net>
46175 Date: Fri Sep 21 17:14:44 2007 -0700
46177 Merge branch 'buffer-objects'
46179 This branch changes i830_memory.c's allocator to use TTM when available to
46180 allocate memory, which also allows TTM to control almost the entire aperture.
46181 As a result, our front/back/depth buffers are created as real buffer objects,
46182 which may be used by the DRI driver instead of the fake buffer type in TTM.
46183 The updated DRM with bo_set_pin ioctl is required, to allow us to pin and
46184 unpin our buffers as needed.
46186 commit 126031655c3c1db9881ad17d712461533eb59632
46187 Merge: c343835 bda3c69
46188 Author: Eric Anholt <eric@anholt.net>
46189 Date: Thu Sep 20 08:57:58 2007 -0700
46191 Merge remote branch 'origin/master' into buffer-objects
46198 commit b73235f40497cfb10792ba191d4f6eac3a5df009
46199 Author: Hong Liu <hong.liu@intel.com>
46200 Date: Thu Sep 20 11:28:52 2007 +0800
46204 Use consistent interface for counting pixmap offset.
46206 commit 798448e4641acf241b7cbae0d6c243ae383da9f4
46207 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46208 Date: Wed Sep 19 13:29:31 2007 +0800
46210 Add media state definition on i965
46212 This adds VFE state and interface descriptor struct define.
46214 commit 36b81ba20d20e7b235cb701c5135cbf3f3eedb12
46215 Merge: 02ad9ce bda3c69
46216 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46217 Date: Wed Sep 19 11:44:41 2007 +0800
46219 Merge branch 'master' into xvmc-i915
46226 commit bda3c69a7a044869ea63514b97f8aa9260d6b5a7
46227 Author: Dmitry Babrovich <dmitry.babrovich@densitrongaming.com>
46228 Date: Tue Sep 18 09:34:34 2007 -0700
46230 Bug #12457: Let the ch7xxx driver probe the 7301
46232 Looking at the 7304, this chipset likely requires slightly different
46233 programming. However, this change allows for at least some functionality.
46235 commit 445a59d262b93bd9d2c16cee9934ce77a103776b
46236 Author: Jesse Barnes <jesse.barnes@intel.com>
46237 Date: Wed Sep 12 09:32:41 2007 -0700
46239 Remove unused plane->pipe mappings from SAREA private
46241 Turns out we can get away without this, so remove it, fix a crash, and only
46242 swap planes/pipes if the DRM can support it.
46244 commit d02336290bea30de3c390b8121046c38fd6b0f62
46245 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
46246 Date: Tue Sep 11 04:28:14 2007 -0700
46248 Fix plane/pipe mapping compat code
46250 Move plane->pipe mapping adjustment to ScreenInit so we can check
46251 against the DRM driver version accurately.
46253 commit d9f2b3c0d2d08a4b7fad865dab7deb6224b57999
46254 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
46255 Date: Mon Sep 10 23:49:34 2007 -0700
46257 Fix crash in ScreenInit
46259 Use pScreen directly when getting at the SAREA private, since
46260 pScrn->pScreen may not be initialized yet.
46262 commit 4c7542ef43a5267e470ca1608a2ae57abf9783ec
46263 Author: Jesse Barnes <jesse.barnes@intel.com>
46264 Date: Mon Sep 10 23:30:50 2007 -0700
46266 Only swap planes and pipes if DRM supports it
46268 We want to associate plane A with pipe B on pre-965 mobile chips, since that's
46269 the only way to get framebuffer compression on the builtin LVDS on those
46270 platforms. However, if we do this swapping and DRM isn't aware of it, we may
46271 end up requesting vblank events for the wrong pipe, or setting up SAREA buffer
46272 swap state incorrectly.
46274 This mod checks whether DRM supports the new plane->pipe swapping behavior, and
46275 only enables the swapping if so. This should fix the bugs Lukas found and
46276 debugged. Reviewed by Michel Danzer.
46278 commit 286f5df0b62f571cbb4dbf120679d3af029b8775
46279 Author: Keith Packard <keithp@koto.keithp.com>
46280 Date: Thu Sep 6 14:31:55 2007 -0700
46282 Switch to pci_device_map_range/pci_device_unmap_range APIs.
46284 With the libpciaccess change that added these new APIs, use them
46287 commit 2a8592f2ebcba86b1127aa889155d58a3dc186ca
46288 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46289 Date: Wed Sep 5 14:52:56 2007 +0800
46291 Fix G33 GTT stolen mem range
46293 G33 GTT table lives in seperate stolen mem with
46294 graphics data stolen mem.
46296 commit 7fd9a98178cdebda4213796fdc452a8a265a1197
46297 Author: Keith Packard <keithp@koto.keithp.com>
46298 Date: Tue Aug 28 16:00:01 2007 -0700
46300 Don't set supported TV formats until after RandR initialized.
46302 The TV format property cannot be configured until RandR has been
46305 commit c6e637cd683dc60567b3b4f69b7f2b4c338c89ea
46306 Author: Keith Packard <keithp@koto.keithp.com>
46307 Date: Tue Aug 28 12:17:15 2007 -0700
46309 Limit TV formats to those supported by current connection
46311 commit ddd6053987b9ca9bd3722ddbdfd412a3d8d252cf
46312 Author: Keith Packard <keithp@koto.keithp.com>
46313 Date: Tue Aug 28 12:30:46 2007 -0700
46315 Add register defines for hw binning
46317 commit 3fbbd0afde49c53a5a8661f75c8c8c4be3020c30
46318 Author: Michel Dänzer <michel@tungstengraphics.com>
46319 Date: Tue Aug 28 17:48:20 2007 +0200
46321 Fix build against pre-pci-rework xserver.
46323 commit 0fdbf64b34e4114c2b89d696b268b9c7464f1efd
46324 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46325 Date: Tue Aug 28 21:56:21 2007 +0800
46327 Fix i915 a8 color buffer blending
46329 From spec, i915 engine uses green channel when reading from 8bit
46330 color buffer for blending, and also writes back green channel.
46331 Fix blend factor in dest alpha case by using dest color instead.
46332 Now rendercheck can pass a8 tests.
46334 commit 3411eb0dbae470b910af3116a4ab960c821b9b20
46335 Author: Keith Packard <keithp@koto.keithp.com>
46336 Date: Sun Aug 26 23:09:01 2007 -0700
46338 i830_driver.c changes for libpciaccess.
46340 Change to use libpciaccess APIs, including computing and using BAR indices
46341 for various mapping activities.
46343 commit 387fed6daa7426e4a85d30ba7cf608b5f41d24bb
46344 Author: Keith Packard <keithp@koto.keithp.com>
46345 Date: Sun Aug 26 23:06:57 2007 -0700
46347 i810_driver.c changes for libpciaccess.
46349 This includes new probe code (intel_pci_probe) and changes for i810 to
46350 use BAR indices to refer to suitable portions of the device mappings.
46352 commit 2c794192052ca55c3263e27e13d16aafe8caa92c
46353 Author: Keith Packard <keithp@koto.keithp.com>
46354 Date: Sun Aug 26 22:46:19 2007 -0700
46356 Mechanical API conversions for libpciaccess.
46358 Uncomplicated API transistions for libpciaccess usage:
46360 Legacy xf86 API libpciaccess API
46361 --------------- ----------------
46362 xf86ReadPciBIOS pci_device_read_rom
46363 pciReadWord pci_device_cfg_read_u16
46364 pciWriteByte pci_device_cfg_write_u8
46366 And, more use of the API-independent DEVICE_ID/SUBVENDOR_ID/SUBSYS_ID macros
46367 to pull PCI identification data from the underlying structure.
46369 commit 70e8e5957200401474967a467663ae049e9080f2
46370 Author: Keith Packard <keithp@koto.keithp.com>
46371 Date: Sun Aug 26 22:40:25 2007 -0700
46373 Change DRI interface to fill in PCI data from new libpciaccess structure.
46375 The DRI interface requires bus identification for each DRI object; pull that
46376 data from the libpciaccess structures as necessary.
46378 commit 5516cc781bd488c936af225123812a61ed5874b8
46379 Author: Keith Packard <keithp@koto.keithp.com>
46380 Date: Sun Aug 26 22:37:38 2007 -0700
46382 Add libpciaccess declarations to I810Rec and I830Rec.
46384 Using libpciaccess requires a different type for PciInfo (struct pci_device
46385 instead of pciVideoPtr) and it requires knowing which BAR each memory region
46386 needs to be mapped from. Add these definitions to the driver private record
46387 along with the includes necessary to use libpciaccess.
46389 commit daada59b5f8c2294b524a4b5920dc6b1c213642f
46390 Author: Keith Packard <keithp@koto.keithp.com>
46391 Date: Sun Aug 26 22:32:39 2007 -0700
46393 Change IS_Ixxx tests to work with or without libpciaccess.
46395 libpciaccess has a new structure that holds the PCI identifier data; borrow
46396 macros from the mga driver to work with either the old xf86-specific
46397 structure or the new libpciaccess structure.
46399 commit 3d3bf493a3973f4067433d27a4d7ddfecaa18f1c
46400 Author: Keith Packard <keithp@koto.keithp.com>
46401 Date: Sun Aug 26 22:27:34 2007 -0700
46403 Intel driver configuration (only) changes for X server libpciaccess usage.
46405 Detect whether the target X server uses libpciaccess, using it in the driver
46406 compilation as necessary. This change means that utilities that used to use
46407 libpciaccess will not do so unless the driver itself uses libpciaccess. Yes,
46408 that could be fixed, but it doesn't seem that important.
46410 This patch does not include any code changes necessary to actually have the
46411 driver build against an X server using libpciaccess.
46413 commit 5faf9cc6afe1c30fa88bc6446088a6fa47fc5d0b
46414 Author: Keith Packard <keithp@koto.keithp.com>
46415 Date: Sun Aug 26 09:34:06 2007 -0700
46417 Sort quirk table, add Dell Latitude X1
46419 commit a9e1d42a47cef79d8bbde2afd89d26aed964e344
46420 Author: Keith Packard <keithp@koto.keithp.com>
46421 Date: Sat Aug 25 12:54:11 2007 -0700
46423 Lenovo 201a is x60s, not x61s
46425 commit ffc2907f7f2eb039004eff0014c5563a01463fb0
46426 Author: Keith Packard <keithp@koto.keithp.com>
46427 Date: Sat Aug 25 12:31:21 2007 -0700
46429 Thinkpad X61s has no TV out
46431 commit 0c20fbabd18c19b2753cb60280f89e240ce5645f
46432 Author: Keith Packard <keithp@koto.keithp.com>
46433 Date: Fri Aug 17 22:21:47 2007 -0700
46435 Make sure XV_PIPE is used whenever possible.
46437 The code was not consistently using XV_PIPE when the desired crtc contained
46438 any portion of the video output.
46440 commit e443f83dd6f110156743c93f7d793cdddb8195a1
46441 Author: Eric Anholt <eric@anholt.net>
46442 Date: Fri Aug 17 18:13:49 2007 -0700
46444 Tune acceleration architecture allocator sizes down.
46446 commit bd874b11bbfe582aebd3115771f90807e75afc31
46447 Author: Eric Anholt <eric@anholt.net>
46448 Date: Fri Aug 17 17:49:21 2007 -0700
46450 Replace AA allocator usage with i830_memory.c for RandR rotation.
46452 This requires EXA 2.2 (server 1.3) for rotated performance with EXA, because
46453 the i830_memory.c allocation may not fall within what EXA considers the
46454 offscreen area, so the PixmapIsOffscreen hook is needed.
46456 commit 9ad33dd65a79277ef75a6e95373614852725f5a9
46457 Author: Eric Anholt <eric@anholt.net>
46458 Date: Fri Aug 17 16:46:48 2007 -0700
46460 Use i830_memory.c instead of the AA's allocator for XV buffers.
46462 This should fix issues with XV being allocated into XAA's tiled pixmap
46463 cache and resulting bad rendering. Its also brings us closer to being able
46464 to shrink the size of the pixmap cache on XAA, which is of limited utility.
46466 commit 3655a1ecb62f6c387a16fa87cf6f00bf7835dce4
46467 Merge: e5c336e 2231cdc
46468 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
46469 Date: Thu Aug 16 12:04:20 2007 -0700
46471 Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
46473 commit e5c336eaa32be8f9379a2c1dd51006b85bc8b270
46474 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
46475 Date: Thu Aug 16 12:04:02 2007 -0700
46477 Disambiguate plane and pipe mapping, use plane A on pipe B on pre-965 LVDS
46479 Add a new 'plane' field to the intel_crtc private structure for tracking
46480 planes separate from pipes. This allows pre-965 chips to use plane A
46481 on pipe B, enabling framebuffer compression for builtin LVDS displays.
46483 commit c3438356d6b7b63ad9d0086efb084be8e048980e
46484 Author: Eric Anholt <eric@anholt.net>
46485 Date: Thu Aug 16 11:34:02 2007 -0700
46487 Add tiling information to BO layout description.
46489 commit 0c9e4aeea84e20a18e3b76d8cf8e802af004df57
46490 Merge: 79d9a30 2231cdc
46491 Author: Eric Anholt <eric@anholt.net>
46492 Date: Thu Aug 16 11:30:16 2007 -0700
46494 Merge branch 'master' into buffer-objects
46501 commit 2231cdcd8f1ee81b3e59cc5e3a325c22ee0f40e4
46502 Author: Dave Airlie <airlied@redhat.com>
46503 Date: Thu Aug 16 17:15:54 2007 +1000
46505 i915: add support for render to a8
46507 commit a69db6f7fe1703b473e5c1d1e0088ccc203f4d5a
46508 Author: Dave Airlie <airlied@redhat.com>
46509 Date: Wed Aug 15 18:28:50 2007 +1000
46511 intel: don't setup texOffsetStart unless using EXA
46513 commit 5126a71f82767b9e23cd590453718f3364789740
46514 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46515 Date: Tue Aug 14 13:54:55 2007 +0800
46517 Fix seg fault introduced in tiling patch when TV detect
46519 When TV does load detect, fb hasn't been setup, so we should check
46520 that in i830_display_tiled(). Caught by Nanhai.
46522 commit d9f89a1af7e7ff4056727060cdf2e35c15a4dcdd
46523 Author: Keith Packard <keithp@koto.keithp.com>
46524 Date: Fri Aug 10 17:59:33 2007 -0700
46526 Save/restore tile-mode offset registers DSPATILEOFF and DSPBTILEOFF
46528 Now that the driver sets these registers, they must be saved and restored.
46530 commit 5bc194d3d3c87bb0128d9ac10f090f031345eb37
46531 Author: Keith Packard <keithp@koto.keithp.com>
46532 Date: Fri Aug 10 17:54:32 2007 -0700
46534 Set DSPATILEOFF/DSPBTILEOFF to handle 965 tiled frame buffers.
46536 DSPATILEOFF and DSPBTILEOFF replace DSPASURF and DSPBSURF when the frame
46537 buffer is in tiled mode.
46539 commit 64b943c79cf957a4c54482720195d7f27b7f0c0d
46540 Author: Eric Anholt <eric@anholt.net>
46541 Date: Fri Aug 10 15:48:05 2007 -0700
46543 Add #if 0-ed fence debugging code. It's noisy, and of little use to most.
46545 commit ba9a503ba2099025e393f3382bb453985ef23497
46546 Author: Eric Anholt <eric@anholt.net>
46547 Date: Fri Aug 10 15:06:58 2007 -0700
46549 Don't force tiling on if it is disabled in configuration but fbc is possible.
46551 commit b7751c7d1d6bcf310824295c3bab4ff36760c791
46552 Author: Eric Anholt <eric@anholt.net>
46553 Date: Fri Aug 10 17:42:09 2007 -0700
46555 Fix stack-smashing in the last commit.
46557 commit cb36635a053d4ac3971fea05060d31dbd3d382d2
46558 Author: Eric Anholt <eric@anholt.net>
46559 Date: Fri Aug 10 14:16:59 2007 -0700
46561 Attempt to fix several front buffer tiling failure cases.
46563 Front buffer tiling is now disabled with G965 and XAA. Some of the acceleration
46564 that i830_xaa.c does can't be supported on tiled buffers.
46566 Adds a tiling field to struct i830_memory, and uses it instead of separate
46567 variables for each potential tiled buffer.
46569 commit ed1b106fabf3a18489bdb3083326f27387a9cb72
46570 Author: Keith Packard <keithp@koto.keithp.com>
46571 Date: Fri Aug 10 14:31:16 2007 -0700
46573 Clean up tv mode name allocation and copy.
46575 TV mode names used to contain the signalling standard along with the pixel
46576 size. The signalling has been moved to the TV_FORMAT property, but the
46577 allocation and initialization of the mode name was left a bit messy as a
46580 commit f71b9358b4157a8cfdc694ddef8ca3f98926ca91
46581 Author: Jesse Barnes <jesse.barnes@intel.com>
46582 Date: Fri Aug 10 15:53:04 2007 -0400
46584 Cleanup tiling and FBC driver output.
46586 Remove an extra "FBC enabled" message from i830_memory.c (only report errors
46587 if they occur), and don't print the "forcing FBC on" message if tiling was
46588 already enabled, as it's redundant and confusing.
46590 commit e6746d0f286ef9d9a87f748d40e5421c268f2f7d
46591 Author: Jesse Barnes <jesse.barnes@intel.com>
46592 Date: Fri Aug 10 15:48:15 2007 -0400
46594 Enable tiling by default on 965.
46596 commit e0fcf645a228094620b8f7fdd580963611bdd6ef
46597 Author: Jesse Barnes <jesse.barnes@intel.com>
46598 Date: Fri Aug 10 15:43:06 2007 -0400
46600 Tiling fixes for 965
46602 This should be close to the last set of tiling fixes for 965 chipsets.
46603 Prior to this commit, the 965 composite hook didn't take tiling into
46604 account, nor did 965 textured video, which caused display corruption.
46605 However, there seems to be at least one last bug to squash--on occasion,
46606 a configuration with tiling enabled won't properly display text. This
46607 is likely another tiling related problem with the composite hook.
46609 commit 02ad9cee330fb5fb0008ad6f74e74e8d8a1fc36e
46610 Merge: 105f8a1 7b143e5
46611 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46612 Date: Fri Aug 10 16:45:56 2007 +0800
46614 Merge branch 'master' into xvmc-i915
46616 commit 105f8a183eb64623d9a9e2833dbe27156a551f69
46617 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46618 Date: Fri Aug 10 16:44:46 2007 +0800
46620 disable subpicture initial
46622 until we really implement it, OSD can't work for now.
46624 commit 7b143e5c8397da077c0e02455c21c5a99cf50942
46625 Author: Dave Airlie <airlied@redhat.com>
46626 Date: Thu Aug 9 12:14:44 2007 +1000
46628 i965: increase composite vertex buffer size and alignment to be safe
46630 commit 14691b24da5aa29d8c41ac7b7c61828e3cd9eab7
46631 Author: Dave Airlie <airlied@linux.ie>
46632 Date: Thu Aug 9 09:41:32 2007 +1000
46634 i965: fix memcpy of the sf_kernel when a mask is needed
46636 commit 5e18c6af9051da654d2a6a97553ef4fe777bb61e
46637 Author: Carl Worth <cworth@cworth.org>
46638 Date: Wed Aug 8 11:13:37 2007 -0700
46640 Allow 965 composite acceleration to A8 destinations.
46642 Note that this is a slowdown in text rendering due to the high overhead of our
46643 compositing setup, but appears to be correct according to rendercheck.
46645 commit b0ec670cdb0b6ca6fc0f4f165fa3ee5a20d7c985
46646 Author: Eric Anholt <eric@anholt.net>
46647 Date: Wed Aug 8 11:03:51 2007 -0700
46649 Bug #11593: Remove dead struct vch_bdb_20 which was angering the sun compiler.
46651 commit 87cc72ef509b85635bf0e84c47dfed6b7242170b
46652 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46653 Date: Wed Aug 8 09:39:55 2007 +0800
46655 check early if we have slot left for new context or surface
46657 commit fec6744b763f2b85b4e592c51326b9ee979bef5d
46658 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46659 Date: Wed Aug 8 09:26:10 2007 +0800
46661 I missed to remove port attributes wrapper
46663 commit 76c084d41999f13497b80d4ac6799cb974adaa03
46664 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46665 Date: Wed Aug 8 09:22:42 2007 +0800
46667 explicit say "unsigned int"
46669 commit cfc614b1f9143ed1407733aaa81c4a980c613a63
46670 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46671 Date: Wed Aug 8 09:06:47 2007 +0800
46673 remove xvmc attributes
46675 We don't have extra attributes than Xv port.
46677 commit 92af2f4bbcb395cbde097776718449d99843ad67
46678 Merge: da82a47 e0be352
46679 Author: Eric Anholt <eric@anholt.net>
46680 Date: Tue Aug 7 15:18:17 2007 -0700
46682 Merge branch 'origin'
46688 commit da82a47a558597f3653e2b33bc6adbab18574b57
46689 Author: Eric Anholt <eric@anholt.net>
46690 Date: Mon Aug 6 17:01:37 2007 -0700
46692 Fix EXA rendering with tiled front buffer on pre-965.
46694 The 915 and earlier appear to respect the fence registers, while only the 965
46695 requires the per-operation tiling setting and pitch shifting. This will also
46696 fix issues with rendering on the 965 involving multiple cliprects, where the
46697 pitch would get divided repeatedly.
46699 This removes the offset < 4096 fallback, which essentially resulted in no
46700 acceleration to tiled buffers, hiding the issues.
46702 commit e0be352f5017f0e645a4ff8a40961d9c2b98863a
46703 Author: Jesse Barnes <jesse.barnes@intel.com>
46704 Date: Tue Aug 7 12:37:25 2007 -0700
46706 Fixup pitch in Prepare* functions, since actual hooks may
46707 be called many times for the same pixmap, and we don't want
46708 to keep dividing the pitch by 4.
46710 commit b29a932bec2a3cc5a38e9c09c012cfdc1b6209c1
46711 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46712 Date: Tue Aug 7 16:47:13 2007 +0800
46716 and put wrap function in driver xvmc priv instead of per xv port priv
46718 commit 7431abee5fb971d1f8bc7ac4bea137f6ece9418b
46719 Author: Brice Goglin <bgoglin@debian.org>
46720 Date: Tue Aug 7 09:13:00 2007 +0200
46722 Define INTEL_VERSION_MAJOR/MINOR/PATCH using PACKAGE_VERSION_*
46724 commit 9e1914270a0978ec4dfae757d3dd57ca7ffe17e5
46725 Author: Jesse Barnes <jesse.barnes@intel.com>
46726 Date: Mon Aug 6 17:55:00 2007 -0700
46728 Remove 4k offset checks from Copy & Solid hooks.
46729 Reading the docs too literally can cause you to hide bugs with false fixes...
46731 commit 3510d5728fa972b36d022b4f9189d46ff98d7b16
46732 Author: Eric Anholt <eric@anholt.net>
46733 Date: Mon Aug 6 16:44:39 2007 -0700
46735 Fix accumulated whitespace nits in i830_exa.c
46737 commit 5ff05dffe229e35da7619762628fdd0f125585e8
46738 Author: Jesse Barnes <jesse.barnes@intel.com>
46739 Date: Mon Aug 6 16:01:10 2007 -0700
46741 More tiled rendering fixes: - check for tiling, not just offset in PrepareSolid - combine pI830->tiling and frontbuffer checks into new exaPixmapTiled function for readability
46743 commit ba90d944329dd8c79a757c38128964fbbe4ab898
46744 Author: Eric Anholt <eric@anholt.net>
46745 Date: Fri Aug 3 16:46:09 2007 -0700
46747 Add the file mode for bios_dumper output so it doesn't have 000 permissions.
46749 commit 322a163cfbda885adc6bb09c1f976d36617ea83b
46750 Author: Eric Anholt <eric@anholt.net>
46751 Date: Fri Aug 3 10:46:39 2007 -0700
46753 Quirk away the nonexistent TV connector on the Panasonic CF-Y4.
46755 commit ffbab2ee5dc227b2a8a5ffd1717ae00e8e37f956
46756 Author: Jesse Barnes <jesse.barnes@intel.com>
46757 Date: Fri Aug 3 21:27:52 2007 -0700
46759 Limit Solid & Copy offsets to 4k when rendering to tiled targets
46761 commit 019dbfda294aaafb28d8bea0fe2f5dadc2ea3e0b
46762 Merge: 3d3c0e8 15f71ed
46763 Author: Jesse Barnes <jesse.barnes@intel.com>
46764 Date: Fri Aug 3 20:45:14 2007 -0700
46766 Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
46768 commit 3d3c0e8c55f639a501c0756948b518abd903d7d0
46769 Author: Jesse Barnes <jesse.barnes@intel.com>
46770 Date: Fri Aug 3 20:40:45 2007 -0700
46772 Tiled rendering & fbc fixes:
46773 - actually enable tiling in DSP(A|B)CNTR if needed
46774 - add logic to EXA routines for tiled case (still needs work)
46775 - enable/disable fbc on DPMS events (meant moving functions higher in file)
46776 - fix fence register pitch programming (use correct pitch instead of kludged value)
46778 commit 79d9a309b19e22561e000a47b732c67479c2e6d4
46779 Author: Eric Anholt <eric@anholt.net>
46780 Date: Thu Aug 2 15:50:42 2007 -0700
46782 Mark DRI buffers as shareable, and pass their buffer handles through the SAREA.
46784 commit 0da4f2b0cd7203377ad10407928a367b8c6d310e
46785 Author: Jesse Barnes <jesse.barnes@intel.com>
46786 Date: Tue Jul 31 16:22:36 2007 -0700
46788 Legacy backlight changes:
46789 - add support for 965GM
46790 - make sure legacy enabled systems don't reduce the range of backlight values we can present to the user
46792 commit 15f71edba37738f8ba279fa07452fda10cc65298
46793 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46794 Date: Sat Jul 28 17:43:29 2007 +0800
46796 Update Lenovo TV quirk info
46798 commit 857b4a7bcb69ca43b866b4283fe075abbafb1d22
46799 Author: Eric Anholt <eric@anholt.net>
46800 Date: Fri Jul 27 18:03:12 2007 -0700
46802 Pin some buffer objects at creation time, which can't be moved yet.
46804 A number of other interfaces of ours don't allow buffer offsets to be updated
46805 after screeninit. This attempts to catalog why for each one, so that they
46806 can be fixed one by one.
46808 This happens to restore the EXA offscreen allocator for now, as a fixed-offset
46811 commit d7b5b595ad29ae0ac3adbd3a176be3fe7ffa6474
46812 Author: Eric Anholt <eric@anholt.net>
46813 Date: Fri Jul 27 12:48:08 2007 -0700
46815 Delay the first screen pixmap update to CreateScreenResources.
46817 The return value of GetScreenPixmap before CreateScreenResources is not, in
46820 commit 2b1ec0f51479c1d7e5e62803c68feca29375519d
46821 Author: Eric Anholt <eric@anholt.net>
46822 Date: Fri Jul 27 10:00:35 2007 -0700
46824 Use the dontMapFramebuffer option available with DRIINFO 5.4.
46826 If not available, AIGLX init will fail. While here, simplify DRIINFO tests
46827 since we refuse to init with a version queried less than the version we
46828 compiled against, anyway.
46830 commit f403a50afbcef1e54f554481c72037338bd5357c
46831 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
46832 Date: Fri Jul 27 09:24:24 2007 +0800
46834 Add another Lenovo TV output quirk
46836 From issue report http://lists.freedesktop.org/archives/xorg/2007-July/026644.html
46838 commit 34c82ad7ce83394db47588693b578cf91991bf1c
46839 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
46840 Date: Fri Jul 27 09:14:13 2007 +0800
46844 This one trys to use a flag for possible quirks. It adds a quirk
46845 for my Lenovo T61 TV output, and ports some origin LVDS quirks to it.
46847 commit 1a585d03972394cb97dec2462937b0104bdf4de2
46848 Author: Eric Anholt <eric@anholt.net>
46849 Date: Thu Jul 26 10:46:35 2007 -0700
46851 Use new drmBOSetPin interface instead of NO_EVICT/NO_MOVE buffers.
46853 To do this, we have to deal with buffer offsets being set at EnterVT time
46854 instead of screen init time. We've wanted to move this direction for a long
46855 time, but there are repercussions. The EXA offscreen memory manager has to
46856 be disabled, because it can't be moved. That will be replaced by BO-backed
46857 pixmaps soon. Also unresolved is whether our moving
46858 front/back/depth/texture buffers will break the classic-mode DRI driver.
46860 This code doesn't actually work yet.
46862 commit 0fd3ba0518b3cde9ca0e4e2fc1854c00d8a43d5c
46863 Author: Brice Goglin <Brice.Goglin@ens-lyon.org>
46864 Date: Wed Jul 25 20:11:32 2007 +0200
46866 Fix typo in intel.man
46868 Reported by A. Costa" <agcosta@gis.net> in
46869 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=432061
46871 commit 18c707a8c13ac5fffdd30d55e6f5926a68b367df
46872 Author: Eric Anholt <eric@anholt.net>
46873 Date: Wed Jul 25 10:11:16 2007 -0700
46875 When TTM is available, use it instead of manual AGP allocations when possible.
46877 This is a step towards being able to expose buffer objects through the screen
46878 private to DRI clients, instead of having them have to use the fake buffer
46881 This fails in two ways. First, the kernel memory manager is not currently
46882 suitable for doing the physical allocations we need, so we still use AGP for
46883 those. Additionally, the DRI lock can't be initialized early enough for us, so
46884 these buffer object allocations fail. This will be fixed by improving the
46887 commit c4deefa80a6349db968e25bd6778cdf7a514ca7f
46888 Author: Xiang, Haihao <haihao.xiang@intel.com>
46889 Date: Mon Jul 23 10:35:32 2007 +0800
46891 only enable xvmc for 915/945/G33
46893 commit 45962eed51120ff77326c29d72cf8b6cd8a934b5
46894 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46895 Date: Mon Jul 23 09:50:17 2007 +0800
46897 Fix a typo in i915 render
46899 Fence setting is in mapstate actually. This fixes rotation in
46900 tiled fb case, thanks Keith to report this.
46902 commit a52c53b8603495ab05e07e3da5092fe122fcb6f2
46903 Author: Xiang, Haihao <haihao.xiang@intel.com>
46904 Date: Fri Jul 20 17:10:27 2007 +0800
46906 don't initialize XvMC if xv is disabled.
46908 commit b1af2c0e01c54ef1d40fd0ca1ede29a1dd7ed97b
46909 Author: Zhenyu Wang <zhenyu.z.wang@intel.com>
46910 Date: Fri Jul 20 15:18:48 2007 +0800
46912 Fix device id info for 945GME, 965GME
46914 which do have new host bridge ids
46916 commit 6b4d43dbf4bbb86d479d8fca480c8bec39e2714b
46917 Author: Xiang, Haihao <haihao.xiang@intel.com>
46918 Date: Fri Jul 20 09:59:40 2007 +0800
46920 lock display when creating context
46922 commit 42e91c7092f9cf7f5c0b98e9ae3d92a5b30821bf
46923 Author: Xiang, Haihao <haihao.xiang@intel.com>
46924 Date: Thu Jul 19 17:30:54 2007 +0800
46926 move i830_allocate_xvmc_buffer out of #if/#endif
46928 commit 36ea26f2779d951fa173da785aca100564e5ecf8
46929 Author: Xiang, Haihao <haihao.xiang@intel.com>
46930 Date: Thu Jul 19 17:21:05 2007 +0800
46934 commit 9d03873423c0968b2d6700b7d8d44c7e23b8d7f3
46935 Author: Xiang, Haihao <haihao.xiang@intel.com>
46936 Date: Thu Jul 19 10:03:19 2007 +0800
46938 block some signals within critical section to avoid
46942 commit 25658f838bd1bcd7559d7a032d127c626833fb7c
46943 Author: Xiang, Haihao <haihao.xiang@intel.com>
46944 Date: Mon Jul 16 13:03:58 2007 +0800
46946 1. The pitches for y/u/v surfaces are 1024bytes aligned.
46950 commit 8f316a4584d0676e5dae051698bde0831f018cc4
46951 Author: root <root@xhh-bl32.(none)>
46952 Date: Thu Jul 12 16:24:03 2007 +0800
46956 2. fix an error in map_state
46958 commit 4844254431695158287167e3b0fad25d9efb7a6c
46959 Author: root <root@xhh-bl32.(none)>
46960 Date: Tue Jun 26 16:32:30 2007 +0800
46962 a fix for 945G/945GM
46964 commit 8c249765ce788f8adb0325a8e3016a0eae74b13b
46965 Author: root <root@xhh-bl32.(none)>
46966 Date: Mon Jun 25 14:07:01 2007 +0800
46968 1. using batch buffer
46970 2. implement macroblock_ipicture instruction
46971 3. 16bit INTRA block
46973 commit 2b20b395541f786e4009211e4c56042257b8d114
46974 Author: Xiang, Haihao <haihao.xiang@intel.com>
46975 Date: Mon Jun 25 10:43:49 2007 +0800
46977 fix a bug related to display surface
46979 commit 6e47601ac7427390fa4cb6e228dfb566ff95db09
46980 Author: Xiang, Haihao <haihao.xiang@intel.com>
46981 Date: Mon Jun 25 10:40:08 2007 +0800
46985 commit d038ffdfffb846e22f360faac5d469c8e59795ad
46986 Author: Xiang, Haihao <haihao.xiang@intel.com>
46987 Date: Mon Jun 25 10:37:12 2007 +0800
46989 Release resource allocated for surface/subpicture
46991 commit a9bf7e28b91692c7c7a4915cdf9603cd2ef025ea
46992 Author: Xiang, Haihao <haihao.xiang@intel.com>
46993 Date: Mon Jun 25 10:33:35 2007 +0800
46997 commit ccac60bf035bb54661c13f306031def81e105a32
46998 Author: Xiang, Haihao <haihao.xiang@intel.com>
46999 Date: Mon Jun 25 10:23:28 2007 +0800
47003 commit 934da9f2b83716bb37d0719a61b8e6b179fff2cd
47004 Author: Xiang, Haihao <haihao.xiang@intel.com>
47005 Date: Mon Jun 25 10:17:08 2007 +0800
47007 enable XvMC for i915
47009 commit c7920a0e819308762fca3d6fc7ab194bd565b06a
47010 Author: Dave Airlie <airlied@linux.ie>
47011 Date: Thu Jul 19 15:09:54 2007 +1000
47013 strip out remainder of drmmm code in driver
47015 commit 37652b68880f1881b90bd22218cfe86eca7e5974
47016 Author: Dave Airlie <airlied@linux.ie>
47017 Date: Tue Jul 17 14:03:21 2007 +1000
47019 intel: oops I commited pixman local workaround - undo it
47021 commit e40f6a4923d1323702406266e90eae3218a8a44e
47022 Author: Dave Airlie <airlied@linux.ie>
47023 Date: Tue Jul 17 09:20:07 2007 +1000
47025 intel: actually 2.3.1 should be good enough
47027 commit 1e169be25b2e4ab34afd4b8ae8ae0041f6069125
47028 Author: Dave Airlie <airlied@linux.ie>
47029 Date: Tue Jul 17 09:17:31 2007 +1000
47031 intel: don't try and use TTM memory manager with old libdrm interface
47033 I probably need to release a libdrm with this interface in it now..
47035 commit ff2be3995d33f9e4b7f63b380f166b6168c9b9c6
47036 Author: Keith Packard <keithp@neko.keithp.com>
47037 Date: Fri Jul 13 12:47:18 2007 -0700
47039 Remove hard-coded CRT blanking frobbing for load detection.
47041 CRT blanking needn't be adjusted to perform load detection on 9xx chips, and
47042 the 8xx load detection path now adjusts blanking just during load detection.
47043 Adjusting the blanking interval turned out to cause many monitors to fail to
47046 commit 00f4587025a3879626623135b0a153fcdb906719
47047 Author: Keith Packard <keithp@neko.keithp.com>
47048 Date: Fri Jul 13 10:58:06 2007 -0700
47050 Ensure pipe/output active before doing load detection.
47052 If the pipe or output have been set to DPMSOff, then load detection will not
47053 work correctly. Also, share the load detection configuration code between
47054 crt and tv outputs.
47056 commit 6f18300aed1340348c6d395f326061b5315be643
47057 Author: Keith Packard <keithp@neko.keithp.com>
47058 Date: Mon Jul 9 21:29:55 2007 -0700
47060 Eliminate bogus (and harmful) blanking adjustment for load detect.
47062 Instead of always adding blanking to mode lines, use the FORCE_BORDER option
47063 on i9xx hardware where it works, and dynamically add a bit of border if
47064 necessary on i8xx hardware to make load detection work. This may cause
47065 flashing when a usable crtc is not otherwise idle when load detection is
47068 commit 04130ac6b705aa49161fb6dae83ad0bdd76e89d9
47069 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47070 Date: Wed Jul 11 11:42:56 2007 +0800
47072 Fix i915 rendering for tiled buffer
47074 Make it to check fence register for dest buffer.
47076 commit 88f8b688e2316ae4a1f7485f0010ce90de54783a
47077 Author: Eric Anholt <eric@anholt.net>
47078 Date: Mon Jul 9 12:56:13 2007 -0700
47080 Fix some physical address handling for >4GB addresses.
47082 The upper bits would have been inappropriately dropped on G33-class hardware,
47083 and on G965-class hardware in a 32-bit environment. The only use of physical
47084 addresses on these should be for FBC, though, and FBC requires addresses
47085 below 4GB. This is unresolved.
47087 commit bf831117b4659cc4f2774098dee938505f780a9b
47088 Author: Jesse Barnes <jbarnes@hobbes.virtuousgeek.org>
47089 Date: Sat Jul 7 10:15:32 2007 -0700
47092 - allow FBC and Tiling to be forced off if configured to do so
47093 - only touch FBC registers if pI830->fb_compression is true
47095 commit b426866fe1be2ad3861559beff69186379a6afad
47096 Author: Jesse Barnes <jesse.barnes@intel.com>
47097 Date: Fri Jul 6 20:48:40 2007 -0700
47099 Fix manpage to reflect default behavior.
47101 commit 377c58373daa6bef5d37ead2b6f9a769a905b6fa
47102 Author: Jesse Barnes <jesse.barnes@intel.com>
47103 Date: Fri Jul 6 20:39:19 2007 -0700
47105 Fix naming of FBC plane enable bits (mistakenly called them pipes earlier).
47107 commit 9c0388dc8d4c6495fae21af6da644b34e20173d1
47108 Author: Jesse Barnes <jesse.barnes@intel.com>
47109 Date: Fri Jul 6 20:38:41 2007 -0700
47111 Update man page with current behavior.
47113 commit cecbc71fdc9af832cef23427696f6f654f7d6104
47114 Author: Jesse Barnes <jesse.barnes@intel.com>
47115 Date: Fri Jul 6 16:36:34 2007 -0700
47117 Fix debug output in fbc enable/disable routines. Add logic to make sure fbc
47118 isn't enabled twice on two different pipes.
47120 commit 4359df9419d2d02a2f9d9adc7f5a49ecf07ddd30
47121 Author: Jesse Barnes <jesse.barnes@intel.com>
47122 Date: Fri Jul 6 16:17:45 2007 -0700
47124 Fix tiling and fb compression defaults for 965 (not yet fully supported).
47126 commit ca593a5219549df94a6d234ebbcf9e7c44723c9b
47127 Author: Jesse Barnes <jesse.barnes@intel.com>
47128 Date: Fri Jul 6 16:10:52 2007 -0700
47130 FBC and tiling changes
47131 - change framebuffer option name to "FramebufferCompression"
47132 - add new "Tiling" option (controls all tiling, not just front buffer)
47133 - add debug message to fb compression enable/disable routines
47134 - update man page with new options
47136 commit 8798ef11321ee6957919279076758d47ad956cf3
47137 Merge: 8919b22 3c552af
47138 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47139 Date: Thu Jul 5 12:21:31 2007 -0700
47141 Merge branch 'master' into fbc
47143 commit 8919b2292147add41a1c1c6e5e673257cb6c6c6e
47144 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47145 Date: Thu Jul 5 12:21:06 2007 -0700
47147 Re-add tiling kludge, but only for 965.
47149 commit 407b124af8f7bb42abe4eecc87476c4c3e555cd0
47150 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47151 Date: Thu Jul 5 11:31:34 2007 -0700
47153 Remove tiling kludge. May need more fixes for 965.
47155 commit 7a87b9d2a2eb4d281dce67586756ff5653b2805a
47156 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47157 Date: Thu Jul 5 11:23:06 2007 -0700
47159 Revert discard alpha change, requires other fixes to work.
47161 commit fecf964534f5ba6d40480cb13adc89094946a51e
47162 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47163 Date: Thu Jul 5 10:59:23 2007 -0700
47166 - properly check several FBC enablement constraints
47167 - don't use alpha discard if FBC is in use
47169 commit 60ee7b6a91b2b8c447130c60cd8b19eb68119777
47170 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47171 Date: Tue Jul 3 14:20:34 2007 -0700
47173 Fixup line length buffer padding, add kludge for front buffer tile
47176 commit 3c552af65d28fafec1d09484a8914b690b961349
47177 Author: Eric Anholt <eric@anholt.net>
47178 Date: Mon Jul 2 18:33:47 2007 -0700
47180 Update documentation and bump driver version to 2.1.0.
47182 commit f02036aedcd7866c567a6adc070eda3dad872105
47183 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47184 Date: Mon Jul 2 15:42:02 2007 -0700
47186 Framebuffer compression changes:
47187 - move FBC register definitions to i830_reg.h
47188 - add fix from Arjan for 965 depth buffer tiling
47189 - add VT switch and clear-at-server-start code for FBC registers
47191 commit 2b9961eb9ce8734565ecdb01cb11610714d7f610
47192 Author: Eric Anholt <eric@anholt.net>
47193 Date: Mon Jul 2 15:16:33 2007 -0700
47195 Fix reversed LVDS dither enabling logic on GM965.
47197 commit 3d9ee8b2991ec0da8cc21b8455ff7f00fd0335b5
47198 Author: Eric Anholt <eric@anholt.net>
47199 Date: Mon Jul 2 14:38:28 2007 -0700
47201 Bug #11365: Disable the panel fitter unless it's needed for the chosen mode.
47203 The automatic panel scaling appears to choose bad sampling on some GM965
47204 hardware for 1:1 mapping modes, and there's no real sense in having it on
47205 if we just want 1:1.
47207 commit b384c608978dcd3d2ea6c0018179673cb4735f4c
47208 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
47209 Date: Mon Jul 2 09:32:28 2007 -0700
47211 Enable framebuffer compression (use Option "FrameBufferCompression"
47212 "true" in your xorg.conf). Should save ~0.5W during typical 2D usage.
47214 commit 1e2e301348b4168aeed38b3fdc6b0e43d5678a86
47215 Author: Keith Packard <keithp@dulcimer.keithp.com>
47216 Date: Sat Jun 30 12:45:24 2007 -0700
47218 Fix load detection to use border region instead of blanking.
47220 Make sure there is some border area to use by changing how the pipe is
47221 configured, then pick a scanline in the middle of the border for load
47222 detection. This lets the load detect code use an active pipe instead of
47223 requiring an idle one.
47225 commit 11862c2e1f23b77b56d7bd8b384579b5e3ae377b
47226 Author: Alan Coopersmith <alan.coopersmith@sun.com>
47227 Date: Thu Jun 28 23:31:28 2007 -0700
47229 Add *~ to .gitignore to skip emacs & patch backup files
47231 commit 6503eb45023d0db9a94cb9d1e14a26af07a6628d
47232 Author: Alan Coopersmith <alan.coopersmith@sun.com>
47233 Date: Thu Jun 28 23:30:35 2007 -0700
47235 Add AM_PROG_CC_C_O to configure.ac
47237 Clears automake-1.10 warning: src/bios_reader/Makefile.am:8: compiling
47238 `bios_dumper.c' with per-target flags requires `AM_PROG_CC_C_O' in
47241 commit 5257e36f502676fd6a44bbb8e747d9138ed3bc5c
47242 Author: Keith Packard <keithp@neko.keithp.com>
47243 Date: Thu Jun 28 15:29:52 2007 -0700
47245 Handle dual-channel LVDS on i855.
47247 Just as with i9xx LVDS, the i855 LVDS can operate in dual-channel mode with
47248 a modified P2 divisor value (7 instead of 14). Just using the existing 9xx
47249 code for 855 appears to work fine.
47251 commit 16bfcb8042519f24b4494fd621814f39949ceeb6
47252 Author: Keith Packard <keithp@neko.keithp.com>
47253 Date: Thu Jun 28 15:27:56 2007 -0700
47255 Decode PLL registers in LVDS mode a bit better in debug code.
47257 LVDS mode changes how the PLL works in fairly dramatic ways; the debug code
47258 wasn't properly accounting for those differences resulting in fairly bogus
47261 commit 9675ccb30818bf831ac4c634751ab4bfe35f7bfe
47262 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47263 Date: Wed Jun 27 09:23:33 2007 +0800
47265 EXA: fallback mask transform on i965
47267 It needs to fix shader programs which hasn't been done yet.
47269 commit 7a2300c88ae59f5b7c3ce89d33147e3f0ca23c18
47270 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47271 Date: Wed Jun 27 09:19:22 2007 +0800
47273 EXA: don't have to check offscreen size
47275 DDX will check it for EXA_OFFSCREEN_PIXMAPS flag
47277 commit fff4a3b58fa18ee2ad91f998d190e90b77c051ab
47278 Author: Carl Worth <cworth@cworth.org>
47279 Date: Mon Jun 25 23:15:58 2007 -0700
47281 Use local structures for vs_state, sf_state, and wm_state
47283 commit 0a8a4afd3c59011d6b1f5b39aedfb9bce0e55c48
47284 Author: Carl Worth <cworth@cworth.org>
47285 Date: Mon Jun 25 23:12:23 2007 -0700
47287 Use local structure for src_sampler_state and mask_sampler_state
47289 commit 499166a60fcbf16021bd9ec233790ba55803aa44
47290 Author: Carl Worth <cworth@cworth.org>
47291 Date: Mon Jun 25 23:09:17 2007 -0700
47293 Use local structure for mask_surf_state
47295 commit a418ef7316808b239884a90c3fe890220bcc0242
47296 Author: Carl Worth <cworth@cworth.org>
47297 Date: Mon Jun 25 23:07:19 2007 -0700
47299 Use local structure for src_surf_state
47301 commit 0e3c0b17826b7b5a21ee2c1d789b084fc167f1ed
47302 Author: Carl Worth <cworth@cworth.org>
47303 Date: Mon Jun 25 23:05:50 2007 -0700
47305 Use local structure for dest_surf_state
47307 commit 41a2c0f15446d59678461648f476fa71d40d44e0
47308 Author: Carl Worth <cworth@cworth.org>
47309 Date: Mon Jun 25 23:03:47 2007 -0700
47311 Use local structure for cc_state
47313 commit 59f2150caca7eb374a2db43a472ba85f50d23274
47314 Author: Carl Worth <cworth@cworth.org>
47315 Date: Mon Jun 25 11:28:57 2007 -0700
47317 Remove redundant i830WaitSync from i965_prepare_composite
47319 There were two calls to i830WaitSync, and between them no state was
47320 being changed---just offsets were being computed.
47322 commit 66aa0e61e1e8d2216a9c0555be5be004ed0a3192
47323 Author: Dave Mueller <dave.mueller@gmx.ch>
47324 Date: Fri Jun 22 16:45:27 2007 -0700
47326 Bug #11171: Add support for the Ti TFP410 DVO TMDS transmitter.
47328 commit f8d7cbc6e1322acad3351591336cefcfba7d9aaf
47329 Author: Eric Anholt <eric@anholt.net>
47330 Date: Fri Jun 22 16:29:21 2007 -0700
47332 Move the ivch's fixed panel mode support to i830_dvo.c for other LVDS drivers.
47334 This also results in removal of the setup hook, which was being called
47335 unconditionally and breaking non-ivch dvo drivers.
47337 commit ec236c76b93aea5f2ee1e8b8509cde4625974fcb
47338 Author: Keith Packard <keithp@neko.keithp.com>
47339 Date: Fri Jun 22 16:32:46 2007 +0100
47341 I830 needs to have plane/pipe/pll started in mode_set.
47343 The patch for the i855 to stop enabling plane/pipe/pll in mode_set broke the
47344 i830. Revert that just for the i830, leaving it enabled for the i855.
47346 commit d957c6b8e1dde8e11c1db3431e0ff58c5d984880
47347 Author: Keith Packard <keithp@neko.keithp.com>
47348 Date: Fri Jun 22 01:32:02 2007 +0100
47350 Increase vblank wait timeout from 20ms to 30ms. 49.6Hz < 20ms.
47352 The x40 LVDS mode has a 49.6Hz vertical refresh. Waiting for only 20ms can
47353 sometimes cause the driver to start programming the hardware before the
47354 vblank has occurred, which will lock up the i855 chipset. Extend this to
47355 30ms (the maximum timeout used by the BIOS) to ensure this doesn't happen.
47357 Detecting actual vblank occurance using the various status registers should
47358 also be possible but isn't yet working.
47360 commit a67c2965385001bcb8987265f698ff0f5809cd11
47361 Author: Keith Packard <keithp@neko.keithp.com>
47362 Date: Thu Jun 21 23:59:38 2007 +0100
47364 Follow BIOS configuration for Legacy Backlight Brightness.
47366 The backlight control in the LVDS controller can either operate in 'normal'
47367 mode or 'legacy' mode. In legacy mode, it uses the PCI config space register
47368 0xf4 which can range from 0 to 0xff. In normal mode, it reads the range and
47369 current value from the BLC_PWM_CTL register.
47371 commit d6e46f67ab3af1ad3bfa72acb0efd9fe79dbf1dc
47372 Author: Keith Packard <keithp@neko.keithp.com>
47373 Date: Thu Jun 21 20:16:36 2007 +0100
47375 Eliminate some uninitialized variable warnings
47377 commit 9d104634cf03bea82d1467f01e577cb8d2e4b554
47378 Author: Keith Packard <keithp@neko.keithp.com>
47379 Date: Thu Jun 21 01:15:39 2007 +0100
47381 Add 3DSTATE_CLEAR_PARAMETERS bits
47383 commit 3bbf313ba541526a893915f8b6c64b1eccf325e0
47384 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47385 Date: Tue Jun 19 09:33:50 2007 +0800
47387 Fix left G33 issues
47389 Be sure to check G33 chip type in:
47393 - and xaa composite
47394 Sorry for that I should have fixed them very earlier...
47396 commit acef342c870f3b5b781e48c8bf44739aa5ee8ffa
47397 Author: Eric Anholt <eric@anholt.net>
47398 Date: Mon Jun 18 11:57:48 2007 -0700
47400 Bug #11295: Disable textured video on i915 with framebuffer width too large.
47402 commit fbbb41bc5e03478cb46ee8f64ef68b23ff3fc14b
47403 Author: Keith Packard <keithp@neko.keithp.com>
47404 Date: Sun Jun 17 14:59:24 2007 +0100
47406 Let DPMS functions enable plane/pipe/output on 8xx hardware.
47408 On 855, letting crtc_mode_set enable the plane and pipe will occasionally
47409 hang the chip. Instead, wait for crtc_enable to light things up. For 9xx,
47410 leave things alone.
47412 commit d5ca000ece145a35fd6df0dcf3fb3460bd2d64e3
47413 Author: Rémi Cardona <remi@gentoo.org>
47414 Date: Sat Jun 16 13:17:54 2007 +0100
47416 Include stdint.h to get uint64_t
47418 commit 6b2ae93506d6795f87d6993bebfcb4e6632508ee
47419 Author: Dave Airlie <airlied@linux.ie>
47420 Date: Fri Jun 15 23:30:04 2007 +1000
47422 sdvo: add support for RGB outputs on SDVO
47424 This lights up my monitor VGA-1 - it doesn't look the best though
47426 commit 671ba03befebfdd7256855858987aabc28b2e8cd
47427 Author: Eric Anholt <eric@anholt.net>
47428 Date: Wed Jun 13 16:30:26 2007 -0700
47430 Fix and enable the 915-class planar textured video path.
47432 commit 6c29e0bae5f1e7cee02b678418394abb971594eb
47433 Author: Eric Anholt <eric@anholt.net>
47434 Date: Wed Jun 13 13:40:39 2007 -0700
47436 Improve the drm_i915_flip_t check.
47438 commit 420e41e7921d3cc07c784fd17936ec8a675f3b20
47439 Author: Eric Anholt <eric@anholt.net>
47440 Date: Wed Jun 13 13:34:26 2007 -0700
47442 Revert "Replace failure-prone configure test for fresh libdrm with a simple ifndef."
47444 This reverts commit c2b130354aecffbeb2a2d23c7371461feaf5766a.
47446 Sadly, a non-working DRM_IOCTL_I915_FLIP already existed.
47448 commit 51612e5ac3ddfb2bb172c58f2dfff9631093b69c
47449 Author: Eric Anholt <eric@anholt.net>
47450 Date: Tue Jun 12 16:09:54 2007 -0700
47452 On hang, dump up to the head pointer, not just up to the tail.
47454 commit ceb6dd72443c094212b0281c42cbe92e9a29f682
47455 Author: Eric Anholt <eric@anholt.net>
47456 Date: Mon Jun 4 16:37:53 2007 -0700
47458 Fix context switching between DRI and X.
47460 Now, all 3D pipeline consumers in the driver just call
47461 IntelEmitInvariantState(), which handles basic state setup, the caching of that
47462 state setup, and notifying DRI clients. This also removes a mistaken idle
47463 wait in the Render code which was papering over the brokenness in the context
47466 commit c2b130354aecffbeb2a2d23c7371461feaf5766a
47467 Author: Eric Anholt <eric@anholt.net>
47468 Date: Tue Jun 12 08:49:21 2007 -0700
47470 Replace failure-prone configure test for fresh libdrm with a simple ifndef.
47472 commit 0e1deb607f94e4aa3ec4b9df8ff7a07a1c95e31d
47473 Author: Eric Anholt <eric@anholt.net>
47474 Date: Mon Jun 11 17:33:33 2007 -0700
47476 Fix a typo in _3DSTATE_DEPTH_SUBRECT_DISABLE definition.
47478 This is already fixed in the definition in the 3d driver.
47480 commit 8d7a0ccd4f674659eb781def2cfdc3a6e5a219ce
47481 Author: Eric Anholt <eric@anholt.net>
47482 Date: Mon Jun 11 17:33:07 2007 -0700
47484 Clean up some nits in i915_video.c setup.
47486 - The screen dimensions were used for the clipping despite drawing being done
47487 to any pixmap, not necessarily the screen.
47488 - One piece of state setup was not documented anywhere, and isn't used in other
47489 3d hardware paths that also work.
47490 - A 3DSTATE_MODES_1 command (830-class only) was issued even though it no
47493 commit 15caa64a497dcc0eacb0f91166d9b70206a8db35
47494 Author: Keith Packard <keithp@neko.keithp.com>
47495 Date: Fri Jun 8 18:44:28 2007 -0700
47497 Add description for how to use the frame and pixel counter registers.
47499 The 24-bit frame and pixel counters were not described in detail and
47500 will be useful for DRM.
47502 commit 404fd47573f855b0442d49a383542fc093825ad0
47503 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47504 Date: Wed Jun 6 11:01:48 2007 +0800
47506 Enable overlay on G33 class chipsets
47508 Which have to use gfx vm offset fot setup overlay regs.
47510 commit f4c05973d391bdb0a9b0eadb155548310baa98fd
47511 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47512 Date: Tue Jun 5 11:07:41 2007 -0700
47514 Add support for the G33, Q33, and Q35 chipsets.
47516 These chipsets require that the hardware status page be referenced by an offset
47517 in the GTT rather than a physical memory address, so the X Server allocates it
47518 rather than the DRM.
47520 commit 36fcaeb2ef94db5399071540bba106dec3db81d5
47521 Author: Eric Anholt <eric@anholt.net>
47522 Date: Mon Jun 4 15:58:31 2007 -0700
47524 Fix misplaced merge of 1280x768 panel fixup.
47526 commit 8a19e7d57bc23dd163b45e0ab7deca4f074c934d
47527 Author: Keith Packard <keithp@dulcimer.keithp.com>
47528 Date: Tue Jun 5 00:09:57 2007 -0700
47530 Always ensure the pipe A is lit when activating overlay on pipe B.
47532 Ok, so moving video from pipe A to pipe B still requires that pipe A be
47533 active during the transition. Instead of trying to be fancy, just ensure
47534 that pipe A is running on each transition to pipe B.
47536 commit e986f6cb62f8644c5fa835bd7dfb7b014c2677c5
47537 Author: Keith Packard <keithp@dulcimer.keithp.com>
47538 Date: Mon Jun 4 23:52:23 2007 -0700
47540 Automatically switch overlay when crtcs are reconfigured.
47542 As crtcs are disabled and enabled, make sure the automatic crtc selection
47543 mechanism drives overlay configuration at each request to display an image.
47545 commit 0984c1fc0963f1ebab31f5b8fce5ad4c387fbd2c
47546 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47547 Date: Thu May 31 10:17:26 2007 +0800
47549 Add pci ids for 945GME
47551 commit 88ee25ebad78e54d243d728b775a69365359b5fb
47552 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47553 Date: Thu May 31 10:13:30 2007 +0800
47555 Add pci ids for 965GME/GLE chip
47557 commit 92e4deb50e049cc83cbde4995ba0b901feceb15a
47558 Author: Keith Packard <keithp@neko.keithp.com>
47559 Date: Wed May 30 11:49:07 2007 -0600
47561 Extend XV_PIPE range to include new -1 value
47563 commit 888a4a5f469bf955e3ee3e184b628808ae8a4498
47564 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47565 Date: Wed May 30 14:42:35 2007 +0800
47567 Fix i965 render's draw clip rectangle
47569 Use scrn's virtual size is not correct in rotation rendering.
47570 This fixes initial rotation problem on i965.
47572 commit 516fb73ffee0aea7cf892e6703d37f8ecf52b812
47573 Author: Eric Anholt <eric@anholt.net>
47574 Date: Tue May 29 10:22:25 2007 -0700
47576 Remove README statement that 830M panels are unsupported.
47578 commit 72462568da589054828b72ace83232a71636ee73
47579 Author: Eric Anholt <eric@anholt.net>
47580 Date: Tue May 29 10:21:12 2007 -0700
47582 Expand manpage description of outputs supported.
47584 commit 906b974bfeeed18d79c244ad3db4f5d30e13e4c8
47585 Author: Eric Anholt <eric@anholt.net>
47586 Date: Tue May 29 09:49:08 2007 -0700
47588 Add a fixup to LVDS panel mode detection for 1280x768 panel from text mode.
47590 Apparently some BIOSes will program a small mode with large blanking instead of
47591 using the pannel fitter.
47593 commit 4b2781291844b61b397e257a0fdb43e964e5f603
47594 Author: Keith Packard <keithp@neko.keithp.com>
47595 Date: Sat May 26 10:09:11 2007 -0700
47597 Mark IVCH as connected when detected
47599 commit 2a365eab0178c28782fba97bdd22365f30ce8963
47600 Author: Keith Packard <keithp@dulcimer.keithp.com>
47601 Date: Sun May 27 12:35:55 2007 -0700
47603 On i830, Pipe B cannot be lit the first time unless Pipe A is running.
47605 I don't understand it, but just like the video overlay, if Pipe A is not
47606 running, Pipe B will not turn the first time it is activated. This
47607 patch restructures the code used for the video overlay to share it
47608 with the crtc commit function.
47610 commit ff8c8cb869a3c780dbd826f7c94f06e4f3fda6af
47611 Author: Keith Packard <keithp@dulcimer.keithp.com>
47612 Date: Fri May 25 20:29:59 2007 -0700
47614 Compute and clip to crtc before call to xf86XVClipVideoHelper.
47616 By clipping to the crtc ahead of time, xf86XVClipVideoHelper will
47617 correctly clip to the bounds of the crtc, eliminating the need for any
47618 custom crtc clipping.
47620 Also, replace the broken xf86XVFillKeyHelper with a private version that
47621 doesn't end up stuck with the wrong clip list when the root window changes
47624 commit 33f635d79fe891079558fd909d564f3cf424c482
47625 Author: Keith Packard <keithp@dulcimer.keithp.com>
47626 Date: Fri May 25 17:32:06 2007 -0700
47628 Ensure Pipe A is active when enabling overlay the first time.
47630 The overlay on the i830 appears to be clocked by Pipe A when being enabled.
47631 If pipe A is not running, it will freeze the overlay and blank the screen.
47632 Setting a random mode on the Pipe and turning it on fixes this problem
47635 commit ff0ac8ea63dd6e55573652c5826c482881da0d62
47636 Author: Keith Packard <keithp@dulcimer.keithp.com>
47637 Date: Fri May 25 17:26:15 2007 -0700
47639 Xv does not require directRendering
47641 commit 109e5d597b76f7414601cb39c07b133ebf1b0e61
47642 Author: Keith Packard <keithp@dulcimer.keithp.com>
47643 Date: Thu May 24 11:40:50 2007 -0700
47645 Clean up CRTC selection. Remove more overlay blocking.
47647 Create separate CRTC selection function, use ints instead of floats for
47648 coverage measurement. Remove pipe stalls waiting for overlay update.
47650 commit dfb1ec9a07f74125cb1724d41ed4342c4714e12b
47651 Author: Keith Packard <keithp@dulcimer.keithp.com>
47652 Date: Thu May 24 11:23:44 2007 -0700
47654 Eliminate blocking for video overlay.
47656 No need to block for the overlay; just use the idle buffer. This will
47657 always work because the buffer switch occurs at vblank time, so there is
47658 always plenty of time to get the next buffer contents in place before it
47659 starts getting scanned to the screen.
47661 commit d172344599585e11388e59659dc9aaa86d7a31c1
47662 Author: Keith Packard <keithp@dulcimer.keithp.com>
47663 Date: Wed May 23 20:00:16 2007 -0700
47665 Reformat i830_video.c to four-space indents.
47667 Yes, I can't stand it anymore. it's a huge patch, but
47668 git-diff -w shows no changes.
47670 commit 02935ced3fba598a01d908ae49ccc30cbcc765a8
47671 Author: Keith Packard <keithp@dulcimer.keithp.com>
47672 Date: Wed May 23 19:24:25 2007 -0700
47674 Clean up overlay management.
47676 Create separate on/continue/off functions for overlay.
47677 Manage overlayOn boolean within those functions.
47678 Eliminate redundant management code in other routines.
47680 commit 08753f9b79f3f09879a18b552d90d88dbf52d4be
47681 Author: Keith Packard <keithp@dulcimer.keithp.com>
47682 Date: Wed May 23 18:59:10 2007 -0700
47684 Use FLIP_CONTINUE with ~OVERLAY_ENABLE to turn overlay off.
47686 This makes the overlay work on i830 with the modesetting driver. I don't
47687 know why the pre-modesetting driver worked without this, but it did.
47688 A more 'correct' fix would be welcome, but this does seem to do the trick.
47690 commit aa187186dc4f2d770a642060fe54f547ea8952b3
47691 Author: Keith Packard <keithp@dulcimer.keithp.com>
47692 Date: Mon May 21 10:49:08 2007 -0700
47694 Automatically select crtc based on coverage.
47696 By default, select crtc based on which one covers more of the video output.
47697 pipe property can be used to override selection when both have partial
47700 commit f5017a06a271bba0ace3c5415b78e78bc0c96f22
47701 Author: Keith Packard <keithp@dulcimer.keithp.com>
47702 Date: Sun May 20 17:25:33 2007 -0700
47704 Use CRTCs instead of pipe indices for video pipe selection.
47706 Replace pipe indices with crtc indices and store references to the xf86Crtc
47707 objects in the video private structure.
47709 commit 2df87256df755e972eb884bc742832038a020b2c
47710 Author: Keith Packard <keithp@dulcimer.keithp.com>
47711 Date: Sun May 20 17:23:26 2007 -0700
47713 Dump pending ring on crash.
47715 When the hardware locks up, dump the pending commands in the ring for
47718 commit 9971fac87622c93503540196e1756fded3d4869e
47719 Author: Michel Dänzer <michel@tungstengraphics.com>
47720 Date: Tue May 22 10:46:39 2007 +0200
47722 i830: Provide new DRI texOffsetStart hook when available with EXA.
47724 commit 5390a2e2611950d3f48cc735df4a0c37bc5377a5
47725 Author: Michel Dänzer <michel@tungstengraphics.com>
47726 Date: Fri May 18 17:51:36 2007 +0200
47728 Update vblank pipes when a pipe gets disabled.
47730 commit 4120a20626998272424225261f2cf7960b7ec0ca
47731 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47732 Date: Fri May 18 10:10:34 2007 +0800
47734 EXA: add render enter helper function
47736 That notify mesa rendering is smashing the state, and check last 3d
47737 operation to do sync after we're swapped in or others.
47739 commit 8db28aeaa6e908017b40bd9180f144a2972f6278
47740 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47741 Date: Fri May 18 09:54:34 2007 +0800
47743 Fix ring debug code
47745 Use proper unsigned type for timer variables, and try to dump 965G state.
47747 commit 16e50a91dd8b3676e8ce06052c549ab27e6843b7
47748 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47749 Date: Fri May 18 09:53:36 2007 +0800
47751 EXA: remove a flush cmd in i915 render code
47753 commit 12a9fcfe1b25cee850380d8ce11ef11cde9aaacb
47754 Merge: b930bb9 e89d5f2
47755 Author: Keith Packard <keithp@neko.keithp.com>
47756 Date: Thu May 17 15:11:29 2007 -0700
47758 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
47760 commit e89d5f275442915cc7777e75d3fcf7e7ed0f2084
47761 Author: Keith Packard <keithp@dulcimer.keithp.com>
47762 Date: Thu May 17 15:00:12 2007 -0700
47764 Make each output control clones/crtcs. Split DVO into LVDS, TMDS, TV.
47766 Move clone/crtc config into each output where it's easier to understand (no
47767 need for a switch statement in I830PrepareOutputs. Also, split DVO into
47768 three sub-types (TMDS, LVDS, TVOUT) as those have different cloning
47771 commit a441954630c6cdabbf463bfc3404160f97a04b4f
47772 Author: Keith Packard <keithp@dulcimer.keithp.com>
47773 Date: Thu May 17 14:11:49 2007 -0700
47775 Enable panel fitter on ivch DVO.
47777 Using BIOS source code as a guide, set up the panel fitter on the ivch. This
47778 involves setting the pipe to the panel fixed mode, the DVO to the source
47779 size and assigning vertical and horizontal scaling factors in the ivch
47782 commit c0daa0a982e7074af4b50653b4a45b0a6352b43d
47783 Author: Keith Packard <keithp@dulcimer.keithp.com>
47784 Date: Wed May 16 14:02:00 2007 -0700
47786 Change DVO module interface to pass more state across. Fix IVCH display.
47788 The DVO module interface reflected most of the xf86Output API to the
47789 underlying functions; finish that work given the changes that have since
47790 occurred in the xf86Output API.
47792 Move the LVDS-specific code into the IVCH module and make that work on the
47793 Thinkpad X30 (an i830-based laptop). Panel scaling does not work yet.
47795 commit b28817a87a1608e849e4a9a736dda43533a84b0c
47796 Author: Keith Packard <keithp@dulcimer.keithp.com>
47797 Date: Wed May 16 13:59:36 2007 -0700
47799 Add i830_bios_get_aim_data_block to read AIM data from BIOS
47801 Add-in modules have per-module data in the BIOS which contains configuration
47802 information which cannot be entirely discovered.
47804 commit b31bef1a8effa9acb6de7edd206b9d8c48d88144
47805 Author: Keith Packard <keithp@work.jf.intel.com>
47806 Date: Sat May 12 20:04:31 2007 -0700
47808 Deal with i830 CRT load detection which cannot use FORCE_BORDER.
47810 Chips newer than the i830 can force the border color for the active period
47811 of the screen, allowing the load detection to easily see the right data. In
47812 addition, newer chips appear to have more sensible load detection hardware
47813 which either ignores inactive periods on the screen or performs some
47814 longer-term averaging. The i830 appears to provide unfiltered samples of the
47817 For the i830, then, emit a border at the bottom of the screen and, for load
47818 detection, simply turn it purple and wait for the current line to be within
47819 the border. Sample an entire scanline, counting the number of times the load
47820 detection sees a monitor. In my testing, the presence of a monitor will
47821 cause the detection to succeed every time, while the absense will cause it
47822 to fail about 75% of the time. The code here, checks for presence at least
47823 75% of the time, which should be adequate.
47825 Also, as the new mode configuration code has already taken care to enable
47826 the CRT output, eliminate much of the load detection code which is simply
47827 duplicating functionality from the general mode setting code. This should
47828 result in faster load detection as this code will now run in no more than
47829 one frame time. It does burn the CPU the whole time though, polling the
47830 displayed scanline register.
47832 commit b930bb9d6da8c24dbe0949afb7bb2aa4bcb24687
47833 Author: Eric Anholt <eric@anholt.net>
47834 Date: Thu May 3 13:44:12 2007 -0700
47836 Disable vblank interrupts when no DRI clients are running.
47837 (cherry picked from commit 6621dd71ada839f4c1742e9e5b272e924cee21d9)
47839 commit 3b769af53e0ef6ef9b56afd679446c73a0e63ea5
47840 Author: Eric Anholt <eric@anholt.net>
47841 Date: Thu May 3 13:44:12 2007 -0700
47843 Disable vblank interrupts when no DRI clients are running.
47845 commit 775fc125aa7ecd0f054959ef210be2df4dc54345
47846 Author: Dave Airlie <airlied@airlied2.(none)>
47847 Date: Thu May 3 20:58:50 2007 +1000
47849 i810: be a bit more verbose about disabling DRI
47851 commit 34f362d099d255f8f0bb34e9de30f953ee770163
47852 Author: Eric Anholt <eric@anholt.net>
47853 Date: Wed May 2 15:40:49 2007 -0700
47855 Fix typo s/i/index/ in LoadPalette for depth 16.
47857 Reported by: Haihao Xiang <haihao.xiang@intel.com>
47859 commit f850d4727a2ad55c2116d0788f6684b2a0192d24
47860 Author: Eric Anholt <eric@anholt.net>
47861 Date: Wed May 2 14:16:21 2007 -0700
47863 Make up a fixed panel timing for DVO LVDS, and use DVOA for DVO LVDS.
47865 The fixed panel timing will only be available when the LVDS is already on
47868 So far, our only mostly-working LVDS driver is for the i830, and on i830 the
47869 LVDS is always on DVOA, so use that for all LVDS chips. This may need to
47870 change if we support the ch7017 I've seen used on embedded i845, for example.
47872 commit f3168e3b0c5664a322ca6bb1c81fc94844cb30ab
47873 Author: Eric Anholt <eric@anholt.net>
47874 Date: Wed May 2 14:08:30 2007 -0700
47876 Disable non-working GTT decoding on i830, and fix map/unmap of GTT.
47878 commit 1fc630f24f8ad9e304cb0761f9cacca2224203c4
47879 Author: Eric Anholt <eric@anholt.net>
47880 Date: Wed May 2 13:29:21 2007 -0700
47882 Add DVO[ABC] register debugging.
47884 commit d0ec37e9c0ceab1080700cd7be4a7cc58552d465
47885 Author: Eric Anholt <eric@anholt.net>
47886 Date: Tue May 1 15:56:37 2007 -0700
47888 Make the DVO output name LVDS if it's an LVDS chip.
47890 commit 490d05f99d2b62dd612d514d9ae0badbac9285ce
47891 Author: Eric Anholt <eric@anholt.net>
47892 Date: Tue May 1 15:47:01 2007 -0700
47894 Fix typo in previous commit with s/XF86_DRI/XF86DRI/
47896 commit c7bb34e83d7c459d932d01070cfeffbbf6c703ac
47897 Author: Dave Airlie <airlied@linux.ie>
47898 Date: Wed May 2 14:25:39 2007 +1000
47900 disable all outputs on EnterVT
47902 This disables all outputs on EnterVT as the SDVO output can confuse
47903 the VGA output if the BIOS has enabled it on the same pipe but X
47904 isn't going to use the SDVO.
47906 Worked out on irc with keithp
47908 commit cae0ae237b79fa7d3a82dfc8d3fb595ccb6c63e1
47909 Author: Samuel Thibault <samuel.thibault@ens-lyon.org>
47910 Date: Tue May 1 12:41:18 2007 -0700
47912 Bug #10714: Fix build without DRI.
47914 commit 6748d620fbf39dd98982856c09256bdec0fc82a1
47915 Author: Eric Anholt <eric@anholt.net>
47916 Date: Mon Apr 30 17:27:23 2007 -0700
47918 Ignore VideoRam now that its original purpose is obsolete.
47920 It had been necessary to allow more than a small amount of memory to be
47921 allocated, but now those old small allocations people had configured are
47922 getting in the way.
47924 commit a4f1a7872f6f959bb4bc6568face710bee3589de
47925 Author: Eric Anholt <eric@anholt.net>
47926 Date: Mon Apr 30 17:13:09 2007 -0700
47928 Allow physical-memory allocations within stolen memory.
47930 Because stolen memory happens to be a contiguous block of high system memory,
47931 we can just read the GTT entries for it to get physical addresses for our
47932 allocations there if needed. This reduces fragmentation of the aperture space,
47933 and will often reclaim up to 7 MB of memory that had been left unused since the
47934 simplified aperture manager was put in place, but without reintroducing the
47935 complexities of the old aperture manager.
47937 commit 7d0d34cfdcc67d07e7667e13a9413743853134f8
47938 Author: Eric Anholt <eric@anholt.net>
47939 Date: Mon Apr 30 10:39:54 2007 -0700
47941 Disable some clock gating functions documented to work incorrectly.
47943 commit 138ac8f36cb4e4b3776f313955372522646acbb2
47944 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47945 Date: Sun Apr 29 14:43:19 2007 +0800
47947 Alloc state mem buffer on 965G for xaa rotation
47949 965G needs state mem buffer to setup render pipeline.
47950 Thanks Barry Scrott for report this.
47952 commit 0cd524e5411e35c8483c02ecc5062625809e6fc6
47953 Author: Kristian Høgsberg <krh@redhat.com>
47954 Date: Wed Apr 25 18:09:10 2007 -0400
47956 Implement the custom I2C protocol used by the ivch DVO.
47958 The ihch DVO uses a modified I2C addressing scheme as described
47959 in section 5.2 of the data sheet. Implement this by over-riding
47960 the I2C read and write word routines.
47962 commit 880314aabe6326ed56517034940f0e10fb16e866
47963 Author: Keith Packard <keithp@guitar.keithp.com>
47964 Date: Tue Apr 24 11:37:08 2007 -0700
47966 CRTC Rotation under XAA wasn't hitting accelerated path.
47968 The server rotation code is now using the root window in IncludeInferiors
47969 mode rather than using the screen pixmap. Change the XAA Composite code
47970 to check for this case now.
47972 commit b23eae55c8cdd73e0aba1bf7ced283d402ee6470
47973 Merge: 31bf269 cebdb8b
47974 Author: Keith Packard <keithp@neko.keithp.com>
47975 Date: Thu Apr 19 20:38:18 2007 -0700
47977 Merge branch 'origin'
47979 commit 31bf269afed0a830e79cbbd9d4b1ee9843af326c
47980 Author: Keith Packard <keithp@neko.keithp.com>
47981 Date: Thu Apr 19 20:03:41 2007 -0700
47983 Update version to 2.0.0
47985 commit cebdb8bfc6170a0fb441039f4422917fd0c77e70
47986 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
47987 Date: Fri Apr 20 10:54:34 2007 +0800
47989 EXA: set enabling bits properly for i830
47991 This was found when debug exa on a 865GV, we should set
47992 pipeline state bits properly, otherwise the engine will hang.
47994 commit 60e891915af7d0f522c9c3f966599fa07779f7aa
47995 Author: Keith Packard <keithp@neko.keithp.com>
47996 Date: Thu Apr 19 14:02:23 2007 -0700
47998 Eliminate LinearAlloc option and code.
48000 With the fixes to the 2D frame buffer allocation that allows up to 65536
48001 lines of 2D frame buffer in XAA mode, the old linear allocation hacks are no
48004 commit cca389769001c657435f056e1f1c26b0f52a48bd
48005 Author: Keith Packard <keithp@neko.keithp.com>
48006 Date: Thu Apr 19 13:15:23 2007 -0700
48008 Clean up 1.2 xserver build support.
48010 Convert relative X server source path to absolute. Check for local copies
48011 of needed header files before building, rather than requiring server source.
48012 Remove extra duplicate -I elements in AM_CFLAGS in sub directories.
48014 commit 07797fee88d6be0dfb30394a419dd86f8a3c9095
48015 Author: Keith Packard <keithp@neko.keithp.com>
48016 Date: Thu Apr 19 13:01:37 2007 -0700
48018 Fix mismatching braces when XF86DRI_MM is not defined.
48020 A closing brace was left inside #ifdef XF86DRI_MM while the matching
48021 open brace was outside.
48023 commit 163c565527e8cda1f5a47c7fd63f04c80feaf3c7
48024 Author: Keith Packard <keithp@neko.keithp.com>
48025 Date: Thu Apr 19 13:00:03 2007 -0700
48027 Use I2C delay function instead of usleep.
48029 usleep isn't always available, and we have an existing delay mechanism
48032 commit 378ceea3d9ddbec7a08ac2f07f9a8cd9cf3cef36
48033 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48034 Date: Thu Apr 19 17:30:28 2007 +0800
48036 Fix mem list order and remove extra unbind call when free memory
48038 commit db4b9e18810990e8900bdf54aa3091b876ea2658
48039 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
48040 Date: Wed Apr 18 13:52:08 2007 +0100
48044 commit 3f5111940e35989d334aa99cd1b0eb26293ebf1b
48045 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
48046 Date: Wed Apr 18 13:27:59 2007 +0100
48048 Update read_response to include the try on PENDING status
48050 and remove it from get attached displays call.
48052 commit 902388fa06f85486fe8010807ab53e4926cc979a
48053 Author: Eric Anholt <eric@anholt.net>
48054 Date: Tue Apr 17 14:21:25 2007 -0700
48056 Don't try to init the XAA linear region unless we allocated memory for it.
48058 Reported by JM Ibanez
48060 commit 8abecae202b609375b6754dbd5ecce3d59036daf
48061 Author: Eric Anholt <eric@anholt.net>
48062 Date: Tue Apr 17 12:28:43 2007 -0700
48064 Fix i852 EXA Composite acceleration setup.
48066 Reported by JM Ibanez.
48068 commit 2dbe8d678b02b724c4f06255383f49bb4c2708b0
48069 Author: Timo Aaltonen <tjaalton@cc.hut.fi>
48070 Date: Mon Apr 16 14:14:19 2007 -0700
48072 Fix build against xserver 1.2.
48074 commit a089ac11beb4c801928c17780401e913bc0d5257
48075 Author: Eric Anholt <eric@anholt.net>
48076 Date: Tue Apr 17 11:01:05 2007 -0700
48078 Add all the possible ivch slave addresses (still commented out).
48080 commit 37ee68a95ca8c86ebe9abafaaf55b060dd2a2f73
48081 Author: Eric Anholt <eric@anholt.net>
48082 Date: Tue Apr 17 10:50:02 2007 -0700
48084 Bug #10438: Fix 965 XV when sourcing from less than the full image.
48086 Bob deinterlacing in MythTV, and the zoom options in totem would result in
48087 attempting to source from outside the video instead of scaling appropriately.
48089 commit ac9181c014638dbeb334b40b4029d0ccb2b7a0fc
48090 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
48091 Date: Tue Apr 17 16:30:17 2007 +0100
48093 Check for the PENDING message when reading the attached
48094 displays. Ensures the command has completed before continuing.
48096 (probably need to check PENDING in other SDVO calls too)
48098 commit ab5bdee8a62c842ae32aaef57eb841ebcb644d2b
48099 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48100 Date: Tue Apr 17 16:23:46 2007 +0800
48102 EXA: fix i830 render
48104 Fix tex blend pipeline in case that src/mask pict has no
48105 alpha. Unmask color buffer write disable bits. These make
48106 rendercheck run fine on 855GM.
48108 commit 1a29750b8dba1371d7d0802744cdf2f3bfa83c13
48109 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48110 Date: Mon Apr 16 16:21:39 2007 +0800
48112 EXA: fix 830/845G pict format
48114 Fallback in 830/845G when pict format is a8, x8r8g8b8 or
48115 x8b8g8r8. The hw doesn't support them.
48117 commit 3a634bbd198650c1597dec4306d99928374c30f3
48118 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48119 Date: Mon Apr 16 15:14:49 2007 +0800
48121 EXA: Add i830 supported pict format XRGB8888, XBGR8888
48123 commit 3bcb9a0b4ba7f3df346b5708617a7aafcbe2490a
48124 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48125 Date: Mon Apr 16 14:27:49 2007 +0800
48127 EXA: i830 render misc fix and cleanups
48129 Try to map texture stream when setup texture map, and use
48130 correct order in load_immediate_1 cmd, which fixed crash on
48131 845GV. Also remove some flush cmds.
48133 commit 64c30cf896f8bde3ee74c92b970132ab91b418cd
48134 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48135 Date: Mon Apr 16 13:58:50 2007 +0800
48137 Update intel.man with 965GM chipset support
48139 commit b67adb6de34cede0e31f02f26cd5ec7b1adfa586
48140 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48141 Date: Fri Apr 13 22:34:52 2007 +0800
48143 Fix crash on G965 XAA with LinearAlloc option
48145 We should alloc xaa_linear mem in LinearAlloc case, otherwise
48146 we get crash when initializing xf86 fb manager.
48148 commit b5b243e4120d6a048fb6bbe8814fe3184271e9d9
48149 Author: Eric Anholt <eric@anholt.net>
48150 Date: Thu Apr 12 12:34:23 2007 -0700
48152 Add license headers to 965 programs.
48154 commit 08cd5f9b0f086e51112008d50de48556372899f9
48155 Author: Eric Anholt <eric@anholt.net>
48156 Date: Tue Apr 10 11:26:46 2007 -0700
48158 Bug #10520: i810 manpage not installed despite users expecting it to be there.
48160 Install it as an alias to intel.4x, since we're letting people load the driver
48163 commit bf9771e9711361632afe3abeeedca7ce03497005
48164 Author: Eric Anholt <eric@anholt.net>
48165 Date: Mon Apr 9 15:26:05 2007 -0700
48167 Don't use extended regexps in sed when not necessary.
48169 Some seds require a flag to enable extended regexps.
48171 commit 656fbd952542ba5ddc9b018071008ceb38b1bd19
48172 Author: Brice Goglin <brice.goglin@ens-lyon.org>
48173 Date: Mon Apr 9 13:06:46 2007 -0700
48175 Bug #10515: Fix module version number with more than one digit per field.
48177 commit aa850d3c633a510d787ee59e538e26e5c3acbf4a
48178 Author: Eric Anholt <eric@anholt.net>
48179 Date: Mon Apr 9 11:44:25 2007 -0700
48181 Turn off ALWAYS_SYNC in EXA.
48183 This slipped in as a debugging aid, and never got turned off. The driver
48184 appears to work fine without it on an i915 system, and for the non-default EXA
48185 option, we'd rather see issues found than continue running with debugging aids
48186 and hiding them behind bad performance.
48188 commit d2e75d8acde2541b85a0050d9e47182c1db7fc2c
48189 Author: Eric Anholt <eric@anholt.net>
48190 Date: Mon Apr 9 11:34:57 2007 -0700
48192 Avoid a magic number and correct a (currently harmless) ifdef in ResetState().
48194 commit 9ed446e12bd21b7c8222c6e63a03a8e2ec6b97e2
48195 Author: Eric Anholt <eric@anholt.net>
48196 Date: Tue Apr 3 11:39:48 2007 -0700
48198 It's 2007 now. Didn't I get the memo?
48200 commit d9d05b553ca1f98efc230ef13f34576122311ac1
48201 Author: Eric Anholt <eric@anholt.net>
48202 Date: Tue Apr 3 00:15:02 2007 -0700
48204 Bump version to 1.9.94 for 2.0 RC4.
48206 commit eb45315b07624cb0506d539f523728a651efb798
48207 Author: Eric Anholt <eric@anholt.net>
48208 Date: Mon Apr 2 18:20:33 2007 -0700
48210 Move README to being a plain old text file.
48212 commit 9b78208f1ef1a8ee98f2aa139956659169e234ab
48213 Author: Eric Anholt <eric@anholt.net>
48214 Date: Mon Apr 2 18:20:15 2007 -0700
48216 Move a bunch of debugging verbosity under Option "ModeDebug".
48218 commit 9ea83d440d6b9e30d4627e34168226a6b4b2a841
48219 Author: Eric Anholt <eric@anholt.net>
48220 Date: Mon Apr 2 11:39:27 2007 -0700
48222 Update README source for new release notes.
48224 commit c6d0bf9fb6033339159d98dc2b9fe44759ad9de5
48225 Author: Eric Anholt <eric@anholt.net>
48226 Date: Mon Apr 2 10:09:43 2007 -0700
48228 Remove long-stale TODO file.
48230 commit c59e5895a0ea1f219c8f63a02b429a024bd1ed66
48231 Author: Dave Airlie <airlied@linux.ie>
48232 Date: Mon Apr 2 07:18:04 2007 +1000
48234 make warning info only so ppl don't go reporting stupid bugs
48236 commit e119eaabb2ad1ffc1d83f18199f49c52ec71b0fe
48237 Author: Eric Anholt <eric@anholt.net>
48238 Date: Fri Mar 30 17:09:18 2007 -0700
48240 Warnings cleanup in ch7xxx.
48242 commit fd98e88d1a6e0e8d35bda868e7afcb78d2d11524
48243 Author: Eric Anholt <eric@anholt.net>
48244 Date: Fri Mar 30 17:07:29 2007 -0700
48246 Finish off some function renaming that was missed before.
48248 Thanks to alanc for catching this.
48250 commit 274b99e5e750228eba9a67773725a40ebaa9bf56
48251 Author: Eric Anholt <eric@anholt.net>
48252 Date: Fri Mar 30 17:06:43 2007 -0700
48254 Don't forget the warnings CFLAGS in subdirectories.
48256 commit aa6a9abb34e93780c07563ab5b21d8c064ea0a1c
48257 Merge: 670c852 107ac12
48258 Author: Keith Packard <keithp@neko.keithp.com>
48259 Date: Fri Mar 30 12:10:17 2007 -0700
48261 Merge branch 'crestline-qa'
48263 commit 670c8521f18f01b9ea899ceb68d2a3dbb884b199
48264 Author: Eric Anholt <eric@anholt.net>
48265 Date: Thu Mar 29 13:43:31 2007 -0700
48267 Add missing file for old server compatibility build.
48269 commit 89c84f939f2dbed1d03a5f0e1daae23a229e47bc
48270 Author: Eric Anholt <eric@anholt.net>
48271 Date: Thu Mar 29 13:40:52 2007 -0700
48273 Whitespace cleanup in Makefile.
48275 commit 6cd3b2c2de4596886aab85988b16cf63e8887add
48276 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48277 Date: Tue Mar 27 15:34:32 2007 +0800
48279 EXA: i830 fix blend action
48281 i830_get_blend_cntl() has already added S8 offset.
48283 commit 35b3dd881a333569ce20ddf36b915b095757a1ba
48284 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48285 Date: Tue Mar 27 15:33:43 2007 +0800
48287 EXA: fix i830 componentAlpha support
48289 Pick fix from i915 render, change tex blend pipeline for CA.
48291 commit df96d9a11d038ad64673022ce14c24223b3bd34c
48292 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48293 Date: Tue Mar 27 15:32:53 2007 +0800
48295 EXA: fix i830 texture setup
48297 Use LOAD_IMM_2 helper cmd for tex setup. Enable RepeatNormal
48298 support. Fix A8 format, i830 can support it now.
48300 commit aa78e67553ffe8ca986330f0ee49b5414c71a71f
48301 Merge: 768821b 2191634
48302 Author: Eric Anholt <eric@anholt.net>
48303 Date: Mon Mar 26 22:00:36 2007 -0700
48305 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
48307 commit 107ac12867eda6b86212159db15c640d3490f2da
48308 Merge: 6548817 2191634
48309 Author: Nian Wu <nian.wu@intel.com>
48310 Date: Tue Mar 27 12:51:45 2007 +0800
48312 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48314 commit 768821bd3832aad01e44856199e60c6b7726a83b
48315 Author: Eric Anholt <eric@anholt.net>
48316 Date: Mon Mar 26 21:27:54 2007 -0700
48318 Bump version to 1.9.93 for 2.0RC3.
48320 commit 2191634dd67b3219bb88f365bcf951d5a58140e8
48321 Author: Alan Coopersmith <alan.coopersmith@sun.com>
48322 Date: Mon Mar 26 14:58:41 2007 -0700
48324 Add Solaris to list of OS'es with agpgart in man page
48326 commit 1c809f92467f97b098bcc89d66e60f6f88d9bade
48327 Author: Eric Anholt <eric@anholt.net>
48328 Date: Mon Mar 26 13:03:28 2007 -0700
48332 commit f7befe50af4c13554d1f7aee6b05848ac312411b
48333 Author: Eric Anholt <eric@anholt.net>
48334 Date: Mon Mar 26 13:01:11 2007 -0700
48336 Power on the LVDS B-channel pairs only when we've chosen dual-channel mode.
48338 It was basing off of the clock rate, but we have an override to use the
48339 existing dual channel state when we can detect it, so the two settings were
48342 commit 827dc457a4fd0b97c95ec696ec9c3cd2f4a4becb
48343 Author: Eric Anholt <eric@anholt.net>
48344 Date: Mon Mar 26 12:45:23 2007 -0700
48346 Update manpage to remove Rotate option and point at xorg.conf instead.
48348 While here, update a few other bits as well.
48350 commit 654881794ae1ad7214e85091b9015ae0fbdc5ddc
48351 Merge: 6de3edc d572771
48352 Author: Nian Wu <nian.wu@intel.com>
48353 Date: Mon Mar 26 17:00:11 2007 +0800
48355 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48357 commit d5727717c9141be28a69b1154ccd23c23207f8f6
48358 Author: Keith Packard <keithp@guitar.keithp.com>
48359 Date: Sun Mar 25 23:44:59 2007 -0700
48361 Switch TV_FORMAT property to format XA_ATOM
48363 This allows the driver to report the set of valid formats in the property
48366 commit 6de3edcb52e6258f1af75e4f4bef73de1698445d
48367 Merge: d874aa3 1e6e675
48368 Author: Nian Wu <nian.wu@intel.com>
48369 Date: Sat Mar 24 17:00:13 2007 +0800
48371 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48373 commit 1e6e675524461ef0eb1983de89e2877426571a55
48374 Author: Keith Packard <keithp@neko.keithp.com>
48375 Date: Fri Mar 23 23:42:52 2007 -0700
48377 Eliminate calls to RRPostPendingProperty.
48379 RRPostPendingProperty has been removed in favor of RRPostPendingProperties,
48380 and that call is now managed outside of the driver.
48382 commit 28da6f4e307880326dd553f50fe3fff3b9be9f4f
48383 Author: Keith Packard <keithp@neko.keithp.com>
48384 Date: Fri Mar 23 14:36:42 2007 -0700
48386 Record 3D state loss at EnterVT in last_3d value.
48388 last_3d set to LAST_3D_OTHER indicates that the 3D hardware has unknown
48391 commit ab19439cf4592e4607dc0bfc602aba3d56645a42
48392 Author: Keith Packard <keithp@neko.keithp.com>
48393 Date: Fri Mar 23 14:34:27 2007 -0700
48395 Don't call AdjustFrame on EnterVT.
48397 AdjustFrame is strictly for legacy compatibility; calling it on EnterVT
48398 wrecks crtc positions.
48400 commit d874aa31599da4777438cc51469afe9b66601f55
48401 Merge: 94dbc37 26f32ef
48402 Author: Nian Wu <nian.wu@intel.com>
48403 Date: Fri Mar 23 17:00:12 2007 +0800
48405 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48407 commit 26f32ef680a19e63af4b7c8c84141fe32263f298
48408 Author: Keith Packard <keithp@guitar.keithp.com>
48409 Date: Fri Mar 23 01:17:58 2007 -0700
48411 Property size is in units, not bytes.
48413 Several places were using byte lengths instead of unit lengths for
48416 commit f48dc501fbf74e0ac348f0b77750016597849ef4
48417 Author: Keith Packard <keithp@guitar.keithp.com>
48418 Date: Fri Mar 23 01:08:08 2007 -0700
48420 TV output properties: TV_FORMAT, LEFT, TOP, RIGHT, BOTTOM.
48422 Remove TV format from mode name, instead use an explicit output property and
48423 split the input resolution from the tv format. Add properties to set the
48424 blank area on all four sides of the image.
48426 commit 94dbc3725358d63fe0ac8e6749489c993d24ede2
48427 Merge: 28af380 20b2685
48428 Author: Nian Wu <nian.wu@intel.com>
48429 Date: Thu Mar 22 17:00:15 2007 +0800
48431 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48433 commit 20b26854abdacb6dc45cba2d81d515b2e47e25f1
48434 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48435 Date: Thu Mar 22 10:33:15 2007 +0800
48437 EXA: don't smash vertex header in G965
48439 Originally we smashed vertex header to store texture
48440 coordinates, this is working as we only use sf/wm kernel
48441 and disable all other stages on pipeline. But better to
48442 not do this. This also cleans up vertex elements state
48443 and makes vertex buffer order looks "normal".
48445 commit 28af380ab133eb14d21dc650c77bdbab66576255
48446 Merge: 300e893 e06c5f7
48447 Author: Nian Wu <nian.wu@intel.com>
48448 Date: Wed Mar 21 17:00:04 2007 +0800
48450 Merge branch 'crestline' of git://otc-graphics.jf.intel.com/git/xorg/driver/xf86-video-intel into crestline
48452 commit e06c5f727fb2e2de111ac9e691a877c56d2ca1b4
48453 Merge: 0202399 3025fa0
48454 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48455 Date: Wed Mar 21 15:04:21 2007 +0800
48457 Merge branch 'master' into crestline
48459 commit 02023998663cc7f0735fadfb1719d93dc2e5a112
48460 Author: Eric Anholt <eric@anholt.net>
48461 Date: Wed Mar 21 00:00:56 2007 -0700
48463 Whitespace and symbolic reg names cleanup in i830_panel_fitter_pipe().
48465 commit 3e9ec78b4f54defb9986e11e6f2ac3475755849d
48466 Author: Eric Anholt <eric@anholt.net>
48467 Date: Tue Mar 20 23:58:48 2007 -0700
48469 Set the panel fitter to the right pipe on Crestline.
48471 commit 3025fa0fb2bf5ace7076796e45e2560fe8410e8d
48472 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48473 Date: Wed Mar 21 14:50:45 2007 +0800
48475 EXA: try to enable rotation for G965
48477 The new sf/wm should handle the texture sampling only in
48478 rotated case. Also fix possible hole in VUE slot.
48480 commit a50009604152bf9237c40bd098b3b1da3d018929
48481 Merge: 0a612e7 2239448
48482 Author: Eric Anholt <eric@anholt.net>
48483 Date: Tue Mar 20 23:32:19 2007 -0700
48485 Merge branch 'master' into crestline
48492 commit 300e893cec19dca48e00ee25014b8714dc13b278
48493 Merge: 96e8699 0202399
48494 Author: Nian Wu <nian.wu@intel.com>
48495 Date: Wed Mar 21 08:55:50 2007 +0800
48497 Merge branch 'crestline' of git://otc-graphics.jf.intel.com/git/xorg/driver/xf86-video-intel into crestline
48499 commit 223944878cf38f86580df5a7d3102d86cfc061b9
48500 Author: Eric Anholt <eric@anholt.net>
48501 Date: Tue Mar 20 14:33:53 2007 -0700
48503 Attempt to fix single/dual-channel issues on i9xx LVDS panels.
48505 - Use the existing single/dual-channel state when available, as changing it
48506 doesn't appear to work out.
48507 - Set the power state of the CLKB and B0-B3 pairs according to whether
48508 choose to go dual-channel or not.
48509 - Restore the LVDS register at the appropriate point (before DPLLs are
48512 commit 96e86994f3d1b4938e99a751454ee99bebfe40d3
48513 Merge: f465c23 0a612e7
48514 Author: Nian Wu <nian.wu@intel.com>
48515 Date: Tue Mar 20 13:11:09 2007 +0800
48517 Merge branch 'crestline' of git://otc-graphics.jf.intel.com/git/xorg/driver/xf86-video-intel into crestline
48519 commit 0a612e7115ff993bb8e9a00df13c0b0d20122fd6
48520 Merge: 8bb6778 4c4faf2
48521 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
48522 Date: Tue Mar 20 11:34:40 2007 +0800
48524 Merge branch 'master' of git://proxy.ims.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48530 Change LVDS output and postread like upstream. This might
48531 need to be retested on 965GM LVDS.
48533 commit 4c4faf260eb4dad1b1919c6168fa9ef477b98a39
48534 Author: Eric Anholt <eric@anholt.net>
48535 Date: Mon Mar 19 13:36:37 2007 -0700
48537 Set the CURSOR_SIZE register when present.
48539 Failure to do so gets you a lot of pretty colors.
48541 commit 64c14204453bea3f98d19861c450612e718e6c69
48542 Author: Eric Anholt <eric@anholt.net>
48543 Date: Mon Mar 19 13:35:43 2007 -0700
48545 Print the mode actually being set per pipe.
48547 commit 3ce802414a20ca8af128a00e6925a099dd90ceb4
48548 Author: Eric Anholt <eric@anholt.net>
48549 Date: Mon Mar 19 11:35:11 2007 -0700
48551 Add debug output for ADPA.
48553 commit c21b88d838fda1f00f6f6bcfe7855d32543c6f3f
48554 Author: Michel Dänzer <michel@tungstengraphics.com>
48555 Date: Mon Mar 19 10:45:58 2007 +0100
48557 Fix build when DAMAGE is not defined.
48559 commit 991719c21a6cc1b5d9b7cbe30d4b333718b3e686
48560 Author: Keith Packard <keithp@neko.keithp.com>
48561 Date: Sun Mar 18 23:05:33 2007 -0700
48563 Make i830_sdvo_write_sdvox write everything twice.
48565 For some reason, certain chips don't correctly enable the SDVO hardware when
48566 this register is written only once. We're following what the BIOS code does
48567 and writing it twice now, but with extra posting reads to boot. Yes, this is
48568 cult-and-paste, but it fixes problems found on deployed hardware.
48570 commit 9118122a232d4cf7760bcb0874fe970c25251378
48571 Author: Keith Packard <keithp@guitar.keithp.com>
48572 Date: Sat Mar 17 21:34:03 2007 -0700
48574 Allocate 4 separate buffers for HW Cursors on Linux.
48576 Linux cannot allocate a large fixed buffer for the HW cursors as needed for
48577 FreeBSD; instead, allocate four separate buffers. The code now prefers to
48578 allocate one buffer (less overhead) and falls back to separate buffers only
48581 commit 62a5399d70ac3f8579441d617f8d80c94942a32a
48582 Author: Keith Packard <keithp@guitar.keithp.com>
48583 Date: Sat Mar 17 21:32:36 2007 -0700
48585 Elide I830DRIClipNotify for older DRI versions.
48587 I830DRIClipNotify is passed to newer versions of DRI; don't include it in
48588 the server when building against older versions.
48590 commit 05e0021147a89254182c277007236448f315231c
48591 Author: Keith Packard <keithp@guitar.keithp.com>
48592 Date: Sat Mar 17 21:31:04 2007 -0700
48594 Cast ARGB cursor address to CARD32 * to eliminate warning.
48596 While we're just doing a memcpy, it's nice for the two argument types to
48599 commit d05bb5362e986c9d27bc03c7e1a939ba28824810
48600 Author: Keith Packard <keithp@neko.keithp.com>
48601 Date: Sat Mar 17 20:21:59 2007 -0700
48603 Increase DDC I2C RiseFallTime to handle older monitors
48605 Changing this value slows the entire I2C bus down, making it far more
48606 reliable on older monitors. Note the same change has been made in the core X
48607 server code; this change is included here to ensure that older X servers
48608 work reliably with this driver.
48610 commit 44d1b544cec2e75735d2e27d66a9240317b962c2
48611 Author: Keith Packard <keithp@neko.keithp.com>
48612 Date: Sat Mar 17 17:15:43 2007 -0700
48614 Build fix for pre-XF86DRI_MM environments.
48616 Misplaced brace broke builds with older DRM libraries.
48618 commit a58befe9d243bd562cb4b2f08ec5c9f754148c20
48619 Author: Keith Packard <keithp@neko.keithp.com>
48620 Date: Sat Mar 17 17:12:37 2007 -0700
48622 Remove extra (and incorrect) I2C ByteTimeout setting.
48624 Setting the value correctly and then immediately breaking it caused many I2C
48625 transactions to timeout with slow monitors. Oops.
48627 commit 9d6d9ace4bd3180a4484321c3b96a83bc4adaf84
48628 Author: Eric Anholt <eric@anholt.net>
48629 Date: Fri Mar 16 19:41:54 2007 -0700
48631 Move vendor ID check in the utils to after pci_device_probe.
48633 Even current libpciaccess seems to require this.
48635 commit a117bc9e3eb824f2056ee6416859ea3ab1fdcc30
48636 Author: Eric Anholt <eric@anholt.net>
48637 Date: Fri Mar 16 19:39:34 2007 -0700
48639 Add the (afaict) correct ch7017 I2C slave address.
48641 commit 29446cdb4f72e12c5249b0d4b79ea56d9fe19934
48642 Author: Eric Anholt <eric@anholt.net>
48643 Date: Fri Mar 16 19:18:37 2007 -0700
48645 Fix compile failure due to needed types being in inttypes.h on Linux.
48647 commit 213394fbaaf353404cbb3aaa4c20860f48ee1079
48648 Author: Eric Anholt <eric@anholt.net>
48649 Date: Thu Mar 15 19:13:28 2007 -0700
48651 Fix crashes and other failures when a cursor allocation fails.
48653 Now, we allocate one single block of memory for cursors, and either succeed or
48654 fail once, rather than trying to support partial fallback modes that generally
48655 resulted in pain due to being untested. In particular, this fixes cursors on
48656 FreeBSD, which only allowed one large physically-contiguous allocation.
48658 commit 316ee682d342556b65cbd60409201591e916aac5
48659 Author: Eric Anholt <eric@anholt.net>
48660 Date: Thu Mar 15 18:53:55 2007 -0700
48662 Don't reload the cursors if we haven't set up the screen yet.
48664 This avoids a crash during preinit if we set a mode for load detecting.
48666 commit 8b06ab50bbd79dfaf4c90a6f76116ace64b85b77
48667 Author: Eric Anholt <eric@anholt.net>
48668 Date: Thu Mar 15 10:12:14 2007 -0700
48670 Fix sparse warnings about using 0 for NULL.
48672 commit 8ae6ad93329e2842c6f2d5b20ffeb0c14d10c0de
48673 Author: Keith Packard <keithp@guitar.keithp.com>
48674 Date: Thu Mar 15 00:00:51 2007 -0700
48676 Use new driver-independent CRTC-based cursor layer.
48678 This eliminates all of the cursor rotation code and other cursor management
48679 infrastructure, leaving a fairly simple hardware layer in its place.
48681 commit 9fbef2de9edbbed9c2f6a80c4074b9b245547c45
48682 Author: Eric Anholt <eric@anholt.net>
48683 Date: Wed Mar 14 12:34:34 2007 -0700
48685 Bump to 1.9.92 for RC2.
48687 commit 555b801a75cafa082808bc9bb683e700fc97d79a
48688 Author: Michel Dänzer <michel@tungstengraphics.com>
48689 Date: Wed Mar 14 11:41:50 2007 +0100
48691 Make sure the legacy texture area is there when needed.
48693 This currently only matters when the DRM memory manager is not available and
48694 Option "Legacy3D" "off" is specified, but that hasn't always been the case and
48695 might change again in the future.
48697 commit 66fdb08c83d353fbe4e917900c54b555c869eb80
48698 Author: Eric Anholt <eric@anholt.net>
48699 Date: Tue Mar 13 17:07:10 2007 -0700
48701 Refine the i855 LVDS clock code. In particular, p2 is always 14.
48703 This gets correct clocks detected on most harware. The SSC is always assumed
48704 to be 66Mhz, which may not be true, but we'll fix that when we find example
48707 commit 44708bdd9ebfef0328302c9a964b80deb46e57c6
48708 Author: Eric Anholt <eric@anholt.net>
48709 Date: Tue Mar 13 16:55:38 2007 -0700
48711 Get SDVO DPMS working on the Mac Mini by writing SDVOB and SDVOC together.
48713 Also, add code for setting the encoder power state like the BIOS does, but this
48714 doesn't appear to work. We do much more than the BIOS does in powering things
48715 down, so perhaps that's interfering somehow.
48717 commit 5135b3a79f9c30ebce78c84c49846bba83607fed
48718 Author: Eric Anholt <eric@anholt.net>
48719 Date: Tue Mar 13 10:21:06 2007 -0700
48721 Use a POSTING_READ(reg) macro instead of using the (void)INREG(reg) pattern.
48723 commit 578da7ca705c5a58c0bd397b0831e2f95140f8ae
48724 Author: Eric Anholt <eric@anholt.net>
48725 Date: Tue Mar 13 10:11:29 2007 -0700
48727 Add PCI write posting protection to i2c putbits.
48729 commit cf33abe43bd95c9437fad8e6201a24084ff96cb8
48730 Author: Eric Anholt <eric@anholt.net>
48731 Date: Tue Mar 13 10:08:57 2007 -0700
48733 Add write posting protection for the SDVO DPMS-on path.
48735 commit 9d30f0007203157e6b82fa0ffc57324490eb2ca0
48736 Author: Eric Anholt <eric@anholt.net>
48737 Date: Tue Mar 13 10:04:06 2007 -0700
48739 Replace #if 0ed LVDS register setting with updated comment.
48741 commit 9c17c6e9c63563cad5edff837519a73fe0afe313
48742 Author: Eric Anholt <eric@anholt.net>
48743 Date: Tue Mar 13 09:55:49 2007 -0700
48745 Add write posting protections to relevant register writes in the mode-set path.
48747 commit 2824ec7ccbf44ba413a6133f735f4a548c73b3cd
48748 Author: Eric Anholt <eric@anholt.net>
48749 Date: Tue Mar 13 09:53:35 2007 -0700
48751 Fix uninitialized string use in SDVO non-TMDS case.
48753 commit f465c23750adf908c0ea874f95aad98ebd2f1015
48754 Merge: d33e8da 1ed3843
48755 Author: Nian Wu <nian.wu@intel.com>
48756 Date: Tue Mar 13 17:00:44 2007 +0800
48758 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48760 commit 1ed3843f73a0d8efa405daff3483ebe70bf6134f
48761 Author: Eric Anholt <eric@anholt.net>
48762 Date: Mon Mar 12 17:47:32 2007 -0700
48764 Make the 965 use Y-major tiling for the depth buffer, as required by the spec.
48766 An example of the failure can be seen with the reflect demo when set to
48769 Reported by: Haihao Xiang <haihao.xiang@intel.com>
48771 commit 7aa257154685bd2520649ce87a3a84e55644d02c
48772 Author: Michel Dänzer <michel@tungstengraphics.com>
48773 Date: Mon Mar 12 19:00:14 2007 +0100
48775 Fix build against released libdrm.
48777 commit 7c561956a28e90667fef140bc3cfa0edca464f15
48778 Author: Michel Dänzer <michel@tungstengraphics.com>
48779 Date: Mon Mar 12 18:55:27 2007 +0100
48781 Defer flipping pages back to normal until the end of the DRI block handler.
48783 Doing it earlier can result in the wrong page being visible, giving the
48784 appearance of a frozen X server.
48786 commit fe59ab9f562fd10118563d80eb3351a4d3b48b3f
48787 Author: Michel Dänzer <michel@tungstengraphics.com>
48788 Date: Mon Mar 12 13:03:47 2007 +0100
48790 Disable page flipping if the DRM is older than 1.9.
48792 Older versions don't support the functionality we need.
48794 commit ba55ff15df974197bebd871e28bb96d817ae41c7
48795 Author: Michel Dänzer <michel@tungstengraphics.com>
48796 Date: Mon Mar 12 13:01:00 2007 +0100
48798 Fix attempt to flip pages back to normal when the last 3D window disappears.
48800 When this succeeds, 2D rendering does not have to be synchronized to back
48801 buffers until the next 3D window appears.
48803 commit c4a23c5ef8ce56ee0fe547fbc7c6623c021f801b
48804 Author: Michel Dänzer <michel@tungstengraphics.com>
48805 Date: Sat Mar 10 16:15:33 2007 +0100
48807 Remove unused have3DWindows from pI830.
48809 commit d33e8daa6874ced978d2a1f687c48922555f2524
48810 Merge: 58aeb87 797aa6f
48811 Author: Nian Wu <nian.wu@intel.com>
48812 Date: Mon Mar 12 09:03:52 2007 +0800
48814 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48816 commit 797aa6fcb1231587bde1efb47bc8430c4c8d8110
48817 Author: Dave Airlie <airlied@linux.ie>
48818 Date: Sun Mar 11 12:58:50 2007 +1100
48820 fixup missing assert includes
48822 commit 43a80ef9094efcb49027c83f0e726f907fecfbb2
48823 Author: Dave Airlie <airlied@linux.ie>
48824 Date: Sun Mar 11 12:58:26 2007 +1100
48826 fixup brace alignment for older X.org
48828 commit 0d33fd3d03cef3a7e63d88ae441354390b37a937
48829 Author: Dave Airlie <airlied@linux.ie>
48830 Date: Sun Mar 11 12:58:02 2007 +1100
48832 add XMODES flags to the i2c drivers
48834 commit 3c2d6e07bdf8daef6486b594aef0d22460eb2585
48835 Author: Michel Dänzer <michel@tungstengraphics.com>
48836 Date: Fri Mar 9 23:49:46 2007 +0100
48838 Don't crash when the SAREA pointer is NULL.
48840 commit 7518b8959ee7598f3526365a83ea7e143a5d6a4e
48841 Author: Michel Dänzer <michel@tungstengraphics.com>
48842 Date: Fri Mar 9 19:50:03 2007 +0100
48844 Revert change accidentally included in previous commit.
48846 commit 1991a90ae90b388c914985d20d6f8c3637856e9a
48847 Author: Michel Dänzer <michel@tungstengraphics.com>
48848 Date: Fri Mar 9 19:47:13 2007 +0100
48850 Update SAREA pipe sizes in i830_crtc_dpms instead of i830PipeSetBase.
48852 This allows setting the size to 0 when a pipe gets disabled.
48854 commit aef161853033907c6916337cebe88c8f111dd542
48855 Merge: 1b54ee4 14ee919
48856 Author: Michel Dänzer <michel@tungstengraphics.com>
48857 Date: Fri Mar 9 15:14:48 2007 +0100
48859 Merge remote branch 'origin/master' into i830-pageflip
48861 commit 1b54ee493fb96f944aa81e56fa6f00a8067f87fc
48862 Merge: e972265 81722a2
48863 Author: Michel Dänzer <michel@tungstengraphics.com>
48864 Date: Fri Mar 9 15:13:50 2007 +0100
48866 Merge remote branch 'origin/master' into i830-pageflip
48872 commit 14ee9195d203192d3f613919f230b20b900ffdba
48873 Author: Eric Anholt <eric@anholt.net>
48874 Date: Wed Mar 7 13:00:03 2007 -0800
48876 Bug #10157: Fix cursor corruption on server regen.
48878 commit 58aeb87f2e335d15eba73ce9dd1982e867c52403
48879 Merge: a4e7e81 81722a2
48880 Author: Nian Wu <nian.wu@intel.com>
48881 Date: Wed Mar 7 16:02:03 2007 -0500
48883 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48885 commit a5f0522b1d34236278861fe15bac2df099f0a2c7
48886 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
48887 Date: Wed Mar 7 15:49:47 2007 +0000
48889 Add some additional checks when XAA is enabled.
48891 commit 81722a21d232fa6cfb11fbe3d984abab50e89bcc
48892 Author: Keith Packard <keithp@guitar.keithp.com>
48893 Date: Tue Mar 6 23:16:53 2007 -0800
48895 Remove usage of 'shadow' module. Use xf86CrtcScreenInit.
48897 With the new mode setting code, rotation is handled outside of the driver,
48898 so the old usage of the 'shadow' module is no longer needed. Code to
48899 initialize the crtc structures has been moved out of the driver and into the
48902 commit 04f50961e2f1610c39e7e4b45811f2a6b517cad6
48903 Author: Eric Anholt <eric@anholt.net>
48904 Date: Tue Mar 6 14:23:06 2007 -0800
48906 Bug #9898: Fix a crash with NoAccel set.
48908 commit b07dfbba5df7728232b38211c623185116dcea5c
48909 Author: Eric Anholt <eric@anholt.net>
48910 Date: Tue Mar 6 13:59:14 2007 -0800
48912 Remove leftover code that was disabling tiling after we set it up.
48914 commit 94c37f35872487c04136fb659526bffefd9c46ad
48915 Author: Eric Anholt <eric@anholt.net>
48916 Date: Tue Mar 6 13:57:04 2007 -0800
48918 Make the depth buffer X tiled instead of Y, and fix a Y tiling nit on 945.
48920 The previous code claimed to set the depth buffer up as Y tiled, but due to
48921 lack of implementation in SetFence, it ended up being X tiled. Actually
48922 setting the Y tiling flag in the new version broke the depth buffer, so just
48923 switch the depth buffer to X tiling, which appears to work fine.
48925 commit a4e7e814a24dbe30a33e6ad45baeb41d190a2724
48926 Merge: 862088b 4042b27
48927 Author: Nian Wu <nian.wu@intel.com>
48928 Date: Tue Mar 6 16:01:40 2007 -0500
48930 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48932 commit 30bb719ca0abc2599ffb89e59f297fa9a0a00c3c
48933 Author: Eric Anholt <eric@anholt.net>
48934 Date: Tue Mar 6 12:23:43 2007 -0800
48936 Continue to allocate the legacy texture pool by default.
48938 This is a partial revert of 7358642e64ab6d13bc1dc1a44703ee66d715ff61
48940 If we don't allocate it now, when the DRM version is too low there won't be any
48941 memory allocated and DRI will fail. Instead, waste the memory in the i915tex
48942 case for now, and leave fixing it right (check DRM version up front and decide
48943 which memory manager to set up) to later.
48945 commit e972265261c421268e4fb806e587378d0adec577
48946 Author: Michel Dänzer <michel@tungstengraphics.com>
48947 Date: Tue Mar 6 16:54:16 2007 +0100
48951 Not sure how I missed this before... Thanks to Todd Merrill for reporting.
48953 commit 862088ba6a91d3e7cf8d37126b1d9f4ee03a1f73
48954 Merge: 11e9158 d717d9d
48955 Author: Nian Wu <nian.wu@intel.com>
48956 Date: Tue Mar 6 07:43:16 2007 -0500
48958 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
48960 commit 1e1b45fa6ed1683cba4ae73ac98933e74c3ab9d6
48961 Author: Michel Dänzer <michel@tungstengraphics.com>
48962 Date: Tue Mar 6 10:28:41 2007 +0100
48964 Fix reduced DRI memory manager size.
48966 pI830->mmSize is in kB.
48968 commit 3c08bc7d6974a1a6cf5f9cb81898617032966c52
48969 Author: Michel Dänzer <michel@tungstengraphics.com>
48970 Date: Tue Mar 6 10:14:47 2007 +0100
48972 Fix some conditionals related to triple buffering.
48974 Guard code that dereferences pI830->third_buffer with tests for that instead of
48975 pI830->TripleBuffer. It could happen that we want to enable triple buffering
48976 but (temporarily) can't because the third buffer couldn't be allocated.
48978 commit c25cfafbe1eb380b58b5fc16e94f5cc6f422f0cd
48979 Author: Michel Dänzer <michel@tungstengraphics.com>
48980 Date: Tue Mar 6 10:00:12 2007 +0100
48982 Unify allocation of back buffers.
48984 commit e787d7b698d320a7c45df35d58c5113413561fe2
48985 Author: Michel Dänzer <michel@tungstengraphics.com>
48986 Date: Tue Mar 6 09:35:42 2007 +0100
48988 Remove warnings about potential artifacts with page flipping and mixed 2D/3D.
48990 The artifacts only seemed to occur when EXA was falling back to software for
48993 commit 6c299aea8e87e72c68c96c03594706c976601ede
48994 Merge: 3308152 4042b27
48995 Author: Michel Dänzer <michel@tungstengraphics.com>
48996 Date: Tue Mar 6 09:28:52 2007 +0100
48998 Merge remote branch 'origin/master' into i830-pageflip
49005 commit 4042b27f01fdb94e7fc0d4e991e054fff88479ea
49006 Author: Keith Packard <keithp@guitar.keithp.com>
49007 Date: Mon Mar 5 22:32:52 2007 -0800
49009 Move EnterVT mode setting code to xf86SetDesiredModes.
49011 Make the application of crtc desiredModes generic code instead of
49012 per-driver by creating xf86SetDesiredModes from the code that was in EnterVT
49015 Also, move the frame buffer clear until just before mode setting to make
49016 sure things are mapped correctly.
49018 commit 55ee46aebbf1ec1a8ce914fbd0c8894fc857db8f
49019 Merge: 5c72014 50aa094
49020 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49021 Date: Tue Mar 6 12:45:08 2007 +0800
49023 Merge branch 'master' of git://proxy.ims.intel.com:9419/git/xorg/driver/xf86-video-intel
49025 commit 5c720147e2b86ca4046b7c3812c1ca6b0fb78c9d
49026 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49027 Date: Tue Mar 6 12:44:04 2007 +0800
49029 EXA: try to always alloc exa i965 state buffer in stolen mem
49031 I think this is a safe way to work around any possible chip
49034 commit 11e91586169b4f4bf2b41c0e0e620a69670c2f07
49035 Merge: a24962a 0150270
49036 Author: Nian Wu <nian.wu@intel.com>
49037 Date: Mon Mar 5 16:00:20 2007 -0500
49039 Merge git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49041 commit a24962af9ed39fabca0152cae1265a29fe6237d3
49042 Merge: 35e9310 8bb6778
49043 Author: Nian Wu <nian.wu@intel.com>
49044 Date: Mon Mar 5 16:00:11 2007 -0500
49046 Merge branch 'crestline' of git://otc-graphics.jf.intel.com/git/xorg/driver/xf86-video-intel into crestline
49048 commit 50aa09425f54c4eeca7f8b0fae9579209b10b9c4
49049 Merge: d717d9d c2c6255
49050 Author: Keith Packard <keithp@neko.keithp.com>
49051 Date: Mon Mar 5 11:28:46 2007 -0800
49053 Merge branch 'modesetting'
49055 commit c2c62559e702e7de1fa2ef309fa647ab13564dc3
49056 Author: Keith Packard <keithp@neko.keithp.com>
49057 Date: Sat Mar 3 23:12:54 2007 -0800
49059 Move single mode setting code to X server.
49061 Code to drive the global configuration from a single mode setting operation
49062 (from RandR 1.1, XFree86-VidModeExtension or XFree86-DGA) has been included
49063 in the X server now, so remove it from this driver.
49065 commit 35e9310ef59873877422dcaf9a65b38789fc8ad6
49066 Merge: c8e6f0b b27fa2c
49067 Author: Nian Wu <nian.wu@intel.com>
49068 Date: Mon Mar 5 09:01:55 2007 -0500
49070 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49072 commit 8bb677889d3f71cde671f17a3589939acad2c3b3
49073 Author: Eric Anholt <eric@anholt.net>
49074 Date: Mon Mar 5 05:08:51 2007 -0800
49076 Fix 965GM SDVO by not setting fields in SDVO[BC] which have new meanings.
49078 commit 330815251dee808754b328cd2fd60db39f2460eb
49079 Merge: 50ba1ff 0150270
49080 Author: Michel Dänzer <michel@tungstengraphics.com>
49081 Date: Mon Mar 5 13:05:20 2007 +0100
49083 Merge remote branch 'origin/master' into i830-pageflip
49094 commit d717d9d566fe3c0866b06840114e1c1990bd7be0
49095 Author: Michel Dänzer <michel@tungstengraphics.com>
49096 Date: Mon Mar 5 12:57:21 2007 +0100
49098 Fix DRM memory manager initialization.
49100 It takes the offset and size in pages, not bytes.
49102 commit a0c83af3430b6705ab2ecae59085d1c74e890c19
49103 Merge: c0f99b4 bc20b54
49104 Author: Eric Anholt <eric@anholt.net>
49105 Date: Mon Mar 5 03:37:53 2007 -0800
49107 Merge branch 'modesetting' into crestline
49109 commit 7358642e64ab6d13bc1dc1a44703ee66d715ff61
49110 Author: Michel Dänzer <michel@tungstengraphics.com>
49111 Date: Mon Mar 5 11:53:09 2007 +0100
49113 Fix handling of new vs. old texture pools.
49115 Only one of them can be active.
49117 commit 50ba1fff886a7f51b178ac6d3a1ba79a3014b214
49118 Author: Michel Dänzer <michel@tungstengraphics.com>
49119 Date: Mon Mar 5 10:22:07 2007 +0100
49121 Be more verbose when page flipping can't be enabled for various reasons.
49123 commit 015027034e970f1e3bb6ab239f7e0119235e404f
49124 Merge: 9a51064 bc20b54
49125 Author: Eric Anholt <eric@anholt.net>
49126 Date: Sun Mar 4 21:49:00 2007 -0800
49128 Merge branch 'modesetting'
49137 commit af565872a49a1a464ee4154c27136660b184c4c1
49138 Author: Keith Packard <keithp@neko.keithp.com>
49139 Date: Sun Mar 4 21:20:33 2007 -0800
49141 Set version number to 2.0 RC1 (1.9.91).
49143 commit bc20b54c34088356a277beaebcc90bb4a7063e19
49144 Author: Keith Packard <keithp@neko.keithp.com>
49145 Date: Sun Mar 4 19:35:44 2007 -0800
49147 Support new CRTC/Output prepare/commit hooks.
49149 New hooks replace explicit invocation of DPMS functions to manage mode
49150 setting sequencing.
49152 commit b27fa2c257ccc49c6f29a20a3e672ebaaf58e7aa
49153 Author: Eric Anholt <eric@anholt.net>
49154 Date: Sun Mar 4 17:15:45 2007 -0800
49156 Fix a crash with XAA and DRI disabled after the allocation rework.
49158 commit fc7d43c2b6b98be597152fbd88024273edd931d0
49159 Author: Eric Anholt <eric@anholt.net>
49160 Date: Sun Mar 4 17:14:30 2007 -0800
49162 Use -f for ln of server source. Otherwise, changing paths didn't take effect.
49164 commit b7b6063e62927dd135e118c433f48b4d0b5fe246
49165 Author: Eric Anholt <eric@anholt.net>
49166 Date: Sun Mar 4 15:53:50 2007 -0800
49168 Add a little BIOS dumper program.
49170 This isn't actually tested yet as libpciaccess doesn't do rom access on
49173 commit 6aeb855b9d4e03d458773240a9c0c9c1edd3e403
49174 Author: Keith Packard <keithp@neko.keithp.com>
49175 Date: Sun Mar 4 17:11:40 2007 -0800
49177 Shadow pixmaps use fixed offset of 0.
49179 Shadow pixmap contents are copied from the main frame buffer and contain
49180 only the bits displayed by the crtc, not the whole frame buffer. The crtc
49181 origin is always interpreted by the shadow copying code, the frame buffer
49182 itself doesn't move.
49184 commit c8e6f0bac13e0e407f9771b1bbf126f023c1511b
49185 Merge: 0ed5a2f c0f99b4
49186 Author: Nian Wu <nian.wu@intel.com>
49187 Date: Sun Mar 4 09:00:12 2007 -0500
49189 Merge branch 'crestline' of git://otc-graphics.jf.intel.com/git/xorg/driver/xf86-video-intel into crestline
49191 commit 7d08f720a95499e8d00dcd7174f06731de420017
49192 Merge: dd24d2a 435d355
49193 Author: Keith Packard <keithp@neko.keithp.com>
49194 Date: Sat Mar 3 23:13:03 2007 -0800
49196 Merge branch 'modesetting-origin' into modesetting
49198 commit 435d35558d8135a2c1724dfd813fb4a9a4e14178
49199 Author: Keith Packard <keithp@guitar.keithp.com>
49200 Date: Sat Mar 3 22:46:31 2007 -0800
49202 Update for Aaron's xserver changes.
49204 Add xf86CrtcConfigFuncs to xf86CrtcConfigInit
49205 Add canGrow option to xf86InitialConfiguration
49207 commit c0f99b4962553e560a5cb882a5060d95db5477a4
49208 Author: Keith Packard <keithp@gamba.jf.intel.com>
49209 Date: Sat Mar 3 22:36:46 2007 -0800
49211 LVDS dither control moved from PFIT to LVDS register for Crestline
49213 The LVDS register now contains lots of new controls for dual-channel LVDS control
49214 along with dither enabling. The PFIT register has a lot fewer controls as a result.
49216 commit 0fa3d4f51b5ee0dba3882fd74b6ac4e7da708f8f
49217 Merge: 10655c4 fd52d63
49218 Author: Keith Packard <keithp@gamba.jf.intel.com>
49219 Date: Sat Mar 3 21:02:17 2007 -0800
49221 Merge branch 'modesetting' into crestline
49223 commit 0ed5a2fdcfe9674175aa9af155c20695ef41cadc
49224 Merge: 7ba80cc fd52d63
49225 Author: Nian Wu <nian.wu@intel.com>
49226 Date: Sat Mar 3 09:00:20 2007 -0500
49228 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49230 commit fd52d635603b7093c5a2b7fa9c987cf59f9be27c
49231 Author: Eric Anholt <eric@anholt.net>
49232 Date: Fri Mar 2 13:44:57 2007 -0800
49234 Add a WIP UploadToScreen implementation. This almost displays right.
49236 commit ca0fa875e8bb5cb778d4db7d8053ec0a5ae34ef4
49237 Author: Eric Anholt <eric@anholt.net>
49238 Date: Fri Mar 2 11:46:45 2007 -0800
49240 Move EXA offscreen memory back after framebuffer to avoid fallbacks.
49242 Without this, the EXA code decides that the framebuffer is not in video
49243 memory because it's not located between the start of video memory and the end
49244 of EXA allocator memory. We should either have EXA let the driver decide if a
49245 pointer is in video memory, or add a new field for "end of video memory" (not
49246 just end of EXA offscreen allocation), but this fixes things for now.
49248 commit 5a73624b9bcce0004e5b72c74667c9dd504b9c3e
49249 Author: Eric Anholt <eric@anholt.net>
49250 Date: Fri Mar 2 11:03:16 2007 -0800
49252 Fix a crash when an ARGB cursor allocation had failed.
49254 commit 7ba80cc3a5e0f469e3ee55c7537fdc952cd85911
49255 Merge: 9796cb7 d5df52b
49256 Author: Nian Wu <nian.wu@intel.com>
49257 Date: Fri Mar 2 09:01:46 2007 -0500
49259 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49261 commit d5df52be59a7dd950e73336ce5698e73480108b0
49262 Author: Eric Anholt <eric@anholt.net>
49263 Date: Thu Mar 1 08:35:13 2007 -0800
49265 Actually add the files for reg_dumper.
49267 commit 9796cb7ac4bd74d5336986bb194ff5875b028121
49268 Merge: 42deb1e 1f5d166
49269 Author: Nian Wu <nian.wu@intel.com>
49270 Date: Thu Mar 1 09:02:22 2007 -0500
49272 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49274 commit 1f5d1666c8386ca4597c6f2c1ec239f9d821da4c
49275 Author: Eric Anholt <eric@anholt.net>
49276 Date: Wed Feb 28 16:27:55 2007 -0800
49278 Add a non-installed command line tool using libpciaccess to dump registers.
49280 This reuses the i830_debug.c code, so we can run that from the console or from
49281 the BIOS-based X server to debug some remaining issues.
49283 commit cd1d4b398ec91d551bdaaa26c769e5e6a9442df1
49284 Author: Eric Anholt <eric@anholt.net>
49285 Date: Wed Feb 28 09:47:38 2007 -0800
49287 Many fixes to mode_get, mode_set, clock limits, and register dumps on i855.
49289 This should fix a number of issues with i855s, particularly with integrated
49292 commit 36cad3fcb65e3dcd88e58e301cd60adb121cb96b
49293 Author: Michel Dänzer <michel@tungstengraphics.com>
49294 Date: Wed Feb 28 17:52:42 2007 +0100
49296 Update vblank pipe setup when setting a mode.
49298 commit b5316fb2623e9630cbd58020e0a7c95bf354c587
49299 Author: Michel Dänzer <michel@tungstengraphics.com>
49300 Date: Wed Feb 28 17:50:56 2007 +0100
49302 I830DRIClipNotify: Fix initialization of crtcBox[1].[xy]2.
49304 commit 42deb1e10fc445d152b74e2b225daeaee2abb3d1
49305 Merge: 76aac38 3e8e75e
49306 Author: Nian Wu <nian.wu@intel.com>
49307 Date: Wed Feb 28 09:00:24 2007 -0500
49309 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49311 commit 76aac382659bc1dab7dca078f703b9cf48b4ad74
49312 Merge: 42fc06c 1ac83f5
49313 Author: Nian Wu <nian@tinderbox.sh.intel.com>
49314 Date: Tue Feb 27 14:42:42 2007 -0500
49316 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49318 commit 3e8e75e5d83a2fa7e9fc6e9a3fbb07dac548ea5a
49319 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49320 Date: Tue Feb 27 16:48:48 2007 +0800
49322 Revert "EXA: enlarge 965 exa state buffer size"
49324 This reverts commit 1ac83f51ee46d65237eae1b5c767e106ac9e29bc.
49325 oops, current exa state + kernel program size + scratch space
49328 commit 1ac83f51ee46d65237eae1b5c767e106ac9e29bc
49329 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49330 Date: Tue Feb 27 13:58:47 2007 +0800
49332 EXA: enlarge 965 exa state buffer size
49334 Origin size seems break all render, be sure to assign enough
49335 mem. 80k works fine for me.
49337 commit acd9f3e28fcf5e91e09d8bdca4cc58681b1fc92a
49338 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49339 Date: Tue Feb 27 13:56:31 2007 +0800
49341 Fix typo when assign overlay reg mem pointer
49343 commit 5b765065c0bdf671b60d8a2fe512f6307f753eab
49344 Author: Eric Anholt <eric@anholt.net>
49345 Date: Mon Feb 26 17:25:57 2007 -0800
49347 Consolidate shared structure allocation to the start of ScreenInit.
49349 Previously, the structures would get allocated unnecessarily once, and
49350 wouldn't get reallocated in time at server regen.
49352 commit 6ed28c5aa34ff1d8b99a8d4721ddac7b9cac41ed
49353 Author: Eric Anholt <eric@anholt.net>
49354 Date: Mon Feb 26 17:12:44 2007 -0800
49356 Fix XV issues on 965 after the allocation rework
49358 commit 36bfeb890aeb17a8ee058790dd69b4e9aba644a1
49359 Author: Eric Anholt <eric@anholt.net>
49360 Date: Mon Feb 26 16:58:33 2007 -0800
49362 Move the initial EnterVT to after we've set up acceleration.
49364 This fixes a crash with initial rotation and EXA due to EXA not being
49365 initialized in time for the shadow pixmap setup.
49367 commit 732885c08daaf17034da8f4855d0b957ec3df9d7
49368 Author: Eric Anholt <eric@anholt.net>
49369 Date: Mon Feb 26 15:04:56 2007 -0800
49371 Bug #9604: Align the sizes of allocations to page increments as well.
49373 Without this, the 965 DRI driver fell over when pitch * height wasn't
49374 page-size aligned. Since the allocator only allocates at page-aligned offsets
49375 anyway this shouldn't hurt us at all.
49377 commit 0bfaeaab2838184827236c5c0fcc17f06d9e1372
49378 Author: Eric Anholt <eric@anholt.net>
49379 Date: Mon Feb 26 11:11:56 2007 -0800
49381 Fix physical memory allocations, including description in log.
49383 commit 5732dd7221b7bb24bf21e16066f014a0c7cb6b2a
49384 Author: Eric Anholt <eric@anholt.net>
49385 Date: Mon Feb 26 09:54:04 2007 -0800
49387 Fix some texture memory allocation issues.
49389 commit aeeae1b40fc012bcc01fcd2243d4993de819b7cf
49390 Author: Eric Anholt <eric@anholt.net>
49391 Date: Mon Feb 26 09:13:17 2007 -0800
49393 Fix a crash in startup monitor detect.
49395 commit 10655c4674cdac8a231c50dd9afc5d43fe43b4bd
49396 Author: Zou Nan hai <nanhai.zou@intel.com>
49397 Date: Mon Feb 26 09:23:14 2007 +0800
49399 for merge with external tree
49401 commit 9417af8630fce0476273b693ebac12129415e0cf
49402 Author: Zou Nan hai <nanhai.zou@intel.com>
49403 Date: Mon Feb 26 08:02:50 2007 +0800
49405 fix conflict with external branch
49407 commit 42fc06c85a0c487cfa26b9d4aa396402dbc4c2fe
49408 Merge: 10655c4 3bce8bf
49409 Author: Nian Wu <nian@graphics.(none)>
49410 Date: Sun Feb 25 12:53:47 2007 -0800
49412 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
49414 commit 84a7429730ddd29072d696accb9fdd8fea40731a
49415 Author: Eric Anholt <eric@anholt.net>
49416 Date: Fri Feb 23 14:59:24 2007 -0800
49418 Always allocate logical_context, since we use it in emitting invariants.
49420 commit 7bb725dee9ef0ccdc478ed4931dc267bcb80203f
49421 Author: Eric Anholt <eric@anholt.net>
49422 Date: Fri Feb 23 14:56:58 2007 -0800
49424 Improve logging of memory allocation passes, and clean the code up a bit.
49426 commit 54198b26e87baa180208932b97f10b27d9c0fddc
49427 Author: Eric Anholt <eric@anholt.net>
49428 Date: Fri Feb 23 14:17:07 2007 -0800
49430 Move memory allocation to ScreenInit rather than PreInit.
49432 This lets us do memory allocation just once rather than having several passes
49433 (as long as things succeed), avoids trouble with zaphod mode, and will let us
49434 do better automatic sizing of allocations soon.
49436 commit 828deff6453872ef0f9fa35c08c9f95f464b1785
49437 Author: Eric Anholt <eric@anholt.net>
49438 Date: Fri Feb 23 13:16:02 2007 -0800
49440 Fix fence allocations on 965.
49442 commit d3ac440e127c8a7716062852cd3b5e7943e289e4
49443 Author: Eric Anholt <eric@anholt.net>
49444 Date: Thu Feb 22 10:41:09 2007 -0800
49446 Rework the video memory allocation.
49448 The previous allocator worked in multiple passes, with (at least) one of
49449 setting up allocations, another to attempt to adjust those for tiling, and
49450 then a pass to set up the offsets and fix them in memory.
49452 The new allocator is simpler, allocating memory immediately if possible,
49453 setting up tiling up front, and choosing offsets immediately. AGP memory
49454 is only allocated to back actual memory used, saving some memory that would
49455 have been allocated for padding previous. It will also allow dynamic freeing
49456 and reallocation of memory, which will be useful for framebuffer resizing.
49458 commit a61a6b1db610a07060d0dcca54b66a4b2b3686cc
49459 Author: Eric Anholt <eric@anholt.net>
49460 Date: Wed Feb 21 09:43:24 2007 -0800
49462 Reformat to 4-space indents, no trailing whitespace, and 80 columns.
49464 I'm in the process of rewriting this file, and wanted to separate my whitespace
49467 commit dd24d2a082b41abc466391c37a9b3d066586dc55
49468 Author: Keith Packard <keithp@neko.keithp.com>
49469 Date: Thu Feb 22 11:46:44 2007 -0800
49471 Remove copied X server headers from driver.
49473 These two headers were copied during mode setting development; as that code
49474 is now inside the server, these files are no longer needed here.
49476 commit 681966ce6548c6f47b087eb7e369674b7ddbaade
49477 Author: Keith Packard <keithp@neko.keithp.com>
49478 Date: Thu Feb 22 11:45:43 2007 -0800
49480 Disable unused functions after configuring the remaining resources.
49482 This eliminates an unsightly flash of the hardware cursor during X server
49485 commit f5ab9e0ad3a65c972861dd53be6e33e1aac13191
49486 Author: Michel Dänzer <michel@tungstengraphics.com>
49487 Date: Thu Feb 22 18:57:23 2007 +0100
49489 I830DRISwapContext: Some cleanups.
49491 Also call I830EmitFlush() when entering/leaving the server context. Doesn't
49492 seem to help for artifacts with page flipping and mixed 2D/3D unfortunately.
49494 commit 8cce74f195e5ef6ce5599ecd52f35bcfcb0d7913
49495 Author: Michel Dänzer <michel@tungstengraphics.com>
49496 Date: Thu Feb 22 18:27:59 2007 +0100
49498 Do not sync DRI windows between pages when possible.
49500 commit f346549183a407a4bff6521ec2450dacb97442e9
49501 Author: Michel Dänzer <michel@tungstengraphics.com>
49502 Date: Wed Feb 21 09:19:54 2007 +0100
49504 intel.man: Fix typo.
49506 commit 3bce8bf0e95e5842399959a5d6f6413e96f03adb
49507 Author: Keith Packard <keithp@guitar.keithp.com>
49508 Date: Tue Feb 20 23:10:17 2007 -0800
49510 Pre-965 hardware had fixed pipe 1 for panel fitter.
49512 The pipe mode setting code needs to disable the panel fitter when using the
49513 pipe for things other than LVDS output. The driver was checking for panel
49514 fitter conflicts using bits that the 965 chipset defines for selecting which
49515 pipe the panel fitter is connected to. However, on pre-965 hardware, the
49516 panel fitter works only with pipe 1 and those bits returned 0.
49518 The result was that when pipe 1 was using the panel fitter, configuring pipe
49519 0 would disable the panel fitter.
49521 The fix provided uses a model-specific test for the panel fitter pipe.
49523 commit 0bee64f4bc7581de7ab28ca438581d215e85c610
49524 Author: Michel Dänzer <michel@tungstengraphics.com>
49525 Date: Tue Feb 20 19:09:37 2007 +0100
49527 Add support for triple buffering using a third static buffer.
49529 Need to bump the DRI DDX version minor for the added SAREA fields.
49531 commit 2212baa8454abb4c7948c3f2e20e337f831d1b86
49532 Author: Michel Dänzer <michel@tungstengraphics.com>
49533 Date: Tue Feb 20 18:39:13 2007 +0100
49535 Don't mark page flipping as active when it really isn't.
49537 commit fac8f2eb49de22abd853ca75f3c985d4d94e3633
49538 Author: Michel Dänzer <michel@tungstengraphics.com>
49539 Date: Tue Feb 20 18:38:40 2007 +0100
49541 Document Option "PageFlip".
49543 commit 5a1a04649c62aa4b1c0617560b813642ce0c67b5
49544 Author: Keith Packard <keithp@guitar.keithp.com>
49545 Date: Sat Feb 17 17:01:04 2007 -0800
49547 Make distcheck fixes. Distribute all sources needed even for older servers.
49549 Always allow (but do not require) link to server sources so that needed
49550 files can be included in the generated tar files.
49552 Add remaining .g4a files and assembly output to distributed file lists.
49554 commit 0ca926e63c7fa9122a8cba1424ce4cdee1c7b215
49555 Author: Keith Packard <keithp@guitar.keithp.com>
49556 Date: Fri Feb 16 02:19:10 2007 -0800
49558 Check for presence of xf86Modes.h file.
49560 xf86Modes.h file signals the availability of the new modes API in the
49561 server; use that instead of counting on X server version numbers.
49562 Also, finish eliminating use of local copies of those header files.
49564 commit 972f1109d23c0a3f821fa573e2aa40cd17ce2795
49565 Author: Keith Packard <keithp@guitar.keithp.com>
49566 Date: Fri Feb 16 01:07:20 2007 -0800
49568 Remove common code that has moved to xserver/hw/xfree86/modes
49570 commit 20e146a09e3b4907a6800588ebfa787bfc5b0d89
49571 Author: Keith Packard <keithp@guitar.keithp.com>
49572 Date: Fri Feb 16 00:57:44 2007 -0800
49574 Change rotation pixmap creation API to permit start-time rotation.
49576 Start time rotation requires that the pixmap be created after the server has
49577 initialized the screens. Delay the pixmap creation until the first block
49578 handler invocation.
49580 commit 17e38e39a49b2a96e31cca4a0b922631c5f7879e
49581 Author: Keith Packard <keithp@guitar.keithp.com>
49582 Date: Thu Feb 15 23:09:43 2007 -0800
49584 Enable mode setting function renames from shared code.
49586 commit 8e7bc4b1f95f5d62ff18cd94f14f5e3949ef6a29
49587 Author: Keith Packard <keithp@guitar.keithp.com>
49588 Date: Thu Feb 15 11:17:53 2007 -0800
49590 Use server sources for crtc/output generic code.
49592 As sources for these have been merged to the server, use them from the
49593 server source code instead of keeping copies in this driver.
49595 commit 6e33c4c44030e05d6d07ca9274cfe4d6df2c0387
49596 Author: Eric Anholt <eric@anholt.net>
49597 Date: Thu Feb 15 09:54:43 2007 -0800
49599 Print the correct meaning of bit 30 of pipeconf for 965 in debug output.
49601 commit 4541150cdce3f2f83b8530d0863aecf28c6610a9
49602 Author: Eric Anholt <eric@anholt.net>
49603 Date: Thu Feb 15 09:51:04 2007 -0800
49605 Detect core clock frequencies, to avoid double-wide mode when possible.
49607 Additionally, don't attempt to set double-wide on the 965, where there is
49608 no such thing any more (not that we'd ever see modes high enough to trigger
49611 commit d1be0dc4e9d443dcbd218a083cfb21a0b5d634ee
49612 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49613 Date: Thu Feb 15 11:03:12 2007 +0800
49615 EXA: fix i965 default color (0,0,0,1.0) -> (0,0,0,0)
49617 (0,0,0,1.0) is actually black...
49619 commit c010d28f0a91a9831d366ea490878ef3d958021f
49620 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49621 Date: Thu Feb 15 11:00:06 2007 +0800
49623 EXA: mask componentAlpha support
49625 we're getting load of multiple sf/wm programs for i965, might
49626 need to be cleanup in future.
49628 commit ec55dd16c4c5f7612d33ae22a6249b3b1c60e2b6
49629 Author: Keith Packard <keithp@guitar.keithp.com>
49630 Date: Wed Feb 14 14:37:02 2007 -0800
49632 Global namespace cleanups, filename fixes.
49634 Ensure all xf86 symbols created here are protected with XF86NAME.
49635 Remove accidentally exported symbols from namespace.
49636 Make all to-be-DI files prefixed with i830_xf86.
49638 commit 22a271555a46267c40448fa926d45692498ef7c6
49639 Author: Keith Packard <keithp@guitar.keithp.com>
49640 Date: Wed Feb 14 14:17:14 2007 -0800
49642 Create driver-independent DGA hooks, remove i830 DGA code.
49644 As RandR needs to poke at DGA code, and we want the RandR code to be
49645 driver-independent, it seemed easier to just make the DGA code
49646 driver-independent as well.
49648 commit a7c64d5e28f608d54c94fa7a5c92bd5b471179c7
49649 Author: root <root@guitar.keithp.com>
49650 Date: Wed Feb 14 10:20:28 2007 -0800
49652 WIP DGA generic support.
49654 commit 55c7b017ed718107b9b11467030addf225728b84
49655 Author: Michel Dänzer <michel@tungstengraphics.com>
49656 Date: Wed Feb 14 16:45:22 2007 +0100
49658 Use new DRI ClipNotify hook to track page flipping per CRTC when available.
49660 commit 3bd8edb95fcbbb49e97a1a2bdd5c8eae4f83d3a8
49661 Author: Michel Dänzer <michel@tungstengraphics.com>
49662 Date: Fri Feb 2 17:41:18 2007 +0100
49664 Revive i830 page flipping support.
49666 Use the damage layer directly instead of via shadowfb.
49668 commit 1d22bad33b634aaf61976907cb2ceeec8a9ebe4d
49669 Author: Michel Dänzer <michel@tungstengraphics.com>
49670 Date: Wed Feb 14 12:57:12 2007 +0100
49672 Update SAREA pipe fields in i830PipeSetBase.
49674 commit 0df490938af10d3a748221e2fb5467b0f244f9eb
49675 Author: Michel Dänzer <michel@tungstengraphics.com>
49676 Date: Thu Dec 21 11:27:28 2006 +0100
49678 If the DRM can handle it, enable vertical blank interrupts for both pipes.
49680 commit 1623b47c4863f5362587cdac855e3e11c109b007
49681 Author: Eric Anholt <eric@anholt.net>
49682 Date: Tue Feb 13 17:37:56 2007 -0800
49684 Switch the RECTLIST vertex order to (x2, y2), (x1, y2), (x1, y1).
49686 This is the documented correct ordering, and while the previous ordering
49687 (reversed) worked on some hardware, it failed on others.
49689 Reported by: Wang Zhenyu <zhenyu.z.wang@intel.com>
49691 commit 4bd4058dfdca84d6ebab82a1dfba99adf3852c9c
49692 Author: Eric Anholt <eric@anholt.net>
49693 Date: Tue Feb 13 14:29:52 2007 -0800
49695 Bug #9941: ViewSonic VX2025wm sync polarity quirk to fix mode distortion.
49697 commit 0ec4ffcb4e8e5e707e68bef1b13a324677f19f8f
49698 Author: Eric Anholt <eric@anholt.net>
49699 Date: Tue Feb 13 13:18:15 2007 -0800
49701 Fix double free in the case of LVDS disabled through quirks.
49703 commit 699fa88d2570de6173e2d866e11b7437f4842d74
49704 Merge: 681b919 6641aec
49705 Author: Eric Anholt <eric@anholt.net>
49706 Date: Tue Feb 13 13:10:08 2007 -0800
49708 Merge branch 'modesetting-origin' into modesetting
49710 commit 6641aec0a1cbc869fba1956c556cdd204631545a
49711 Author: Eric Anholt <eric@anholt.net>
49712 Date: Tue Feb 13 10:21:12 2007 -0800
49714 Attempt to detect panel fixed mode from EDID or current programmed mode.
49716 These two sources are placed in higher priority to the BIOS data when
49717 available, since the BIOS data has proven unreliable. The BIOS data is still
49718 read, and warnings printed if it doesn't match what we probe. The BIOS data
49719 remains useful for the situation where we want to turn on LVDS but there is no
49720 EDID available and no current mode programmed (i.e. booting with VGA or TV
49723 commit 991439d4c78cf5b2a8f6bb8f5b36fffbfcc4e4fc
49724 Author: Eric Anholt <eric@anholt.net>
49725 Date: Tue Feb 13 10:07:47 2007 -0800
49727 Remove dead #if 0-ed structure.
49729 commit c3aed56d46baba057d83dc6ea12c6b4e705e54ba
49730 Author: Eric Anholt <eric@anholt.net>
49731 Date: Fri Feb 9 13:16:37 2007 -0800
49733 Fix comments on 965 pitch alignment restrictions.
49735 commit 98318dfcc99264fb7b7ef34568712812aa6300b6
49736 Author: Daniel Stone <daniel@fooishbar.org>
49737 Date: Fri Feb 9 16:20:57 2007 +0200
49739 add missing files to fix distcheck
49740 A bunch of files were missing from _SOURCES, and the build was also broken
49741 for missing files (from EXTRA_DIST) when building without gen4asm.
49743 commit fce4390c4f1a76c65423114bb479814787612cd7
49744 Author: Zou Nan hai <zhenyu.z.wang@intel.com>
49745 Date: Fri Feb 9 11:32:22 2007 +0800
49749 commit 00dadaa115951c36de9caf2a9d78909811ef34bc
49750 Author: Zou Nan hai <nanhai.zou@intel.com>
49751 Date: Fri Feb 9 11:14:50 2007 +0800
49753 set DPLL before detect
49755 commit 10712c7a781e50475ee7e8ccd5bfd2d820f863e2
49756 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49757 Date: Thu Feb 8 08:56:42 2007 +0800
49759 EXA: i965 pixmap pitch align
49761 This seems like a typo, which causes screen artifacts.
49763 commit 44eacf2323454e26b535cc5a4f0789cb0ff0e7fb
49764 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49765 Date: Wed Feb 7 17:30:51 2007 +0800
49767 EXA: fix render issue with i965
49769 Fix SF kernel with corrent coeffient work, and correct
49770 VUE storage in multi texture case.
49772 commit 920aeaf8478f706ab3b1de4f20f5d351e64a0197
49773 Author: Zou Nan hai <nanhai.zou@intel.com>
49774 Date: Wed Feb 7 13:15:52 2007 +0800
49776 Make TV code same with external tree.
49777 Fallback to old filter table.
49778 4 spaces indent reformat.
49779 Auto scaling seem to not work well on some low res interlace mode when input width > 1024, filter those modes.
49780 Fix some subcarriar number
49782 commit 6d9757e466863594300c83f3806fd4376bea0504
49783 Author: Zou Nanhai <nanhai.zou@intel.com>
49784 Date: Mon Feb 5 14:59:53 2007 +0800
49786 Fix hang when TV is not connect.
49788 commit 681b91924c364a1c00732f548539f2767929ba0e
49789 Author: Eric Anholt <eric@anholt.net>
49790 Date: Fri Feb 2 17:40:34 2007 -0800
49792 Add subpixel offsets to fix accelerated rotated rendering on i915.
49794 This fixes the rendercheck "transformed src/mask coords 2" tests. Previously,
49795 the source pixels chosen would be off by one in some cases.
49797 The particular values were taken from Mesa, which uses .125 offsets (except
49798 apparently broken for y), but the signs are changed. I would be happier if
49799 I had better justification for why this worked.
49801 commit 76a4f2eea826036e2ac16b7e8c61ab946038c120
49802 Author: Eric Anholt <eric@anholt.net>
49803 Date: Fri Feb 2 17:34:41 2007 -0800
49805 Fix bad offsets in "right" rotation of the CRTC in cursor and update handling.
49807 commit 785a59ead0e8d1d681b2cb6827ee58ad2c51f8c6
49808 Author: Keith Packard <keithp@neko.keithp.com>
49809 Date: Thu Feb 1 22:48:44 2007 -0800
49811 Rename driver from i810 to intel. Permit old name for compatibility.
49813 Driver installs as intel_drv.so with symlink to i810_drv.so to ensure
49814 existing configurations continue to work. Updated manual page to reflect
49815 name change and add attributions for recent work.
49817 commit 1dde7a15a3a42b881c57ece95feceffadf412cff
49818 Merge: 16f30b2 1ba45a3
49819 Author: Keith Packard <keithp@neko.keithp.com>
49820 Date: Thu Feb 1 21:31:27 2007 -0800
49822 Merge branch 'modesetting-origin' into modesetting
49824 commit 16f30b2e6c72e228c66946ff9efadd7278379745
49825 Author: Keith Packard <keithp@neko.keithp.com>
49826 Date: Thu Feb 1 21:30:19 2007 -0800
49828 Add "Ignore" option to per-output monitor, removes output from server.
49830 Setting option "Ignore" "Yes" will cause the server to pretend as if the
49831 specified output does not exist at all. It will not be listed by the
49832 RandR1.2 extension, and the server will not attempt to detect monitors at
49835 commit b3633506dcd4fd58f0f9519cd67f6e75d6fb2042
49836 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49837 Date: Fri Feb 2 11:35:44 2007 +0800
49839 Add Intel 965GM chipset support
49841 commit 380129a17c61c783d77a87685ad5333e8b1c253c
49842 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49843 Date: Fri Feb 2 11:33:46 2007 +0800
49845 Revert original crestline pci id patch
49847 commit f7489bbec150349bf03bea8d9d55451dbc92bfc2
49848 Merge: f1285ec ae6d358
49849 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49850 Date: Fri Feb 2 11:26:57 2007 +0800
49852 Merge branch 'modesetting' into crestline
49858 src/i830_xf86Crtc.c
49860 commit ae6d3585a4d96aa4dc05ca8563cebfa7796324ef
49861 Merge: 6d549ed 1ba45a3
49862 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
49863 Date: Fri Feb 2 11:21:59 2007 +0800
49865 Merge branch 'modesetting' of git://proxy.ims.intel.com:9419/git/xorg/driver/xf86-video-intel into modesetting
49867 commit 9f0acf1eb01ad8320f4da4cc5e498af25c0ecc5e
49868 Author: Keith Packard <keithp@neko.keithp.com>
49869 Date: Thu Feb 1 15:44:18 2007 -0800
49871 Check for rotation change in xf86RandR12CrtcSet.
49873 xf86RandRCrtcSet was ignoring changes to only rotation, so the screen would
49874 be left incorrectly rotated.
49876 commit 1ba45a3fcd40aaf4d373447fb9f252ca053f9105
49877 Author: Eric Anholt <eric@anholt.net>
49878 Date: Thu Feb 1 13:26:38 2007 -0800
49880 Improve register debugging output.
49882 This includes not reporting some fields on hardware where those bits are
49883 reserved, correcting one of the hardware error bit numbers, and reducing
49884 the severity of the debugging output warnings.
49886 commit 9bfeadbd5b37acb56cb646009239a94727ec52bc
49887 Author: Eric Anholt <eric@anholt.net>
49888 Date: Thu Feb 1 12:04:58 2007 -0800
49890 Remove some more dead options.
49892 commit d6bc03c379c46842676f640b8edb13ea828bf20c
49893 Author: Eric Anholt <eric@anholt.net>
49894 Date: Thu Feb 1 11:28:34 2007 -0800
49896 Bug #9819: Add two headers that are apparently necessary on older X Servers.
49898 commit 351866b0328b1d7afdc6b414c77d9ab4da4bd085
49899 Author: Eric Anholt <eric@anholt.net>
49900 Date: Wed Jan 31 15:48:52 2007 -0800
49902 Sync before overwriting Render acceleration state on the 965.
49904 commit 96783a458957ca280ebb6275fa692dd415df6931
49905 Author: Eric Anholt <eric@anholt.net>
49906 Date: Wed Jan 31 15:48:02 2007 -0800
49908 Fix copy-and-paste-o in the render code cleanup that broke XAA.
49910 commit 5238ff9c80718f9f8baa06123aa1bb6e3193fce6
49911 Author: Eric Anholt <eric@anholt.net>
49912 Date: Wed Jan 31 14:01:41 2007 -0800
49914 Naming and formatting cleanup in Render acceleration code.
49916 commit c47210d94ed17ad9e8e489a1e1886732e6dc7220
49917 Author: Eric Anholt <eric@anholt.net>
49918 Date: Wed Jan 31 12:50:31 2007 -0800
49920 Move the render code to non-EXA-specific filenames.
49922 commit 72ea0e514dcbebd009833c21f40e9bfb84074628
49923 Author: Eric Anholt <eric@anholt.net>
49924 Date: Wed Jan 31 12:43:38 2007 -0800
49926 Remove custom accelerated rotation code now that it's all done through Render.
49928 commit e62751db8b1a631c22ba0f77c932be4ab39ba741
49929 Author: Eric Anholt <eric@anholt.net>
49930 Date: Wed Jan 31 12:15:38 2007 -0800
49932 Add an accelerated path for rotation Render operations in XAA.
49934 The now-generic (kind of) EXA code will be cleaned up and moved to generic
49935 files in a later commit.
49937 commit f1285ecb72e807ac10db398fa983bf69d4154d97
49938 Author: Zou Nan hai <nanhai.zou@intel.com>
49939 Date: Wed Jan 31 14:58:08 2007 +0800
49943 commit 0620a7650f758652d05a0c25d159587f641aaf3e
49944 Author: Zou Nan hai <nanhai.zou@intel.com>
49945 Date: Wed Jan 31 11:21:11 2007 +0800
49949 commit 12441e217a1968dea3a6b0646b6c5d0e2ede0e7e
49950 Author: Zou Nan hai <nanhai.zou@intel.com>
49951 Date: Wed Jan 31 11:20:23 2007 +0800
49954 however 720p in this freq seems to not support src width,
49955 hactive > 800, so skip those inputs.
49957 commit 726c7dd89627409820ee5cf9bf657d5a926b3648
49958 Author: Zou Nan hai <nanhai.zou@intel.com>
49959 Date: Wed Jan 31 10:34:39 2007 +0800
49963 commit 4cd552e8f4851e029e43bf778cd8340f6c2c4881
49964 Author: Eric Anholt <eric@anholt.net>
49965 Date: Tue Jan 30 16:34:50 2007 -0800
49967 Fix accelerated Render transformations.
49969 Previously, we tried to use 2 points instead of 3 to describe the source
49970 rectangles, which mostly just worked for scaling.
49972 commit 6a628ae12b0568d656059891c5bca4415d8a735f
49973 Author: Eric Anholt <eric@anholt.net>
49974 Date: Tue Jan 30 14:19:05 2007 -0800
49976 Simplify EXA acceleration transform handling.
49978 commit c8581254cb811f76aa6eae49d74489b543b3eb05
49979 Merge: 9da1791 d329fa1
49980 Author: Keith Packard <keithp@neko.keithp.com>
49981 Date: Mon Jan 29 21:25:13 2007 -0800
49983 Merge branch 'modesetting-rotation' into modesetting
49985 This adds reasonably driver-independent rotation support to the common
49986 layer. The piece required in the driver is to allocate and redirect the crtc
49987 to a shadow frame buffer. The driver uses Render to perform the actual
49988 rotation operation (which leaves us free to do fun projective transforms at
49989 some point in the future :-).
49991 commit 9da1791eeca446cd59e2e7d0803e8a7920dbbea5
49992 Author: Eric Anholt <eric@anholt.net>
49993 Date: Mon Jan 29 17:10:14 2007 -0800
49995 Restore a few important lines from 89ace3e81fe72075b1709b073c05532f13b7f0d1
49997 Typical results were failure to sync, and a black screen.
49999 commit 67c6a490e63f6997f8198bc3ab180a1af097fe86
50000 Author: Eric Anholt <eric@anholt.net>
50001 Date: Mon Jan 29 09:53:58 2007 -0800
50003 Remove dead code to undo the damage of xf86SetModeCrtc() blank limiting.
50005 commit 89ace3e81fe72075b1709b073c05532f13b7f0d1
50006 Author: Eric Anholt <eric@anholt.net>
50007 Date: Mon Jan 29 09:43:35 2007 -0800
50009 Bug #9680: Remove bogus blank length limiting in xf86SetModeCrtc().
50011 Our modes typically come from EDID or default modes, and when the monitor
50012 asks for a specific mode, deciding to tweak it usually results in incorrect
50013 display. And if the user is specifying a mode by hand, tweaking it then is
50018 commit 6eb69e737864d06dc6d4fbb4cd85a027cbb7643a
50019 Author: Dave Airlie <airlied@linux.ie>
50020 Date: Sun Jan 28 19:40:12 2007 +1100
50022 ch7xxx: actually power up the chip around connection detection
50024 commit d329fa121b2401cadb991c2854e077cfa68e15e8
50025 Author: Keith Packard <keithp@guitar.keithp.com>
50026 Date: Sat Jan 27 18:36:15 2007 -0800
50028 Fetch from root window instead of screen pixmap during rotation.
50030 The XAA Composite wrapper doesn't check for sourcing from the root pixmap
50031 when deciding whether to sync the hardware drawing. Fix the rotation code to
50032 reference the root window in IncludeInferiors mode to avoid reading without
50033 correct synchronization.
50035 commit 20419d664c77e9a1a2de82987838b45f44774b47
50036 Author: Keith Packard <keithp@guitar.keithp.com>
50037 Date: Sat Jan 27 18:34:12 2007 -0800
50039 Paint ARGB cursor data to ARGB cursor memory space in frame buffer.
50041 When converting from global cursor to per-crtc cursor, the ARGB cursor data
50042 was accidentally painted to the wrong location, overwriting some critical
50043 data and causing the hardware to lock up. (along with generating a garbage
50044 cursor image from uninitialized data).
50046 commit 00decd3b6cea0de22c88b9504dbe26d680e8ab16
50047 Author: Krzysztof Halasa <khc@pm.waw.pl>
50048 Date: Sat Jan 27 12:29:31 2007 -0800
50050 Don't zero out mode structure after setting name field.
50052 Instead of using memset to zero the structure (at the wrong time, no less),
50053 use xcalloc to gather pre-initialized memory.
50055 commit 7a5f17087bda2833e84fa7e7ff0cb168943b89b6
50056 Author: Keith Packard <keithp@guitar.keithp.com>
50057 Date: Sat Jan 27 00:11:21 2007 -0800
50059 Mark crtc region as damaged when rotation is set.
50061 When shadow frame buffer is allocated for rotation, it needs to be
50062 initialized by copying from the frame buffer. Do this by simply marking the
50063 entire screen as damaged which will force an update.
50065 commit 2d95cb6f041653f7e530b1f32cf007929c23ef3b
50066 Author: Keith Packard <keithp@guitar.keithp.com>
50067 Date: Sat Jan 27 00:04:13 2007 -0800
50069 Make rotated pixmap size of mode, not rotated mode. Update only damage.
50071 Rotated pixmap should match size of displayed mode, not the rotated version
50074 Take damaged region, rotate each box and display them instead of displaying
50075 the whole screen after each damage update. This exposes a bug where the
50076 screen is not correctly damaged after rotation.
50078 commit 5c1e27cdd243dc24dd2bfdeb46d757bbef6ba6af
50079 Author: Eric Anholt <eric@anholt.net>
50080 Date: Thu Jan 25 15:31:22 2007 -0800
50082 Make rotated shadow buffer allocation dynamic.
50084 For EXA, this requires version 2.1 of EXA to do rotation, as the VT switching
50085 issues were too complicated otherwise.
50087 commit 5a67f3d3690903a181ca854d060cfa8b4c2aca26
50088 Merge: 3cc583e 9811059
50089 Author: Nian Wu <nian.wu@intel.com>
50090 Date: Thu Jan 25 13:40:50 2007 -0800
50092 Merge branch 'crestline' of /git/xorg/driver/xf86-video-intel into crestline
50094 commit 3cc583ec8e79ec65103308220dd6047a4a7acddf
50095 Merge: 566e1d3 effe579
50096 Author: Nian Wu <nian.wu@intel.com>
50097 Date: Thu Jan 25 13:37:05 2007 -0800
50099 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into crestline
50101 commit 6d549ed280d3fcf3fe611b095d9f8adc0196bfb6
50102 Merge: 75f4df2 effe579
50103 Author: Nian Wu <nian.wu@intel.com>
50104 Date: Thu Jan 25 13:36:33 2007 -0800
50106 Merge branch 'modesetting' of git://proxy01.pd.intel.com:9419/git/xorg/driver/xf86-video-intel into modesetting
50112 commit 83cc4601b27d871484a2408f31154e9387064b9e
50113 Author: Eric Anholt <eric@anholt.net>
50114 Date: Wed Jan 24 16:52:22 2007 +0800
50116 Fix a possible failure to misalign video allocation in XAA.
50118 commit e21332419581eff1d7651741bae0b640c84d0ecd
50119 Author: Eric Anholt <eric@anholt.net>
50120 Date: Wed Jan 24 12:32:38 2007 +0800
50124 commit effe579e691b044e3ce59b41b5c0eaaac4368dda
50125 Author: Dave Airlie <airlied@linux.ie>
50126 Date: Sun Jan 21 20:23:13 2007 +1100
50128 ch7xxx: fix output detection as per datasheet, have no DVI to test this.
50130 commit c68dae2705d6e5b05dba7d40b6da112b98b926b4
50131 Author: Dave Airlie <airlied@linux.ie>
50132 Date: Sun Jan 21 20:22:39 2007 +1100
50134 fix building with 7.1 server by hiding stuff inside RANDR_12_INTERFACE
50136 commit 8eb861fb61fda78e6c142ffa66c586357aaa2e44
50137 Author: Eric Anholt <eric@anholt.net>
50138 Date: Mon Jan 22 15:50:14 2007 +0800
50140 Replace the custom SaveScreen function with a generic function calling DPMS.
50142 With this, additional power saving occurs with the core screensaver, for
50143 example by turning off LVDS backlight.
50145 commit 7161e824e89a6a69f1277e510f2344ad31721f62
50146 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50147 Date: Fri Jan 19 15:15:19 2007 +0800
50149 Fix DRI issue when EXA enabled
50151 tell DRI we are smashing the context state.
50153 commit 74ebff6732b9bfcf8c865b52cbebfd9bf6b73eb2
50154 Author: Eric Anholt <eric@anholt.net>
50155 Date: Thu Jan 18 14:26:21 2007 -0800
50157 Allocate separate cursor memory per CRTC and rotate cursors appropriately.
50159 Also, add bind/unbind of the shadow rotate buffers, which was missed in a
50162 commit 8759f4d89a141f7d69f1894ec3059ba6bd1cf86b
50163 Author: Eric Anholt <eric@anholt.net>
50164 Date: Thu Jan 18 13:13:25 2007 -0800
50166 Partially deal with cursor rotation.
50168 The remaining issue is having a cursor image per CRTC so that the cursors can
50169 be rotated on the independently rotated CRTCs.
50171 commit b296cd9b8f63ab80e8fe46fcfcdba2e9af846468
50172 Author: Eric Anholt <eric@anholt.net>
50173 Date: Thu Jan 18 13:02:26 2007 -0800
50175 Fix transformation matrices for rotation at 90 and 270 degrees.
50177 commit d2ae2e2ccc3a8e7dc84745c235836cba150754d8
50178 Author: Eric Anholt <eric@anholt.net>
50179 Date: Thu Jan 18 11:37:14 2007 -0800
50181 Update CRTC values first so they can be used by the driver in mode setting.
50183 With this, the screen now gets rotated, though there are major issues.
50185 commit 92b0f3d2013c32d94fe57fe674da1b195288f347
50186 Author: Eric Anholt <eric@anholt.net>
50187 Date: Thu Jan 18 11:23:22 2007 -0800
50189 Disable the clipping for now, since it won't work.
50191 commit 14d265a02312e9cda7a21614d77525c7140c0ca8
50192 Author: Eric Anholt <eric@anholt.net>
50193 Date: Thu Jan 18 11:21:58 2007 -0800
50195 Store fixed values, not integers, in the transformation matrix.
50197 Also, add a bit more error checking.
50199 commit cc09dc3efe7279bf67512babb0f6cdaeba654772
50200 Merge: 53e765d 3094100
50201 Author: Eric Anholt <eric@anholt.net>
50202 Date: Thu Jan 18 10:20:36 2007 -0800
50204 Merge branch 'modesetting' into modesetting-rotation
50206 commit 53e765d1cf5016961c25a798a27f0fe09bd1f237
50207 Author: Eric Anholt <eric@anholt.net>
50208 Date: Wed Jan 17 16:56:06 2007 -0800
50210 Fix broken last-minute change in determining whether we're rotated or not.
50212 commit f256243fff195f73bc8aa72e8dd2c6f7b33446cf
50213 Author: Eric Anholt <eric@anholt.net>
50214 Date: Wed Jan 17 16:01:18 2007 -0800
50216 Allocate per-crtc rotation buffers and hook up RandR 1.2 per-CRTC rotation.
50218 It currently displays garbage.
50220 commit d17c386aee55bd223297adf78cf0fc1db43baf2d
50221 Author: Eric Anholt <eric@anholt.net>
50222 Date: Wed Jan 17 15:49:08 2007 -0800
50224 Fix some segfaults: crtc->rotatedPixmap not set and dangling damage pointer.
50226 commit 67a152c43cb7231695b99030c952c52d4edd1e65
50227 Author: Eric Anholt <eric@anholt.net>
50228 Date: Wed Jan 17 14:35:31 2007 -0800
50230 Set all CRTCs' allowable rotations when xf86RandR12SetRotations() is called.
50232 commit 98110591001dfaf2b8b5ac440bea0c38456da718
50233 Author: Zou Nan hai <nanhai.zou@intel.com>
50234 Date: Wed Jan 17 10:45:59 2007 +0800
50236 enable all TV modes
50238 commit 3094100cdd04560af1b098bbe8024ae5270a10f4
50239 Author: Gary Wong <gtw@gnu.org>
50240 Date: Tue Jan 16 13:25:24 2007 -0800
50242 Bug #9550: Add a quirk for too-high clock from EDID on the EN-7100e.
50244 commit 66c6427cc83e5e381932b0c10743cba4cb9ac883
50245 Author: Eric Anholt <eric@anholt.net>
50246 Date: Tue Jan 16 12:54:29 2007 -0800
50248 Add a settable backlight property for LVDS.
50250 This extends the output funcs to have a callback for when it's time to
50251 configure properties, and another for when the server has changed a property
50252 whose value isn't pending a mode set. (Pending properties are to be picked up
50253 by the driver at mode_set time).
50255 commit a52df90ffcb19e7c0a3b8018db3cd1391a3ec343
50256 Merge: 37946c9 45696aa
50257 Author: Keith Packard <keithp@neko.keithp.com>
50258 Date: Tue Jan 16 12:20:26 2007 +1100
50260 Merge branch 'modesetting' into modesetting-rotation
50267 Just a few sync changes.
50269 commit 37946c9c8a4afa4a4fc1fe9864a8eec1c4571d2c
50270 Author: Keith Packard <keithp@neko.keithp.com>
50271 Date: Tue Jan 16 12:14:55 2007 +1100
50273 Interim work on rotation support with new xf86 code.
50275 Yes, there are lots of bits tied together here, and we should split this
50278 Move I830 mode setting to xf86 mode setting.
50280 Make mode setting function also set the base address. This should fix
50281 problems where the base gets set incorrectly at times.
50283 Add driver-independent rotation support, requires driver-specific hooks for
50284 shadow pixmap allocation, otherwise it uses Render for painting.
50286 commit 9a5106401a65c90df32cb71987fca7126dc22e81
50287 Author: Emilio Scalise <emisca.ml@gmail.com>
50288 Date: Mon Jan 15 17:22:27 2007 +0000
50290 patch to control mergedfb options, bug #9664
50292 commit bfeda3bfc59d309772398f688d29452ba97f3973
50293 Author: Keith Packard <keithp@neko.keithp.com>
50294 Date: Mon Jan 15 08:05:32 2007 +1100
50296 Convert I830PipeSetMode to xf86CrtcSetMode. Add rotation structures.
50298 Makes the mode setting logic device independent.
50299 Magic rename hooks allow multiple drivers to provide the
50300 generic code without name conflicts.
50301 Rotation code requires special pixmap creation hook, and uses
50302 Render to perform the actual rotation.
50304 commit 319be199ba657d2b82ba034edf3581aea6dcc3f0
50305 Author: root <root@jetpack.demon.co.uk>
50306 Date: Sat Jan 13 21:34:00 2007 +0000
50310 commit 85de57947570de6b3dc9197647be3e9237480880
50311 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
50312 Date: Sat Jan 13 19:18:24 2007 +0000
50314 Fix bug #9639, i810 XvMC needs linking against libdrm.
50316 commit 45696aa29124e2852f94880642e70bb2e0cee827
50317 Author: Eric Anholt <eric@anholt.net>
50318 Date: Thu Jan 4 12:34:47 2007 -0800
50320 EXA: Use PRIM3D_RECTLIST instead of TRIFAN so we don't get diagonal tearing.
50322 A side effect is the reduction in vertex dispatch, which is nice.
50324 commit 6874a6f25ac87783d3770f77b9192e2d36d083a3
50325 Author: Eric Anholt <eric@anholt.net>
50326 Date: Thu Jan 11 12:34:38 2007 -0800
50328 Restore legacy overlay size limits, due to card hangs at larger sizes.
50330 commit c288aea40775a9cf561fda9912187c3cb5baa419
50331 Author: Eric Anholt <eric@anholt.net>
50332 Date: Thu Jan 11 10:39:00 2007 -0800
50334 Bug #8845: Unify all our XV ports to 1920x1088 maximum size.
50336 This increases the "legacy" size, but that was checked against the 845
50337 documentation which claims support for 2048x2048 as well. It decreases the
50338 textured video size, which was running into the limits of the bigrequests
50339 extension. The new limits should fit within bigrequests while still supporting
50342 commit f5d528f8ea27de31054e7f1843e34d8379f811ea
50343 Author: Eric Anholt <eric@anholt.net>
50344 Date: Thu Jan 11 10:36:33 2007 -0800
50346 Don't limit cachelines to a vertical of 2048, and increase default allocation.
50348 The cachelines are used for two things: XAA pixmap cache and XV memory.
50349 Only XAA pixmap cache is referred to using an offset pointing at the
50350 beginning of the front buffer in rendering, and XAA only uses the 2d BLT
50351 engine, which actually has a vertical limit of 65536. So, pixmap cache is now
50352 limited to that much vertical.
50354 Additionally, the previous cachelines allocation was too small for our
50355 advertised XV limits, so video at the limits would fail with BadAlloc. Now,
50356 XAA allocates the same approximate amount of offscreen memory as EXA:
50357 3 times the screen size, plus one packed HD video.
50359 commit d13bc016c0723f1df633ddaf5610ad73003b7c96
50360 Author: Eric Anholt <eric@anholt.net>
50361 Date: Wed Jan 10 15:06:56 2007 -0800
50363 Correct x/y/pitch limitations in several cases, and detail them in i830_exa.c.
50365 This reduces max framebuffer width and increases max framebuffer height on
50366 965, reduces max X/Y on pre-965 EXA (could have caused mis-rendering), and
50367 increases max X/Y on 965 EXA (would have prevented acceleration).
50369 commit fa383289ac8a6dd1cb359e6f1991cc42beb6ff02
50370 Author: Eric Anholt <eric@anholt.net>
50371 Date: Thu Jan 4 14:12:45 2007 -0800
50373 EXA: Wait for sync before we set up new state in our static state buffers.
50375 commit 5857b4a1693085b8b42dd9560a7c4f5c3c82f862
50376 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50377 Date: Wed Jan 10 15:09:47 2007 +0800
50379 Formalize sync interface
50381 We should just call i830MarkSync/i830WaitSync in places we need,
50382 which care for both XAA and EXA.
50384 commit efb75f56053ee06f1dbd4edfaea9986b27162afb
50385 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
50386 Date: Tue Jan 9 14:05:43 2007 +0000
50390 commit 33c0fac6d63d9a509a24c7a5e5a46a0d6ee4dc5d
50391 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
50392 Date: Tue Jan 9 14:02:47 2007 +0000
50394 Fix bug #8536, i915 BIOS fails when restarting Xserver.
50396 commit ee52c0ec4e95fa6e5f35c9cd75005a0c0003fd97
50397 Author: Eric Anholt <eric@anholt.net>
50398 Date: Mon Jan 8 16:53:07 2007 -0800
50400 Correct typo resulting in a crash with ivch. = != ==.
50402 commit 566e1d397744d0b477de4da6a206919906176f49
50403 Merge: b13d638 da6a00f
50404 Author: Keith Packard <keithp@neko.keithp.com>
50405 Date: Mon Jan 8 12:37:49 2007 -0800
50407 Merge branch 'modesetting' into crestline
50409 commit da6a00f787e4d13e6b75768c1976f1c44ae5bf72
50410 Author: Keith Packard <keithp@neko.keithp.com>
50411 Date: Mon Jan 8 12:36:54 2007 -0800
50413 Move EXA function declarations from source to header.
50415 Several new global functions were not put into the header file leading to
50416 potential mismatches between declaration and definition.
50418 commit 66546d8a8006e1c828e18ccab850214fd8d56b63
50419 Merge: 0fd2752 40af0ee
50420 Author: Keith Packard <keithp@neko.keithp.com>
50421 Date: Mon Jan 8 10:14:24 2007 -0800
50423 Merge branch 'modesetting-origin' into modesetting
50425 commit b13d6386dfb3a00bd9d21ac0695cdce9f812d1f3
50426 Author: Keith Packard <keithp@neko.keithp.com>
50427 Date: Sun Jan 7 22:56:37 2007 -0800
50429 Back out rotation changes for RandR 1.2 which won't work.
50431 RandR 1.2 requires a new rotation structure for per-CRTC rotation.
50433 commit 7473e7d5297f94164e0f35822166713fb21a11bd
50434 Merge: cdde9e7 a87801f
50435 Author: Keith Packard <keithp@neko.keithp.com>
50436 Date: Sun Jan 7 22:50:27 2007 -0800
50438 Merge branch 'crestline-otc' into crestline
50440 commit cdde9e7f4a0645ab1ee3e124de54433c1a250097
50441 Merge: 3110630 0fd2752
50442 Author: Keith Packard <keithp@neko.keithp.com>
50443 Date: Sun Jan 7 22:50:12 2007 -0800
50445 Merge branch 'modesetting' into crestline
50447 commit 0fd2752f199928f846fe03c9087f7b6d48cc28d9
50448 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50449 Date: Thu Jan 4 14:35:55 2007 +0800
50451 minor fix on last exa mem binding commit
50453 commit 3110630e04f37e184609c91494fa7f9f4d59b93f
50454 Merge: be9b635 736d82a
50455 Author: Keith Packard <keithp@neko.keithp.com>
50456 Date: Sun Jan 7 22:44:36 2007 -0800
50458 Merge branch 'modesetting' into crestline
50464 Pull in upstream changes to crestline branch leaving only
50465 the PCI-IDs as local changes.
50467 commit 40af0ee6ba4ab7596fbc7fcc3ad04d109746ca6f
50468 Author: Keith Packard <keithp@mandolin.keithp.com>
50469 Date: Sat Jan 6 18:19:34 2007 -0800
50471 Numerous symbol scope issues.
50473 I830EntityIndex is shared between 810 and newer driver.
50474 Move most EXA rendering state into I830 structure.
50475 Declare shared variables in shared header files rather than .c.
50477 commit 736d82a6b43f174cb95b425faacd4b0b889916fa
50478 Merge: 53b42f5 4c790f6
50479 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50480 Date: Mon Dec 4 15:48:04 2006 +0800
50482 Merge branch 'modesetting-origin' into modesetting
50484 commit 53b42f5bc7a58d02106436486e5bb56e56dbbfa1
50485 Author: Zou Nan hai <nanhai.zou@intel.com>
50486 Date: Sat Jan 6 14:59:14 2007 -0800
50488 support NTSC 480i M-J, PAL 576i for 640x480-1280x1024 sizes
50490 I still have problem with non-interlace mode and Hi Res mode.
50491 also I don't know how to pickup those mode in xorg.conf
50493 commit 71946bcdc3c68c220996afac944698eea1974a36
50494 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50495 Date: Wed Jan 3 22:37:32 2007 -0800
50497 [PATCH] Add rotation support for 965.
50499 commit 35cebed70827999812f8343ac97ad0dffda20786
50500 Author: Eric Anholt <eric@anholt.net>
50501 Date: Thu Nov 16 15:12:43 2006 -0800
50503 [PATCH] Replace broken PCI resource size detection with pciGetBaseSize() call.
50505 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50507 commit 33e912aca08fa11ef588eb386e16ba5f9ea13727
50508 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50509 Date: Thu Jan 4 11:25:31 2007 +0800
50511 [PATCH] Fix EXA mem binding
50513 We should check if EXA is really enabled.
50515 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50517 commit fa54a3c08301e59558ab0493b3d22324f4162496
50518 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50519 Date: Wed Dec 6 13:24:44 2006 +0800
50521 [PATCH] fix Makefile.am
50523 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50525 commit 2f2c443e971faa54ffcf751b6acb444e8e7875ce
50526 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50527 Date: Wed Dec 6 10:43:29 2006 +0800
50529 [PATCH] Formats fixes
50531 We should use card_fmt for src/mask picture, and use dest color
50532 buffer format helper. Also fix wrong name for G965 texture formats,
50533 and pict_x1r5g5b5 isn't supported by sampler engine.
50535 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50537 commit 0bf04fe78a8a915310ef8a90f5c7872be7476e2e
50538 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50539 Date: Mon Dec 4 15:48:04 2006 +0800
50541 [PATCH] set correct default border color
50543 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50545 commit 5c461063cde68092e778c44ac6abd9129cd8019e
50546 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50547 Date: Mon Dec 4 15:47:31 2006 +0800
50549 [PATCH] fix typo in ps kernel
50551 fix corrupt in some subspans
50553 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50555 commit 89a42d489bd370b89e5ff4e01f026b4d64723cd8
50556 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50557 Date: Wed Nov 29 17:16:46 2006 +0800
50559 [PATCH] shut up warning
50561 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50563 commit 01bfa4fa6fc0ceec8581676e5d72c68dd71efa96
50564 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50565 Date: Wed Nov 29 17:14:55 2006 +0800
50567 [PATCH] fix alpha blending state
50569 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50571 commit 79018bb47c43510d59c592592f06204189bd12dc
50572 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50573 Date: Wed Nov 29 17:05:32 2006 +0800
50575 [PATCH] Add in sf/wm program for mask picture without CA
50577 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50579 commit a5b9b438469f171b002fa0b99d8cab83e51ec968
50580 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50581 Date: Wed Nov 29 16:52:44 2006 +0800
50583 [PATCH] misc cleanup for G965 vs/sf/wm states
50585 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50587 commit b7c1e1656f45e43ea2f9a47f1a487050c0884c22
50588 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50589 Date: Wed Nov 29 16:40:15 2006 +0800
50591 [PATCH] WM kernel needs scratch space
50593 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50595 commit db9cfaa35adaf79ea57bc06b27c7e37935d3e1a7
50596 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50597 Date: Wed Nov 29 16:37:06 2006 +0800
50599 [PATCH] Setup default border color for our samplers
50601 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50603 commit 70276e4e9a8a5026ec436d2be5bf5eab868aa178
50604 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50605 Date: Wed Nov 29 16:30:53 2006 +0800
50607 [PATCH] fix vertex buffer size
50609 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50611 commit e8a4cbdeff4125e28d807d0a563efc0606d21a75
50612 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50613 Date: Wed Nov 29 16:24:24 2006 +0800
50615 [PATCH] clean up issue cmd to ring buffer
50617 Make it easy to track different part of ring state, and
50618 use rectlist primitive instead.
50620 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50622 commit 42534474fd2556e5987205626cca8f30e25855a8
50623 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50624 Date: Wed Nov 29 16:04:16 2006 +0800
50626 [PATCH] remove wrong scale_units
50628 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50630 commit 05e8a7989db1b8b9dfa7b688629af65d9910cc53
50631 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50632 Date: Wed Nov 29 16:02:16 2006 +0800
50634 [PATCH] Rename exa sf/wm program files
50636 Also fix some minors in wm program.
50638 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50640 commit 1329c86f2a4b3664920ded970a984a745a530da4
50641 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50642 Date: Wed Nov 29 15:47:19 2006 +0800
50644 [PATCH] EXA state mem for G965
50646 Not split offscreen mem for exa, but alloc a dedicated one
50649 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50651 commit 4c790f614ecba1f6468e51779cfaf0e36b6b17ad
50652 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
50653 Date: Sat Jan 6 08:21:36 2007 -0800
50655 More sparse cleanups:
50657 make some variables static
50658 remove redundant variables
50660 There are only a few sparse warnings left now: some bitfield warnings
50661 and a few 'mixing code and declarations' warnings from the ring macros.
50663 commit a87801f73a73e53524237be7835b8cd8b3eb282c
50664 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50665 Date: Thu Jan 4 14:47:48 2007 +0800
50669 Actually load TV, otherwise we don't have expected
50670 reg state to probe.
50672 commit 9cf759a2cb2580c4ac5f8cdc983389642f3cfce0
50673 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50674 Date: Thu Jan 4 14:35:55 2007 +0800
50676 minor fix on last exa mem binding commit
50678 commit f9c3ef40100acb85fca6e49c5c0e98f5c9ac97ca
50679 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50680 Date: Tue Oct 10 15:50:10 2006 +0800
50682 [PATCH] Mark current ps kernel is experimential with little test.
50685 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50687 commit ef2f25e5c6ce04c202698c5122bc8627130073c0
50688 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50689 Date: Tue Oct 10 14:11:35 2006 +0800
50691 [PATCH] Use sf_prog.h instead
50694 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50696 commit ca608028c5301700444d39a1c631cc0d5648e1a2
50697 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50698 Date: Thu Sep 28 13:55:52 2006 +0800
50700 [PATCH] Fix picture's transform checking
50703 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50705 commit 848368d5d0b90e03d3ec447cb5bd39fc87aea8df
50706 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50707 Date: Thu Sep 28 11:15:33 2006 +0800
50709 [PATCH] Fallback in mask picture for now
50711 Do it later after finish wm kernel program.
50713 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50715 commit df23624eebe938fa444c80cbedcd61919ec1aeda
50716 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50717 Date: Thu Sep 28 11:09:52 2006 +0800
50719 [PATCH] Fix compile, add wm header file.
50722 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50724 commit fc944859b1b9605c748162bad1c93a6303c84aae
50725 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50726 Date: Thu Sep 28 10:36:00 2006 +0800
50728 [PATCH] Add simplest wm kernel program for no mask picture composite
50730 This is a try to use new gen4asm language, and will finish
50731 composite program for mask picture with or without CA case later.
50733 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50735 commit acdc2da77b445e9347a4c6e53e35c81763cbb0b8
50736 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50737 Date: Wed Sep 27 16:48:43 2006 +0800
50739 [PATCH] Add mask sampler state
50742 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50744 commit 926d7fb09aaaabf050949ce7c6127c68441c8801
50745 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50746 Date: Wed Sep 27 13:54:14 2006 +0800
50748 [PATCH] change some src sampler states
50750 sampler for mask should also be set up, and fix
50751 default border texel.
50753 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50755 commit ed73bbaf5c2e9d555c884037a249cf03e7f60fa0
50756 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50757 Date: Mon Sep 25 14:35:51 2006 +0800
50759 [PATCH] Add file for i965 exa composite
50761 This does not include ps program, which will be added
50764 Signed-off-by: Keith Packard <keithp@neko.keithp.com>
50766 commit be9b635b82c0c77ccc0555f178f94de6e5338e27
50767 Merge: 9e48808 5a793b0
50768 Author: Keith Packard <keithp@neko.keithp.com>
50769 Date: Wed Jan 3 22:15:53 2007 -0800
50771 Merge branch 'modesetting' into crestline
50773 commit 5a793b0dcf2d5de408b55073858fcfba6d99f994
50774 Merge: f188525 394124c
50775 Author: Keith Packard <keithp@neko.keithp.com>
50776 Date: Wed Jan 3 21:54:05 2007 -0800
50778 Merge branch 'modesetting-origin' into modesetting
50780 commit f188525030a8fac59e41520449b1aec9b123e4ea
50781 Author: Keith Packard <keithp@neko.keithp.com>
50782 Date: Wed Jan 3 21:53:07 2007 -0800
50784 Add interlace defines for pipeconf regs
50786 commit 9e48808fccf5b153c2fb4027f9a09944574fdace
50787 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50788 Date: Thu Jan 4 11:25:31 2007 +0800
50790 Fix EXA mem binding
50792 We should check if EXA is really enabled.
50794 commit ebd636dbd62cb69ed7defbd86f297c6bc7dac5dd
50795 Merge: 75f0033 3d9ea6f
50796 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50797 Date: Thu Jan 4 10:13:55 2007 +0800
50799 Merge branch 'modesetting' into crestline
50806 commit 3d9ea6f3c8a274f930d286fc15d3665dfaa8327a
50807 Merge: 75f4df2 394124c
50808 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50809 Date: Thu Jan 4 10:12:50 2007 +0800
50811 Merge branch 'modesetting' of git+ssh://zhen@git.freedesktop.org/git/xorg/driver/xf86-video-intel into modesetting
50817 commit 394124ceaadb46d976ad5c3bdeb1b77d351c57f6
50818 Author: Eric Anholt <eric@anholt.net>
50819 Date: Tue Jan 2 16:57:33 2007 -0800
50821 Add per-CRTC gamma control support for RandR.
50823 This makes the CRTCs now always run in gamma-enabled mode, rather than having
50824 flaky logic for switching modes. Also, it should clear up issues with the LUTs
50825 being uninitialized when outputs are first brought up.
50827 commit 69f250af60220a875f4a04c6d682bffa352281e4
50828 Merge: 232e209 d960dea
50829 Author: Keith Packard <keithp@mandolin.keithp.com>
50830 Date: Wed Jan 3 00:39:15 2007 -0800
50832 Merge branch 'modesetting-origin' into modesetting
50834 commit 232e2094321dbcdd6a67ef230eb50494a1c7d6df
50835 Author: Keith Packard <keithp@mandolin.keithp.com>
50836 Date: Wed Jan 3 00:38:34 2007 -0800
50838 Allow initial position to be set in config file. Increase 965 max size.
50840 Add relative and absolute position configuration code, using per-output
50841 monitor sections. Options include:
50843 PreferredMode selects a preferred mode for this output by name
50844 Position absolute position, x and y in a single string.
50846 Below relative positions; argument names other monitor.
50851 Enable force the monitor to be disabled by setting
50852 Disable enable to no or disable to yes.
50854 MinClock Set valid clock ranges
50857 Monitor sections can also include sync ranges, physical size and mode lines
50858 as documented in xorg.conf(5).
50860 Monitors are associated with outputs through options in the Device section:
50862 Option "monitor-VGA" "My VGA Monitor"
50864 Output named 'VGA' will use monitor section "My VGA Monitor".
50866 commit 58e797b2caa6effa5455fc1f13dc4c58d0658744
50867 Author: Keith Packard <keithp@mandolin.keithp.com>
50868 Date: Wed Jan 3 00:04:58 2007 -0800
50870 Sync dspbase/dspsurf registers by re-reading them.
50872 This seems to eliminate base/surf value confusion during EnterVT.
50874 commit d960deab39eef91fb82b9f23118323aeb4c9c63e
50875 Author: Jesse Barnes <jbarnes@nietzche.virtuousgeek.org>
50876 Date: Tue Jan 2 20:23:37 2007 -0800
50878 Various sparse cleanups:
50880 remove redundant pScrn in I830DRISwapContext
50881 make I830DRISwapContext static
50883 commit 5057769d3a7c1b3a94f49bbff47b9697f368d975
50884 Author: Dave Airlie <airlied@linux.ie>
50885 Date: Tue Jan 2 19:25:08 2007 -0800
50887 Use correct variable to initialize DVO I2C bus.
50889 intel_output->pI2CBus isn't initialized until the output is all happy and
50890 running, so using it before that is rather unlikely to do much good.
50892 commit 2bb9bb951e664d704bdf7a808edbe8b2f50c4366
50893 Author: Keith Packard <keithp@neko.keithp.com>
50894 Date: Tue Jan 2 14:04:10 2007 -0800
50896 Typo disabled new DRI MM on all chips.
50898 Detection for new DRI MM wasn't converted to #define correctly because of a
50901 commit 5080307230d534b84090bfefc13ac437d53f915d
50902 Author: Keith Packard <keithp@mandolin.keithp.com>
50903 Date: Tue Jan 2 00:16:41 2007 -0800
50905 Fix driver to build on RandR 1.2 systems again.
50907 Changes to move initial config code to RandR 1.1/1.2 common path broke
50908 code required only for RandR 1.2.
50910 commit 237847007afb5a429a50b6aa0b1acc239326571b
50911 Author: Keith Packard <keithp@mandolin.keithp.com>
50912 Date: Tue Jan 2 00:09:35 2007 -0800
50914 Make driver build and run on Xorg 7.1. Delay DSP*BASE setting until last.
50916 Autodetect libdrm version, disable new memory manager on older libraries.
50917 Move new M_T_ defines from i830.h to i830_xf86Crtc.h. Add many system
50918 headers to define functions. Use i830PipeSetBase at end of mode setting
50919 code to set DSP*BASE and flush changes. Don't duplicate PipeSetBase call
50920 from screen init function. Make initial RandR configuration code usable on
50921 older versions of extension so the server doesn't start in a panning mode.
50922 Use xfree instead of free in i830_tv.c.
50924 commit 25d5a892319b02dc6eb81390dea29cd88a1e7da4
50925 Author: Keith Packard <keithp@guitar.keithp.com>
50926 Date: Sun Dec 31 15:39:20 2006 -0800
50928 Elide identical modes from reported list.
50930 Where two modes would produce precisely the same crtc settings and have the
50931 same name, remove the latter mode from the mode list.
50933 commit feeefc92e450e9de58da51147325300ffabd2059
50934 Author: Keith Packard <keithp@neko.keithp.com>
50935 Date: Sun Dec 31 14:22:53 2006 -0800
50937 Replace author-specific license with generic license.
50939 This file retained the old MIT license with the embedded author name.
50940 Replace that with the version that uses generic 'copyright holders' instead.
50942 commit 5e456251f0d6c8bef3241267f6c1b3bca4670c0d
50943 Author: Keith Packard <keithp@neko.keithp.com>
50944 Date: Sun Dec 31 14:21:12 2006 -0800
50946 Replace author-specific license with generic license.
50948 This file retained the old MIT license with the embedded author name.
50949 Replace that with the version that uses generic 'copyright holders' instead.
50951 commit 75f0033ec41a34bbe695cfd4a9bc9aa60e919c99
50952 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
50953 Date: Sun Dec 31 15:16:59 2006 +0800
50955 Issue current virtual size to mode valid.
50957 Revert change in rotation patch.
50959 commit 2330b341c925e60bc2c991b4ca8450badbc459a4
50960 Author: Keith Packard <keithp@neko.keithp.com>
50961 Date: Sat Dec 30 22:47:58 2006 -0800
50963 Prune default modes to EDID clock limit.
50965 Pick out the EDID max clock value and use that to elide default modes which
50968 commit c70e7bc701cba3092acfdd66ed4d63487ac8ff9f
50969 Merge: 8b6c456 4ed79a2
50970 Author: Keith Packard <keithp@neko.keithp.com>
50971 Date: Sat Dec 30 21:58:36 2006 -0800
50973 Merge branch 'modesetting-origin' into modesetting
50975 commit 8b6c456e16c81f719bb68178e3e4d26f4108a9c2
50976 Author: Keith Packard <keithp@neko.keithp.com>
50977 Date: Sat Dec 30 21:57:42 2006 -0800
50979 Detect in ProbeModes. Let get_modes set edid. Prune interlace/dblscan.
50981 Move output connection status detection from RandR code up to ProbeModes so
50982 it is done before mode sets are built. Otherwise, the mode building code
50983 will elide all modes the first time through as it ignores outputs that are
50986 Most get_modes functions fetch EDID data; make sure that any
50987 EDID changes are used in the ProbeModes filtering of default modes.
50988 Otherwise, stale EDID data will be used.
50990 Allow outputs to advertise support for interlaced and double scan modes;
50991 prune such modes from the default mode list when outputs do not support them.
50993 commit 4ed79a2ba6250354ffc24fa1f7a21ca914ad157e
50994 Author: Eric Anholt <eric@anholt.net>
50995 Date: Fri Dec 29 12:19:41 2006 -0800
50997 Relax tiling requirements on G965.
50999 For the 965, we can tile with the pitch at any integer multiple of a tile size
51000 (128 or 512B), up to 128KB.
51002 commit b8692e646227e56c9ae4f72b9aaa75457b4c0f5f
51003 Author: Keith Packard <keithp@neko.keithp.com>
51004 Date: Thu Dec 28 22:23:16 2006 -0800
51006 Lack of configured monitor implies no configured modes.
51008 Missing check for missing monitor configuration would result in segfault.
51010 commit 973da654219ea43916b0b44acfa09a415bed3d7a
51011 Author: Keith Packard <keithp@neko.keithp.com>
51012 Date: Thu Dec 28 22:22:22 2006 -0800
51014 Sync ranges from EDID/default should not limit configured modelines.
51016 Limit the effect of sync ranges so that sync ranges found via EDID will not
51017 eliminate modes explicitly added by the user. Limit default sync range to
51018 eliminating only default modes, not configured or EDID modes.
51020 commit f7b1d4c1f7d17a811e17c6a17861ff70be9fbdd7
51021 Author: Keith Packard <keithp@neko.keithp.com>
51022 Date: Thu Dec 28 22:18:57 2006 -0800
51024 Belinea 10 15 55 monitor quirk - override preferred mode with largest @60Hz
51026 Belinea 10 15 55 model monitor reports a preferred mode of 640x350, when in
51027 fact it wants a 1024x768 mode @ 60Hz. Add an edid quirk that selects the
51028 largest size mode, preferring those closer to 60hz among equal sized modes.
51030 commit fbd3f588f0d412c20cc3f86dfad73700734a9a76
51031 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
51032 Date: Fri Dec 29 10:47:52 2006 +0800
51034 rotation support for randr-1.2
51036 This patch trys to resolve rotation regression with randr-1.2.
51037 Tested with Crestline VGA output. See bug 83.
51039 commit a1796bfb51d90fd545973909404e1a77ad27c9cc
51040 Merge: 84915ac fabfccd
51041 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
51042 Date: Fri Dec 29 10:10:26 2006 +0800
51044 Merge branch 'crestline' of git+ssh://zhen@otc-graphics.jf.intel.com/git/xorg/driver/xf86-video-intel into crestline
51046 commit bedab1654e2dfcf7800bd0101e6991800a544019
51047 Author: Eric Anholt <eric@anholt.net>
51048 Date: Thu Dec 28 16:34:40 2006 -0800
51050 Remove 8MB/32MB default VideoRam limits.
51052 With modern monitors and increased XV and EXA memory requirements, these small
51053 limits were resulting in DRI and other initialization failures because we
51054 wouldn't allow them enough memory. Instead, allow each piece of the system
51055 (DRI, EXA, XAA, etc) to request as much memory as it wants, and choose the
51056 actual videoRam to be used for laying out the memory afterwards.
51058 With this change, in the absence of a VideoRam option, 32MB will be allocated
51061 commit 4cc5dcf6d7423239f0181f246638366edcb8be87
51062 Author: Keith Packard <keithp@neko.keithp.com>
51063 Date: Thu Dec 28 16:43:39 2006 -0800
51065 Oops, dont use symlinks for stolen xf86 header files
51067 commit 86da0d05e4e338e7f4c841577002eb1ba9f6b2ac
51068 Author: Keith Packard <keithp@neko.keithp.com>
51069 Date: Thu Dec 28 16:40:54 2006 -0800
51071 Add stolen xf86 header files to driver source
51073 commit 2b0a997e5b4b51267fa4f6725f8a965093392434
51074 Merge: bb238a8 9cfbf1c
51075 Author: Keith Packard <keithp@neko.keithp.com>
51076 Date: Thu Dec 28 16:34:38 2006 -0800
51078 Merge branch 'modesetting-origin' into modesetting
51080 commit bb238a8fc234a8e5e86cd2f42c58c9816a15732c
51081 Author: Keith Packard <keithp@neko.keithp.com>
51082 Date: Thu Dec 28 16:27:22 2006 -0800
51084 Configuration support: per-output mode lines, preferred mode.
51086 Add the modelines specified in the per-output monitor and all of the default
51087 modes to the list to each output. Prune the resulting list to specified sync
51088 limits and virtual sizes. Sort the resulting mode list on
51089 preferred/size/refresh.
51091 commit 9cfbf1ceda8f153438df1ba4b1712c55cc872017
51092 Author: Eric Anholt <eric@anholt.net>
51093 Date: Thu Dec 28 12:59:23 2006 -0800
51095 In EXA mode, force backing pixmaps into memory when doing XV to them.
51097 commit e889bde13dcc2438181a03319e204ae7b9235a78
51098 Author: Eric Anholt <eric@anholt.net>
51099 Date: Thu Dec 28 12:49:24 2006 -0800
51101 Draw textured video to the backing pixmap in the composited case.
51103 Currently, when the backing pixmap is not in framebuffer, we just BadAlloc
51104 rather than drawing garbage to the front buffer. This can be fixed with EXA.
51106 commit c7083a6f30fdf1859f60beba4c352bd790af4773
51107 Author: Eric Anholt <eric@anholt.net>
51108 Date: Thu Dec 28 11:53:33 2006 -0800
51110 Add WIP CH7017 LVDS driver. This is disconnected because it's untested.
51112 commit 778db49616b233f5a2f4f007ceb62a1a878a6e25
51113 Author: Eric Anholt <eric@anholt.net>
51114 Date: Thu Dec 28 11:36:15 2006 -0800
51116 Add WIP IVCH (i82807aa) driver.
51118 It is not currently hooked up as it hasn't been tested. The hardware that this
51119 was written for decided to stop working.
51121 commit 5806f670eb3a5171ffb0b75701640d5311b34c80
51122 Author: Eric Anholt <eric@anholt.net>
51123 Date: Thu Dec 21 16:22:14 2006 -0800
51125 Reformat the DVO drivers table to be a bit more normal style.
51127 commit b928cef9cd5b3f88dc0880f406300d246a59e0fe
51128 Author: Eric Anholt <eric@anholt.net>
51129 Date: Thu Dec 21 16:21:28 2006 -0800
51131 Move dvo driver detection into i830_dvo_init(), and use GPIOB for LVDS drivers.
51133 The documentation states that GPIOB is (generally) used for devices on DVOA
51134 on the motherboard, which appears to be the case on the laptop we have with
51135 LVDS on the motherboard.
51137 This patch is probably not entirely accurate, as there was apparently an LVDS
51138 DVO card sold that could be put in desktop machines, which would likely be on
51139 GPIOE like other ADD cards. Given that we couldn't find one of these cards for
51140 purchase, I'm not worrying about it.
51142 commit fabfccd3d8a55a83138d361f66c53cb7e6d371a8
51143 Author: Zou Nan hai <nanhai.zou@intel.com>
51144 Date: Thu Dec 28 14:48:54 2006 +0800
51148 commit ddefce4f02eb3272ad37e890350f73f03eff73c9
51149 Author: Zou Nan hai <nanhai.zou@intel.com>
51150 Date: Thu Dec 28 14:45:53 2006 +0800
51152 support NTSC 480i M-J 1280x1024 1024x768 800x600 640x480
51153 PAL 576i 1280x1024 1024x768 800x600 640x480
51155 I still have problem with non-interlace mode and Hi Res mode.
51156 also I don't know how to pickup those mode in xorg.conf
51158 commit c87462ded20904dbc6c6cbdb9547523b75fe5471
51159 Author: Eric Anholt <eric@anholt.net>
51160 Date: Wed Dec 27 19:14:14 2006 -0800
51162 Remove the nasty #define away of ErrorF in i830_video.c.
51164 Every new person touching this file hits this trap, so it's now removed.
51166 commit 6dc3387219b2b31e0c7dbaa88b568fbd54d72878
51167 Author: Eric Anholt <eric@anholt.net>
51168 Date: Wed Dec 27 18:58:40 2006 -0800
51170 Fix operator precedence issue that stuck 965 t-v state into the front buffer.
51172 commit 210f30bd75529d0064bc46af15464a2a47f9d4c6
51173 Author: Eric Anholt <eric@anholt.net>
51174 Date: Wed Dec 27 18:57:50 2006 -0800
51176 debugging for location of various 965 textured-video state.
51178 commit 82738e60769a4798d7425ffc67882d1f74043b76
51179 Author: Eric Anholt <eric@anholt.net>
51180 Date: Wed Dec 27 18:35:35 2006 -0800
51182 Fix discrepancy between XAA/EXA linear allocations that broke EXA XV.
51184 The XAA allocator returned an offset from the beginning of the frontbuffer
51185 allocation (where the linear allocations come from), while EXA returned one
51186 from the start of framebuffer. Now, both of them use the start of framebuffer
51189 commit 8165e043670f6a7fee5e7e8f33644adda5757591
51190 Author: Eric Anholt <eric@anholt.net>
51191 Date: Wed Dec 27 18:17:37 2006 -0800
51193 Bump the size of the extra space for 965 so it actually fits.
51195 commit f830c55e5f4efc97451240f8f18b49f46e8bce0f
51196 Author: Eric Anholt <eric@anholt.net>
51197 Date: Wed Dec 27 18:14:59 2006 -0800
51199 Make the assert() in i965_video.c actually work and expose an error.
51201 commit a75b416c91085df1db4027a25eafcd04fef1e57c
51202 Author: Eric Anholt <eric@anholt.net>
51203 Date: Wed Dec 27 17:58:57 2006 -0800
51205 Replace XAA mark/waitsyncs with the XAA/EXA wrappers.
51207 commit 150010d2f20f4a57f8507eba8762e36966b0f426
51208 Author: Eric Anholt <eric@anholt.net>
51209 Date: Wed Dec 27 17:57:34 2006 -0800
51211 reformat i965_video.c to 4-space indents, 80 columns, no trailing whitespace.
51213 commit 05066632b068ae91d7be77e3a199ea0d337d975b
51214 Author: Eric Anholt <eric@anholt.net>
51215 Date: Wed Dec 27 17:41:45 2006 -0800
51217 Move 965 textured-video out to a separate file.
51219 commit 46df75ccd4647ea033583130253a2a2218b34a20
51220 Author: Eric Anholt <eric@anholt.net>
51221 Date: Wed Dec 27 13:45:06 2006 -0800
51223 Bug #7524: Major improvements to EXA/XAA static memory allocation.
51225 With this, we no longer allocate XV through the XF86 linear allocator in the
51226 EXA case. We also no longer allocate extra space for the XF86 2D allocator
51227 that we don't use in EXA mode, or space for the EXA allocator in XAA mode.
51228 The EXA offscreen allocator now gets enough space for several screenfuls of
51229 pixmaps plus one 1920x1088 movie.
51231 A large duplicated section of code for allocating each framebuffer in the old
51232 dual-screen mode was also factored out.
51234 commit 84915ac8afeb4bbc03df8f94ab3ba351788d6501
51235 Author: Keith Packard <keithp@guitar.keithp.com>
51236 Date: Thu Dec 21 23:42:53 2006 -0800
51238 If DDC detailed modes are missing physical size, pull from features.
51240 Monitors without detailed modes, or those which do not bother to set a
51241 physical size in their detailed modes may still have physical size in the
51244 commit 1e9105395a7900521206b28bae4bb4f85669a0cd
51245 Merge: df9ecf8 c237e93
51246 Author: Keith Packard <keithp@guitar.keithp.com>
51247 Date: Thu Dec 21 23:34:30 2006 -0800
51249 Merge branch 'modesetting-origin' into modesetting
51251 commit df9ecf8ab93f0aeffde11d438513d41220ba75cb
51252 Author: Keith Packard <keithp@guitar.keithp.com>
51253 Date: Thu Dec 21 23:32:35 2006 -0800
51255 Change belinea edid quirk to cover the 10 20 30W model as well.
51257 This larger model reported different (but still incorrect)
51258 sync polarities, so instead of flipping them, just set them to the right
51261 commit c237e930615f1b048895efe988d28503f5d67c37
51262 Author: Keith Packard <keithp@ukulele.keithp.com>
51263 Date: Thu Dec 21 23:30:37 2006 -0800
51265 Use Vesa DDC timeouts for all I2C busses.
51267 The default I2C timeouts are very short, and while most
51268 modern monitors have no trouble responding at that rate,
51269 some older ones cannot manage.
51271 commit d8c5dba4d797fc50d7b2b5855f34e2d2e2ad3e4f
51272 Author: Keith Packard <keithp@guitar.keithp.com>
51273 Date: Thu Dec 21 21:20:43 2006 -0800
51275 When cleaning duplicate modes, make sure ->Last is reset correctly.
51277 When removing the very last mode for a monitor, move the ->Last pointer to
51278 the previous list element.
51280 commit fab9a6b6210daea423b609208ef57fa26571f5d3
51281 Author: Keith Packard <keithp@guitar.keithp.com>
51282 Date: Thu Dec 21 21:19:32 2006 -0800
51284 Add EDID quirk support for broken EDID data.
51286 For EDID with known errors, add a quirk mechanism to automatically
51287 compensate. The first quirk is for a Belinea 1440x900 monitor which
51288 incorrectly specifies sync polarities in the detailed mode.
51290 commit d9b27667e6cc6c7e171b0f513d40be7658cf4574
51291 Author: Keith Packard <keithp@guitar.keithp.com>
51292 Date: Thu Dec 21 21:17:03 2006 -0800
51294 PLL computations missed one possible 'm2' value.
51296 m2 was ranging from min <= m2 < max instead of <= max resulting in
51297 inaccurate PLL frequencies for some modes.
51299 commit c5205595f2abacc9f736e1f53666302ec551cbc4
51300 Merge: b8f3ec7 4c0c1aa
51301 Author: Keith Packard <keithp@neko.keithp.com>
51302 Date: Thu Dec 21 02:51:33 2006 -0800
51304 Merge branch 'modesetting' into crestline
51306 commit 4c0c1aa882cfec77b2183baec93cbc4cfaf4abe0
51307 Author: Keith Packard <keithp@guitar.keithp.com>
51308 Date: Thu Dec 21 02:33:39 2006 -0800
51310 Computed corred color conversion values.
51312 Extract correct color conversion values for all video formats from
51313 documentation. Use those, with appropriate conversions, for the color
51314 conversion register values.
51316 commit 98fd44d681220aa31200e4262f1a7ec952a09530
51317 Author: Keith Packard <keithp@guitar.keithp.com>
51318 Date: Thu Dec 21 01:24:24 2006 -0800
51320 TV subcarrier was computed from wrong clock value.
51322 The constants provided in the documentation for the subcarrier DDA values
51323 assumed the clock was programmed to precisely 108MHz, but the PLL can't hit
51324 that value exactly (or our PLL computation can't, in any case). The
51325 result was an incorrect subcarrier frequency which resulted in synthetic
51326 subcarrier phase shift and a lovely rainbow effect on the screen.
51328 Unfortunately, the documentation didn't exactly describe the function
51329 performed by the subcarrier clock hardware, so a bit of detective work was
51330 needed. New constants were computed using the code in tv.5c and those, along
51331 with lots of other values from the documentation were inserted into the
51334 The result appears to generate stable NTSC video on the svideo connector.
51336 commit b8f3ec77104882fc755ddc88f2583cd5d15387a4
51337 Author: Zou Nan hai <nanhai.zou@intel.com>
51338 Date: Thu Dec 21 10:46:17 2006 +0800
51340 This fix should be better, I830SetMode will call i830_sdvo_dump,
51341 if the sdvo is not initialized, X will crash. So any application which
51342 SetMode will crash X if there is no SDVO output.
51344 commit b83cbec34a8d1ff39bd303fcd27fc03dcdb1d6c3
51345 Author: Zou Nan hai <nanhai.zou@intel.com>
51346 Date: Wed Dec 20 17:32:05 2006 +0800
51348 fix null pointer reference,
51349 I hit this when running doom3-demo
51351 commit 4ba72fc408e7fab7c384aff5e73f0cfb8fd86cf3
51352 Author: Keith Packard <keithp@guitar.keithp.com>
51353 Date: Tue Dec 19 22:39:57 2006 -0800
51355 Reinitialize DGA mode list whenever we update the global list.
51357 DGA has a copy of the current mode list (yes, this is broken).
51358 Regenerate it whenever the ddx mode list changes.
51360 commit 07d3f141712e2abf704cb68a59cd5d5c3390e2ca
51361 Merge: 0cd1a11 c28075e
51362 Author: Keith Packard <keithp@neko.keithp.com>
51363 Date: Tue Dec 19 22:01:02 2006 -0800
51365 Merge branch 'modesetting' into crestline
51367 commit c28075e1d70c98b39fc9829a6a01da171a9b9426
51368 Author: Keith Packard <keithp@guitar.keithp.com>
51369 Date: Tue Dec 19 21:53:10 2006 -0800
51371 TV output I830OutputPrivate had wrong type (crashed in SetMode).
51373 TV output private type field was set to I830_OUTPUT_SDVO instead of
51374 I830_OUTPUT_TVOUT. When DGA mode setting occurred, the sdvo output debug
51375 code would be invoked and attempt to dereference sdvo-specific bits of the
51376 private structure leading to a segfault.
51378 commit 2ef4c5e8f6444aad192304e5a2f7a0c77bfb917d
51379 Author: Keith Packard <keithp@guitar.keithp.com>
51380 Date: Tue Dec 19 16:45:39 2006 -0800
51382 Re-initialize physical screen size only if -dpi was passed.
51384 Physical screen size from the config file or DDC will already
51385 be set correctly in the screen structure, unless it was computed from the
51386 virtual size using the -dpi command line option. Recompute physical size as
51387 we reset the screen size if the -dpi option was used.
51389 commit b00951a5c87ca8c596ee9cc73d96d598eec60751
51390 Merge: 05cd921 c0a0ddc
51391 Author: Keith Packard <keithp@guitar.keithp.com>
51392 Date: Tue Dec 19 16:22:04 2006 -0800
51394 Merge branch 'modesetting-origin' into modesetting
51396 commit 05cd921140fae2746c3421715d7f18248a66735a
51397 Author: Keith Packard <keithp@guitar.keithp.com>
51398 Date: Tue Dec 19 16:21:46 2006 -0800
51400 Align textured video dither matrix to window
51402 commit c0a0ddc0ce69b60ac3a58d45cb32ccd0e0bfb0c9
51403 Author: Eric Anholt <eric@anholt.net>
51404 Date: Tue Dec 19 16:06:13 2006 -0800
51406 Bug #9382: Save the CRTC's desired mode in the old RandR 1.0 mode set path.
51408 commit 3c86fdda1d09fd22d4c15a8e412611db4cea7dcf
51409 Author: Keith Packard <keithp@guitar.keithp.com>
51410 Date: Tue Dec 19 15:53:07 2006 -0800
51412 Turn on dithering for 915 textured video
51414 commit 293ad158ece941eeb9894f5af4d663464b9954e5
51415 Author: Eric Anholt <eric@anholt.net>
51416 Date: Tue Dec 19 14:35:51 2006 -0800
51418 Remove 8k framebuffer stride restriction on 965.
51420 Also, clean up the message if the limitation is hit.
51422 commit ff1223142aedcf888665de6e4d7d5d03a55b2688
51423 Author: Eric Anholt <eric@anholt.net>
51424 Date: Tue Dec 19 12:49:55 2006 -0800
51426 Only erase curMode in i830DisableUnusedFunctions when the CRTC is disabled.
51428 This bug resulted in broken xinerama, among other issues.
51430 commit 3b5703c7fa08e63f2a31be7496ded9c70015946d
51431 Author: Daniel Stone <daniel@fooishbar.org>
51432 Date: Tue Dec 19 09:49:49 2006 -0800
51434 Respect the srcdir.
51436 commit d89e211b12ff715d7475421af69157b102accaca
51437 Author: Eric Anholt <eric@anholt.net>
51438 Date: Fri Dec 15 14:39:14 2006 -0800
51440 Disable configure-disabled outputs before mode setting.
51442 We're not supposed to turn off pipes/planes/dplls with outputs attached, which
51443 we could have done before this.
51445 commit 3fc9feec7b29e0b908e3b90e0f6634117d66ba6d
51446 Author: Eric Anholt <eric@anholt.net>
51447 Date: Fri Dec 15 14:38:23 2006 -0800
51449 Simplify i830DisableUnusedFunctions() by using the CRTC dpms routines.
51451 commit 7b8056e3e347b15b37be0116315d7d3a23aef99f
51452 Author: Keith Packard <keithp@ukulele.keithp.com>
51453 Date: Tue Dec 19 01:32:54 2006 -0800
51455 Writing 1 to I2C line means to tristate the bus so others can manipulate it.
51457 We were forcing bus lines to 1 which was breaking DDC for some monitors.
51458 Instead, make the PutBits function just tristate when writing 1 bits.
51460 commit 0cd1a11d4493ef4a2d05ccf8aadec743504dba5f
51461 Author: Zou Nan hai <nanhai.zou@intel.com>
51462 Date: Tue Dec 19 14:03:19 2006 +0800
51466 commit eedef7adc53dd8337d27c02551c5778fb43bae05
51467 Author: Eric Anholt <eric@anholt.net>
51468 Date: Mon Dec 18 15:57:44 2006 -0800
51470 Add a detect() function for DVO chips, and implement it on sil164.
51472 commit 8983845f91cacf8110c70121e0f5f293fe443e6d
51473 Author: Eric Anholt <eric@anholt.net>
51474 Date: Mon Dec 18 15:57:08 2006 -0800
51476 Fix crash in xf86SetScrnInfoModes when pScrn->modes ends up empty.
51478 commit fa4642048b183134544fc5ee47558446d27f6194
51479 Author: Eric Anholt <eric@anholt.net>
51480 Date: Mon Dec 18 15:32:35 2006 -0800
51482 Clean up i2c_vid interface, including de-StudlyCapsing and removing dead code.
51484 The old Init() function is removed and the previous Detect() function is now
51485 init(). This leaves us room in the namespace for a detect() like other
51486 outputs have ("is the monitor connected?"). Also, Power() became dpms(),
51487 taking a DPMSMode*. In general, the mode setting path now matches the intel
51488 internal path, except for the lack of mode_fixup().
51490 commit dcb069a1da6b3300b0772843dd9b67efdf319637
51491 Author: Dave Airlie <airlied@linux.ie>
51492 Date: Tue Dec 19 10:29:56 2006 +1100
51494 ch7xxx: add lookup table for name
51496 Don't load if we don't match name
51498 commit 0305298405bd7018e5c520bdd5ea46a75be09f37
51499 Author: Eric Anholt <eric@anholt.net>
51500 Date: Mon Dec 18 14:49:01 2006 -0800
51502 Limit modes to 165Mhz on CH7009, according to spec.
51504 commit 23114fbccad849249b4b7200e3d085a9ac0e2e99
51505 Author: Eric Anholt <eric@anholt.net>
51506 Date: Mon Dec 18 14:47:44 2006 -0800
51508 Align the driver to the CH7009[AB] spec.
51510 With this, the driver works at 640x480 and 1280x1024 on my panel.
51512 commit 633d718f6fb019844304228a81c84a8605da004f
51513 Author: Eric Anholt <eric@anholt.net>
51514 Date: Mon Dec 18 13:59:17 2006 -0800
51516 Clean up a bunch of log noise from ch7xxx startup.
51518 commit 3167bc6b24bd74c50ca4c4d361b14aa24d77cfa3
51519 Author: Eric Anholt <eric@anholt.net>
51520 Date: Mon Dec 18 13:45:55 2006 -0800
51522 Fix DVO mode valid function to not throw out all modes.
51524 commit 4ca92da5b93b07f0bff51843f8fb5d04a2a29443
51525 Author: Eric Anholt <eric@anholt.net>
51526 Date: Mon Dec 18 13:45:34 2006 -0800
51528 Add save/restore to ch7xxx to avoid segfaults on server start.
51530 commit 31849edb3b503ade659981dfa36babf5ad8757c5
51531 Author: Eric Anholt <eric@anholt.net>
51532 Date: Mon Dec 18 13:32:21 2006 -0800
51534 Move the ch7xxx code to the style I've been settling on for output code.
51536 commit 1975fa5b010100196af201e40f43b30a149b7750
51537 Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
51538 Date: Mon Dec 18 14:08:55 2006 -0500
51540 Revert "Convert callers of LookupWindow() to dixLookupWindow()."
51542 This reverts commit 08575331ac75783c9910cfb6e78db701a29983ac.
51544 commit 2d0ca2202cb0d780bb4387bbe04e5caa6512e3b1
51545 Author: Eric Anholt <eric@anholt.net>
51546 Date: Mon Dec 18 10:37:33 2006 -0800
51548 Bug #8786: Treat pictures with no color data as non-component alpha.
51550 This is an un(der?)-documented part of the render protocol: If the mask is
51551 alpha only, then the component alpha flag is ignored on it.
51553 commit 65cd18b97bc533b24031d6eb882fd3074779ceba
51554 Merge: bf43f8c 6823ca8
51555 Author: Keith Packard <keithp@neko.keithp.com>
51556 Date: Sun Dec 17 17:32:24 2006 -0800
51558 Merge branch 'modesetting' into crestline
51560 commit 6823ca87f3b1ef3b28ed167254dcfce2a80467df
51561 Author: Keith Packard <keithp@mandolin.keithp.com>
51562 Date: Sat Dec 16 21:12:47 2006 -0800
51564 Follow mode setting order in RestoreHWState.
51566 Add delays after output and CRTC disable. Restore panel fit register before
51567 PLLs are restarted. Move all VGA restore code last. Shuffle various register
51568 writes around and add delays to match PipeSetMode code.
51570 commit 86558cc622b516b568cc26efdf9b64d4b660f50f
51571 Author: Keith Packard <keithp@mandolin.keithp.com>
51572 Date: Sat Dec 16 21:09:31 2006 -0800
51574 Prefer earliest CRTC when mapping to outputs.
51576 For some reason, the code was preferring the last possible output when
51577 mapping outputs to crtcs. Use the earlier CRTC instead to make the i830
51578 driver consistent with BIOS usage.
51580 commit 8e6ab99b3195325f9fe5432725fe328591c0c7e2
51581 Author: Keith Packard <keithp@mandolin.keithp.com>
51582 Date: Sat Dec 16 21:06:36 2006 -0800
51584 Not restoring active outputs. Wait for input sync before enabling outputs.
51586 Oops--looks like a typo to me; the code was callint set_target_output
51587 instead of set_active_outputs.
51589 BIOS loops waiting for the SDVO input to sync before enabling outputs, this
51592 commit bffd611b0a1cb05868e0f93e6ff9357a3116eaa6
51593 Author: Keith Packard <keithp@mandolin.keithp.com>
51594 Date: Sat Dec 16 21:03:15 2006 -0800
51596 Follow BIOS order in writing DPLL/DPLL_MD registers.
51598 965 BIOS writes DPLL and then DPLL_MD.
51599 945 BIOS writes DPLL twice.
51601 commit 9b1a1b170befae2e705c23ce295837d0d13b60c0
51602 Author: Keith Packard <keithp@mandolin.keithp.com>
51603 Date: Sat Dec 16 16:53:49 2006 -0800
51605 reorder restore writes for 965 VGA
51607 commit 08575331ac75783c9910cfb6e78db701a29983ac
51608 Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
51609 Date: Fri Dec 15 17:44:44 2006 -0500
51611 Convert callers of LookupWindow() to dixLookupWindow().
51613 commit 60411bc4d0b3c53850c73b7246d5f7ed5c2d4084
51614 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51615 Date: Thu Dec 14 18:36:44 2006 -0800
51617 Follow BIOS PLL write protocol (disable,enable,enable).
51619 This gets VGA output and text restore working, but SDVO is still not working.
51621 commit 082519f34b468d0c5aa08db74ff2b51cf411d743
51622 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51623 Date: Thu Dec 14 16:51:48 2006 -0800
51625 Turn pll/pipe/plane on in crtc_set_mode
51627 Instead of delaying pll/pipe/plane enables to the dpms function, turn them
51628 on right away in the crtc_set_mode function. To avoid rewriting these
51629 registers in the subsequent dpms function, check each register written there
51630 to see if the enable bit is already on and don't rewrite.
51632 commit 55e7a32096a58f7bb2380b04df6b3dd9dc8e5b6a
51633 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51634 Date: Thu Dec 14 16:50:26 2006 -0800
51636 Respect Virtual in xorg.conf
51638 Instead of growing virtual to fit our desired sizes, if it is set in the
51639 config file, use the value as the virtual size and make the default
51640 configuration fit within that size (if possible).
51642 commit d4142abc68888178a8f6c54fdcece91e898bc9f5
51643 Author: Keith Packard <keithp@neko.keithp.com>
51644 Date: Thu Dec 14 13:49:08 2006 -0800
51646 Set frame buffer size ranges in xf86_config.
51648 Initialization and default configuration code needs to know the range of
51649 legitimate sizes for the frame buffer.
51651 commit 2b5a5542ad7d1c558d1684a0a69b66bab6e9d001
51652 Author: Keith Packard <keithp@neko.keithp.com>
51653 Date: Thu Dec 14 13:43:50 2006 -0800
51655 Change TV modes from Preferred to Driver.
51657 This avoids having the artificial TV mode size drive overall screen size.
51659 commit 3fe802453a85183a69c36a098639895f49b17df1
51660 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51661 Date: Wed Dec 13 13:15:14 2006 -0800
51663 Move xf86CrtcConfig to ScrnInfo private.
51665 Pull xf86CrtcConfig out of the driver private structure and allocate a
51666 ScrnInfo private index for it. Also, make the arrays of outputs and crtcs
51667 dynamic instead of fixed.
51669 commit 0f6addc8a6aeb9bd041d0f8e8e5850e76764ba51
51670 Merge: 001e272 3fa5b39
51671 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51672 Date: Wed Dec 13 12:10:31 2006 -0800
51674 Merge branch 'modesetting-origin' into modesetting
51676 commit 001e272437a0247ffbc4b9ff8a3f2b437cf4c533
51677 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51678 Date: Wed Dec 13 12:08:58 2006 -0800
51680 Auto-detect working TV output by checking TV regs functionality.
51682 We can't figure out which chips are supposed to have TV out, so instead we
51683 prod the TV_DAC register to see if it will hold the value written to it, if
51684 not, we assume the chip doesn't have TV out.
51686 commit 3fa5b3998353518c40e2fb0c28b425ee22c8d625
51687 Author: Eric Anholt <eric@anholt.net>
51688 Date: Wed Dec 13 11:54:36 2006 -0800
51690 Add magic double-write of the dpll register to fix mac mini cold boot.
51692 commit e27372e85a007d8a7e31678dbdb62755c22918bf
51693 Author: Eric Anholt <eric@anholt.net>
51694 Date: Wed Dec 13 00:26:27 2006 -0800
51696 Add some caution with PCI write posting and DPLL delays to i830_crtc_dpms.
51698 While it doesn't specifically help/hurt my test case, we've seen enough
51699 mysterious behavior that caution is probably warranted.
51701 commit df0a5a25aa72563a33731f8063602dd7faad8459
51702 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51703 Date: Wed Dec 13 11:38:18 2006 -0800
51705 Rename outputs to TMDS-1/TMDS-2, LVDS, TV and VGA.
51707 These names are reasonably short and describe the connector rather than the
51708 target media, but we don't appear to have any way to determine what is on
51709 the other end of the wire. More importantly, they're all unique now.
51711 commit e3604fc63243ab0f31673a923d20e23131b607f6
51712 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
51713 Date: Wed Dec 13 15:42:53 2006 +0000
51717 commit 4cdcaac26b3f425ec8fcc2fd3eaa9bfb7a137422
51718 Author: Keith Packard <keithp@mandolin.keithp.com>
51719 Date: Tue Dec 12 23:26:00 2006 -0800
51721 Write dpll_md register while updating dpll as that appears to be required.
51723 SDVO multiplier on 965 is in the dpll_md register; for some reason, that
51724 needs to be written along with the dpll value or the multiplier doesn't get
51727 commit 9dd1520421dcb8c7f36a684f3d2ffb6b42a6671a
51728 Merge: d57a258 63bf484
51729 Author: Keith Packard <keithp@guitar.keithp.com>
51730 Date: Tue Dec 12 22:52:59 2006 -0800
51732 Merge branch 'modesetting-origin' into modesetting
51734 commit d57a25815398ae83eae8bdcb3a1b607760aa30b6
51735 Author: Keith Packard <keithp@guitar.keithp.com>
51736 Date: Tue Dec 12 22:48:21 2006 -0800
51738 Elide duplicate modes in pScrn->monitor.
51740 xf86DDCMonitorSet dumps all of the DDC-discovered modes
51741 into the monitor mode list without checking to see if they
51742 are already present. This provides an ever-changing list of
51743 modes for outputs which have no DDC and which simply duplicate the monitor
51746 commit 63bf48481d4e5e285fdfe7037ea4150283163a54
51747 Author: Eric Anholt <eric@anholt.net>
51748 Date: Tue Dec 12 22:42:41 2006 -0800
51750 Correct typo in mac mini hack, and return modes when the hack is unnecessary.
51752 commit 41444183b59ed84c09749ca89afbef036d42ec5f
51753 Author: Eric Anholt <eric@anholt.net>
51754 Date: Tue Dec 12 18:08:57 2006 -0800
51756 Replace custom, partially broken DPMS implementation with a generic one.
51758 commit 7ed1b05922c07ff45a5794a992fd3d59ab55aa73
51759 Author: Eric Anholt <eric@anholt.net>
51760 Date: Tue Dec 12 16:03:52 2006 -0800
51762 Fix mac mini SDVO output: write the SDVO[BC] enabled register state twice.
51764 commit 9776f6c68b3cdd5585e58e677c1b1318d9aedaf4
51765 Author: Eric Anholt <eric@anholt.net>
51766 Date: Tue Dec 12 16:02:47 2006 -0800
51768 Flush the plane changes in i830_crtc_dpms()
51770 Otherwise, the changes may not have taken effect.
51772 commit 54823ac39ce9666099d69196694643f04123cb4d
51773 Author: Eric Anholt <eric@anholt.net>
51774 Date: Tue Dec 12 14:38:08 2006 -0800
51776 Move the SDVO sync detection to after we've turned the port on.
51778 commit ec45d7274333cccd20ad080facc8e995ad7b26b0
51779 Author: Eric Anholt <eric@anholt.net>
51780 Date: Tue Dec 12 13:54:49 2006 -0800
51782 Extend the error state reporting to cover ESR and decode PGTBL_ERR for 945.
51784 commit 838af10b85ee8c085c6d2c4fbe18fd90b2be9d3f
51785 Author: Eric Anholt <eric@anholt.net>
51786 Date: Tue Dec 12 13:45:41 2006 -0800
51788 Move PrintErrorState and CheckInheritedErrors code to i830_debug.c.
51790 commit 668ae6ce3c3c4b93a58749fd44ef32451adb6f9c
51791 Author: Eric Anholt <eric@anholt.net>
51792 Date: Tue Dec 12 11:39:40 2006 -0800
51794 Set the SDVO port to on before sending the SDVO output on command.
51796 commit 91a538beadc17155be395ea2c8109a0958a969bd
51797 Author: Eric Anholt <eric@anholt.net>
51798 Date: Sat Dec 9 15:58:47 2006 -0800
51800 More debugging output for SDVO.
51802 commit ce54538b905eb329f45c1b9b15e95ddce4a7927f
51803 Author: Eric Anholt <eric@anholt.net>
51804 Date: Fri Dec 8 08:26:30 2006 -0800
51806 Add a hack for DDC on my Mac Mini.
51808 commit c8c1fb64ccecfb88f9923ad65c9898aa44692260
51809 Author: Eric Anholt <eric@anholt.net>
51810 Date: Mon Dec 11 16:02:29 2006 -0800
51812 Move the overlay mode switch status updates to when we enable/disable crtcs.
51814 This should help if we mode switch with randr during video playback, and also
51815 simplifies the code by moving the protection all to one place.
51817 commit 3f3791ba2ece8e8a7aa3a5676f5f6361862a292c
51818 Author: Eric Anholt <eric@anholt.net>
51819 Date: Mon Dec 11 15:20:15 2006 -0800
51821 Remove remnants of old savedCurrentMode hack. Man, that was gross.
51823 commit c21baa5a3044eb66f2518b018dae07eadb989aa1
51824 Author: Eric Anholt <eric@anholt.net>
51825 Date: Mon Dec 11 15:15:01 2006 -0800
51827 Remove dead remnants of old clone mode support.
51829 commit b3865f34669319f449f2da8f87a4eef6ec8e7804
51830 Author: Joshua J. Berry <des@condordes.net>
51831 Date: Mon Dec 11 14:46:20 2006 -0800
51833 Include i830.h to get M_T_* compat defines for older X Servers.
51835 Not duplicating the defines ensures that we won't let the compat bits slip
51836 into the X Server when we move this code over there.
51838 commit 24d9521c7ae72e146025f10ea85b07ebd7041716
51839 Author: Eric Anholt <eric@anholt.net>
51840 Date: Mon Dec 11 09:34:50 2006 -0800
51842 Add error messages for previous silent DRI initialization failures.
51844 commit bf43f8ce736d91659c2757fb39b13e63bc5891a4
51845 Merge: 684ea26 9452321
51846 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
51847 Date: Fri Dec 8 14:17:43 2006 +0800
51849 Merge branch 'modesetting' into crestline
51851 commit 94523215127a3458a28d0f1217224a27c8870884
51852 Merge: 75f4df2 fde52de
51853 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
51854 Date: Fri Dec 8 14:15:06 2006 +0800
51856 Merge branch 'modesetting' of git+ssh://zhen@git.freedesktop.org/git/xorg/driver/xf86-video-intel into modesetting
51858 commit 684ea26b740d07ec8f6b0d223c1171e70c073bfa
51859 Merge: 64269de 6b977f0
51860 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
51861 Date: Fri Dec 8 14:03:10 2006 +0800
51863 Merge branch 'exa-i965' of git+ssh://zhen@git.freedesktop.org/git/xorg/driver/xf86-video-intel into crestline
51865 commit 99f6f84076b25af69dc226e6d4704c74d207e657
51866 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
51867 Date: Thu Dec 7 09:47:34 2006 +0000
51871 commit fde52de870c84821ab457e17634c334a10cf71ab
51872 Merge: 7dbe6c3 85de8a0
51873 Author: Keith Packard <keithp@guitar.keithp.com>
51874 Date: Wed Dec 6 23:40:13 2006 -0800
51876 Merge branch 'modesetting-origin' into modesetting
51878 commit 7dbe6c378d1bc0c06ed6be4c1f9ca38ac31a5571
51879 Author: Keith Packard <keithp@guitar.keithp.com>
51880 Date: Wed Dec 6 23:39:14 2006 -0800
51882 Move vga restore to end of RestoreHWState.
51884 I was having consistent system lockups when the vga restore
51885 was first. Moving it to the end has reduced them to an infrequent
51886 occurrence (but, alas, it has happened once since). This does not
51889 commit 28c7002b1cf81e286a083384575d53021d3ce906
51890 Author: Keith Packard <keithp@guitar.keithp.com>
51891 Date: Wed Dec 6 23:35:15 2006 -0800
51893 Use DOUBLE_WIDE mode on all hardware.
51895 DOUBLE_WIDE mode is needed when the pixel clock is > 90% of the core
51896 clock rate. The code guesses what the core clock rate is based on
51897 the bus (AGP -> 133MHz, PCI-E -> 200MHz).
51899 commit 85de8a05804eaebcd61430793ef5406f196f9a95
51900 Author: Eric Anholt <eric@anholt.net>
51901 Date: Wed Dec 6 20:26:53 2006 -0800
51903 Include render.h to get SubPixelUnknown define.
51905 commit 64269de3c34db047e78b788dc5681abdc009f8aa
51906 Merge: 8ddbc26 182d2f5
51907 Author: Keith Packard <keithp@neko.keithp.com>
51908 Date: Wed Dec 6 18:09:28 2006 -0800
51910 Merge branch 'modesetting' into crestline
51912 commit 182d2f503a7a1be90d93a4a15059585a564fd3ce
51913 Merge: 677c454 6777d80
51914 Author: Keith Packard <keithp@neko.keithp.com>
51915 Date: Wed Dec 6 17:18:58 2006 -0800
51917 Merge branch 'modesetting-origin' into modesetting
51919 commit 6777d8044d5cc063698e05afb7a93a58a7a3b64a
51920 Author: Eric Anholt <eric@anholt.net>
51921 Date: Wed Dec 6 17:10:06 2006 -0800
51923 Harmless warning fix.
51925 commit c6ce8a3c3355e8a56e86856caa56baf1cd3c0231
51926 Author: Eric Anholt <eric@anholt.net>
51927 Date: Wed Dec 6 17:07:48 2006 -0800
51929 Reduce dumpregs to X startup, after a mode set, and LeaveVT.
51931 While here, remove some other register dumping that is better done by
51934 commit b23dec37b28a76433ad5d537ab508294e843cabe
51935 Author: Eric Anholt <eric@anholt.net>
51936 Date: Wed Dec 6 16:58:24 2006 -0800
51938 Save/restore registers around calls in PreInit with register side-effects.
51940 commit d3c21c09e3904b3d32aca7563044194b061ac2fa
51941 Author: Eric Anholt <eric@anholt.net>
51942 Date: Wed Dec 6 16:50:52 2006 -0800
51944 Remove stale i830_tv_pre_set_mode. Replaced by generic dpms hook call.
51946 Also replaced by not spamming unrelated ADPA register. At least, it had sure
51947 better not be related.
51949 commit 9c3e733aaa2068fcb0164577237ed70d177e9b5a
51950 Author: Eric Anholt <eric@anholt.net>
51951 Date: Wed Dec 6 16:36:42 2006 -0800
51953 Save/restore more TV registers.
51955 commit d5ec9d2160f47f21a6015c1cc05b57274cbb0471
51956 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51957 Date: Wed Dec 6 14:44:40 2006 -0800
51959 Cannot modify DPLL register in output mode_set function.
51961 DPLL register contains clock setup, so it cannot be written
51962 from the output mode_set function or 'bad things' happen.
51964 commit 5d345c13a949caef384a63769d7185885de6d26d
51965 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51966 Date: Wed Dec 6 13:19:58 2006 -0800
51968 Clean up SDVO multiplier debug output
51970 commit 6ee63364f5fabbc5578bcc9ded38c778595f5a6d
51971 Merge: 56f7119 b75ecdb
51972 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
51973 Date: Wed Dec 6 12:53:22 2006 -0800
51975 Merge branch 'modesetting' into nonrandr-setup
51977 Also, fix buffer overflow in i830_debug.c
51979 commit b75ecdb48309a15eb7c52b279c7f8523a95bcd48
51980 Merge: 1cb6311 b3bb10e
51981 Author: Eric Anholt <eric@anholt.net>
51982 Date: Wed Dec 6 11:59:33 2006 -0800
51984 Merge branch 'generic-mode-set' into modesetting
51986 commit b3bb10e33e44c78f132d239e30931f97065a9fd6
51987 Author: Eric Anholt <eric@anholt.net>
51988 Date: Wed Dec 6 11:53:40 2006 -0800
51990 Fix LVDS: Don't change bits in LVDS other than port enable/pipe select.
51992 This behavior matches what we did before when LVDS worked.
51994 commit 47d07b1073c162ec00a8b173b8b6389dc4adbe88
51995 Author: Eric Anholt <eric@anholt.net>
51996 Date: Wed Dec 6 10:56:38 2006 -0800
51998 More LVDS fixed mode fixing: use the EDID DTD's blank length.
52000 commit 727bf1cbf72362edcbdd68001a3558fd2b2b4eca
52001 Author: Eric Anholt <eric@anholt.net>
52002 Date: Wed Dec 6 10:12:49 2006 -0800
52004 Add missing newlines to new debug output.
52006 commit 236c53be5d94798d55219651a6885fee32ce175b
52007 Author: Eric Anholt <eric@anholt.net>
52008 Date: Wed Dec 6 10:10:27 2006 -0800
52010 Fix LVDS fixed mode code after generic-mode-set.
52012 commit 1cb6311c1182fe98d2b8d237cef42509c3178f9a
52013 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52014 Date: Wed Dec 6 16:31:00 2006 +0800
52016 fix typo in checking xf86CrtcScreenInit return
52018 commit 56f71194157ef929b62fe34a89c840bd59e56301
52019 Author: Keith Packard <keithp@mandolin.keithp.com>
52020 Date: Wed Dec 6 00:13:57 2006 -0800
52022 Use xf86 structures for default config instead of RandR.
52024 To avoid requiring RandR 1.2 in the X server, use the
52025 xf86 Crtc and Output structures as the basis for the default configuration
52026 computation (and, eventually, the config-file based configuration as well).
52028 commit 6b977f0eca388a295ca7348f5960bd8e0192cf1d
52029 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52030 Date: Wed Dec 6 13:24:44 2006 +0800
52034 commit ffd8aacbe7c72b696ff7257609e3c1d45c057609
52035 Author: Keith Packard <keithp@mandolin.keithp.com>
52036 Date: Tue Dec 5 19:48:39 2006 -0800
52038 new server APIs are not in server-1.2 (version 7.1.99.2)
52040 commit a73649611414cb92697575efa90addfc3cf91cf9
52041 Author: Keith Packard <keithp@mandolin.keithp.com>
52042 Date: Tue Dec 5 19:47:04 2006 -0800
52044 Capture VGA registers sooner so checks are more informative
52046 commit 62fba80d3ea9c185ec8ae1e06c099c02bf844d8b
52047 Author: Keith Packard <keithp@mandolin.keithp.com>
52048 Date: Tue Dec 5 19:43:41 2006 -0800
52050 fix up misleading comment in i830_crt.c about border register
52052 commit 041f0029c78620ca1b50e7e5d11f5d022e189641
52053 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52054 Date: Wed Dec 6 10:43:29 2006 +0800
52058 We should use card_fmt for src/mask picture, and use dest color
52059 buffer format helper. Also fix wrong name for G965 texture formats,
52060 and pict_x1r5g5b5 isn't supported by sampler engine.
52062 commit 1d94ec7de3387b70815679977cfc5d88200efa99
52063 Author: Eric Anholt <eric@anholt.net>
52064 Date: Tue Dec 5 18:22:59 2006 -0800
52066 Fix ordering of PIPE[AB]SRC debug output to be (x, y).
52068 commit eee23fdd0dadd38a2b80525ec886120150f91077
52069 Author: Eric Anholt <eric@anholt.net>
52070 Date: Tue Dec 5 18:19:49 2006 -0800
52072 Fix copy'n'paste-o of the ordering of h/v fields in DSP[AB]SIZE.
52074 commit 58c247c6de56508b00eadf77b4c145c94134270d
52075 Merge: 0b4c3e7 2e8c927
52076 Author: Eric Anholt <eric@anholt.net>
52077 Date: Tue Dec 5 15:28:21 2006 -0800
52079 Merge branch 'modesetting-origin' into modesetting
52085 commit 0b4c3e7bff790e1e99ace5036a41e96046335b04
52086 Author: Eric Anholt <eric@anholt.net>
52087 Date: Tue Dec 5 14:39:47 2006 -0800
52089 Add a bunch of per-register debug code to i830DumpRegs().
52091 commit 7844e576e71c004ed495bbc31f7bbe890d18b8d9
52092 Author: Eric Anholt <eric@anholt.net>
52093 Date: Tue Dec 5 14:39:09 2006 -0800
52095 Move reg dump from i830PipeSetMode to after we set up the modes with RandR.
52097 commit e777d38ce98d7220621b049b09df1deca5a5df42
52098 Author: Eric Anholt <eric@anholt.net>
52099 Date: Tue Dec 5 10:01:31 2006 -0800
52101 WIP code to move mode set sequencing to XFree86 handlers.
52103 It compiles. It definitely doesn't run.
52105 commit 2e8c927f9308069a82f25b65bb0c62bc5a156832
52106 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
52107 Date: Mon Dec 4 14:02:30 2006 -0800
52109 Re-create RandR Crtc/output structures on server regen.
52111 RandR structures must be re-created when the server reinitializes,
52112 but the driver PreInit function is not re-invoked. Recreate them
52113 manually in this case during ScreenInit.
52115 commit 583619cdb2b6a469299dd5bf658bf5b51ee999fe
52116 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52117 Date: Mon Dec 4 15:48:04 2006 +0800
52119 set correct default border color
52121 commit f28d5fc30ccf60d72d5832bb6542522dc7f1c463
52122 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52123 Date: Mon Dec 4 15:47:31 2006 +0800
52125 fix typo in ps kernel
52127 fix corrupt in some subspans
52129 commit 677c454469a6223460778033758e384e989cea47
52130 Merge: 16f9a56 8fcf9a8
52131 Author: Keith Packard <keithp@neko.keithp.com>
52132 Date: Sat Dec 2 23:28:38 2006 -0800
52134 Merge branch 'modesetting-origin' into modesetting
52136 commit 8fcf9a81179ee8577ddab5e904c58fbfd14cf59c
52137 Author: Keith Packard <keithp@ukulele.keithp.com>
52138 Date: Sat Dec 2 22:58:31 2006 -0800
52140 DOUBLE_WIDE mode for high pixel clock 8xx. Rewrite PLL search.
52142 High pixel clock modes on pipe A of an 8xx chip require
52143 DOUBLE_WIDE mode. It's supposed to be modes > 180MHz or so,
52144 but the board I have requires DOUBLE_WIDE mode for clocks > 108MHz
52145 or so. The limit is related to the core clock speed of the chip, which
52146 can be found indirectly through PCI config space. None of the possible
52147 values explain why this board needs this mode for these relatively low
52148 clock rates though.
52150 Also, create tables of data for the PLL computation and use them
52151 instead of code. I think it's cleaner looking. It is also untested on
52152 9xx. It'll work. Really.
52154 commit 81dde11d419c8f9198ab3502d9813d66d0bc6d6d
52155 Author: Eric Anholt <eric@anholt.net>
52156 Date: Fri Dec 1 10:37:19 2006 -0800
52158 Fix copy'n'paste-o from restructure-outputs that resulted in crashes.
52160 Our driver private is stored in the bus that we set up, not the DDC device
52161 that xf86 code does.
52163 commit a504e31f0fefdec802ae403573f52fac3d577b65
52164 Author: Eric Anholt <eric@anholt.net>
52165 Date: Fri Dec 1 12:49:43 2006 -0800
52167 Don't try to write the read-only PP_STATUS register.
52169 Also, don't bother reading the PP_CONTROL register to try to get LVDS status --
52170 that's what PP_STATUS is for.
52172 commit f21230d3e1cc5470e8f88404b5d9f1702bec44bb
52173 Author: Eric Anholt <eric@anholt.net>
52174 Date: Fri Dec 1 12:48:57 2006 -0800
52176 Add the [ax]4r4g4b4 source picture formats.
52178 commit 8ddbc26b276016b922487f7b1e8997998b655a7d
52179 Merge: 93975a5 16f9a56
52180 Author: Keith Packard <keithp@neko.keithp.com>
52181 Date: Fri Dec 1 10:57:16 2006 -0800
52183 Merge branch 'modesetting' into crestline
52185 Use new TV output condition (all 9xx except 915g)
52187 commit 16f9a56070a01c7ada8e566d9b73fcf7900136e0
52188 Author: Keith Packard <keithp@neko.keithp.com>
52189 Date: Fri Dec 1 10:50:27 2006 -0800
52191 TV output exists on all 9xx chips except for 915G
52193 commit 93975a526086432cdeb8d7ead721a862d632c1e7
52194 Merge: 08b6569 e603cd0
52195 Author: Keith Packard <keithp@neko.keithp.com>
52196 Date: Fri Dec 1 10:28:58 2006 -0800
52198 Merge branch 'modesetting' into crestline
52200 commit e603cd0c73344ef137d3276b5cfcbcf4df340778
52201 Author: Keith Packard <keithp@neko.keithp.com>
52202 Date: Fri Dec 1 10:08:36 2006 -0800
52204 Leave detected TV status alone when no free CRTC is available.
52206 Yes, this means not detecting TV hotplug when two outputs are
52207 already running. An alternative would be to turn off one of the other
52208 outputs temporarily, but that would cause flashing. Something to consider.
52210 commit ec30356d950199903978265076b8a4e3960cf84f
52211 Author: Keith Packard <keithp@neko.keithp.com>
52212 Date: Fri Dec 1 10:06:29 2006 -0800
52214 Do output detection before any crtc allocation.
52216 Some output detection requires a crtc for load detection, perform all of the
52217 output detection before allocating any crtcs so that there will be a free
52218 crtc for any load detection. Avoids losing TV detection when two monitors
52221 commit 7642da82781826cb3b45078750cc54d8f592893a
52222 Author: Keith Packard <keithp@ukulele.keithp.com>
52223 Date: Thu Nov 30 23:21:39 2006 -0800
52225 When CRTC is disable through RandR, the mode is NULL. Don't dereference it.
52227 Correctly interpret mode == NULL as CRTC disable, make sure the CRTC is
52228 disabled in this case.
52230 commit de50ca9ed9e78c17be1badec62e35ed5e4d62b05
52231 Author: Eric Anholt <eric@anholt.net>
52232 Date: Thu Nov 30 17:56:47 2006 -0800
52234 Tristate the clock/data pins during GPIO when released while getting values.
52236 While the register is laid out suggesting that you can read a low value while
52237 driving the output high, and the I2C spec seems to indicate that you should be
52238 able to as well, and on some hardware this works successfully, on the i865 and
52239 perhaps some other chips it doesn't. So, if we're not holding the clock or
52240 data pin low during GetBits, tristate the pin so that we can successfully read.
52242 This fixes i865 analog (VGA) DDC so it successfully sees slave acks.
52244 Also, improve the I2C bit-banging debugging.
52246 commit 8c44556408601db2be799b5ac5c4da1d92886d31
52247 Author: Eric Anholt <eric@anholt.net>
52248 Date: Thu Nov 30 15:56:52 2006 -0800
52250 Preserve some GPIO bits that the docs tell us to.
52252 commit 28401b25cc5c1afb272e506cef196ba3a1258bd3
52253 Author: Eric Anholt <eric@anholt.net>
52254 Date: Thu Nov 30 15:45:13 2006 -0800
52256 Add debug code for the bit-banging I2C implementation.
52258 commit ff64bc7397cafbec94e388e1625b3a1999f1aca7
52259 Merge: b94b7c4 16e01b1
52260 Author: Keith Packard <keithp@neko.keithp.com>
52261 Date: Thu Nov 30 14:09:31 2006 -0800
52263 Merge branch 'restructure-outputs' into modesetting.
52265 Outputs and Crtcs now have a driver-independent representation which should
52266 permit generic code to control RandR 1.2 and startup configuration.
52268 commit 16e01b117bb7ae90b150dd4f25a887dd895cf473
52269 Author: Keith Packard <keithp@neko.keithp.com>
52270 Date: Thu Nov 30 11:55:29 2006 -0800
52272 Oops. Don't try to use xf86RandR names, use xf86RandR12 instead.
52274 Attempts to override functions in core server fail, so use
52277 commit bcf206ff17c6eca8212db72dd86fcd8fdc0162da
52278 Merge: 27ca1c1 2c8c310
52279 Author: Keith Packard <keithp@neko.keithp.com>
52280 Date: Thu Nov 30 11:39:56 2006 -0800
52282 Merge branch 'restructure-outputs-origin' into restructure-outputs
52284 commit 27ca1c17aab30670dd1d07e0809fd58cdecfc3b6
52285 Author: Keith Packard <keithp@neko.keithp.com>
52286 Date: Thu Nov 30 11:39:37 2006 -0800
52288 Rename I830 randr functions to xf86 as they are becoming generic.
52290 Change function names, create separate i830_randr.h header file
52291 to contain definitions.
52293 commit b94b7c4bcfdb7ba59ed818f72309b5060a2ab7ee
52294 Merge: b6fc8df 25e6e49
52295 Author: Eric Anholt <eric@anholt.net>
52296 Date: Thu Nov 30 09:15:30 2006 -0800
52298 Merge branch 'exa' of ../xf86-video-intel into modesetting
52309 commit b6fc8df9a52f5fe1b4d26ae06bc4d48235b44a67
52310 Author: Eric Anholt <eric@anholt.net>
52311 Date: Wed Nov 29 15:06:32 2006 -0800
52313 Properly detect the GTT size on the G965.
52315 In the past, the GTT has always been sized just large enough to map the whole
52316 graphics aperture. However, apparently on the G965 that isn't the case, and
52317 it is actually 512KB on hardware with a 256MB aperture. This resulted in X
52318 not bothering to allocate memory for 256KB that it thought was already mapped
52319 into stolen memory, and thus garbage rendering (particularly visible in large
52320 video modes that displayed this unallocated memory). The kernel happens to
52321 get the right answer by hardwiring a 512KB GTT size already, but that may not
52322 be true on future hardware.
52324 Instead, we use a convenient field in PGETBL_CTL that's specifically for the
52325 GTT size rather than the aperture size, which gets us the answer we want.
52327 commit 359dc81c07901665da0f86c573c096fa1661cdd2
52328 Author: Eric Anholt <eric@anholt.net>
52329 Date: Wed Nov 29 15:01:39 2006 -0800
52331 Revert "Don't allocate stuff in the first 256K of video memory (GATT?)"
52333 This reverts commit 997e8c9bb4235cab1fff4738387df9afcbea0a03.
52335 The GTT is definitely located at the end of stolen memory. This commit
52336 apparently worked around mis-estimation of the GTT size.
52338 commit db391e8e4c4d87bfe3ccad0de14dd5b47b69b8fe
52339 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52340 Date: Wed Nov 29 17:16:46 2006 +0800
52344 commit 290f15cd4cda97727ebcaadacbbbf7650278934b
52345 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52346 Date: Wed Nov 29 17:14:55 2006 +0800
52348 fix alpha blending state
52350 commit 3d4edd325f3859c749ee42df102bb4239eac5287
52351 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52352 Date: Wed Nov 29 17:05:32 2006 +0800
52354 Add in sf/wm program for mask picture without CA
52356 commit a704120b15efae47344a90d972e7f3da64a202a6
52357 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52358 Date: Wed Nov 29 16:52:44 2006 +0800
52360 misc cleanup for G965 vs/sf/wm states
52362 commit e3c70c68e39183226e498271c44e98ef1b96a681
52363 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52364 Date: Wed Nov 29 16:40:15 2006 +0800
52366 WM kernel needs scratch space
52368 commit aa515c54f0cfd9025fc38dc4b7938ff17a8a13fb
52369 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52370 Date: Wed Nov 29 16:37:06 2006 +0800
52372 Setup default border color for our samplers
52374 commit b6eba96584bcd2c024f6443d9f3728eb65b234fb
52375 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52376 Date: Wed Nov 29 16:30:53 2006 +0800
52378 fix vertex buffer size
52380 commit 453842c9ff733af45fa665d9db6a35164f45c60a
52381 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52382 Date: Wed Nov 29 16:24:24 2006 +0800
52384 clean up issue cmd to ring buffer
52386 Make it easy to track different part of ring state, and
52387 use rectlist primitive instead.
52389 commit 18ad7d5cf04081d89a9f978ccc7794116f7c498b
52390 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52391 Date: Wed Nov 29 16:04:16 2006 +0800
52393 remove wrong scale_units
52395 commit 3d5bd0c14eea7951540f7a12eee257428f78e2d1
52396 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52397 Date: Wed Nov 29 16:02:16 2006 +0800
52399 Rename exa sf/wm program files
52401 Also fix some minors in wm program.
52403 commit a06beb5f80f097fac3b718e742742bb32f1c1194
52404 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52405 Date: Wed Nov 29 15:47:19 2006 +0800
52407 EXA state mem for G965
52409 Not split offscreen mem for exa, but alloc a dedicated one
52412 commit 9e4e7d4fa25a64a2494e7531967599142e60e716
52413 Author: Eric Anholt <eric@anholt.net>
52414 Date: Tue Nov 28 16:36:09 2006 -0800
52416 Conditionally compile old-server compatibility code in.
52418 This also replaces calls to compat code with the real names of the functions,
52419 and slips #defines to an i830-namespaced version in when doing compat.
52420 The current server version (7.1.99.2) is still left as requiring compat code,
52421 since the version hasn't been bumped yet.
52423 This also fixes some failures to call the compat code, and some failures to
52424 actually compile the compat code. Oops.
52426 commit 8d9ba2405d1f1e58a0cfb570e83e9206af8525b5
52427 Author: Eric Anholt <eric@anholt.net>
52428 Date: Tue Nov 28 15:44:06 2006 -0800
52430 sparse cleanups (use NULL instead of 0 for pointers)
52432 commit dbb8135e10e943a4d4e65f95853eeccc764ccf24
52433 Author: Eric Anholt <eric@anholt.net>
52434 Date: Tue Nov 28 13:21:31 2006 -0800
52436 Remove printing of the XFree86 screen mode list, and some other detritus.
52438 commit e8fb40ab1063b18372be2b5996257eb69d3f15c4
52439 Author: Eric Anholt <eric@anholt.net>
52440 Date: Tue Nov 28 11:17:21 2006 -0800
52442 Replace custom DDC mode list code with a copy of the X Server's.
52444 The X Server now has most of the mode helper functions we want, and we can
52445 start removing compilation of our copies when the server is new enough.
52447 commit 2c8c310bd8a105dbeda5c989dca31cc766eb6bb8
52448 Author: Keith Packard <keithp@guitar.keithp.com>
52449 Date: Mon Nov 27 22:18:59 2006 -0800
52451 Driver-independent code tracks num_crtc itself, don't do it in driver.
52453 Now that driver-independent code manages array of crtcs, let it track
52454 the count and just have the driver compute how many there are in
52455 a separate variable.
52457 commit c421724d8151d1d4578468abdbe8fe1120e373ca
52458 Author: Keith Packard <keithp@guitar.keithp.com>
52459 Date: Mon Nov 27 22:15:40 2006 -0800
52461 Move #include "randrstr.h" from i830.h to i830_xf86Crtc.h
52463 As part of the separation of driver-dependent from driver-independent code,
52464 the interactions with RandR are moving entirely out of the driver, which
52465 means the driver shouldn't even know about RandR, while the
52466 driver-independent code does.
52468 commit e8d1db326388c9a112f2b8e2eff5a81210b8d6e4
52469 Author: Keith Packard <keithp@neko.keithp.com>
52470 Date: Mon Nov 27 12:01:47 2006 -0800
52472 Move crtc/output config to sub-structure.
52474 Place crtc/output in separate structure at head
52475 of driver private structure. Use this from the config code
52476 to make it driver-independent. Still lots of effectively driver
52477 independent code that continues to use driver dependent stuff,
52478 but that will change.
52480 commit 72692ba2e0254460c4f5a8cd476f5748383390ae
52481 Author: Eric Anholt <eric@anholt.net>
52482 Date: Mon Nov 27 11:23:55 2006 -0800
52484 Postpone SDVO DDC bus creation until we've detected the SDVO device.
52486 This reduces log noise for those of us with no SDVO devices.
52488 commit 5f38bc3e2a685da5c03efa8103a353f79c842aa7
52489 Author: Eric Anholt <eric@anholt.net>
52490 Date: Mon Nov 27 11:06:50 2006 -0800
52492 Reduce the severity of many informational log messages.
52494 commit 85f404bc67d5ab54175c19e75aad406abfcf2133
52495 Author: Eric Anholt <eric@anholt.net>
52496 Date: Mon Nov 20 09:32:26 2006 -0800
52498 Re-indent ch7xxx driver.
52500 commit a0518f5a440630365b1d935b7c2d0725f326ad51
52501 Author: Keith Packard <keithp@neko.keithp.com>
52502 Date: Sun Nov 26 19:44:38 2006 -0800
52504 Remove mode setting from load detect CRTC allocation.
52506 To share load-detect CRTC allocation with TV driver,
52507 move it to the output driver
52509 commit a47c549df036990e29f05bc3df80e1a2ab9f3b3c
52510 Author: Keith Packard <keithp@neko.keithp.com>
52511 Date: Sun Nov 26 18:26:26 2006 -0800
52513 Clean up reworked data structure code so the server actually starts.
52515 Use i830GeLoadDetectPipe again (instead of missing xf86AllocCrtc). Actually
52516 create new Crtc structures. Fix a few other NULL pointer dereferences.
52518 commit 2529863a1ade782819d76be2d0dc16e89028c1e3
52519 Author: Keith Packard <keithp@neko.keithp.com>
52520 Date: Sun Nov 26 16:44:17 2006 -0800
52522 Start output/crtc restructuring work. It compiles.
52524 Outputs and Crtcs are now split between 'generic'
52525 and 'driver specific' pieces in the hope that more code
52526 will be able to migrate to the xf86-generic layer.
52528 Right now, the code remains tangled together, significant
52529 work remains to tease the pieces apart. First the code
52530 needs to be made to actually work as-is though.
52532 commit 9aca4e207440119f4280b78199a221f85d50c511
52533 Author: Keith Packard <keithp@neko.keithp.com>
52534 Date: Wed Nov 22 15:52:32 2006 -0800
52536 Remove custom configuration code (which was #ifdef'd out).
52538 Multi-screen configuration is moving to generic code.
52540 commit 850e3652f4d5bb57f3c4ccb9cf29e2c75fecba9d
52541 Author: Keith Packard <keithp@neko.keithp.com>
52542 Date: Wed Nov 22 15:36:43 2006 -0800
52544 Note which pipe has failed to be configured for VBlank interrupt
52546 commit 659e3db925de91954421f7f33c072576dfefcc4a
52547 Author: Keith Packard <keithp@neko.keithp.com>
52548 Date: Wed Nov 22 15:35:59 2006 -0800
52550 Set up maximum screen size for clone instead of panorama.
52552 This avoids heading past the 8Kbyte stride limit of DRI so
52553 that 3D will be available by default.
52555 commit 03c12f866529230df7a1503de61954cd5ade242e
52556 Author: Keith Packard <keithp@neko.keithp.com>
52557 Date: Wed Nov 22 15:34:03 2006 -0800
52559 Limit DRI use to cases when pitch < 8Kbytes, not 4096 pixels.
52561 3D rendering pipeline is limited to pitch of 8Kbytes and not 4096
52562 pixels, which rather makes a difference at 32bpp.
52564 commit 6781575f734f05547d7d5ceef4116fc157bba44d
52565 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
52566 Date: Wed Nov 22 16:49:57 2006 +0000
52568 Be a little more verbose when rejecting modes.
52570 commit e065324661ad08b3b359136f48090232f6138959
52571 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
52572 Date: Wed Nov 22 16:49:36 2006 +0000
52574 update port attributes immediately when overlay is on.
52576 commit a0a9d1ea2ac9a0c665e5a1de5b8b8358dee3b9b4
52577 Author: Keith Packard <keithp@guitar.keithp.com>
52578 Date: Mon Nov 20 17:25:00 2006 -0800
52580 Set configured values for screen virtual size and initial frame.
52582 Computation for virtual size and initial frame origin is quite
52583 broken in xf86 common code.
52585 commit d6a0f917e601ea36643c6ad857756e19d24ecd73
52586 Author: Keith Packard <keithp@bouzouki.jf.intel.com>
52587 Date: Mon Nov 20 15:17:32 2006 -0800
52589 Enable second SDVO channel.
52591 Rework SDVO support so that it can deal with two channels correctly,
52592 also save/restore all connected output timings.
52594 commit 08b6569dc663ddf38cb36a6875de6d4ab55acac9
52595 Author: Eric Anholt <eric@anholt.net>
52596 Date: Sun Nov 19 16:15:45 2006 -0800
52598 Enable TV output on Crestline (untested).
52600 commit 05202cabbd23f15330b811ae6b8d708ad042bc40
52601 Merge: ee502dd b945a65
52602 Author: Eric Anholt <eric@anholt.net>
52603 Date: Sun Nov 19 16:14:18 2006 -0800
52605 Merge branch 'modesetting-origin' into crestline
52607 commit b945a650e952f98c2d101b71bd3ec0f390478da5
52608 Author: Keith Packard <keithp@guitar.keithp.com>
52609 Date: Sun Nov 19 00:54:30 2006 -0800
52613 Subcarrier defines were incorrect in header file leaving one of the
52614 DDA phases disabled.
52616 commit 28224af3d90a1a08d54a865dfaf20184330fe8a4
52617 Author: Keith Packard <keithp@guitar.keithp.com>
52618 Date: Sun Nov 19 00:40:46 2006 -0800
52620 Preliminary 945 TV output. Color key is broken. Fixed mode.
52622 TV output is generating video with this patch, but the color burst
52623 signal is incorrect somehow.
52625 commit ee502dd92a3dfccdc2efcfb76fc652694bc89e3b
52626 Merge: 15ff17c 2fe6107
52627 Author: Eric Anholt <eric@anholt.net>
52628 Date: Fri Nov 17 14:57:41 2006 -0800
52630 Merge branch 'crestline-origin' into crestline
52632 commit 15ff17c756e42f392306820e3f7ffbdcc56b9892
52633 Merge: 9cc2f33 816fc1a
52634 Author: Eric Anholt <eric@anholt.net>
52635 Date: Fri Nov 17 11:48:53 2006 -0800
52637 Merge branch 'modesetting' into crestline
52639 This works for analog, but SDVO output appears to not work yet.
52645 commit 816fc1a76a5ac738e41b172ba8f43137c1521328
52646 Merge: 9948d83 7a7bb33
52647 Author: Eric Anholt <eric@anholt.net>
52648 Date: Fri Nov 17 00:05:53 2006 -0800
52650 Merge branch 'modesetting-origin' into modesetting
52652 commit 7a7bb331e10498e5b8ccec58130bb23334d36562
52653 Author: Keith Packard <keithp@neko.keithp.com>
52654 Date: Thu Nov 16 21:19:20 2006 -0800
52656 Don't dereference null DisplayModePtr on disabled output.
52658 During initial configuration, outputs which are disabled have null
52661 commit 2fe6107ac68e86ed183d8602436633348340678d
52662 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52663 Date: Fri Nov 17 13:11:36 2006 +0800
52665 Rotation support for 965GM
52667 commit c4508c1cadf323e9ef1d0e69dd77d5e841a6a978
52668 Author: Keith Packard <keithp@neko.keithp.com>
52669 Date: Thu Nov 16 21:09:23 2006 -0800
52671 RandR-based initial output configuration.
52673 Using pre-init computed RandR information, make reasonable
52674 default choices for the output configuration at startup time.
52675 Either some preferred size or a size which yields 96dpi is chosen,
52676 from which other monitors are set to a similar size. The largest
52677 size sets the screen size.
52679 This needs to be extended to respect config file settings, but
52680 those have not been defined yet.
52682 commit 9948d8377d70e898260c12210151b952b3fb8bf1
52683 Merge: 45a27f8 b649f95
52684 Author: Eric Anholt <eric@anholt.net>
52685 Date: Thu Nov 16 19:47:24 2006 -0800
52687 Merge branch 'master' into modesetting
52697 commit 9cc2f3313d77487dea372e6ab32d6d9c06617ae3
52698 Author: Eric Anholt <eric@anholt.net>
52699 Date: Thu Nov 16 15:12:43 2006 -0800
52701 Replace broken PCI resource size detection with pciGetBaseSize() call.
52703 commit 45a27f80e1c783627f570c309e7a853dcc9af0c1
52704 Author: Keith Packard <keithp@neko.keithp.com>
52705 Date: Thu Nov 16 13:38:35 2006 -0800
52707 Remove output options. Let outputs be connected to other crtcs.
52709 Output options will be replaced by properties.
52710 Permits outputs to be connected to arbitrary CRTCs (within hardware limits).
52713 commit 9aea79d1e954fe4fb5c101edcb7c2d0f706a5c4c
52714 Author: Keith Packard <keithp@neko.keithp.com>
52715 Date: Thu Nov 16 11:40:26 2006 -0800
52717 Inverted boolean sense when selecting pipe for CRT detection.
52719 Pipes are available when they are not in use.
52721 commit 4889b9f33336c92f07aac86d75f50316db6ef81a
52722 Author: Eric Anholt <eric@anholt.net>
52723 Date: Wed Nov 15 15:26:31 2006 -0800
52725 Enable the LVDS if we find it and assign it to a pipe (oops).
52727 commit 854ff826c0031a44e874239c0b0a3533f4d9e14b
52728 Author: Eric Anholt <eric@anholt.net>
52729 Date: Tue Nov 14 09:14:24 2006 -0800
52731 Disable setup of the second SDVO device until we fix it.
52733 commit d51555fba4e57c059fd184c1e54822d7e5b62a2f
52734 Author: Eric Anholt <eric@anholt.net>
52735 Date: Fri Nov 10 14:40:40 2006 -0800
52737 Fix clock range for single-channel LVDS.
52739 commit dd1dcfab0ab0f2d0c25077fa663209e2762f26e8
52740 Author: Eric Anholt <eric@anholt.net>
52741 Date: Fri Nov 10 11:08:02 2006 -0800
52743 Fill in some of the high bits of mode timings for SDVO.
52745 commit 3955f044cfe1d592bcb36c43c539fb4a75840b8d
52746 Author: Eric Anholt <eric@anholt.net>
52747 Date: Thu Nov 9 20:28:46 2006 -0800
52749 Expose the DDC-probed EDID data as the EDID_DATA output property.
52751 commit 0f5886689d7ef7dbbef6425d5c855ac6b67d3350
52752 Author: Keith Packard <keithp@mandolin.keithp.com>
52753 Date: Wed Nov 8 23:19:59 2006 -0800
52755 Create RandR 1.2 objects in I830PreInit.
52757 Creating the objects early will allow the driver to use
52758 randr structures to select a reasonable configuration.
52759 That part has not been done yet.
52761 commit 679c7bd82639a09cdce133becb8a08629ce3a4e9
52762 Author: Keith Packard <keithp@mandolin.keithp.com>
52763 Date: Wed Nov 8 21:39:28 2006 -0800
52765 ignore edited man page
52767 commit 81b7b489afa2cab4d8614c64f4906be627f1d07e
52768 Author: Keith Packard <keithp@mandolin.keithp.com>
52769 Date: Wed Nov 8 21:38:00 2006 -0800
52771 Adapt to RandR updates that split object creation from screen association.
52773 RandR DIX code is preparing for xf86 drivers that want to allocate RandR
52774 objects at PreInit time. This patch adapts to that change without taking
52777 commit 81bace0c316c3ed80201a34eca533254d12cd193
52778 Merge: 713c5b0 beb8916
52779 Author: Keith Packard <keithp@mandolin.keithp.com>
52780 Date: Wed Nov 8 20:23:20 2006 -0800
52782 Merge branch 'modesetting-keithp' into modesetting
52784 Conflicts in PipeSetMode were resolved to use the keithp changes
52785 that pushed more modesetting stuff into the per-pipe function.
52787 Switched availablePipes to num_pipes.
52789 Used modesetting default output configuration.
52791 commit 713c5b0899428edfea7cea0780244488115dbe1d
52792 Author: Eric Anholt <eric@anholt.net>
52793 Date: Wed Nov 8 19:55:31 2006 -0800
52795 Change the output and pipe "is it on/off" field name to "enabled".
52797 commit ff77e9d84f2037b99a8e4ac55da0b0ac92ed6b9b
52798 Author: Eric Anholt <eric@anholt.net>
52799 Date: Wed Nov 8 19:53:31 2006 -0800
52801 Clean up i830_crt_detect_load() a bit more.
52803 ADPA might not have been set right in some cases (DPMS-off monitor, for
52804 example), and a wait for vsync that the bios does was missing.
52806 commit 9b267014b55f49d9362c1d432d6ba19ddd0ad95d
52807 Author: Eric Anholt <eric@anholt.net>
52808 Date: Wed Nov 8 19:39:18 2006 -0800
52810 Go back to only setting up outputs that have a display connected.
52812 commit 35ab689bbde5f74752598cd743d735640486b639
52813 Author: Eric Anholt <eric@anholt.net>
52814 Date: Wed Nov 8 19:35:49 2006 -0800
52816 Fix i830DisableUnusedFunctions after pipe structure change.
52818 Using "pipe" instead of the index "i" meant pipe(3) got referenced instead of
52819 a nice small integer. Oops.
52821 commit b649f95ea6fd3555d073fdbf8f2f035dfe1afd33
52822 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
52823 Date: Wed Nov 8 19:01:25 2006 +0000
52825 Disable some debug message
52827 commit 75f4df278e9db360967d77cdba4756cbde622d56
52828 Author: root <root@localhost.localdomain>
52829 Date: Wed Nov 8 13:56:32 2006 +0800
52833 commit f1ff01e31eb8e9dc05190bf1a8b318d4f587f64a
52834 Author: Keith Packard <keithp@neko.keithp.com>
52835 Date: Sat Nov 4 00:46:18 2006 -0800
52837 Eliminate operatingDevices member and PIPE_* values.
52839 operatingDevices and MonType1/MonType2 duplicate information already stored
52840 in the device structures. Eliminate them and replace uses with direct
52841 references to the appropriate other data.
52842 (cherry picked from 3ab7f9693217d8fe993bdc94c376b219b0082961 commit)
52844 commit 0b2d36d4f038c4e8fa08632b6f1368627f010392
52845 Author: Keith Packard <keithp@neko.keithp.com>
52846 Date: Fri Nov 3 23:29:12 2006 -0800
52848 Use pI830->availablePipes instead of MAX_DISPLAY_PIPES everywhere
52849 (cherry picked from e4bcec796e80e9fd66ab0c36394f5946915531f1 commit)
52851 commit d0ef9e99acb9e999e1b6d3eb76edc6355555043b
52852 Author: Eric Anholt <eric@anholt.net>
52853 Date: Mon Nov 6 18:30:46 2006 -0800
52855 Restore PFIT_CONTROL before turning the LVDS back on in the restore method.
52857 commit 94a3731c2b4f2ea2e696a8c87dccc0d214d41e8e
52858 Author: Eric Anholt <eric@anholt.net>
52859 Date: Mon Nov 6 18:26:48 2006 -0800
52861 Move PFIT_CONTROL disable for G965 up before post_set_mode.
52863 Also, remove setting of some other random registers that appears to have
52864 been spammed in at the same time, and don't try to disable on the I830, before
52865 this register existed.
52867 commit beb89163d73376e70870e6e2a6b19863f3a058b1
52868 Author: Keith Packard <keithp@mandolin.keithp.com>
52869 Date: Sun Nov 5 19:06:45 2006 -0800
52871 DSPSURF must be page aligned. Place intra-screen offset in DSPBASE.
52873 DSPASURF/DSPBSURF can only take page aligned values, ignoring
52874 the lower order bits. So, place the offset for the output
52875 within the frame buffer in the DSPABASE/DSPBBASE registers instead.
52877 commit 997e8c9bb4235cab1fff4738387df9afcbea0a03
52878 Author: Keith Packard <keithp@mandolin.keithp.com>
52879 Date: Sun Nov 5 18:56:33 2006 -0800
52881 Don't allocate stuff in the first 256K of video memory (GATT?)
52883 Letting the ring buffer or other objects be allocated within the lowest
52884 portion of memory appears to trash some memory mapping data; I'm assuming
52885 this is the GATT table on the 965. Just marking this out of bounds for
52886 allocation fixes this problem.
52888 commit 5a355c72614ed77f2000e5ede45f3ff5990c79d9
52889 Author: Keith Packard <keithp@mandolin.keithp.com>
52890 Date: Sun Nov 5 18:51:28 2006 -0800
52892 Fix CRT output on 965 chipset.
52894 A few more register settings are needed to get CRT output working on the
52895 965 chipset, in particular the the SDVO/UDI clock multiplier register
52896 needed to get set to the default value (3). No, I really don't know what
52897 this does, but it does get the CRT running at a wide range of sizes.
52899 commit 49a6bea7d969dbfd1dd542c0c3e02abc330d6850
52900 Merge: 64447c7 fbb376b
52901 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
52902 Date: Mon Nov 6 10:25:23 2006 +0800
52904 Merge branch 'master' into crestline
52906 commit 68c3185046b27ab936ca6c92b924b443b3cd6fce
52907 Author: Keith Packard <keithp@mandolin.keithp.com>
52908 Date: Sun Nov 5 13:30:32 2006 -0800
52910 Avoid crashing when disabling sdvo output. XXX
52912 commit 15ef08046bcc3e746453301379f7c5d1bf929ee1
52913 Author: Keith Packard <keithp@mandolin.keithp.com>
52914 Date: Sun Nov 5 13:29:56 2006 -0800
52916 Move remaining pipe mode setting logic to i830PipeSetMode
52918 commit 7fcb555735a58e19ccc10875b211402983170a87
52919 Author: Keith Packard <keithp@neko.keithp.com>
52920 Date: Sat Nov 4 00:52:21 2006 -0800
52922 Rename availablePipes to num_pipes
52924 commit 3ab7f9693217d8fe993bdc94c376b219b0082961
52925 Author: Keith Packard <keithp@neko.keithp.com>
52926 Date: Sat Nov 4 00:46:18 2006 -0800
52928 Eliminate operatingDevices member and PIPE_* values.
52930 operatingDevices and MonType1/MonType2 duplicate information already stored
52931 in the device structures. Eliminate them and replace uses with direct
52932 references to the appropriate other data.
52934 commit e4bcec796e80e9fd66ab0c36394f5946915531f1
52935 Author: Keith Packard <keithp@neko.keithp.com>
52936 Date: Fri Nov 3 23:29:12 2006 -0800
52938 Use pI830->availablePipes instead of MAX_DISPLAY_PIPES everywhere
52940 commit b7262a9a9110dac66e1a92c39dcb3ab59d95d081
52941 Author: Keith Packard <keithp@neko.keithp.com>
52942 Date: Fri Nov 3 23:24:07 2006 -0800
52944 Finish removing persistant vbe data
52946 commit 4625073244d4f521a07e12adcf0609e85658acbe
52947 Author: Keith Packard <keithp@neko.keithp.com>
52948 Date: Fri Nov 3 23:23:38 2006 -0800
52950 Oops, martian memset of randr modes pointer
52952 commit 719ad68515be9b996a6314de5448843de1146b88
52953 Author: Keith Packard <keithp@neko.keithp.com>
52954 Date: Fri Nov 3 19:41:41 2006 -0800
52956 Use VBE only temporarily to fetch BIOS rom image
52957 (cherry picked from 6a9386651785afc70a29e355255e8295b321f28e commit)
52959 commit 27df2ff7908ea7ea2943a5f3445e12dbc24d97c9
52960 Author: Eric Anholt <eric@anholt.net>
52961 Date: Fri Nov 3 15:55:10 2006 -0800
52963 Report pipe status (and status mismatches) in i830DescribeOutputConfiguration()
52965 commit ecbe73b940b2d642115de4b73c2f757eb46ff956
52966 Merge: 561af00 9681602
52967 Author: Eric Anholt <eric@anholt.net>
52968 Date: Fri Nov 3 15:59:59 2006 -0800
52970 Merge branch 'modesetting-origin' into modesetting
52976 commit 561af007974b8cdad1eea907fb73ed9d430c21ac
52977 Author: Eric Anholt <eric@anholt.net>
52978 Date: Fri Nov 3 15:26:14 2006 -0800
52980 Add support for load-based CRT detection.
52982 commit e416b426d83de031441ada7a77b6bd66cec8b5c9
52983 Author: Eric Anholt <eric@anholt.net>
52984 Date: Fri Nov 3 15:25:41 2006 -0800
52986 Print out modelines as info, not error (which had been used for debugging).
52988 commit 282a9e073ea985cbf0d0f3f296d593af1426bad5
52989 Author: Eric Anholt <eric@anholt.net>
52990 Date: Fri Nov 3 13:46:09 2006 -0800
52992 Don't memset the modes pointer on init, which was dereferencing NULL.
52994 commit 9681602177124e84a817a1e1d428f1779f2a45c9
52995 Author: Keith Packard <keithp@neko.keithp.com>
52996 Date: Fri Nov 3 12:55:25 2006 -0800
52998 Create I830PipeRec to hold pipe-specific data. Remove unused I830 members.
53000 I830 contained six parallel arrays for pipe-specific data; these
53001 have been moved to a I830PipeRec structure instead.
53003 I830 also contained several unused members:
53005 unsigned int bios_version;
53006 Bool newPipeSwitch;
53010 These have been removed, along with the code that set them.
53012 commit 0510671a6c5233468ac20f0ec8096e084df03ce6
53013 Author: Eric Anholt <eric@anholt.net>
53014 Date: Fri Nov 3 10:58:23 2006 -0800
53016 Fix a pasteo in I965 register restore.
53018 commit 2c9ab6e0594769274f2dbcdf7c00fe297fc385d5
53019 Author: Keith Packard <keithp@mandolin.keithp.com>
53020 Date: Thu Nov 2 13:44:55 2006 -0800
53022 set the v_sync_off_high to zero. XXX should check docs
53024 commit 56f6d4f1bb67f447500af3f4f7fa557c3e887baa
53025 Author: Keith Packard <keithp@mandolin.keithp.com>
53026 Date: Thu Nov 2 13:42:17 2006 -0800
53028 Disable the panel fitter when not using it. Cleans up SDVO DVI output.
53030 The panel fitter appears to exist on the 965 hardware (at least) and
53031 causes troubles with DVI output over SDVO when enabled. This patch
53032 checks to see if the panel fitter is pointing at the pipe being configured
53033 and disables it unconditionally in that case. The LVDS driver will configure
53034 it correctly if necessary afterwards.
53036 commit f22d9bcc25aea19ba38d35282367b591fd1b7ca0
53037 Author: Keith Packard <keithp@mandolin.keithp.com>
53038 Date: Thu Nov 2 13:34:45 2006 -0800
53040 Add another couple of new registers
53042 commit a9eac38bcdb49df2ce1122b49bd8b1eb19e8cae5
53043 Author: Eric Anholt <eric@anholt.net>
53044 Date: Thu Nov 2 13:24:54 2006 -0800
53046 Remove duplicated register defs that were just added.
53048 commit 7887c76062b7c79e14fb8e4f13486aa592dcbce8
53049 Author: Eric Anholt <eric@anholt.net>
53050 Date: Thu Nov 2 12:27:21 2006 -0800
53052 Add airlied's I2C code, ifdeffed out.
53054 I've gone back to compare our behavior to it several times, so I'll just keep
53055 the code in tree for now.
53057 commit 87b15cfbf762468d4b8728b3e7a39c76654017de
53058 Author: Eric Anholt <eric@anholt.net>
53059 Date: Thu Nov 2 11:30:21 2006 -0800
53061 Remove dead specifiedMonitor field.
53063 commit 2636d68663a02f6d9eaf36971706b67036ebf56c
53064 Author: Keith Packard <keithp@mandolin.keithp.com>
53065 Date: Thu Nov 2 11:57:11 2006 -0800
53067 Dump more registers for debug purposes
53069 commit 786ec54c4c1540f4aced63ef21d567c3b9f3282e
53070 Author: Keith Packard <keithp@mandolin.keithp.com>
53071 Date: Thu Nov 2 11:56:50 2006 -0800
53073 Add a few more registers from the 965 spec
53075 commit 85e32ad2dadcce1134fcadb14ece8ff30f3925f2
53076 Author: Keith Packard <keithp@mandolin.keithp.com>
53077 Date: Thu Nov 2 11:56:12 2006 -0800
53079 ch7xxxSaveRegs receives real type instead of void *
53081 commit ffbd6ca09bc2300bf967d7c248a559d85b8706e0
53082 Author: Eric Anholt <eric@anholt.net>
53083 Date: Wed Nov 1 13:05:44 2006 -0800
53085 Remove dead VESARec struct.
53087 commit 97c3a1b2421031e41f0b2b1630fde1dc4262d264
53088 Author: Eric Anholt <eric@anholt.net>
53089 Date: Wed Nov 1 13:04:08 2006 -0800
53091 Remove the refresh rate appended to some mode names.
53093 This gets the SDVO and CRT outputs I have to have at least 1 common mode
53094 according to RandR.
53096 commit 7971c401554c218c84a8c45335c9b31bbccfece7
53097 Author: Eric Anholt <eric@anholt.net>
53098 Date: Wed Nov 1 12:42:56 2006 -0800
53100 Attempt to pull monitor physical size information out of DDC EDID data.
53102 commit fb94c1210966f7875e5f034f10ea31c06c502c3a
53103 Author: Eric Anholt <eric@anholt.net>
53104 Date: Wed Nov 1 12:23:50 2006 -0800
53106 Move mode lists from per-pipe to per-output.
53108 This should let RandR do the right thing in exposing the modes to userland.
53110 As a side effect of getting this working, the SDVO pixel clock range code
53111 was fixed and the mode valid tests for various outputs got extended. Also,
53112 LVDS grew a get_modes for the fixed panel mode.
53114 Note that we now no longer do automatic enabling of outputs at xrandr -s 0,
53115 hotkey, or VT switch. That will be left to generic RandR code later. Also,
53116 generic modes and user-defined modes are once again not validated into the
53117 lists, so this is a regression there.
53119 commit f30d7f912f36b110c3af7dc795e35456593781ab
53120 Author: Eric Anholt <eric@anholt.net>
53121 Date: Wed Nov 1 11:50:51 2006 -0800
53123 Update for the move of RandR phyiscal size information.
53125 commit 64447c7a059775e7ea8649f4714df7565e932c60
53126 Author: Zou Nanhai <nanhai.zou@intel.com>
53127 Date: Wed Nov 1 14:36:20 2006 +0800
53129 walk around to VBIOS bug in Crestline B0
53131 commit 7195dfabd56239f08cdd8175a2ef3a66ef9600de
53132 Author: Eric Anholt <eric@anholt.net>
53133 Date: Tue Oct 31 17:10:08 2006 -0800
53135 Give each output a get_modes function and expose those modes through RandR.
53137 The get_modes should return the probed modes only. The driver should then
53138 append to the list (for example, compatible modes listed in other outputs,
53139 or standard VESA modes) to create the list to expose through RandR. That
53142 commit cc3728be2481637dda321d3bc2e4e89a220699cd
53143 Author: Eric Anholt <eric@anholt.net>
53144 Date: Tue Oct 31 15:00:36 2006 -0800
53146 Add compat definitions for M_T_PREFERRED and M_T_DRIVER for older X Servers.
53148 commit a71f283650e8cb7b760e5a53c4db79202c4cc5c4
53149 Author: Eric Anholt <eric@anholt.net>
53150 Date: Tue Oct 31 14:46:23 2006 -0800
53152 Connect output detection up to RandR.
53154 commit 68cef9f4e028755bbf3e1862da2ef47d46ddaa6a
53155 Author: Eric Anholt <eric@anholt.net>
53156 Date: Tue Oct 31 14:32:00 2006 -0800
53158 Move output connection detection to a per-output method.
53160 This will be used by RandR, and should let us clean up some of the initial
53161 display configuration, hopefully.
53163 Also, analog hotplug-based detection is now enabled on G965.
53165 commit 9fd719fce27f916ab5120f6e1234affa14eaed9d
53166 Author: Eric Anholt <eric@anholt.net>
53167 Date: Tue Oct 31 14:29:44 2006 -0800
53169 Move SDVOB_PRESERVE_MASK next to SDVOC_PRESERVE_MASK.
53171 commit 4f5d4d8870fc2784192f95a561163cf4fc8737ac
53172 Author: Eric Anholt <eric@anholt.net>
53173 Date: Tue Oct 31 13:37:23 2006 -0800
53175 i830SetLVDSPanelPower is now a static function in i830_lvds.c, so remove it.
53177 commit 49bbdf16c02107c08169f8d2b6e9c6dbd7d8cd95
53178 Author: Eric Anholt <eric@anholt.net>
53179 Date: Tue Oct 31 10:44:45 2006 -0800
53181 Fix many inconsistencies in the SDVO code compared to the spec.
53183 Also, fix some struct padding so that the right bits are sent out.
53185 commit e7d546cac06767ec58325396a3bb5780b2257c53
53186 Author: Eric Anholt <eric@anholt.net>
53187 Date: Mon Oct 30 11:24:43 2006 -0800
53189 Remove some dead code from BIOS modesetting.
53191 commit 2ca57040b0cd24ad3dbe693789091e28be4e69f8
53192 Author: Eric Anholt <eric@anholt.net>
53193 Date: Mon Oct 30 11:19:19 2006 -0800
53195 Remove GetBIOSVersion().
53197 This info hardly useful now that we don't use the BIOS for mode setting.
53199 commit 837b2f632062bc29268f109895a577bd90cabd6d
53200 Author: Eric Anholt <eric@anholt.net>
53201 Date: Mon Oct 30 11:17:55 2006 -0800
53205 commit bca9e6ccbd14eb8f2f103e8e64b28a623113d494
53206 Author: Eric Anholt <eric@anholt.net>
53207 Date: Mon Oct 30 11:17:27 2006 -0800
53209 Remove SetPipeAccess and now-unnecessary VBE reinit.
53211 commit 819a47b27cd4728feb269a08be32403304993ffa
53212 Author: Eric Anholt <eric@anholt.net>
53213 Date: Mon Oct 30 09:50:33 2006 -0800
53215 Use the new fields for SDVO pixel multiply on the G965.
53217 This should fix display at resolutions/refresh rates in a different multiplier
53218 class than the console display (generally, high resolution modes).
53220 commit 71545db4614cfc4650acc4325912474e777c3b36
53221 Author: Eric Anholt <eric@anholt.net>
53222 Date: Mon Oct 30 09:46:10 2006 -0800
53224 Return and use valid status bits for i830_sdvo_get_trained_inputs().
53226 commit c357eca10ca1c535d305e1f3028471a912ae4102
53227 Author: Eric Anholt <eric@anholt.net>
53228 Date: Mon Oct 30 09:44:55 2006 -0800
53230 Clean up whitespace in i830_randr.c.
53232 commit fbb376bd1a4daad4c86e349df98438989ce173f1
53233 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53234 Date: Mon Oct 30 14:15:12 2006 +0800
53236 Bug 8594: Fix Xv hang with G965
53238 commit c0ee50c4ee5ff2c594fdf60c9cb8b952e25644de
53239 Author: Thomas Hellstrom <thomas-at-tungstengraphics-dot-com>
53240 Date: Fri Oct 27 19:13:33 2006 +0200
53242 Add support for the new DRM memory manager.
53243 Some code are duplicated with the new libdrm.
53244 Once this code has been released with xserver,
53247 See the man page for new options and backwards
53248 3D driver compatibility.
53250 commit 25e6e497824a23eb231fc6fd6e483f601d612cee
53251 Author: Eric Anholt <eric@anholt.net>
53252 Date: Thu Oct 26 15:47:49 2006 -0700
53254 Major cleanup of 3D invariant state, fixing hangs with rotation and render.
53256 Now, the generic invariant state is always set while the X Server is active,
53257 and happens automatically when the X Server grabs the DRI lock. More 3D state
53258 is moved to the generic code.
53260 Then, the 3D consumers (video, rotation, render) set last_3d to their enum
53261 entry, and can update their own invariant state when another consumer was
53264 commit 0bdcce2e4541c6e441c44aad49254ad3093cedb2
53265 Author: Eric Anholt <eric@anholt.net>
53266 Date: Thu Oct 26 14:30:46 2006 -0700
53268 Note alignment requirement for i915 3D (texturing).
53270 commit c3666a968b1cfac61b7867874d5fbe2acd894720
53271 Author: Eric Anholt <eric@anholt.net>
53272 Date: Thu Oct 26 14:30:25 2006 -0700
53274 Don't write unused values beyond the end of scale_units array.
53276 commit 3ab9f5a4a8cd62c8a8c2a09d3d105adbe815a83d
53277 Author: Eric Anholt <eric@anholt.net>
53278 Date: Tue Oct 24 00:26:09 2006 -0700
53280 Move vbeInfo out of the driver struct to the one place it's used.
53282 commit df14838eb5d0a056f663d9f12bd8b5c25cf97330
53283 Author: Eric Anholt <eric@anholt.net>
53284 Date: Tue Oct 24 00:18:33 2006 -0700
53286 Remove dead memsize reporting.
53288 The calculation no longer made sense, as we don't use the BIOS for mode
53291 commit 3a6104ab89b159241845314ccf88fa62da14cf7d
53292 Author: Eric Anholt <eric@anholt.net>
53293 Date: Mon Oct 23 19:15:30 2006 -0700
53295 Remove disabled I830DetectMonitorChange().
53297 This used to be called when switching back in to X. It might make some sense
53298 to detect monitors at this time (it happens to occur at resume time, when
53299 monitors are likely to have changed), but it should probably live in either
53300 userland policy with RandR 1.2 or RandR 1.2 XFree86-DDX generic code.
53302 commit 88c12f577f80fa132ec45cdf456d8060f2ece4fd
53303 Author: Eric Anholt <eric@anholt.net>
53304 Date: Mon Oct 23 18:57:57 2006 -0700
53306 Move LVDS initialization and blacklisting damage to the LVDS support file.
53308 commit 2631014e9d5b2e64908ea413729eb5fd819b17fc
53309 Author: Eric Anholt <eric@anholt.net>
53310 Date: Wed Oct 25 12:21:44 2006 -0700
53312 Clean up the SDVO code.
53314 The main change is to send SDVO commands using data passed into the send
53315 command function, and receive responses into memory passed into the read
53316 response function, rather than stuff things in/out through dev_priv->sdvo_regs.
53317 This lets us use structures to represent some arguments, which results in a
53318 nice cleanup (and 100% fewer arguments named magicN as a side effect).
53320 Also, the mode set path is changed to not do any preferred input timing
53321 work. We weren't doing anything legitimate with the results, since we didn't
53322 modify the CRTC timing appropriately, so now we just stuff the CRTC timing into
53323 both and hope for the best. This should probably be revisited later.
53325 commit ddb986e54f5320359abac06f512f2d3f446872db
53326 Author: Eric Anholt <eric@anholt.net>
53327 Date: Tue Oct 24 17:47:58 2006 -0700
53329 Nuke trailing whitespace in SDVO code not already cleaned up.
53331 commit cd2f0d948a6cae61b0d77d2c7184ed8d695be8c3
53332 Author: Eric Anholt <eric@anholt.net>
53333 Date: Tue Oct 24 17:47:47 2006 -0700
53335 Clean up SDVO code, moving the private structure internal and cleaning up caps.
53337 All the SDVO code should now be in lower case rather than StudlyCaps.
53339 This also adjusts the I2C setup to create a bus per SDVO output we set up.
53340 The previous setup with shared buses was failing in some circumstances, which
53341 is probably due to the lack of refcounting in xf86i2c.c.
53343 commit cd9c6e29146e1debaba4b0b9ad0d241f07bdbc14
53344 Merge: c5cca4c eec5580
53345 Author: Keith Packard <keithp@neko.keithp.com>
53346 Date: Tue Oct 24 16:53:46 2006 -0700
53348 Merge branch 'modesetting-origin' into modesetting
53350 commit eec5580cefffc293bf547372ab63b2fedaef4a83
53351 Merge: 53c28b3 a91c0cb
53352 Author: Eric Anholt <eric@anholt.net>
53353 Date: Mon Oct 23 14:52:55 2006 -0700
53355 Merge branch 'overhaul' into modesetting
53357 commit a91c0cbab5de51885bfce7c7dce76f82c1b19553
53358 Author: Eric Anholt <eric@anholt.net>
53359 Date: Mon Oct 23 14:30:38 2006 -0700
53361 Add work-in-progress integrated TV-out support.
53363 This is the TV connector on board for the 915GM and 945GM.
53365 It is currently not hooked up to output initialization as it's entirely
53366 untested. However, I think this is a reasonable starting point for getting
53367 TV-out actually working.
53369 commit 7d67324fa3a5b5cf0227550316c366752fe4abfb
53370 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53371 Date: Mon Oct 23 17:29:17 2006 +0100
53373 Add some MergedFB checks for accelerator limitations
53375 commit 20d263805a365be548279758bc39d027d6e56d63
53376 Merge: 8548fa5 5ac927d
53377 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53378 Date: Mon Oct 23 14:24:41 2006 +0100
53380 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
53382 commit 8548fa53746c9886eccfb67d022136471d8f3316
53383 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53384 Date: Mon Oct 23 14:23:29 2006 +0100
53386 Fix manpage to report 8160 instead of 6144 for LinearAlloc.
53388 commit 5ac927d6f05ba6f9608d7e758cbac66e8b3a9db1
53389 Author: Michel Dänzer <michel@tungstengraphics.com>
53390 Date: Tue Oct 17 13:27:40 2006 +0200
53392 Don't crash if SetCursorPosition is called with pScrn->currentMode == NULL.
53394 This allows the driver to work with servers that don't have the fix for this
53395 that was only recently pushed to git.
53397 commit 8149681f2eac0af3b70a9457c5204e17da56142b
53398 Author: Eric Anholt <eric@anholt.net>
53399 Date: Mon Oct 16 16:51:04 2006 -0700
53401 Add a per-output mode-valid method.
53403 This is currently disconnected, but will be used in more overhaul work.
53404 This should be where any output limitations, such as clocks, resolution,
53405 scaling limits, or other options, are validated. Other limitations, such as
53406 chipset resolution limits, CRTC clock limits, etc. should be elsewhere.
53408 commit c5cca4c20ae6b519e3b021a9d90809c1b3d1facb
53409 Author: Keith Packard <keithp@neko.keithp.com>
53410 Date: Fri Oct 13 15:31:43 2006 -0700
53412 Advertise textured video adapter first
53414 commit adae75ef2f8a655a699e982ca8f26b934dbe950a
53415 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53416 Date: Fri Oct 13 17:10:27 2006 +0100
53418 Make VIDEO_DEBUG into a configure option
53420 commit 1d4cfaa391ac41099de11b8f3b21e5eb56543909
53421 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53422 Date: Fri Oct 13 13:55:10 2006 +0100
53426 commit cd9e51cd3ab0b36d2f97dfc6d8850dafe169c221
53427 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53428 Date: Fri Oct 13 13:54:31 2006 +0100
53430 Ugh. Turn off debugging.
53432 commit 334aa062c333f2c59bc40e1b9239d3ea71c88a8f
53433 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53434 Date: Fri Oct 13 12:59:30 2006 +0100
53438 commit 8045a7a0af1b276cd7f46ce99d034fd5d503e67c
53439 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53440 Date: Fri Oct 13 12:53:53 2006 +0100
53442 Fix bug #8553 & #8542 - suspend/resume regression from 1.5 driver
53444 commit 6596333b077a692c669f8cb4ae32e145f8da6731
53445 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53446 Date: Fri Oct 13 12:48:16 2006 +0100
53450 commit 1838671476875e9f5b3dde235eacf9fb43afb66c
53451 Author: Eric Anholt <eric@anholt.net>
53452 Date: Tue Oct 10 12:41:01 2006 -0700
53454 Only disable the output when marked disabled.
53456 Also, remove a couple of dead variables.
53458 commit 4198f1216eb13b30d1e92d4395e98861f4324c38
53459 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53460 Date: Tue Oct 10 15:50:10 2006 +0800
53462 Mark current ps kernel is experimential with little test.
53464 commit d485c1f2de11db92b515690fbe38b6aeb6207605
53465 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53466 Date: Tue Oct 10 14:11:35 2006 +0800
53468 Use sf_prog.h instead
53470 commit 2591c1fcf15608d96031be4760cf08534461ca34
53471 Author: Eric Anholt <eric@anholt.net>
53472 Date: Mon Oct 9 14:20:49 2006 -0700
53474 Fix LVDS restore path, and move pipe assignment for outputs to the right place.
53476 commit 09e3d10b0ff69d180467fa9099d12da08e4f681b
53477 Author: Eric Anholt <eric@anholt.net>
53478 Date: Mon Oct 9 13:09:18 2006 -0700
53480 Add a function for describing the output connection configuration.
53482 commit 317cc119c575650c1aa8bf992a0f42bdfffcd7ba
53483 Author: Eric Anholt <eric@anholt.net>
53484 Date: Mon Oct 9 11:49:18 2006 -0700
53486 Move per-output mode setting code to per-output methods.
53488 This is not a very clean interface, as a number of outputs require tweaks to
53489 the DPLL registers. When possible, the DPLLs are just adjusted in the
53490 per-output post_set_mode, which happens just after the DPLL is enabled.
53491 However, this seems better than the previous method of having all outputs
53492 programmed in the same function.
53494 commit 53c28b3980d2682de830e8f86553ccad71527ac1
53495 Author: Adam Jackson <ajax@benzedrine.nwnk.net>
53496 Date: Sat Oct 7 11:18:26 2006 -0400
53498 Expand the check for AOpen Mini-PC.
53500 Just match on subsystem vendor, don't bother inspecting the subsystem device
53501 ID, since apparently they're all busted.
53503 commit bf3820f1f505649ac0730add23d97d6de3f6d22c
53504 Author: Keith Packard <keithp@neko.keithp.com>
53505 Date: Fri Oct 6 21:57:26 2006 -0700
53507 Compute LVDS resolution from server DPI and native panel size.
53509 I was unable to find the native LVDS panel physical size in the BDB
53510 information. I would prefer to report accurate information through RandR if
53513 commit c7a23908a0ef45b14b438f656a5f15582bfcbc1c
53514 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53515 Date: Fri Oct 6 13:35:39 2006 +0100
53517 Fix server regeneration problem mapping the aperture
53519 commit d649fb0d964a9b40b8e04314b5fc14bbbfd41bc3
53520 Author: Keith Packard <keithp@guitar.keithp.com>
53521 Date: Thu Oct 5 22:46:07 2006 -0700
53523 Don't require MonitorLayout when two monitors are plugged in.
53525 With randr12 working, we can just leave the second monitor off for now.
53527 commit 9bb7736ab36f172db58703c4664bb1b0cd7f80c3
53528 Author: Eric Anholt <eric@anholt.net>
53529 Date: Thu Oct 5 16:02:02 2006 -0700
53531 Remove checks for output privates in various bits of code.
53533 Now, the output is only set up if it fully initializes, so it will never exist
53534 if the private it requires doesn't.
53536 commit ada8f62da263d1e93e22df4e0b1149bf1dbe24d4
53537 Author: Eric Anholt <eric@anholt.net>
53538 Date: Thu Oct 5 15:55:07 2006 -0700
53540 Give each output type an init method in its file, making other methods static.
53542 commit 0a5504e59f90abecedd81a8435b88a60fe098f29
53543 Author: Keith Packard <keithp@neko.keithp.com>
53544 Date: Thu Oct 5 09:11:29 2006 -0700
53546 Fix DPI at startup in RandR 1.2 code.
53548 Use requested monitor resolution to compute the appropriate screen size when
53549 resizing the screen during RandR initialization.
53551 commit 4ac81d58b7e0fbffbb4981deffe6a576be821a9b
53552 Author: Keith Packard <keithp@neko.keithp.com>
53553 Date: Wed Oct 4 23:22:52 2006 -0700
53555 Remove mode origins, add preferred mode count.
53557 Just tracking changes in the 1.2 protocol spec.
53559 commit 103b4edce7859ddf58f3e1fadeb427a5e85c7acd
53560 Author: Eric Anholt <eric@anholt.net>
53561 Date: Wed Oct 4 15:00:19 2006 -0700
53563 Move the save, restore, and DPMS per-output settings to per-output files.
53565 commit 16988b27258acfe3d56fc2415aa9ade1ae8d03a9
53566 Merge: cc4148e 5a06000
53567 Author: Keith Packard <keithp@neko.keithp.com>
53568 Date: Wed Oct 4 18:44:43 2006 -0700
53570 Merge branch 'modesetting-guitar' into modesetting
53572 commit 5a060002487e16c53dc96e32af72cd1bfcf6a227
53573 Author: Keith Packard <keithp@guitar.keithp.com>
53574 Date: Wed Oct 4 18:43:07 2006 -0700
53576 Add mode origins for randr
53578 commit 07cdc60be8ebe5d0c21149f3cd2aeb8fed05022a
53579 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53580 Date: Wed Oct 4 11:05:30 2006 +0100
53582 Fix bug #5795, VT switching fails because of a bad test on some platforms.
53584 commit 8da10d15f84ccc992b828546af5ba0d848e7263e
53585 Author: Keith Packard <keithp@neko.keithp.com>
53586 Date: Tue Oct 3 22:00:14 2006 -0700
53588 Enable XV_PORT attribute even when not in Clone mode.
53590 As Clone mode is now something that can change after server startup, always
53591 enable the XV_PORT attribute as we cannot change the list of reported
53593 (cherry picked from 1bc1cedbcdf6ea4d2a3e8b07b553ac50b3c126f4 commit)
53595 commit cc4148e25dc226cb1365e090db21f01b77cb3cbe
53596 Merge: 1bc1ced 3e6f81f
53597 Author: Keith Packard <keithp@neko.keithp.com>
53598 Date: Tue Oct 3 22:00:41 2006 -0700
53600 Merge branch 'modesetting-origin' into modesetting
53602 commit 1bc1cedbcdf6ea4d2a3e8b07b553ac50b3c126f4
53603 Author: Keith Packard <keithp@neko.keithp.com>
53604 Date: Tue Oct 3 22:00:14 2006 -0700
53606 Enable XV_PORT attribute even when not in Clone mode.
53608 As Clone mode is now something that can change after server startup, always
53609 enable the XV_PORT attribute as we cannot change the list of reported
53612 commit 33629ed304b64e45d5640397bd1807c5a98907d1
53613 Author: Keith Packard <keithp@neko.keithp.com>
53614 Date: Tue Oct 3 09:39:22 2006 -0700
53616 Always register list of modes for outputs, even when disabled.
53618 commit 3e6f81f70f65a5ba6b5c3a4e0eeaf67776a5f54d
53619 Author: Eric Anholt <eric@anholt.net>
53620 Date: Mon Oct 2 09:53:57 2006 -0700
53622 Move cursor base address register setting to a separate function.
53624 commit f337eea9e73cbf5c71d07df50d2e5b661a1f34cf
53625 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53626 Date: Mon Oct 2 11:16:42 2006 +0100
53630 commit 2b9bdd8a4021231e4b5f4c5831a46c9ff1e38081
53631 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53632 Date: Mon Oct 2 11:12:23 2006 +0100
53636 commit eca082ca0fec12973a4dbb0106b841136b5b6cdb
53637 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53638 Date: Sat Sep 30 13:44:43 2006 +0100
53642 commit cece9a455b54ab184a3277c164ee6dfa65b70f48
53643 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53644 Date: Sat Sep 30 13:27:27 2006 +0100
53646 Fix issues with Xv locking up the engine. bug #7915.
53648 commit 2013b839de3733fd12becb2cc3c1daadde329eb4
53649 Author: Michel Dänzer <michel@tungstengraphics.com>
53650 Date: Thu Sep 28 13:49:44 2006 +0200
53652 Always call I830UpdateXineramaScreenInfo() unconditionally, and document why.
53654 commit 92c5020bbd31e02c7224798a8f094d237afa19a0
53655 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53656 Date: Thu Sep 28 13:55:52 2006 +0800
53658 Fix picture's transform checking
53660 commit 518802843284973f1f86132afd805c0e662e1ba6
53661 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53662 Date: Thu Sep 28 11:15:33 2006 +0800
53664 Fallback in mask picture for now
53666 Do it later after finish wm kernel program.
53668 commit 25ff5baad95c205c7c690da72fc0f252b3ed9289
53669 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53670 Date: Thu Sep 28 11:09:52 2006 +0800
53672 Fix compile, add wm header file.
53674 commit c3a49b5123d094280b5ff358ceef961958319ce4
53675 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53676 Date: Thu Sep 28 10:36:00 2006 +0800
53678 Add simplest wm kernel program for no mask picture composite
53680 This is a try to use new gen4asm language, and will finish
53681 composite program for mask picture with or without CA case later.
53683 commit 6ea16bf6b06c8b3aed4d2c98679ab28304d1b56c
53684 Merge: fdb6de6 b970166
53685 Author: Eric Anholt <eric@anholt.net>
53686 Date: Wed Sep 27 16:38:01 2006 -0700
53688 Merge branch 'master' into modesetting
53690 This reverts most of the mergedfb code. This will instead be done in device-
53691 independent RandR code.
53704 commit fdb6de663579d3b9f31bf9e8a93430b8505ca73f
53705 Author: Eric Anholt <eric@anholt.net>
53706 Date: Wed Sep 27 13:59:41 2006 -0700
53708 Re-disable broken load-based CRT detection.
53710 commit 1407a42c7378706644fd8be554b43b0e7b581011
53711 Merge: 25890ec 4bd3b89
53712 Author: Eric Anholt <eric@anholt.net>
53713 Date: Wed Sep 27 13:33:23 2006 -0700
53715 Merge branch 'randr-1.2' into modesetting
53717 commit 25890ecda9fd00fad9bc53dea83fc58e0013fcdf
53718 Merge: 965609f 117ff04
53719 Author: Eric Anholt <eric@anholt.net>
53720 Date: Wed Sep 27 13:33:17 2006 -0700
53722 Merge branch 'acpi-hotkey' into modesetting
53729 commit 117ff04b504578a24dff70659e2db1b81aaa1177
53730 Author: Eric Anholt <eric@anholt.net>
53731 Date: Thu Sep 21 17:03:34 2006 -0700
53733 Attempt to make the ACPI hotkey support a little more modesetting-compatible.
53735 Previously, we watched for the BIOS to have changed the layout, and repaired
53736 the resulting configuration. Now, we request that the BIOS make no changes,
53737 but leave a note in a register for when the key has been pressed. When we
53738 notice this, we reprobe monitors and turn on/off the things we find.
53740 This is a temporary solution until we can get the hotkey hooked up as an
53741 input key to external applications to control the change using RandR 1.2. It
53742 is also untested as neither of my laptops do anything with the hotkey.
53743 However, this code does result in many fewer BIOS calls.
53745 commit fe691953f08b4d299295f91450412404ba42810b
53746 Author: Michel Dänzer <michel@tungstengraphics.com>
53747 Date: Thu Aug 31 18:35:10 2006 +0200
53749 Store viewport limits of both pipes in SAREA.
53751 This allows DRI clients to determine which pipe they should synchronize buffer
53752 swaps of each window to.
53754 commit 06c5f7bd076e9ba4b925e061a40b837714bc8267
53755 Author: Michel Dänzer <michel@tungstengraphics.com>
53756 Date: Wed Aug 30 19:45:33 2006 +0200
53758 If the DRM can handle it, enable vertical blank interrupts for both pipes.
53760 commit 89cbc6e215a5f313ccc17370424c35630cf75892
53761 Author: Michel Dänzer <michel@tungstengraphics.com>
53762 Date: Thu Aug 24 10:30:27 2006 +0200
53764 Only enable the IRQ after DRIFinishScreenInit.
53766 This makes sure we've been assigned a context ID, so the interrupt context
53767 won't mess things up if it grabs the HW lock.
53769 commit 185df8a18102eea79b8c8e11304dfee148ac2002
53770 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53771 Date: Wed Sep 27 16:48:43 2006 +0800
53773 Add mask sampler state
53775 commit 1422d4f04ac2d57899eee91f6522f8670f7263da
53776 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53777 Date: Wed Sep 27 13:54:14 2006 +0800
53779 change some src sampler states
53781 sampler for mask should also be set up, and fix
53782 default border texel.
53784 commit 6614b4bda2e079fd1e56e0825894481b80e2df58
53785 Merge: d6ca9fe b3ddcf3
53786 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53787 Date: Mon Sep 25 12:04:01 2006 +0100
53789 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
53791 commit d6ca9feb78a9211758510751bbf47013caabff49
53792 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
53793 Date: Mon Sep 25 12:03:27 2006 +0100
53795 Fix bug #7943 where video flickers when changing
53796 attributes. This code was a leftover from testing lockups
53797 and shouldn't be needed anymore.
53799 commit 52a4f2a02b8f38108bfa866bf74b1596b8125512
53800 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53801 Date: Mon Sep 25 14:35:51 2006 +0800
53803 Add file for i965 exa composite
53805 This does not include ps program, which will be added
53808 commit f272f0d811f9ee059e8f8617a516e6c8bff917a1
53809 Merge: ff202f8 b3ddcf3
53810 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53811 Date: Mon Sep 25 14:15:21 2006 +0800
53813 Merge branch 'master' into exa
53815 commit 4bd3b89c73b6c5aa9b0eb553ad5d553ee0e8a489
53816 Author: Keith Packard <keithp@neko.keithp.com>
53817 Date: Sat Sep 23 16:41:38 2006 +0100
53819 Oops, duplicated CRT-redetect code.
53821 commit c34490bbda6604a21809d15c798607806fa6c725
53822 Author: Keith Packard <keithp@neko.keithp.com>
53823 Date: Sat Sep 23 12:00:43 2006 +0100
53825 Construct default monitor description for hotplug non-DDC monitor.
53827 When detecting a monitor that doesn't support DDC, construct a default
53828 monitor with "sensible" values instead of using whatever the builtin LCD
53829 screen uses. Clearly we need a way to set the monitor parameters when we
53830 cannot detect them.
53832 commit 965609f6fa63e28e5a28128f5bc44f8c4d7b9f68
53833 Author: Eric Anholt <eric@anholt.net>
53834 Date: Fri Sep 22 09:51:45 2006 -0700
53836 Restructure i830_bios.c so we don't leak a copy of the BIOS per generation.
53838 commit c52242c22779a51aa12b18a7a589080ce44c8484
53839 Author: Eric Anholt <eric@anholt.net>
53840 Date: Fri Sep 22 09:41:07 2006 -0700
53842 Remove some dead code related to clock ranges.
53844 commit c2446be9b444b16c95f78dab17bf130f9f491ee2
53845 Author: Eric Anholt <eric@anholt.net>
53846 Date: Fri Sep 22 09:38:11 2006 -0700
53848 Remove the GetDevicePresence BIOS call which just printed BIOS information.
53850 Because we aren't using the BIOS to set modes any more, what the BIOS thinks is
53851 present is probably even less important than before.
53853 commit b6ba268d0d5f22c6a18ce45416452fce83438620
53854 Author: Eric Anholt <eric@anholt.net>
53855 Date: Fri Sep 22 09:31:37 2006 -0700
53857 Remove the no-longer-connected VBERestore option.
53859 commit 20956a5d6f1eb518717a680e58938f31461ca5e4
53860 Author: Eric Anholt <eric@anholt.net>
53861 Date: Fri Sep 22 09:27:30 2006 -0700
53863 Remove unused display{Attached,Present} fields.
53865 commit 2cd28be71472d67956f47c7d49283ebabefa089a
53866 Author: Eric Anholt <eric@anholt.net>
53867 Date: Fri Sep 22 08:55:55 2006 -0700
53869 Remove empty SAVERESTORE_HWSTATE code.
53871 commit ff202f8e2ae4117b464c94047001023d5d0531cc
53872 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53873 Date: Fri Sep 22 15:02:19 2006 +0800
53875 Fix a typo to mark sync for XAA.
53877 This should fix the scrolling screen corrupt in
53880 commit d87d1f5bb0475c6f651fcb7e2cab2a7d46edcc69
53881 Author: Keith Packard <keithp@neko.keithp.com>
53882 Date: Fri Sep 22 02:20:35 2006 +0100
53884 Remove BIOS from non-BIOS related names.
53886 Lots of names included BIOS for no apparent reason; as we try to eliminate
53887 BIOS calls from the driver, these only serve to confuse us.
53888 (cherry picked from 8e5d280d94ad3d3ba3c75871c17abec9da62ed34 commit)
53890 commit b3ddcf3348365873aed4a2b8b2750b330fb4cf26
53891 Author: A Costa <agcosta@gis.net>
53892 Date: Thu Sep 21 23:18:43 2006 +1000
53894 Typo fix in man page (Debian bug #364559).
53896 commit 4820caf46e050761d9b347b8a440381e1b1f4727
53897 Author: Keith Packard <keithp@neko.keithp.com>
53898 Date: Thu Sep 21 01:47:27 2006 -0700
53900 Make planeEnabled track pipes controlled by randr.
53902 Also add code to deal with disabling pipes.
53904 commit c11c445bdeac34253b48192a5d406b55ff8b2be7
53905 Author: Keith Packard <keithp@neko.keithp.com>
53906 Date: Thu Sep 21 01:23:10 2006 -0700
53908 Consistently use Cursor A on Pipe 0 and Cursor B on Pipe 1.
53910 Mixing random cursors and pipes didn't work very well. I'm left wondering
53911 whether the palette stuff will work on pre-9xx series hardware though; it is
53912 special cased everwhere else.
53914 commit 4a8b9515e914369e1faf1dca6b11204b233c73ee
53915 Merge: 18a510b bdca869
53916 Author: Keith Packard <keithp@neko.keithp.com>
53917 Date: Wed Sep 20 22:46:55 2006 -0700
53919 Merge branch 'modesetting' into randr-1.2
53921 commit bdca8697cc6ac5a98e0548b2ce2e0032e9bcfa52
53922 Merge: d9db5ec daade50
53923 Author: Keith Packard <keithp@neko.keithp.com>
53924 Date: Wed Sep 20 22:46:42 2006 -0700
53926 Merge branch 'modesetting-origin' into modesetting
53928 commit 18a510bf9af2f288c0e94359f672775d6e9c9ab2
53929 Merge: f6500e9 d9db5ec
53930 Author: Keith Packard <keithp@neko.keithp.com>
53931 Date: Wed Sep 20 22:45:48 2006 -0700
53933 Merge branch 'modesetting' into randr-1.2
53935 commit f6500e94fec0d6db8c1f1350bee1d137bf06a09e
53936 Author: Keith Packard <keithp@guitar.keithp.com>
53937 Date: Wed Sep 20 22:38:55 2006 -0700
53939 Update driver for RandR 1.2 X server API.
53941 This is not entirely what I'd like to see, but it's at least functional.
53944 Can't disable/enable crtcs
53945 Can't move outputs on/off crtcs
53947 But, it does handle monitor hot-plug, detecting changes in VGA and SDVO
53948 status on-the fly. Which makes for good demo material.
53950 commit daade50ca271d1cdf236bbe84afade85d4111ac9
53951 Author: Linus Torvals <torvalds@osdl.org>
53952 Date: Wed Sep 20 12:07:09 2006 -0700
53954 Add standard C headers to fix build on some systems after xf86_ansic.h removal.
53956 This appears to have been hidden for others by header pollution in X Server
53959 commit 1681fe6bb041385e5aefb11baa9afd1f614abc5b
53960 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53961 Date: Wed Sep 20 09:36:02 2006 +0800
53963 Revert "fix default tex wrap mode with clamp"
53965 oops, we've set border color to all zero, which should
53966 give us transparent in RepeatNone case.
53968 commit d9db5ec6866555ec13ba3ddabb1516eb45637afa
53969 Merge: cbaf3cf e3f4caf
53970 Author: Keith Packard <keithp@neko.keithp.com>
53971 Date: Tue Sep 19 16:03:04 2006 -0700
53973 Merge branch 'modesetting-origin' into modesetting
53975 commit 5adb05f0b82ee3a168142d803ab3434f75cf15a3
53976 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53977 Date: Tue Sep 19 17:14:05 2006 +0800
53979 fix default tex wrap mode with clamp
53981 we don't set any 'default' border color,
53982 default tex wrap mode should be clamp to edge.
53984 commit e3ab89b0327ef6ff790ab53bba29c721aef032cd
53985 Author: Eric Anholt <eric@anholt.net>
53986 Date: Tue Sep 19 10:27:36 2006 +0800
53988 shader program fix for component alpha set
53990 If CA is set and blend op needs src alpha, the src
53991 value is not needed and should be (src.A * mask.X).
53992 This is found in handling exa magic two pass composite.
53994 commit 37429c3ecb97df8faf0f3f56128d7a2b22eb8a5e
53995 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
53996 Date: Tue Sep 19 10:08:31 2006 +0800
54000 commit e3f4caf40708478ef327b029d0a75944c51ea905
54001 Author: Adam Jackson <ajax@benzedrine.nwnk.net>
54002 Date: Fri Sep 15 13:51:18 2006 -0400
54004 Add model-specific tweaks for some funky 945GM boards.
54006 For the Aopen Mini-PC, ignore the claimed attached 800x600 LVDS panel.
54007 Likewise for the Apple Mac Mini, but done slightly differently since it
54008 shares PCI IDs with the Macbook Pro.
54010 commit bd758030a6a59af32f9d73655cb691e018e94a7d
54011 Author: Eric Anholt <eric@anholt.net>
54012 Date: Thu Sep 14 11:12:03 2006 -0700
54014 Add a compile flag to enable syncing after each operation in EXA.
54016 This replaces other debug sync options sprinkled around the EXA code. It
54017 doesn't change the mis-rendering of text on the 915.
54019 commit f46c70e877a3432ba23696e1a16d5906183876af
54020 Author: Eric Anholt <eric@anholt.net>
54021 Date: Thu Sep 14 10:05:51 2006 -0700
54023 Add a note on maxX/maxY for the 2D rendering.
54025 commit ce00db054bf0b7f62ff0be312d7e3470141922dc
54026 Author: Eric Anholt <eric@anholt.net>
54027 Date: Thu Sep 14 10:00:06 2006 -0700
54029 Replace a couple of mis-uses of the screen's bpp when the pixmap's was needed.
54031 commit 71c2dd63361be9616c77db80a93445307d63dd41
54032 Author: Eric Anholt <eric@anholt.net>
54033 Date: Thu Sep 14 09:54:19 2006 -0700
54035 Replace duplicated sync code in exa with a call to I830Sync.
54037 The I830Sync version has additional code for the 965.
54039 commit 564c830a91afda76706a8c4c2f3287d0b8d514e9
54040 Merge: bf49084 b970166
54041 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54042 Date: Wed Sep 13 16:34:10 2006 +0100
54044 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
54046 commit bf49084ad700f064b732bf39a163209eee1d6845
54047 Author: root <root@jetpack.demon.co.uk>
54048 Date: Wed Sep 13 16:32:39 2006 +0100
54050 bump max Xvideo height to 1088 to cope with some players.
54052 commit 59ce4180b95270432540624a1fdf7ae8dd18a7f3
54053 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54054 Date: Mon Sep 11 22:57:23 2006 +0800
54056 reimplement Solid function
54058 Replace COLOR_BLT_CMD with XY_COLOR_BLT_CMD, by which
54059 device would care for direction and provide a (x,y) coord
54060 interface. This fixes pixmap artifacts even in no composite
54063 commit 2b9c87bbf8ee5f7f56631114eb98303cd80e4a48
54064 Author: Eric Anholt <eric@anholt.net>
54065 Date: Fri Sep 8 11:14:05 2006 -0700
54067 Remove xf86_ansic.h usage.
54069 commit b970166eab95ac024ff481b0f5fd9aaf3644aabf
54070 Author: Eric Anholt <eric@anholt.net>
54071 Date: Fri Sep 8 11:09:19 2006 -0700
54073 Fix a compile failure from the last commit.
54075 commit 2b7602847aa7ae9b4e88fbea070710695d2cb8a4
54076 Merge: e4cdc42 b45f4b5
54077 Author: Eric Anholt <eric@anholt.net>
54078 Date: Fri Sep 8 11:02:41 2006 -0700
54080 Merge branch 'exa-origin' into exa
54084 src/i915_exa_render.c
54086 commit e4cdc42a98dd1be74e5c2bf5d9566f7449523b14
54087 Author: Eric Anholt <eric@anholt.net>
54088 Date: Thu Sep 7 10:32:02 2006 -0700
54090 Remove another debug printf.
54092 commit 099be52a3bb52516fb5e8d56c154a4d6b6707e09
54093 Author: Eric Anholt <eric@anholt.net>
54094 Date: Thu Sep 7 10:28:09 2006 -0700
54096 Re-remove xf86_ansic.h, and move copyright to the top of the file.
54098 commit 4ad577f83f7190df4d64e66fbf7c515ee6f411ce
54099 Author: Eric Anholt <eric@anholt.net>
54100 Date: Thu Sep 7 10:24:23 2006 -0700
54102 Remove the UTS/DFS implementations that just match the fallback versions.
54104 commit 8b9d7a5f8560ee55ca5cb5376517a4246cd79fb9
54105 Merge: 6500a3a 1e6e288
54106 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54107 Date: Thu Sep 7 17:45:57 2006 +0100
54109 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
54111 commit 6500a3aa134e11a75b7ddaf7fff253003b223241
54112 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54113 Date: Thu Sep 7 17:45:28 2006 +0100
54115 When in MergedFB mode, fix offscreen memory allocation.
54117 commit c378328ece701756ddca966eb108e2b1f5c3ed6c
54118 Author: Eric Anholt <eric@anholt.net>
54119 Date: Wed Sep 6 11:15:17 2006 -0700
54121 Remove a debug printf.
54123 commit 1e6e288b8826789f3b2520d12426ff7852a67ccd
54124 Author: Eric Anholt <eric@anholt.net>
54125 Date: Fri Sep 1 16:47:17 2006 -0700
54127 Add a check for intel-gen4asm, and rules to compile {wm,sf}_prog.h using it.
54129 This adds assembly source for the two programs used in the textured video
54130 implementation on the 965, which should make them easier to modify in the
54131 future. The compiled versions are also included, so that intel-gen4asm isn't a
54132 build requirement for people that aren't modifying these programs.
54134 There are minor differences in the compiled versions of these programs compared
54135 to their previous versions which were compiled with a different tool. I
54136 believe the changes should be harmless, and video continues to work on my
54139 commit cbaf3cf74bd420533d299c4113761ec536097e33
54140 Author: Keith Packard <keithp@neko.keithp.com>
54141 Date: Thu Aug 31 18:25:21 2006 -0700
54143 verbose debug message for panel sync data
54145 commit 1feb733eb8b09a8b07b7a6987add5149c53b0157
54146 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54147 Date: Tue Aug 29 14:54:15 2006 +0100
54149 Fix bug with probing info of DFP2 and LFP2 connected
54150 devices (Aaron Ridout)
54152 commit 26e6c074c17b2d4e9b9f165468ad45a3b0ecc0ef
54153 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54154 Date: Mon Aug 28 11:55:38 2006 +0100
54156 Fix SetVBlank to operate all the time
54158 commit b45f4b54b189a496255ddae12af38fc4c4408a51
54159 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54160 Date: Fri Aug 25 16:23:18 2006 +0800
54162 Making render log a little quiet in normal
54164 commit 6f0d352b83fc9f39dd86edbda9af83243b50c764
54165 Author: Ross Burton <ross@burtonini.com>
54166 Date: Thu Aug 24 18:16:42 2006 -0700
54168 Bug #7957: Fix distcheck.
54170 commit cc79b3ec19a7e858972228a270ef774041eb8946
54171 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54172 Date: Thu Aug 24 10:28:59 2006 +0800
54174 Current 915 render operations for 915G/945G only
54176 commit 078af29fa65653c5322b006404f7ad29aec24cce
54177 Merge: 7458a6a ea63e82
54178 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54179 Date: Thu Aug 24 10:23:22 2006 +0800
54181 Merge branch 'master' into exa
54191 commit ea63e82ef417a9918e7d7105910a8ddeba2994f6
54192 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54193 Date: Tue Aug 22 16:40:08 2006 +0100
54197 commit b7e57deebbda527e878326cf3e6358c0a48d7817
54198 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54199 Date: Tue Aug 22 19:38:13 2006 +0800
54201 pci id update for Crestline
54203 The "Crestline" banner should be replaced with official
54206 commit dc4128ea0781124984dee0cb41e7b6ee87e362a1
54207 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54208 Date: Mon Aug 21 12:13:42 2006 +0100
54212 commit dad0e7fc09fa7794f20278d9d99abd96a21b691e
54213 Merge: 361c49a 43daaec
54214 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54215 Date: Mon Aug 21 12:08:01 2006 +0100
54217 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
54219 commit 361c49ad3cfe58b9ab98dfb6e95de505c0d63730
54220 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54221 Date: Mon Aug 21 12:07:29 2006 +0100
54223 Fix bug #7930, i810 doesn't compile without DRI
54226 commit 43daaec63929c1f0e54a5125375d8147629da4b9
54227 Author: Eric Anholt <anholt@FreeBSD.org>
54228 Date: Thu Aug 17 15:57:31 2006 -0700
54230 Bug #7905: Mark DRI state as dirty in 965 textured video, fixing hangs with 3d.
54232 commit cdcc6f23e8cad797e6cfd57e2ff454ec8509b5ba
54233 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54234 Date: Wed Aug 16 17:16:28 2006 +0100
54236 Disable device check when mergedfb in operation.
54238 commit a94c96abf90c691faecbebf0c8d08f8010bc67ef
54239 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54240 Date: Wed Aug 16 17:15:26 2006 +0100
54242 Fix bug #5150. Disable LVDS usage on i915G, i945G and i965G.
54244 commit 2a4e486d2990af433e915cfa26dc57cae02e01b7
54245 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54246 Date: Fri Aug 11 11:07:08 2006 +0800
54248 Fix undefined alloc type for agp memory
54250 This shuts up the nonfatal warning that type 3 for alloc agp
54251 memory is undefined for intel-agp.
54253 commit 4cfed93df6b23903704348cafe050a6546e84479
54254 Author: Eric Anholt <eric@anholt.net>
54255 Date: Thu Aug 10 15:41:32 2006 -0700
54257 Bump to 1.6.5 for release.
54259 commit 38e7e48418cd48a46e48f5bc8a6547721db8f76d
54260 Author: Eric Anholt <eric@anholt.net>
54261 Date: Thu Aug 10 15:38:14 2006 -0700
54263 Fix README typos from in the generated file.
54265 commit 975e60261088dee124b329d28c64e508bce1f90b
54266 Author: Eric Anholt <eric@anholt.net>
54267 Date: Thu Aug 10 13:49:43 2006 -0700
54269 Bug #7829: Fix reported driver version.
54271 The driver now reports itself as the PACKAGE_VERSION from autoconf. The DRI
54272 DDX -> client interface version is dissociated from this so that we can do
54273 appropriate major/minor versioning of the interface that's not tied to the
54274 package version. Bumped the i830 ddx dri version patchlevel to note the
54275 fix in the previous commit.
54277 commit 820e9a22fdb759cbdaac4a488322825bc0908b0d
54278 Author: Eric Anholt <eric@anholt.net>
54279 Date: Thu Aug 10 13:37:18 2006 -0700
54281 Bug #7835: Restore unused fields in I830DRIRec to avoid i915 DRI breakage.
54283 As it was, 1.6.4 broke compatibility with the released DRI driver.
54285 commit caa3b35cd772fc75d65a7ff791f00addbb39a764
54286 Author: Eric Anholt <eric@anholt.net>
54287 Date: Wed Aug 9 16:11:30 2006 -0700
54289 Correct typos in README source.
54291 commit 2a7426cf138e518a5eafb40f478359160a7ec98b
54292 Author: Eric Anholt <eric@anholt.net>
54293 Date: Wed Aug 9 15:47:09 2006 -0700
54297 commit 24e59a0daa20b7c3e5028c9ca7972052801d02a1
54298 Merge: bb60807 32f1199
54299 Author: Eric Anholt <eric@anholt.net>
54300 Date: Wed Aug 9 14:19:06 2006 -0700
54302 Merge branch 'textured-video', bringing in fixed-up i915 textured video.
54308 commit bb6080735efc40e103e92b65d0c2f1f729156632
54309 Author: Eric Anholt <eric@anholt.net>
54310 Date: Wed Aug 9 12:21:16 2006 -0700
54312 Bump to 1.6.4 for release.
54314 commit 3ca14275d13b1261b69b0e3fda90a112cb567472
54315 Author: Keith Packard <keithp@neko.keithp.com>
54316 Date: Wed Aug 9 10:35:57 2006 -0700
54318 Add missing headers to i810_drv_la_SOURCES
54320 commit 309374f78df35207b1398e14bba986fb891f3643
54321 Author: Keith Packard <keithp@neko.keithp.com>
54322 Date: Wed Aug 9 10:30:41 2006 -0700
54324 Use double quotes to avoid sgml syntax error
54326 commit dca9f856ca21e63abeb87e4ef2c40944c26d4429
54327 Author: Keith Packard <keithp@neko.keithp.com>
54328 Date: Wed Aug 9 10:29:59 2006 -0700
54332 commit b19ea222727ed47b69d28a03242c09d3a6ab2673
54333 Author: Eric Anholt <eric@anholt.net>
54334 Date: Wed Aug 9 10:01:29 2006 -0700
54336 Update the README.sgml file for the upcoming release.
54338 commit f8b47f607297e0591b63e5f25296af07ee74f433
54339 Author: Keith Packard <keithp@neko.keithp.com>
54340 Date: Tue Aug 8 16:19:51 2006 -0700
54342 Update manual page to include i945 and later details.
54343 (cherry picked from 7833d0733bc146cf1ae7f588516c49797886b396 commit)
54345 commit d150b53d102c511f3c9245ef1f6fd36c12b01ca3
54346 Merge: bb65a4d bc5f565
54347 Author: Eric Anholt <eric@anholt.net>
54348 Date: Tue Aug 8 15:51:58 2006 -0700
54350 Merge branch 'i965', adding i965G support.
54357 commit bc5f56568021d8c63313e2b6bf30710e7516c04a
54358 Author: Eric Anholt <eric@anholt.net>
54359 Date: Tue Aug 8 15:48:04 2006 -0700
54361 Intel bug #49: Fix video output at 32bpp by using B8G8R8A8 instead of B8G8R8X8.
54363 While here, don't overallocate video memory for the i965G state.
54365 commit c3b3d479788fcea7e543f29acf83c85b8b148fbe
54366 Author: Alan Hourihane <alanh@tungstengraphics.com>
54367 Date: Tue Aug 8 15:28:14 2006 -0700
54369 Intel bug #35: Fix accelerator syncing with DGA.
54371 Fixes glitches seen with Mark Vojkovich's "texture" demo.
54373 commit bc12208f6e145ec29c3ebe38ae04dc2ebca1b4cc
54374 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54375 Date: Mon Aug 7 13:27:00 2006 +0800
54377 Disable error register dumping in dri TransitionTo2d. This's for
54378 debug which might confuse QA.
54380 commit bb65a4dc55b416437d61ccbee8b6056fe8e08357
54381 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54382 Date: Tue Aug 8 14:51:18 2006 +0100
54386 commit 37644293da751bfc7268c9fc74ecda1044607dee
54387 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54388 Date: Tue Aug 8 14:42:32 2006 +0100
54390 check for xineramaproto
54392 commit 3661d4df86db012682ab4b393287aa02e9fc8453
54393 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54394 Date: Tue Aug 8 12:25:03 2006 +0100
54396 Re-enable monitor detect after a few fixups.
54398 commit 995b142f510d1daab3914d336f3c0d017a043b41
54399 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54400 Date: Tue Aug 8 12:01:39 2006 +0100
54402 Disable detecting of new monitors on VT switch.
54403 It's doesn't always work. The modesetting
54404 branch will deal with hotplug displays correctly.
54406 commit 8e5844bbf9c8880c9eb6eff29b6db2db06b01933
54407 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54408 Date: Tue Aug 8 11:43:10 2006 +0100
54410 bump to 1.6.2 to indicate mergedfb support
54412 commit 421b415e23c1ddc78837cd222167d6ed71a3ef88
54413 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54414 Date: Tue Aug 8 11:28:42 2006 +0100
54416 Fix a build problem.
54418 commit e4f63eaf5efb654121db148f2f8e32cc3b23b6b5
54419 Merge: 633a683 e26f3e3
54420 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54421 Date: Tue Aug 8 11:07:56 2006 +0100
54423 Merge branch 'master' of git+ssh://xorg.freedesktop.org/git/xorg/driver/xf86-video-intel
54425 commit 633a683a4adcb9a44a54519fd7ff66aab2d12f97
54426 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54427 Date: Tue Aug 8 10:23:29 2006 +0100
54429 Ensure palette is updated in mergedfb & clone modes
54431 commit 71e3e2d4e3b2a2c538fe6f1cea41f442fdb8d756
54432 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54433 Date: Fri Aug 4 20:39:50 2006 +0100
54435 Fix a problem creating the I2C bus for the SDVOC
54436 interface, due to a name match with SDVOB.
54438 Bus names must be unique.
54440 commit 0fd4831fdcf4c8f43d80c66e43eff8942f89b324
54441 Author: Eric Anholt <anholt@FreeBSD.org>
54442 Date: Fri Aug 4 00:21:53 2006 -0700
54444 Disable dynamic front buffer mapping on i965.
54446 Moving front buffers should only be necessary for rotation. Currently, the
54447 server isn't ready for it, and the method attempted to work around it caused
54448 crashes with DRI. Since i965 doesn't support rotation yet, this should be
54451 commit 8d0a5138503586cbf980eb9464f2db91b72509c7
54452 Author: Eric Anholt <anholt@FreeBSD.org>
54453 Date: Thu Aug 3 19:16:59 2006 -0700
54455 Add parenthesis so that IS_I965G doesn't make the test pass for 8-bit.
54457 commit aa69018c01d2fa963fb940718dbd653d6ca2c9eb
54458 Author: Eric Anholt <anholt@FreeBSD.org>
54459 Date: Thu Aug 3 19:12:15 2006 -0700
54461 Turn off video debugging now that it appears to work fine.
54463 commit d15f87110807111ab7d71f1254ebe3dfef1fee80
54464 Merge: a79aa0d d56ffa5
54465 Author: Eric Anholt <anholt@FreeBSD.org>
54466 Date: Thu Aug 3 19:08:34 2006 -0700
54468 Merge branch 'broadwater-video-rehash' into i965
54470 The previous merge wasn't done on a synced-up tree, and missed necessary
54473 commit a79aa0d7b27b0b9b032472776d7dda410fd66a98
54474 Merge: 760021e f9e94c1
54475 Author: Eric Anholt <anholt@FreeBSD.org>
54476 Date: Thu Aug 3 17:58:56 2006 -0700
54478 Merge branch 'broadwater-video-rehash' into broadwater
54496 commit 760021e3983f7783900075b8c9603bd4fbe7e0a2
54497 Author: Alan Hourihane <alanh@tungstengraphics.com>
54498 Date: Thu Aug 3 17:08:39 2006 -0700
54500 Add current Tungsten Graphics code drop for i965 support.
54502 commit d56ffa5f35e3cf4262d66469052b2122fdb24027
54503 Author: Eric Anholt <anholt@FreeBSD.org>
54504 Date: Thu Aug 3 16:03:50 2006 -0700
54506 Bump PS_MAX_THREADS to 32 now that the program doesn't fail.
54508 commit 4525379d95ff292d7322e1a7a516c0bedd1f7543
54509 Author: Eric Anholt <anholt@FreeBSD.org>
54510 Date: Thu Aug 3 16:03:15 2006 -0700
54512 Make the sampler's payload be the WM payload rather than uninitialized data.
54514 The sampler's payload happens to be in the same format as the WM payload,
54515 though most of the fields are ignored.
54517 This appears to fix the program in the presence of multiple PS threads.
54519 commit ad2c70b4121121f1fb53190ea49edf2323c804a9
54520 Author: Eric Anholt <anholt@FreeBSD.org>
54521 Date: Thu Aug 3 12:47:19 2006 -0700
54523 Remove some stale XXX-prefixed comments.
54525 commit f9e94c17c55e4c75802d8574c908744e286e7843
54526 Author: Eric Anholt <anholt@FreeBSD.org>
54527 Date: Wed Aug 2 21:18:19 2006 -0700
54529 Set the WM scratch space that we had already allocated.
54531 It appears to be required, even if the kernel doesn't use any scratch space.
54533 commit aefa6fdfc5300546caeb64ace14a7854d3dc7dae
54534 Author: Eric Anholt <anholt@FreeBSD.org>
54535 Date: Wed Aug 2 21:14:14 2006 -0700
54537 Clean up GRF allocation (which was wrong at 16-register boundaries).
54539 Also use PS_MAX_THREADS rather than hard-coding 1 thread, and remove the dead
54540 SF_KERNEL_NUM_URB macro.
54542 commit 7a64e14624514ef31f6fa9f15e8804c45f930212
54543 Author: Eric Anholt <anholt@FreeBSD.org>
54544 Date: Wed Aug 2 20:48:13 2006 -0700
54546 Crank down the SF allocation and comment on why this is a fine lower limit.
54548 commit bc6a2bb7576a7c1e7971f6d1e0b893b2ada1aaa3
54549 Author: Eric Anholt <anholt@FreeBSD.org>
54550 Date: Wed Aug 2 20:34:57 2006 -0700
54552 Remove the clip URB allocation.
54554 Previously, the VS was misconfigured and exceeding its allocation, which the
54555 (unused) clip was providing padding for.
54557 commit defe2795429484ffe4c1438bafb86bb5e5469ba9
54558 Author: Eric Anholt <anholt@FreeBSD.org>
54559 Date: Wed Aug 2 20:32:41 2006 -0700
54561 Correct the VS setup, and allocate a correct, minimal number of URB entries.
54563 The VS number of URB entries and URB entry size are always used, even when
54564 the VS is disabled. Similarly, the cache enable bit is always used.
54566 commit b57ccb682cb3dea3e26c6f1b0c709e63dfde0d31
54567 Author: Eric Anholt <anholt@FreeBSD.org>
54568 Date: Wed Aug 2 19:46:15 2006 -0700
54570 Replace the SF max threads setting with a define for easier tweaking.
54572 Tweak it to 1 for now.
54574 commit 82037a12758c41a304f2e0bbd033d3345cccbe1a
54575 Author: Eric Anholt <anholt@FreeBSD.org>
54576 Date: Wed Aug 2 19:33:28 2006 -0700
54578 Remove CS URB allocation since we don't use any constants.
54580 commit 1d45668d7a42bfa5d7f5bfb68d8bae38bda0936b
54581 Author: Eric Anholt <anholt@FreeBSD.org>
54582 Date: Wed Aug 2 19:18:20 2006 -0700
54584 We only need 3 vertices to fit in the URB, since we only dispatch 3.
54586 commit a076d35bed6f13cf943a0f8948176aa0c999e2da
54587 Author: Eric Anholt <anholt@FreeBSD.org>
54588 Date: Wed Aug 2 19:16:03 2006 -0700
54590 No GS URB allocation is necessary when the function is disabled.
54592 commit befa655168fb8dcb6806592eb44f7ac49f191822
54593 Author: Eric Anholt <anholt@FreeBSD.org>
54594 Date: Wed Aug 2 19:11:38 2006 -0700
54596 Reduce URB_VS_ENTRY_SIZE to 1 as our vertices are under 8 floats.
54598 commit 33acbdca0a0f82725e5bf7887b325726403a6ffd
54599 Author: Eric Anholt <anholt@FreeBSD.org>
54600 Date: Wed Aug 2 19:09:19 2006 -0700
54602 Remove the VS kernel and binding table.
54604 The VS URB entries have to remain as they're used to store the VF output which
54605 isn't modified by a VS program.
54607 commit aafa48cb85cd03c735fb968a4275c19e1a68cd02
54608 Author: Eric Anholt <anholt@FreeBSD.org>
54609 Date: Wed Aug 2 18:26:26 2006 -0700
54611 Fix wm prog to correct the ordering of the Cr and Cb channels.
54613 commit bc2c842d93de04d48c7de60482814db346bd0b78
54614 Author: Eric Anholt <anholt@FreeBSD.org>
54615 Date: Wed Aug 2 18:10:01 2006 -0700
54617 Allocate space for the 965's state at the end of the video buffer.
54619 Fixes corruption in the first few lines of the video.
54621 Based on 1b506798d98d911be733543da2c40cb451a28912
54623 commit 524460ea1f02bb6e8e2239d7763334666012cec4
54624 Author: Eric Anholt <anholt@FreeBSD.org>
54625 Date: Wed Aug 2 17:47:55 2006 -0700
54627 Updated WM kernel to load video and do colorspace conversion.
54629 commit ba896c779c697e1d7458028798ec49013bd9da9f
54630 Author: Eric Anholt <anholt@FreeBSD.org>
54631 Date: Wed Aug 2 17:47:37 2006 -0700
54633 Updated grf/urb state for WM.
54635 commit e5c572f841b626b8b6f21a6966a33956d3b0b35b
54636 Author: Adam Jackson <ajax@benzedrine.nwnk.net>
54637 Date: Wed Aug 2 20:47:12 2006 -0400
54639 Fix a braino in mode list pruning.
54641 Interpreting the size of the display in centimeters as the size in pixels,
54642 and then clipping the modes list based on that, rarely does what you want.
54644 commit 21b62df7c34217be5dd95985c35e33be11c25846
54645 Author: Eric Anholt <anholt@FreeBSD.org>
54646 Date: Wed Aug 2 17:36:49 2006 -0700
54648 Move the WM kernel to a separate file.
54650 commit 5d3424492f9586a4c5a28962a9757f48f2c12e83
54651 Author: Eric Anholt <anholt@FreeBSD.org>
54652 Date: Wed Aug 2 17:34:12 2006 -0700
54654 Replace SF kernel with the one from broadwater-video HEAD.
54656 commit 7458a6adb5ea62f56bb3d4ab19ad7d1aa6ad2d19
54657 Author: Eric Anholt <anholt@FreeBSD.org>
54658 Date: Mon Jul 31 10:09:17 2006 -0700
54660 Move MAPSURF_XBIT settings into the texture formats structure.
54662 commit 7ea74843bbdbf54b6804727be9b41d26832c8f76
54663 Author: Eric Anholt <anholt@FreeBSD.org>
54664 Date: Mon Jul 31 10:06:34 2006 -0700
54666 Whitespace cleanup.
54668 commit e71108f1e05b7a8d8edd174eb64edd6cccacbcdc
54669 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54670 Date: Fri Jul 28 10:32:12 2006 +0100
54672 Fix DGA with MergedFB
54673 Turn off rotation support when MergedFB enabled
54675 commit a91a4f95c664f6905fef61dab251707bf2548bb8
54676 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54677 Date: Thu Jul 27 16:11:48 2006 +0100
54679 Fix pipe reversal for Xv
54681 commit ac3ad32f667b306e771617d784648f7111743f1a
54682 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54683 Date: Thu Jul 27 15:28:42 2006 +0100
54685 Calculate allowable refresh rates on the private
54686 mode data for each independent screen in mergedfb.
54688 Lots of other fixes too.
54690 commit 30952e58ed83e2e18d1007f662d2cc9a773c876c
54691 Author: Eric Anholt <anholt@FreeBSD.org>
54692 Date: Wed Jul 26 13:42:12 2006 -0700
54694 Limit drawing to the destination pixmap's boundaries, not the screen's.
54696 This shouldn't matter, as miComputeCompositeRegion shouldn't giving us
54697 anything that would draw outside the bounds, anyway.
54699 commit cffd2cd36d0437b38ac8164d66ea71be50b19330
54700 Author: Eric Anholt <anholt@FreeBSD.org>
54701 Date: Wed Jul 26 13:22:51 2006 -0700
54703 Fix hangs when compositing with a mask.
54705 The modify bits in the mask field for sampler/map state appear to actually
54706 be enable bits. So, prepare the state values in I915TextureSetup,
54707 then write the sampler/map state out all at once in PrepareComposite.
54709 commit 96754b822df7ac110a70b2d08dda2ebb299772be
54710 Author: Eric Anholt <anholt@FreeBSD.org>
54711 Date: Wed Jul 26 13:19:12 2006 -0700
54713 Fix a couple of typos from code review.
54715 commit e786e2f9f3a4df31702736db6f68a44c9ebba546
54716 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54717 Date: Wed Jul 26 16:45:38 2006 +0100
54719 When detecting new monitors in mergedfb renew the modepool.
54721 commit 8e6e990db34d63174670512f494fa9adb44786f5
54722 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54723 Date: Wed Jul 26 10:48:47 2006 +0100
54725 Update Xvideo to deal with MergedFB modes.
54727 commit e26f3e30b30a57ab4aad0267d689a9a5d7a5e877
54728 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54729 Date: Wed Jul 26 09:17:52 2006 +0100
54731 Fix a build problem
54733 commit b919db75d2f6dc1019f981534b0d5d87c6029727
54734 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54735 Date: Wed Jul 26 09:07:19 2006 +0100
54739 commit 9c93d1498fa2363c52ef7fbe97d781560f67acf5
54740 Author: Eric Anholt <anholt@FreeBSD.org>
54741 Date: Tue Jul 25 21:17:05 2006 -0700
54743 Oops, partial revert of last commit. Some were obviously not NOOPS.
54745 commit 38d1a5e0dbe059f5c01bd5120a108a386ff10718
54746 Author: Eric Anholt <anholt@leguin.anholt.net>
54747 Date: Tue Jul 25 21:09:12 2006 -0700
54749 Replace some OUT_RING(0)s with OUT_RING(MI_NOOP) to remind me why they're there.
54751 commit fd19b12793f09b6714468556ace875ef36ed9e1c
54752 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54753 Date: Tue Jul 25 11:14:11 2006 +0100
54755 Add mergedfb support to the intel driver with
54756 additional pseudo-Xinerama support.
54758 commit aec88cf4c66833f89afa5f8fa4f8bcf30b2b6b4b
54759 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54760 Date: Mon Jul 24 15:46:19 2006 +0800
54762 remove crap blit in UploadToScreen, we should find a more
54763 efficent way and implement blit correctly.
54765 commit 32f1199937e92b9100aba52cbbb97157014e3182
54766 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54767 Date: Mon Jul 24 15:42:15 2006 +0800
54769 remove an extra '-'
54771 commit 02c5c4cd1dd44058d80ec23cccfd45a13bc3bb2c
54772 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54773 Date: Fri Jul 21 12:54:06 2006 +0800
54777 commit 4f0ad337098eb0bacb1c11bc61d780ac18041e1d
54778 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54779 Date: Fri Jul 21 11:59:04 2006 +0800
54781 fixup segfault in xaa when xaa not initialized in i830WaitSync.
54783 commit 914327f40d07a3f7c069752ce005bf8e21352b03
54784 Author: Eric Anholt <anholt@FreeBSD.org>
54785 Date: Thu Jul 20 17:58:07 2006 -0400
54787 Disable drawing to PICT_a8, and turn on support for non-extended repeats.
54789 We don't know what COLOR_BUF_8BIT does, so it's disabled for now while we're
54790 stabilizing the code.
54792 The non-extended repeats appear to be working according to the rendercheck
54793 tests on a8 and a8r8g8b8 sources. Masks continue to hang the card, as before
54796 commit 45cb032e2b7ac1e4c765fc29a2e4dd0cef19066d
54797 Author: Eric Anholt <anholt@FreeBSD.org>
54798 Date: Thu Jul 20 17:34:50 2006 -0400
54800 Use MT_8BIT_A8 so PICT_a8 expands to (0,0,0,a) not (a,a,a,a)
54802 commit fd6d825ae0780f3daf436f4224bf5f177cb8630e
54803 Author: Eric Anholt <anholt@FreeBSD.org>
54804 Date: Thu Jul 20 17:25:31 2006 -0400
54806 Move ss6 setting with other immediate state, and ensure that stencil is off.
54808 commit d6d6a44c66446f34b57394e5ea9a8e32917f7569
54809 Author: Eric Anholt <anholt@FreeBSD.org>
54810 Date: Thu Jul 20 11:23:05 2006 -0400
54812 Try to fix up the basic RepeatNormal and RepeatNone support.
54814 With RepeatNone, access outside of the drawable should return 0 alpha values,
54815 which is done by using the border color. For RepeatNormal, we use the wrap
54816 mode (which requires that we use normalized texture coordinates).
54818 commit 2b2173c618c080b1678990d65fe49b52bd8001b4
54819 Author: Eric Anholt <anholt@FreeBSD.org>
54820 Date: Thu Jul 20 11:15:18 2006 -0400
54822 The pixmap's drawable x/y coordinates are always 0, so don't bother using them.
54824 commit 9f1cec83e0aef36c7d3482e62e8f01595f1fd076
54825 Author: Eric Anholt <anholt@FreeBSD.org>
54826 Date: Thu Jul 20 11:05:49 2006 -0400
54828 The vertex data are all floats, so write the dest coords as floats.
54830 commit 4737955a62c39177e7a7ce7749a2f20e111afc68
54831 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54832 Date: Fri Jul 21 01:02:38 2006 +0800
54834 move i830 default coord set into I830EmitInvarientState
54836 commit a7e30bb051bc03063fe699473610a57054a64973
54837 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54838 Date: Fri Jul 21 00:56:58 2006 +0800
54840 Take from i915, blend ctl code cleanup.
54842 commit ac34a37cb60d763cbd99b6e2f6fdcb639592b99b
54843 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54844 Date: Fri Jul 21 00:52:59 2006 +0800
54846 Take instruction fixes into i830, with indent changes.
54848 commit 3e86bf4273241daa23645f5a8104913e4205a74b
54849 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54850 Date: Thu Jul 20 13:43:27 2006 +0800
54852 remove i915 default ctx setup.
54854 commit 5c945bf9eb9f5bedd34d98dcc7abc3f1ea3cb962
54855 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54856 Date: Thu Jul 20 13:40:39 2006 +0800
54858 Revert "Fix several size error with STATE3D_LOAD_STATE_IMMEDIATE cmd."
54860 This reverts commit 433cb6ba82698676f6f72e09834aba4d64611d54.
54862 commit d5243d1b35ab1db608c81889819e4d4d87048154
54863 Author: Eric Anholt <anholt@FreeBSD.org>
54864 Date: Wed Jul 19 19:27:53 2006 -0400
54866 The height and width of texture maps are one more than the programmed values.
54868 commit 13e372c2095fde94ae41bea959fba0ad95a59c08
54869 Author: Eric Anholt <anholt@FreeBSD.org>
54870 Date: Wed Jul 19 16:42:57 2006 -0400
54872 Fix the texture map pitch setting as documented, and as used by the 3d driver.
54874 commit da5efb09d55291ce5bcabff7db8d1490bb5e838b
54875 Author: Eric Anholt <anholt@FreeBSD.org>
54876 Date: Wed Jul 19 16:36:52 2006 -0400
54878 Fix the alignment padding fix (didn't bump BEGIN_LP_RING count).
54880 commit 36aa43bf73ee1268f0a250788eb637a4123f08eb
54881 Author: Eric Anholt <anholt@FreeBSD.org>
54882 Date: Wed Jul 19 16:22:46 2006 -0400
54884 Force texture alpha channels to 1 if not present.
54886 commit adce6bd0fc9533156da09da7d029ba0d76606b4f
54887 Author: Eric Anholt <anholt@FreeBSD.org>
54888 Date: Wed Jul 19 16:21:49 2006 -0400
54890 Add a trailing MI_NOOP to frag shader output for alignment if necessary.
54892 commit 4e409ea44180f7ef2780d70c9d2f54629bbd0fc3
54893 Author: Eric Anholt <anholt@FreeBSD.org>
54894 Date: Wed Jul 19 15:57:09 2006 -0400
54896 The map bits per pixel must come from the drawable, not the visible screen.
54898 commit 449e4be503c8b938d5bd46ee810244f8bcf54ceb
54899 Author: Eric Anholt <anholt@FreeBSD.org>
54900 Date: Wed Jul 19 13:34:06 2006 -0400
54902 Correct the DWORD count of several 3D instructions.
54904 commit f1b62d890ca22e12d61f7ef67bc4e35d68cde019
54905 Author: Eric Anholt <anholt@FreeBSD.org>
54906 Date: Wed Jul 19 13:02:24 2006 -0400
54908 Replace hand register setting with new i915 fragmet program API.
54910 commit 76a316d0cec92a04c8735926a76e76c21f960d7e
54911 Merge: 433cb6b 8480516
54912 Author: Eric Anholt <anholt@FreeBSD.org>
54913 Date: Wed Jul 19 12:31:36 2006 -0400
54915 Merge branch 'master' into exa
54921 commit baf65ce98abcdd21dff2531a43bb9c5044732c28
54922 Author: Eric Anholt <anholt@FreeBSD.org>
54923 Date: Tue Jul 18 19:42:37 2006 -0400
54925 Re-convert i915 video to new fragment shader API.
54927 Although in the history of this branch it had happened before, this time it's
54930 commit bb81e8d6c777a5e16b8193c07667fbee8e21203e
54931 Merge: 2a1b3cf 8480516
54932 Author: Eric Anholt <anholt@FreeBSD.org>
54933 Date: Tue Jul 18 19:23:21 2006 -0400
54935 Merge branch 'master' into textured-video
54937 This moves the i915 textured video implementation into i915_video.c to avoid
54938 conflicts in register definitions with i830_reg.h when we use i915_reg.h.
54939 This also means that i810_reg.h's i915 3D regs definitions are removed and
54940 replaced with i915_reg.h usage.
54946 commit 84805167ab8a422966355b9753bfcb4dad802413
54947 Author: Eric Anholt <anholt@FreeBSD.org>
54948 Date: Tue Jul 18 18:27:10 2006 -0400
54950 Convert i915 rotate code to the new fragment program API.
54952 commit 5176d62ba58c100c87f75a4f333d00129d780c99
54953 Author: Eric Anholt <anholt@FreeBSD.org>
54954 Date: Tue Jul 18 16:18:18 2006 -0400
54956 Add an API for programming i915 fragment programs.
54958 commit 148ef9bdd9e0ef3e7ac86b56a8662b53a3ea9168
54959 Author: Eric Anholt <anholt@FreeBSD.org>
54960 Date: Mon Jul 17 22:32:25 2006 -0700
54962 Convert magic numbers to symbolic names in i915 rotate code.
54964 This doesn't cover the fragment shader yet, which we need to make a sensible
54965 set of macros for (at least the basic bits).
54969 commit 433cb6ba82698676f6f72e09834aba4d64611d54
54970 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54971 Date: Tue Jul 18 17:00:50 2006 +0800
54973 Fix several size error with STATE3D_LOAD_STATE_IMMEDIATE cmd.
54974 Issue texture coord set with proper scale value.
54976 commit 855a9f13cb7fd4f6c489ff280bf389475bbd9bbe
54977 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
54978 Date: Tue Jul 18 10:02:47 2006 +0800
54980 fallback in 'repeat' case for now
54982 commit a6d438ebe3cf141a0331e0cd55eb9b5e137a5e37
54983 Author: Keith Packard <keithp@neko.keithp.com>
54984 Date: Mon Jul 17 14:11:19 2006 -0400
54986 Disable spread spectrum clock usage.
54988 The spread spectrum clock generator that improves EMI characteristics for
54989 laptop screens lives in an external chip that is programmed over an i2c bus.
54990 Without correct programming, attempts to use this mode for the LVDS can
54991 result in a DC signal being sent to the panel.
54993 Until we find programming information for this external chip, we should
54994 leave this mode disabled.
54996 commit 16d6263e6518a4a05562e2842ff2d0fdb4710304
54997 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
54998 Date: Sun Jul 16 20:39:52 2006 +0100
55000 whoops, reverse part of that.
55002 commit 2f50f6d1b1b3fa4fbec98bd8fa5818df890070e7
55003 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55004 Date: Sun Jul 16 20:17:38 2006 +0100
55006 move ContextMem out of XF86DRI
55008 commit c7083a267209c93b2a91ef00dea2ca840400d160
55009 Author: Eric Anholt <anholt@FreeBSD.org>
55010 Date: Fri Jul 14 13:23:40 2006 -0700
55012 Bug #7404: Only save/restore VGA fonts and not other VGA regs.
55014 This fixes a hang on the i945 during restore. It appears that saving/restoring
55015 the VGA registers is not important, as we're correctly saving/restoring the
55016 registers we touch within the driver anyway.
55018 commit 49b827605628d3e1a6d4d41447cf46c5f38cc0c2
55019 Author: Eric Anholt <anholt@FreeBSD.org>
55020 Date: Fri Jul 14 14:23:04 2006 -0700
55022 Add a register restore implementation so we don't crash on LeaveVT.
55024 We shouldn't ever need to save/restore the hi/lo frequency regs, as they're
55027 commit 6a92a779646ec03a03a3b1f45170b2e705ce8934
55028 Author: Eric Anholt <anholt@FreeBSD.org>
55029 Date: Fri Jul 14 14:20:41 2006 -0700
55031 Don't try to probe on more pipes than we really have.
55033 commit 04d1584737fd0d14e99608a97281fd7b1549ae0e
55034 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55035 Date: Fri Jul 14 16:14:18 2006 +0800
55037 Current exa render implement for i830 and i915, test on
55038 865GM and 915G. There is issue in picture 'repeat' support.
55039 And also stop recursive behavior in I830WaitLpRing to allow
55040 server to abort instead of system hang.
55042 commit de470aaf5c47f4d2b0f477ac678039ef43af773d
55043 Author: Adam Jackson <ajax@benzedrine.nwnk.net>
55044 Date: Thu Jul 13 19:10:11 2006 -0400
55046 In I830xf86SortModes, catch cases where two modes are equal in only one
55047 dimension, by comparing the areas of the modes. Otherwise, 800x600 would
55048 sort before 1024x600 if it was added later.
55050 commit 22843830ebdd14247aa76d19f89494a56e2ac887
55051 Author: Adam Jackson <ajax@benzedrine.nwnk.net>
55052 Date: Thu Jul 13 17:20:17 2006 -0400
55054 Fix a thinko; would only inject the FP native mode if a mode list was already
55055 found, which is never the case when there's no xorg.conf.
55057 commit 05bcbadd130524694e11e372d54cb419cea566cc
55058 Author: Eric Anholt <anholt@FreeBSD.org>
55059 Date: Tue Jul 11 14:05:38 2006 -0700
55061 Avoid NULL dereference if cursor position changes during a mode change.
55063 commit b65f18b05a5fba506b71293b495cab95197037ac
55064 Author: Eric Anholt <anholt@FreeBSD.org>
55065 Date: Tue Jul 11 13:29:57 2006 -0700
55067 Bug #7443: Respect the user's Modes configuration, and make it more useful.
55069 Now, mode names generated by DDC get names of the form "WIDTHxHEIGHTxREFRESH".
55070 The matching for user Modes lines takes the user Modes as the prefix that
55071 needs to match, rather than an exact string match or "WIDTHxHEIGHT" match. So
55072 one can, for example, specify "1024x768" to get any old 1024x768, or
55073 1024x768x60 to get one of the modes named 1024x768x60.
55075 commit 5a2e04bd1b700a8a6e26136b8831ef5e4d11b565
55076 Author: Eric Anholt <anholt@FreeBSD.org>
55077 Date: Tue Jul 11 10:21:51 2006 -0700
55079 Fix crash with DDC when there are no user modes to add.
55081 commit f9499a68da0ce459fed0b29b998678fd81898a51
55082 Author: Luká\u0161 Hejtmánek <xhejtman@mail.muni.cz>
55083 Date: Tue Jul 11 10:13:18 2006 -0700
55085 Make gamma settings apply to the cursor as well, and fix clone-mode gamma.
55087 commit 7068468ac1951bfca0071bb9b1a99df4f37368a0
55088 Author: Luká\u0161 Hejtmánek <xhejtman@mail.muni.cz>
55089 Date: Tue Jul 11 09:51:26 2006 -0700
55091 Add support for adjusting saturation value of overlay video.
55093 commit b1c2ea653502dd8547079e7014b698f241433dff
55094 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55095 Date: Tue Jul 11 08:13:30 2006 +0100
55097 whoops, revert some unnecessary changes
55099 commit 8a44a7acfcadbba2410dca750afc9d32bc83706e
55100 Merge: 584b544 e7723a4
55101 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55102 Date: Tue Jul 11 07:41:27 2006 +0100
55104 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-intel
55106 commit 584b544987be5cf23dce29ddaf3130e59cfe6fa8
55107 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55108 Date: Tue Jul 11 07:40:40 2006 +0100
55110 Add an additional check before rotating
55112 commit b912bf5673e38e03b0b25c2f5d05fe7e26994ba1
55113 Author: Eric Anholt <anholt@FreeBSD.org>
55114 Date: Mon Jul 10 18:35:20 2006 -0700
55116 Clean up warnings in sil164 module.
55118 commit d75490701cdbf2ab6eab82eaa078790a5fe0aea0
55119 Author: Eric Anholt <anholt@FreeBSD.org>
55120 Date: Mon Jul 10 18:21:04 2006 -0700
55122 Hook up SiI164 mode setting (just a matter of turning the chip on).
55124 Also adds register dumping in case this turns out to not be enough, and fixes
55125 a couple of prototypes.
55127 commit 426d26ea446d646fa8f561ea0e03c8e4a2c0c315
55128 Author: Eric Anholt <anholt@FreeBSD.org>
55129 Date: Mon Jul 10 18:19:51 2006 -0700
55131 Fix prototype for SaveRegs.
55133 commit 8d7987d00242020d29a2574ac0c8b6e55cc22112
55134 Author: Eric Anholt <anholt@FreeBSD.org>
55135 Date: Mon Jul 10 17:34:57 2006 -0700
55137 Move to 4-space indents in sil164.
55139 commit f5a01a2ef02125611d5fb74c20d53d52e544701a
55140 Author: Eric Anholt <anholt@FreeBSD.org>
55141 Date: Mon Jul 10 17:17:51 2006 -0700
55143 Make DVO code light up my sil164-based DVI output, when already set up by BIOS.
55145 commit 23a0ee73bce12f9e0b881af420413aeec4c0517f
55146 Author: Eric Anholt <anholt@FreeBSD.org>
55147 Date: Mon Jul 10 17:16:18 2006 -0700
55149 Fix modelist with a configured monitor to not begin with all unvalidated modes.
55151 commit 48ba9273ddfb36d3525e19238b94b18c56667c4d
55152 Author: Eric Anholt <anholt@FreeBSD.org>
55153 Date: Mon Jul 10 15:01:51 2006 -0700
55155 Fix validation when the first mode is thrown out, and print hsync in modelines.
55157 commit 3924ffb7d7b71cb9c6ab9eeb12d0f1b2c26244c0
55158 Merge: 5a8f6a4 f76f94a
55159 Author: Eric Anholt <anholt@FreeBSD.org>
55160 Date: Mon Jul 10 12:12:38 2006 -0700
55162 Merge branch 'dvo-merge' into modesetting
55168 commit f76f94a743505da16e121992eb789c1f74eb7673
55169 Author: Eric Anholt <anholt@FreeBSD.org>
55170 Date: Mon Jul 10 12:04:09 2006 -0700
55172 Supply proper NULL-terminated symbol lists to avoid crashing.
55174 commit 5a8f6a486d79f50d2d659e615283289d59f9caa4
55175 Author: Eric Anholt <anholt@FreeBSD.org>
55176 Date: Sat Jul 8 15:26:19 2006 -0700
55178 Improve output bus setup to include LVDS setup for pre-i915.
55180 commit df333cc9a848bc2299a52a7613fe4ffdff8038a2
55181 Author: Eric Anholt <anholt@leguin.anholt.net>
55182 Date: Fri Jul 7 13:41:33 2006 -0700
55184 Initial add of DVO support code. Probes my sil164.
55186 This is a mostly-untested merge of airlied's work. The I2C modules are intended
55187 to be moved into the core server or a separate driver module when they're
55188 functional and we're happy with the API.
55190 commit e1064f52b0ff69ea7937897b8c951cc3e32cd752
55191 Author: Eric Anholt <anholt@FreeBSD.org>
55192 Date: Wed Jul 5 16:00:03 2006 -0700
55194 Don't try to probe modes on an SDVO device with NULL sdvo_drv.
55196 commit ffa6ecc18bc54151061d9956f1d12575fc057da3
55197 Author: Eric Anholt <anholt@FreeBSD.org>
55198 Date: Wed Jul 5 14:41:08 2006 -0700
55200 More fixes to "choose closest mode for the pipe" code to select correct refresh.
55202 commit dfd7fef457c048c9f0d826e37d91453d9e1485b9
55203 Author: Eric Anholt <anholt@FreeBSD.org>
55204 Date: Mon Jul 3 12:01:57 2006 -0700
55206 Bug #7375: Don't double-free the current XF86 mode after a randr reprobe.
55208 commit 5d07ebdf4f23e16fb8f60eafeadc947701e7877c
55209 Author: Eric Anholt <anholt@FreeBSD.org>
55210 Date: Mon Jul 3 11:59:21 2006 -0700
55212 Only override display size with the XFree86 mode's for actual panel scaling.
55214 commit 48f27ac62128251640a9b1ca54f63376676b47eb
55215 Author: Eric Anholt <anholt@FreeBSD.org>
55216 Date: Wed Jun 28 15:07:01 2006 +0200
55218 Replace i830InjectModes with i830DuplicateModes usage.
55220 The remaining functionality of DuplicateModes was OBE, and the name was bad.
55222 commit ce5bd108c55d2378db072617c380514a39672603
55223 Author: Eric Anholt <anholt@FreeBSD.org>
55224 Date: Wed Jun 28 14:21:49 2006 +0200
55226 Validate and insert user and VESA standard modes for DDC or configured fallback.
55228 This isn't really tested because I lack a good CRT to test against currently.
55230 commit 367f69f8e7710e53dcd286f1b62506a3276e80f9
55231 Author: Eric Anholt <anholt@FreeBSD.org>
55232 Date: Wed Jun 28 13:10:02 2006 +0200
55234 Replace xf86ValidateModes usage with a set of custom validators and pruning.
55236 This moves us to maintaining MonPtrs per pipe instead of using the EDID
55237 structure "xf86MonPtr", which is closer to what we want to be looking at when
55238 doing validation. The new validation isn't enough yet -- particularly, we
55239 aren't importing and validating the custom modelines to the pipes when
55240 applicable, but this will be easier than (for example) trying to make flat
55241 panel modes pass xf86ValidateModes through various gross hacks.
55243 Hotplug turn-on/off also happens at SwitchMode time now, instead of at randr
55246 commit 9fbd3d8f4befb75ed6f6bd9a9ffe0175626e8785
55247 Author: Eric Anholt <anholt@FreeBSD.org>
55248 Date: Tue Jun 27 23:31:40 2006 +0200
55250 If the panel power registers are all zeroes on Mobile parts, disable LVDS.
55252 This is the case on the Mac mini, which is an i945GM but has no LVDS attached.
55253 Powering on with the power timing registers zeroed would probably be a bad idea,
55254 even if there was a panel attached.
55256 commit 56f7aedd0ad1f5645a90a5509b1263ec6b7b7ee1
55257 Author: Eric Anholt <anholt@FreeBSD.org>
55258 Date: Tue Jun 27 00:37:56 2006 +0200
55260 Fix randr current mode reporting and mode loss/mis-configuration on re-randr.
55262 commit b85f268051a785f90e4c55bd1cac80d673388f16
55263 Author: Eric Anholt <anholt@FreeBSD.org>
55264 Date: Mon Jun 26 16:26:58 2006 +0200
55266 Fix remaining warning.
55268 commit 6a3f89f4b6035534c7b93a4c05fd704305349785
55269 Merge: 3cdc537 e7723a4
55270 Author: Eric Anholt <anholt@FreeBSD.org>
55271 Date: Mon Jun 26 16:25:58 2006 +0200
55273 Merge branch 'master' into exa
55279 commit 3cdc53797bf6fdbfbcb53e07b269a5071f5ec97d
55280 Author: Eric Anholt <anholt@FreeBSD.org>
55281 Date: Mon Jun 26 16:23:55 2006 +0200
55283 Fix build with new EXA headers.
55285 commit e7723a4e5725147d3bd9ba22c5a3314b0556e440
55286 Merge: 5111b88 dae9cb7
55287 Author: Eric Anholt <anholt@FreeBSD.org>
55288 Date: Mon Jun 26 16:04:33 2006 +0200
55290 Merge branch 'origin'
55292 commit 7104b915da412a26661bab7b5a940e935e8cbc8a
55293 Author: Eric Anholt <anholt@FreeBSD.org>
55294 Date: Mon Jun 26 15:11:13 2006 +0200
55298 commit 21dc3edfc41987bb8cf0f8d265fba9775cbe8cdb
55299 Merge: f113e90 5111b88
55300 Author: Eric Anholt <anholt@FreeBSD.org>
55301 Date: Mon Jun 26 15:06:42 2006 +0200
55303 Merge branch 'master' into modesetting
55309 commit 5111b883480a5a9cc82200f2684cba67b515aa73
55310 Author: Eric Anholt <anholt@FreeBSD.org>
55311 Date: Mon Jun 26 14:53:10 2006 +0200
55313 Turn on extra warning flags for GCC, and clean up the resulting fallout.
55315 commit f113e9002cf53510e30984f816d44b06f1e71216
55316 Author: Eric Anholt <anholt@FreeBSD.org>
55317 Date: Mon Jun 26 12:54:30 2006 +0200
55319 Fix the plane/pipe disabling and turn off missing outputs when no longer DDCed.
55321 commit 0e5cda3796ba0164496f0814eb57d1dfa7ab9257
55322 Author: Eric Anholt <anholt@FreeBSD.org>
55323 Date: Mon Jun 26 10:30:46 2006 +0200
55325 Fix FP scaling by using the desired mode to get at the real [HV]Display.
55327 commit 51d14f803aae6d0a738520c3cad110289ee444db
55328 Author: Eric Anholt <anholt@FreeBSD.org>
55329 Date: Mon Jun 26 10:21:23 2006 +0200
55331 Add a function to turn off unused outputs, DPLLs, planes, and pipes.
55333 commit d2c18d8d79596513149273e4a0d322f04bf27e80
55334 Author: Eric Anholt <anholt@FreeBSD.org>
55335 Date: Mon Jun 26 10:15:11 2006 +0200
55337 Fix DDC probing after last (untested) commit.
55339 commit 6d1d105d68d9c25890c3fd2ebad1367529c2a991
55340 Author: Eric Anholt <anholt@FreeBSD.org>
55341 Date: Mon Jun 26 10:14:27 2006 +0200
55343 Unset pipe current mode on EnterVT, so we reprogram the mode for sure.
55345 commit 52e8231a19f28bd4744f983aee2197a18c20aa3a
55346 Author: Eric Anholt <anholt@FreeBSD.org>
55347 Date: Mon Jun 26 07:46:28 2006 +0200
55349 Major cleanup of mode reprobing:
55350 - Don't mess with pScrn->monitor->Modes, and instead make our own availModes.
55351 - Don't re-program the pipe with the same values (no flicker at xrandr)
55352 - Move a bunch of stuff that should be exposed through the public API (probably)
55353 to i830_xf86Modes.c
55354 - Use a table with established modes plus GTF to come up with modes from EDID,
55355 instead of trying to walk and find one in pScrn->monitor->Modes. I think
55357 - Reset clone state if we've detected new pipes, which should turn on the
55360 commit f5e5f8aeddb3e0d6d073471aeff6176fb54576e2
55361 Author: Eric Anholt <anholt@FreeBSD.org>
55362 Date: Fri Jun 23 23:29:55 2006 -0700
55364 WIP to allow re-probing and validation of modes for new heads at "xrandr" time.
55366 Now, DDC modes always end up being preferred to custom modelines, even if
55367 smaller. This should probably be fixed by inserting custom modelines into
55368 the probed mode list if they're valid according to the probed parameters of the
55371 Too much code is lifted from static functions in xf86Mode.c, and those should be
55372 made unstatic if possible. Using xf86ValidateModes is also rather hacky, and
55373 I want to break the function down, but this is a first step.
55375 commit 89791914d2a78f19f4f60ca370d387e5b1ccfb46
55376 Author: Eric Anholt <anholt@FreeBSD.org>
55377 Date: Fri Jun 23 18:21:17 2006 -0700
55379 Split probed modes out per pipe, and union them into the available modes.
55381 This is the first stage of getting runtime monitor attachment. The old i830
55382 GTF code is returned to use to provide suitable modelines for xf86ValidateModes
55383 in the LVDS case, even though the LVDS doesn't care about the modeline and just
55384 always programs its fixed values.
55386 commit dae9cb7712d5d8f88697ca83808c59af08364c0e
55387 Author: Alan Coopersmith <alan.coopersmith@sun.com>
55388 Date: Thu Jun 22 15:07:16 2006 -0700
55390 Provide definitions of __FUNCTION__ for non-gcc compilers
55392 commit bb4810521633b6c3db2fc7d01ddc71325583d265
55393 Author: Eric Anholt <anholt@FreeBSD.org>
55394 Date: Thu Jun 22 09:38:27 2006 -0700
55396 Move FP mode validation next to other mode validation code.
55398 commit 66d9a1be302ad34573de98de21cbdf6419592092
55399 Author: Eric Anholt <anholt@FreeBSD.org>
55400 Date: Wed Jun 21 17:11:54 2006 -0700
55402 Detect SDVO display presence at startup and default to displaying to it, too.
55404 commit 72e25a7488c2eabcc92e9e0769a89dee687f52fd
55405 Author: Eric Anholt <anholt@FreeBSD.org>
55406 Date: Wed Jun 21 16:04:18 2006 -0700
55408 Fix SDVO output at low pixel clocks.
55410 I had interpreted the docs as saying that the multiplier setting would further
55411 divide the clock and stuff dummy bytes in. Instead, we have to set the DPLL at
55412 the higher clock rate, and the pixel multiplier just controls the stuffing of
55413 dummy bytes. Also, we have to set the multiplier both in the graphics chip and
55414 on the SDVO device on the other side.
55416 commit 726443309d72134341cff1f6db978aa1d6e3ce52
55417 Author: Eric Anholt <anholt@FreeBSD.org>
55418 Date: Wed Jun 21 15:38:19 2006 -0700
55420 Add decoding of SDVO command names for debug output.
55422 commit 5a1b68993f3a3a2e8dcd428a7118e29c36703cd6
55423 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55424 Date: Wed Jun 21 08:41:16 2006 +0100
55426 Fix build without DRI
55428 commit 16b310823bacab6be4947da234b3a081b0a3cd62
55429 Author: Matthieu Herrb <matthieu@deville.herrb.com>
55430 Date: Wed Jun 21 00:12:27 2006 +0200
55432 Fix build without DRI
55434 commit 896ffe78fe96469cdd3ade77c8e68e1503967223
55435 Merge: 89c2c4b 52243d4
55436 Author: Eric Anholt <anholt@FreeBSD.org>
55437 Date: Tue Jun 20 15:10:35 2006 -0700
55439 Merge branch 'modesetting-origin' into modesetting
55441 commit 89c2c4bc40b8c032915ccb3ed4f3c143c3d8db12
55442 Author: Eric Anholt <anholt@FreeBSD.org>
55443 Date: Tue Jun 20 15:10:19 2006 -0700
55445 Add #if 0-ed code I've been using for CRT detection debugging.
55447 commit be08661e3126907c50c54485042fcde00b0da2b4
55448 Author: Eric Anholt <anholt@FreeBSD.org>
55449 Date: Tue Jun 20 14:48:03 2006 -0700
55451 Only default to enabling CRT or LVDS output if they're actually detected.
55453 Still, if we haven't detected any outputs automatically (including CRT through
55454 DDC), default to CRT anyway.
55456 commit b454c9601f005c69c11556a558150403378d34d9
55457 Author: Eric Anholt <anholt@FreeBSD.org>
55458 Date: Tue Jun 20 14:32:40 2006 -0700
55460 Add support for CRT detection using DDC.
55462 This method is slower (~5ms), but works on older chipsets. Also, load-based
55463 detection is disabled, as it can be fooled by other outputs on the pipe being
55464 active, such as LVDS.
55466 commit 0b76646666e9d330e77c6f81af8b91e34623be92
55467 Author: Eric Anholt <anholt@FreeBSD.org>
55468 Date: Tue Jun 20 13:57:26 2006 -0700
55470 Add CRT detection function by testing for load, and clean up hotplug version.
55472 commit e4584a4f44a70d746396ed48b8e40033504d68b2
55473 Author: Eric Anholt <anholt@FreeBSD.org>
55474 Date: Tue Jun 20 10:39:28 2006 -0700
55476 Remove dead DisplayInfo option.
55478 commit ab60e34dcfc52ab5f22a82145d5b4db51b4c62c5
55479 Author: Eric Anholt <anholt@FreeBSD.org>
55480 Date: Tue Jun 20 10:07:47 2006 -0700
55482 Add debugging info for pipe/display plane size.
55484 commit 8a6edba33213911cc2210b5e903428b81d45862f
55485 Author: Keith Packard <keithp@neko.keithp.com>
55486 Date: Mon Jun 19 13:47:28 2006 -0700
55488 Set vblank interrupt configuration to match pipe configuration
55490 New i915 drm ioctl (in version 1.5) allows the X server to select
55491 which pipe drives vblank interrupts. Use this to drive from the 'preferred'
55492 pipe. Yes, per-window vblanks would be nice in a shared fb environment.
55494 (cherry picked from 2fb375b665f4802819b89f2277fd6154006c11ee commit)
55496 commit 52243d407cad93283956660de4771097ac0b4b2d
55497 Merge: 2fb375b 34f6a82
55498 Author: Keith Packard <keithp@neko.keithp.com>
55499 Date: Mon Jun 19 13:28:09 2006 -0700
55501 Merge branch 'modesetting-origin' into modesetting
55503 commit 2fb375b665f4802819b89f2277fd6154006c11ee
55504 Author: Keith Packard <keithp@neko.keithp.com>
55505 Date: Mon Jun 19 13:24:57 2006 -0700
55507 Set vblank interrupt configuration to match pipe configuration
55509 New i915 drm ioctl (in version 1.5) allows the X server to select
55510 which pipe drives vblank interrupts. Use this to drive from the 'preferred'
55511 pipe. Yes, per-window vblanks would be nice in a shared fb environment.
55514 commit b5acc6b3a3a3a109014d6b971f4722d0f0a4c29a
55515 Author: Keith Packard <keithp@neko.keithp.com>
55516 Date: Mon Jun 19 13:22:17 2006 -0700
55518 Add backlight control to DPMS logic.
55520 Turn backlight on and off in response to DPMS state changes.
55522 commit d7e5d996aa123e52857362fa8a1e69f709bfe218
55523 Merge: 89a5a00 a73ab7f
55524 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55525 Date: Tue Jun 20 01:22:53 2006 +0800
55527 Merge branch 'master' into exa
55529 commit a73ab7f0e6e3b0462e05c0031ffd602ed3e2bcd4
55530 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55531 Date: Mon Jun 19 11:35:42 2006 +0100
55533 additions for rotation fixes
55535 commit 1fe3dd38eb613475d62140850e64767defed7d34
55536 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55537 Date: Mon Jun 19 11:27:28 2006 +0100
55539 Set some invarient state, cures some problems with
55540 rotation at startup.
55542 This mimicks the 3D drivers setup.
55544 commit 89a5a0024a7e077a0726ffe1427e573df3bcbef5
55545 Author: Eric Anholt <anholt@FreeBSD.org>
55546 Date: Sun Jun 18 00:25:34 2006 -0700
55548 Fix crash with EXA during randr, when XAA structures were being accessed. EXA
55549 shouldn't have any issues with there also being pixmaps in framebuffer while
55550 rotation is active.
55552 commit f2ab3aa3199e1c4b0edec72b5acea6d725cc514a
55553 Author: Eric Anholt <anholt@FreeBSD.org>
55554 Date: Sun Jun 18 00:32:01 2006 -0700
55556 Fix DRI in EXA mode: Don't attempt to use XAA symbols (use the driver's internal
55557 ones for EXA support when available). Also, add an abstraction of XAA/EXA
55558 MarkSync and WaitSync functionality so we don't need to sprinkle the ifdefs all
55559 over, and correctly use them.
55561 commit 3592b432b48d51d2273c1e1064f85e656fbba130
55562 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55563 Date: Tue Jun 13 21:42:53 2006 +0100
55565 fix 8bpp & 16bpp rotation modes for i8xx
55568 commit a50610b7719bfe800c3496c17d0ba77739167b35
55569 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55570 Date: Mon Jun 12 13:53:20 2006 +0100
55572 Use 800x600 mode to double check
55574 commit f02268b2091c9a785d26e82bcb35a8b713463072
55575 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55576 Date: Mon Jun 12 12:16:58 2006 +0100
55578 Don't rely on register check to find out
55579 if we're resuming - it's not reliable.
55581 But then, neither is the BIOS, but it's
55582 the best we can hope for until Eric's work
55585 Try setting another mode to cater for some
55588 commit 672c3d18dbb405095e465126053ff887d891409e
55589 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
55590 Date: Mon Jun 12 10:02:06 2006 +0100
55592 Only mark rotation flags after initial screen setup.
55595 commit 34f6a8204f1edec015283fc6b5f196e47897e3de
55596 Author: Keith Packard <keithp@mactel.(none)>
55597 Date: Sun Jun 4 00:15:06 2006 -0700
55599 Get sDVO output working on mac mini.
55600 Add lots of register debugging to track delta from BIOS settings.
55601 Fix various mode settings to mirror BIOS sDVO values.
55602 Disable analog/lvds output on pipe with sDVO.
55603 Borrow Dave Airlie's I830xf86ValidateDDCModes code.
55604 Fix various sDVO I2C messages to mirror Dave's code.
55606 commit 6812b5382077e5d3f421aceeeb2f337e9b3f570e
55607 Author: Dave Airlie <airlied@linux.ie>
55608 Date: Fri Jun 2 12:22:14 2006 +1000
55610 intel: fix VT switch DRI locking
55612 The DRI locking is incorrect at VT switch, due to reference counting
55613 inside the driver. Just call the DRI directly.
55615 commit 3f158fd610a3363a23daa7205bcd9f213686cf1c
55616 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55617 Date: Mon May 29 18:05:57 2006 -0700
55619 Nice texture coordinate gradient, broken slightly in y
55621 commit bb0ad04d46eba2fed57a888ff960d2436ec7d70d
55622 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55623 Date: Sun May 28 22:59:58 2006 -0700
55625 Ok, finally something sensible up on the screen.
55626 Replace PS kernel with constant data source (pink).
55627 Dodge g0/g1 so URB data doesn't land on top of thread data.
55628 Flip source/dest coordinates (dunno why they're fetched this way).
55630 commit ddf3e5b2737399dca6d401f91db51a51f93b6373
55631 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55632 Date: Sun May 28 21:03:39 2006 -0700
55634 Using tiny rectangle, still locks up in pixel shader program somehow
55636 commit 2e16c79dc2f24b0a04111aa6236a44870c6c64df
55637 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55638 Date: Sun May 28 20:31:23 2006 -0700
55640 Lots more debug code. Appears to execute pixel shader thread now though. hurray!
55642 commit 79a514412bda7e38e018c105a603970c4a9d758a
55643 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55644 Date: Sun May 28 16:09:59 2006 -0700
55646 dump out piles of debug. Create VS thread just to see how it works
55648 commit 9c111d89fe19f1773af2eefb000e1c2389b4b6e1
55649 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55650 Date: Sat May 27 19:52:13 2006 -0700
55652 Push all of the obvious Mesa state setting into the video code
55654 commit 01101196b16010ac3dadab647bfe7000a53fa94d
55655 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55656 Date: Sat May 27 01:05:09 2006 -0700
55658 flesh out cc state. set cull mode to none. enable sf kernel
55660 commit 462a860af89ed855fe2b718342fcaf9c169af3fb
55661 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55662 Date: Sat May 27 00:17:25 2006 -0700
55664 Rename BRW instructions, check video instruction generation. Doesnt lock up, but doesnt display anything either
55666 commit 9ec7cf22e3f03c13524bb2d15711699dfcc02984
55667 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55668 Date: Fri May 26 21:30:55 2006 -0700
55670 Use broadwater video code on broadwater hardware. Pad ring to even length. compute state base as address rather than offset
55672 commit f5fe700b9a943c956bcfcc3a0d2de13c23b978bc
55673 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55674 Date: Fri May 26 13:47:39 2006 -0700
55676 Prepare real SF kernel and fake WM kernel
55678 commit 1549accb6f52498fef3dcbd87bb72d89fcd5bccd
55679 Author: Keith his master's voice Packard <keithp@bw.jf.intel.com>
55680 Date: Thu May 25 16:10:31 2006 -0700
55682 Scale video source vertices. Allocate space for kernels
55684 commit c1c46f882f9a11c383c8d1d1ce393be8fda55ed0
55685 Merge: 1e2da24 f2967a2
55686 Author: Eric Anholt <anholt@FreeBSD.org>
55687 Date: Mon May 22 10:42:24 2006 -0700
55689 Merge branch 'master' into modesetting
55695 commit 1e2da2450d7213c0aa4d4bd1fba5723dcda13ddf
55696 Author: Lukáš Hejtmanek <xhejtman@mail.muni.cz>
55697 Date: Mon May 22 09:48:09 2006 -0700
55699 Replace VBE call to do DPMS with native code, and fix screensaver in clone mode.
55701 commit 2a1b3cfccb7de53f7ce8f9e4816e4278afb1fcab
55702 Author: Eric Anholt <anholt@FreeBSD.org>
55703 Date: Mon May 22 10:32:13 2006 -0700
55705 Use RECTLIST instead of TRIFAN for video so we get horizontal shearing instead
55706 of diagonal. Also remove the unnecessary vertex elements that were being
55709 commit bc42dbe07cbd61b357d8eed02608e026c4cf4485
55710 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55711 Date: Mon May 22 11:22:03 2006 +0800
55715 commit bce209cd3f60cb5d51aadc5fc8ec1a4151435ec3
55716 Author: Eric Anholt <anholt@FreeBSD.org>
55717 Date: Fri May 19 17:13:37 2006 -0700
55719 Put in code for idling accelerator on subsequent cliprects.
55721 commit 3640117bd9f2073ff54dc474f0cdefff49742584
55722 Author: Eric Anholt <anholt@FreeBSD.org>
55723 Date: Fri May 19 17:10:04 2006 -0700
55725 Set up the state buffer in framebuffer.
55727 commit e70d99d51edc39ea53536c2f4d62ac622ba3e0d9
55728 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55729 Date: Fri May 19 15:47:19 2006 +0800
55731 revert wrong setting for memorySize, which should
55732 be at the range of pI830->Offscreen.
55734 commit de06cd70a9edb8b56d05d3f505137f7c7f083c2f
55735 Author: Eric Anholt <anholt@FreeBSD.org>
55736 Date: Thu May 18 18:27:11 2006 -0700
55738 Checkpoint for filling out more 3D state.
55740 commit ad7ec6a24b436d5492d38e4fa56845b229cf5fb8
55741 Author: Eric Anholt <anholt@FreeBSD.org>
55742 Date: Thu May 18 15:26:28 2006 -0700
55744 Checkpoint of BW textured video work, filling out vertex submission stuff and
55745 some more other state.
55747 commit 291770efc691a02650e3c580ca40c2f9fce3896c
55748 Author: Eric Anholt <anholt@FreeBSD.org>
55749 Date: Thu May 18 10:43:07 2006 -0700
55751 Start laying out some of the bits that need to be done for BW textured video.
55752 Headers taken from TG code drop.
55754 commit b3d79ad9ecc70e9f841092fc75b7e23576b685dd
55755 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55756 Date: Thu May 18 16:31:40 2006 +0800
55758 oop I830FALLBACK define error
55760 commit 1bc2a8f7a20b2ea9251e93157c0f2342fdbb951c
55761 Author: Dave Airlie <airlied@linux.ie>
55762 Date: Thu May 18 16:58:22 2006 +1000
55764 fixup I830SDVOGetActiveInputs to use return values not args
55766 commit 33ad959323f70f76c494c66ec35a04d7d839f612
55767 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55768 Date: Thu May 18 11:32:59 2006 +0800
55770 Add DEBUG_I830FALLBACK for easier tracking
55772 commit 52af679c94a25c0c6e5df5f3647f62fdd256c826
55773 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55774 Date: Thu May 18 11:31:14 2006 +0800
55776 Check alignment require in PrepareSolid, noticed by Eric,
55777 as pixmap offset align set is violated in Solid.
55779 commit fdb52a8c6da0f3d3f4cfbf1583b2dc4d91f7e933
55780 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55781 Date: Thu May 18 11:24:46 2006 +0800
55783 Solid planemask in PrepareSolid/Copy might be 0xffffffff
55784 or bits up to the depth. This should save some fallback
55785 case, noticed by Eric.
55787 commit bc51d6525a12c748d0a293b7e560f6dcea33eecb
55788 Author: Eric Anholt <anholt@FreeBSD.org>
55789 Date: Wed May 17 13:42:51 2006 -0700
55791 Turn off overlay video on BW until we have stable PCI IDs so we can know whether
55792 the hardware supports overlay.
55794 commit f97895efd5532cca145b6f224f9615739b1e8f26
55795 Author: Dave Airlie <airlied@linux.ie>
55796 Date: Wed May 17 14:46:37 2006 +1000
55798 fixup chipid override
55800 This makes the ChipID override work so that we actually override the pci id
55801 that gets used everywhere in the driver.
55803 commit 011a1c99d49c1c69b5fa81ade6a2ed73e84276ae
55804 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
55805 Date: Wed May 17 11:16:00 2006 +0800
55807 Fix memorySize to enable offscreen mem manage.
55809 commit b0ac5303f33f75dc607cf0c705c23db1da836983
55810 Merge: 4c72725 c2cd10e
55811 Author: Eric Anholt <anholt@FreeBSD.org>
55812 Date: Tue May 16 15:21:17 2006 -0700
55814 Merge branch 'textured-video' into broadwater-video
55820 commit c2cd10e1fba0e75c0ed3db5d17211bddf7ab1e33
55821 Author: Eric Anholt <anholt@FreeBSD.org>
55822 Date: Tue May 16 10:08:58 2006 -0700
55824 Flag the 3D state as dirty when we draw textured video, which should help
55826 rotation (I have other issues with rotation anyway).
55828 commit 63a72e46fa20a4a4ba74efed386f6c3c167be5b5
55829 Author: Eric Anholt <anholt@FreeBSD.org>
55830 Date: Tue May 16 08:54:43 2006 -0700
55832 Turn debugging back off.
55834 commit 01c043de0393170e98515169f8239fef4d3e2053
55835 Author: Eric Anholt <anholt@FreeBSD.org>
55836 Date: Tue May 16 08:53:40 2006 -0700
55838 Use linear min/mag blending.
55840 commit db3683907d15959e79adfb8f0cd94e861fae5c36
55841 Author: Eric Anholt <anholt@FreeBSD.org>
55842 Date: Tue May 16 08:40:53 2006 -0700
55844 For textured video, disable double buffering and sync before uploading new video
55846 data. Allows more videos to play simultaneously.
55848 commit 29a8e88ed01c9e15a2ceba5eb62b19773e14c1f8
55849 Author: Eric Anholt <anholt@FreeBSD.org>
55850 Date: Mon May 15 17:19:33 2006 -0700
55852 Relax the alignment requirements for textured video.
55854 commit c9be11459bc2198b435c97c5a3432425246c4d2d
55855 Author: Eric Anholt <anholt@FreeBSD.org>
55856 Date: Mon May 15 17:04:27 2006 -0700
55858 Enable overlay and/or textured video at runtime according to hardware
55860 capabilities. Sets up 16 textured-video ports. Left in one hack (disconnected
55861 but advertised BRIGHTNESS and CONTRAST atoms) which may actually not be
55864 commit f268979a0c779641c84e8d5b763acbda131474cf
55865 Author: Eric Anholt <anholt@FreeBSD.org>
55866 Date: Mon May 15 10:05:19 2006 -0700
55868 Correct drawing issues with planar formats when top or left != 0, and Y didn't
55872 commit b09fd42d7088ead6c23e040ac4b71114f62de82b
55873 Author: Eric Anholt <anholt@FreeBSD.org>
55874 Date: Mon May 15 09:29:43 2006 -0700
55876 Fix the planar formats to display correctly in textured mode. Still has issues
55878 with clipping, and some sampling differences between ximagesink and xvimagesink.
55880 commit dd48790f4600a880fc4907c6e3b1cd51e9c0f0b7
55881 Author: Eric Anholt <anholt@FreeBSD.org>
55882 Date: Fri May 12 13:27:33 2006 -0700
55884 Divide width by 2 in planar-to-packed conversion loop, since each pass through
55886 the loop writes two source pixels.
55888 commit eec5e996ec9361099bf81d8d3b66933d5981c5a8
55889 Author: Eric Anholt <anholt@FreeBSD.org>
55890 Date: Thu May 11 20:26:26 2006 -0700
55892 Merge textured-video-wip to textured-video-planar-full.
55894 commit 4154a2f74811b91c0ef5bef32a919d6f8baf1a70
55895 Author: Eric Anholt <anholt@FreeBSD.org>
55896 Date: Wed May 10 12:09:00 2006 -0700
55898 Experimental work to use a full pixel shader for planar to YUV conversion, which
55900 also doesn't quite work.
55902 commit 3e0a9c9082942eb6f52612235d84b8408e1e03e9
55903 Author: Eric Anholt <anholt@FreeBSD.org>
55904 Date: Tue May 9 16:57:19 2006 -0700
55906 Do a separate BEGIN/ADVANCE_LP_RING set in the planar vs packed blocks, so I
55908 can adjust the planar code more easily.
55910 commit 3af4a967e73b367bb531f2760b4803db1388bcf9
55911 Author: Eric Anholt <anholt@FreeBSD.org>
55912 Date: Tue May 9 16:50:48 2006 -0700
55914 Add a couple of macros to simplify writing of video pixel shaders.
55916 commit 06e62ec521ed3f7ed232ace8e188891bedb53097
55917 Author: Eric Anholt <anholt@FreeBSD.org>
55918 Date: Tue May 9 12:04:58 2006 -0700
55920 Commit a WIP implementation of the planar video shader that does the
55922 planar-to-packed conversion for us. Unfortunately the documentation is unclear,
55923 and I haven't managed to get any implementation of it working correctly.
55925 commit b1090a42b200710628dd8b0c7ced15db7bbe71a1
55926 Author: Eric Anholt <anholt@FreeBSD.org>
55927 Date: Mon May 8 14:22:00 2006 -0700
55929 More magic number reduction in rotation code.
55931 commit 3a2d8af214a79591322ce6e5546f856a1ee41736
55932 Author: Eric Anholt <anholt@FreeBSD.org>
55933 Date: Mon May 8 09:42:40 2006 -0700
55935 Add initial textured XV support for i915, which can do YUY2 and UYVY, but fails
55937 on I420 and YV12 currently, doesn't support the composite extension, and should
55938 break XV support on non-i915.
55940 commit 66875c1559bc20b531ab72e5d6b921d9f50b29f3
55941 Author: Eric Anholt <anholt@FreeBSD.org>
55942 Date: Fri May 5 10:15:23 2006 -0700
55944 Convert magic numbers in i915 rotation 3D state to symbolic names.
55946 commit 4c727254da354cfd6f35148a334d046d67a50e99
55947 Author: Eric Anholt <anholt@FreeBSD.org>
55948 Date: Fri May 12 16:26:58 2006 -0700
55950 Remove the local, renamed copy of lnx_agp.c. The diff between lnx_agp.c and it
55951 appeared to be removal of bugfixes.
55953 commit a115c4b872a385530dcf94b7e7f2fa9b3b7e3155
55954 Author: Eric Anholt <anholt@FreeBSD.org>
55955 Date: Fri May 12 16:11:29 2006 -0700
55957 Remove the code that changes our behavior based on whether a magic file exists
55958 with magic contents in /tmp (created by some install script).
55960 commit 0cb251fadca1cbb3d4c5b97982cd0d8c2fc3e840
55961 Author: Eric Anholt <anholt@FreeBSD.org>
55962 Date: Fri May 12 16:10:02 2006 -0700
55964 Do a couple of reverts to get the DRI code building. At this point, the whole
55967 commit 78b95386b630039864b31954ebcd02ec8829b0c8
55968 Author: Eric Anholt <anholt@FreeBSD.org>
55969 Date: Fri May 12 16:09:40 2006 -0700
55971 Remove intel_randr.c and stick with the previous code. Broadwater shouldn't be
55972 changing our randr, and the new version didn't compile.
55974 commit 2e58aa401dfbab438752038a9034df571c8f8bde
55975 Author: Eric Anholt <anholt@FreeBSD.org>
55976 Date: Fri May 12 15:54:37 2006 -0700
55978 Make the intel_acpi.c code non-modular and make it compile. I think we'll end
55979 up nuking this code anyway, as keithp (and I, as well) disagree about how ACPI
55980 should be handled, but the goal is to compile at the moment.
55982 commit 2cd6c8fa2321ca217ef89db1027dbe9e716ad7aa
55983 Author: Eric Anholt <anholt@FreeBSD.org>
55984 Date: Fri May 12 15:37:44 2006 -0700
55986 Revert internal shadow module changes back like master, along with RandR
55987 initialization. RandR initialization retains the no-rotation setting for BW.
55989 commit 88558ebeed12d6cefd73bba0ddac3c043861ac89
55990 Author: Eric Anholt <anholt@FreeBSD.org>
55991 Date: Fri May 12 15:00:17 2006 -0700
55993 Start fixing up the build and remove a regression from master (I think) in
55996 commit 9e387ef92be9b38c68bda8a6a28b0d9eb98d53a4
55997 Author: Eric Anholt <anholt@FreeBSD.org>
55998 Date: Fri May 12 13:41:38 2006 -0700
56000 Re-add authorship note in i830_driver.c accidentally left out of last commit.
56002 commit af2432322ba1d561057c34ab185561a8e799e8cd
56003 Author: Eric Anholt <anholt@FreeBSD.org>
56004 Date: Fri May 12 13:32:38 2006 -0700
56006 First pass of integrating the Tungsten Graphics driver for Broadwater. This
56007 patch is based off of diffing from the branchpoint to the supplied code, but
56008 with many chunks containing reversions of commits removed. Won't work yet.
56010 commit fbba4312e7cb326fdcf6a71194c4fc4a7a9b5488
56011 Author: Dave Airlie <airlied@linux.ie>
56012 Date: Fri May 12 18:44:27 2006 +1000
56014 add sdvo capability reading support
56016 This reads the SDVO cap bits and uses them to figure out the input/output
56019 commit f2967a2f5f47b636b2445fa69dbc3ec79e065c90
56020 Author: Keith Packard <keithp@neko.keithp.com>
56021 Date: Tue May 9 13:51:25 2006 -0700
56023 Video overlay gamma bounds checking must be done bytewise.
56024 Also, pend bound computations to register writes to allow
56025 updates to individual values that are 'out of spec' so
56026 the client can update multiple values.
56027 (cherry picked from 190f9ad0606e96e684e0b028d576d822dc9aa3cf commit)
56029 commit 190f9ad0606e96e684e0b028d576d822dc9aa3cf
56030 Author: Keith Packard <keithp@neko.keithp.com>
56031 Date: Tue May 9 13:51:25 2006 -0700
56033 Video overlay gamma bounds checking must be done bytewise.
56034 Also, pend bound computations to register writes to allow
56035 updates to individual values that are 'out of spec' so
56036 the client can update multiple values.
56038 commit fb10966e9f43f44488097b0daf2fbe10c6a804dc
56039 Author: Eric Anholt <anholt@FreeBSD.org>
56040 Date: Thu May 4 18:52:08 2006 -0700
56042 Use xf86int10Addr() when calculating an address to read the video BIOS from, out
56044 of int10's copy. Fixes a crash on FreeBSD.
56046 commit a66f2c01f7d557d860883346671fb864807dbdca
56047 Author: Eric Anholt <anholt@FreeBSD.org>
56048 Date: Mon May 1 12:35:21 2006 -0700
56050 Stop doing the BIOS memory size tweaking now that we don't ask the BIOS about
56051 what modes are available.
56053 commit 99b0e53244aa49f4e7461a1c96aba69453b3516f
56054 Author: Matthieu Herrb <matthieu@reactor.herrb.com>
56055 Date: Mon May 1 10:47:09 2006 +0200
56059 commit 86f0119f47d8be975016ffdea850f0227e32eebc
56060 Author: Matthieu Herrb <matthieu@reactor.herrb.com>
56061 Date: Mon May 1 10:41:10 2006 +0200
56065 commit fcec956807038a2e90287850df8b4410b1f8b3c2
56066 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
56067 Date: Sat Apr 29 10:55:25 2006 +0800
56069 Save initialize EXADriverPtr, and free it in failure.
56071 commit 0be9bc604420e4433697cc31e8ac58c9f36185e1
56072 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
56073 Date: Sat Apr 29 10:54:02 2006 +0800
56075 I830CheckDevicesTimer() sync should be in I830_USE_XAA case.
56077 commit a00a3d198aee7c173c2c36858f4447e13c86eb7f
56078 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
56079 Date: Sat Apr 29 10:52:45 2006 +0800
56081 fix I830BIOSAdjustFrame() sync in I830_USE_EXA case, use exaWaitSync() helper.
56083 commit c94cc67bcc507e8256ae4b51b84dfe3f4aa12ab5
56084 Author: Eric Anholt <anholt@FreeBSD.org>
56085 Date: Fri Apr 28 15:33:57 2006 -0700
56087 Update to EXA 2.0 API.
56089 commit a4a4ced50cc235e79013ce8dd78e35c6eb4c98ca
56090 Author: Eric Anholt <anholt@FreeBSD.org>
56091 Date: Fri Apr 28 14:59:52 2006 -0700
56095 commit 941a49f91eac02e256706d8821a2cb069ff7ffef
56096 Author: Eric Anholt <anholt@FreeBSD.org>
56097 Date: Fri Apr 28 14:57:03 2006 -0700
56099 Commit what applied from Jesse Barnes's i830-exa-latest.patch.
56101 commit a555e28e5afc81969ef7b28482e654cc26b3a446
56102 Author: Eric Anholt <anholt@FreeBSD.org>
56103 Date: Mon Apr 24 16:55:44 2006 -0700
56105 Correct some SDVO-related register definitions.
56107 commit cc70e6b789a1901bbe4e3501b6b654542d3cdc20
56108 Author: Eric Anholt <anholt@FreeBSD.org>
56109 Date: Mon Apr 24 16:49:48 2006 -0700
56111 Save/restore the output's SDVO reg. Note that we might be programming the other
56112 SDVO reg instead of the one assigned to the output. When trying to fix that, I
56113 ended up getting no output at all.
56115 commit b498d2b1d1170123595ada65353428578b59a361
56116 Author: Eric Anholt <anholt@FreeBSD.org>
56117 Date: Mon Apr 24 15:42:46 2006 -0700
56119 Start trying to save/restore SDVO state on VT switches.
56121 commit 56c1f8b0de9b6acc50f51561caf14d3e6bac09ec
56122 Author: Eric Anholt <anholt@FreeBSD.org>
56123 Date: Mon Apr 24 14:10:20 2006 -0700
56125 Clean up some argument passing, and remove extra SetTarget{In,Out}Puts that had
56126 no effect. Note that we are currently trying to program both outputs of any
56127 SDVO device the same way.
56129 commit 9ba5319b36e7286e33cf0dcdd804dfc0458a81f5
56130 Author: Eric Anholt <anholt@FreeBSD.org>
56131 Date: Mon Apr 24 13:55:05 2006 -0700
56133 Only write out as many arguments as the commands need, and fix up the numbers in
56134 some cases. Pretty-print the return status.
56136 commit d32514aee4b00b035652830e8b5e6c0b43cf159c
56137 Author: Eric Anholt <anholt@FreeBSD.org>
56138 Date: Mon Apr 24 12:21:45 2006 -0700
56140 Start trying to implement DDC over SDVO. It's slightly tricky because the
56141 control bus will reset from DDC mode to internal-registers mode after every
56142 Stop afer a Start on the DDC bus. The xf86 DDC code causes multiple Start/Stops
56143 in one probe. So, we create a wrapper bus that does the control bus switch at
56144 every Start. It's not working yet on my hardware, but I'm pretty sure this is
56145 the right way to go.
56147 commit effab21c3d108fac7a4e28ae4dabb0b5f74a5380
56148 Author: Eric Anholt <anholt@FreeBSD.org>
56149 Date: Mon Apr 24 11:42:24 2006 -0700
56151 Set displayWidth to a sufficient value for the modes we come up with for LVDS.
56153 Reported by: Lukáš Hejtmánek
56155 commit bcb441225d1365435bc3373901180de944298e86
56156 Author: Eric Anholt <anholt@FreeBSD.org>
56157 Date: Mon Apr 24 10:54:45 2006 -0700
56159 Simplify the i2c code by using the GetBits/PutBits interface rather than
56162 commit 47bd9059431eadfd8824e496eb91bb50efa0e282
56163 Merge: 729c373 32a0ad5
56164 Author: Eric Anholt <anholt@FreeBSD.org>
56165 Date: Mon Apr 24 10:36:24 2006 -0700
56167 Merge branch 'lukas-resume'
56169 commit 1dbb19059f5dd5cd2ad386c4f73498095987b943
56170 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56171 Date: Thu Apr 20 10:43:19 2006 +0100
56175 commit 729c373121ce2bbb0d813cc923f1254e8b37a025
56176 Author: Dave Airlie <airlied@linux.ie>
56177 Date: Thu Apr 20 13:34:55 2006 +1000
56179 move sdvo output setting
56181 we have to set the sdvo register a lot earlier in order for them to sync
56182 properly otherwise my monitor doesn't sync unfortunately, also
56183 disable the sdvo while tweaking the PLLs.
56185 This also comments out a setting that seems to break my system here for
56186 Eric to look at later.
56188 commit 0ba7b13fb4410c6a48b2fb098d2033e040eca6d2
56189 Author: Dave Airlie <airlied@linux.ie>
56190 Date: Thu Apr 20 12:29:06 2006 +1000
56192 fix type 0 instead of O
56194 commit 2991d81a3b643161babab6b8f44c057aaaf351c3
56195 Author: Dave Airlie <airlied@linux.ie>
56196 Date: Thu Apr 20 12:27:47 2006 +1000
56198 correct height parameter in sdvo packet
56200 commit b5f099e03a9f58b6b99933fb06526bce4db72bd3
56201 Author: Dave Airlie <airlied@linux.ie>
56202 Date: Thu Apr 20 09:50:36 2006 +1000
56204 cleanup sDVO for device on C only
56206 This destroys the i2c device properly if the device isn't detected,
56207 and allows sDVO to work on GM chipsets, and doesn't initialise
56208 the i2c bus twice for sDVO.
56210 commit 2909802de63756972b38651a496b4ff1b36ac8a2
56211 Author: Eric Anholt <anholt@FreeBSD.org>
56212 Date: Wed Apr 19 16:29:06 2006 -0700
56214 Clean up SDVO initialization, include config.h, and make it check the right
56215 slave address on the first device as well. This gets me to the point of
56216 bringing up some modes on my device.
56218 commit 32a0ad570d9c010e7d26d980830f719782d9f2f3
56219 Author: Lukáš Hejtmanek <xhejtman@mail.muni.cz>
56220 Date: Wed Apr 19 19:43:45 2006 -0300
56222 Add more registers to save/restore. Save/restore palette as well
56224 commit 132dc0599cf44389c4cc03919f1da8d3a0762b44
56225 Author: Eric Anholt <anholt@FreeBSD.org>
56226 Date: Wed Apr 19 15:04:17 2006 -0700
56228 Whine if SDVO I2C device init fails, rather than be silent.
56230 commit 88bb4b578857588f34ac84b7a20577139eccab6d
56231 Author: Eric Anholt <anholt@FreeBSD.org>
56232 Date: Wed Apr 19 14:23:45 2006 -0700
56234 Add more SDVO code. It's taken from airlied's driver, but with magic numbers
56235 replaced by symbolic names in many places. I tried to restrain myself from
56236 functional changes in airlied's code in this pass.
56238 commit d8f7dfac769d7b03f069306b1296bb2e1e08b009
56239 Author: Eric Anholt <anholt@FreeBSD.org>
56240 Date: Wed Apr 19 10:45:13 2006 -0700
56242 Start bringing in some SDVO code, mostly from airlied.
56244 commit a371a04a57620b7128e3c4395bc7c2ac55effe19
56245 Author: Eric Anholt <anholt@FreeBSD.org>
56246 Date: Mon Apr 17 14:10:50 2006 -0700
56248 Use the IS_I9XX macro intead of >= i915G, since by PCI ID number, 855GM > 915.
56250 commit 33413a3cf34b06e3207fe1cdb733d586d55a4337
56251 Author: Eric Anholt <anholt@FreeBSD.org>
56252 Date: Mon Apr 17 13:42:57 2006 -0700
56254 Remove some dead code and one particularly useless debug printf.
56256 commit a85be820033af8bf30d1fec4a899349af35a6ef6
56257 Author: Ian Romanick <idr@localhost.localdomain>
56258 Date: Mon Apr 17 12:52:30 2006 -0700
56260 Add missing new-line character in log message.
56262 commit 365b4a53ee965002a5452e6f6016b528e51cee69
56263 Author: Eric Anholt <anholt@FreeBSD.org>
56264 Date: Mon Apr 17 01:12:28 2006 -0500
56266 Disable the BIOS dump-to-file since it's a trivial local DOS, and I can just ask
56267 people to turn it on and recompile when I need to.
56269 commit ee7be006b63b6b1ce7f786b045fb8f26d337433c
56270 Author: Eric Anholt <anholt@FreeBSD.org>
56271 Date: Sun Apr 16 13:13:42 2006 -0500
56273 Bug #6589: Use alternate offsets to successfully get at the panel data for
56274 some broken video BIOSes.
56276 commit 786ee3df1726f08953167d05f7fa1930452703bb
56277 Author: Eric Anholt <anholt@FreeBSD.org>
56278 Date: Sun Apr 16 13:05:35 2006 -0500
56280 Add a standalone program for parsing VBT.
56282 commit 62652127cd12f5a0fc9364285b81d2661372148a
56283 Author: Eric Anholt <anholt@FreeBSD.org>
56284 Date: Sat Apr 15 12:16:46 2006 -0700
56286 Check for LVDS BIOS tables exactly once at startup. While there, dump the BIOS
56287 data to /tmp/xf86-video-intel-VBIOS, for offline debugging.
56289 commit 830fa81792a613fe2127a2b89a3eaa326f56114c
56290 Author: Alan Hourihane <alanh@jetpack.demon.co.uk>
56291 Date: Wed Apr 12 21:55:31 2006 +0100
56295 commit 3c9bc19db63a317db9e2d67a5bc15b1de1a1060a
56296 Author: Alan Hourihane <alanh@jetpack.demon.co.uk>
56297 Date: Wed Apr 12 21:54:50 2006 +0100
56301 commit d6edffee7d987ef551e1a94d9fac21beb72a9598
56302 Author: Eric Anholt <anholt@FreeBSD.org>
56303 Date: Wed Apr 12 12:16:51 2006 -0700
56305 Initial support for pre-i915 PLL programming. Untested.
56307 commit 59f88955f57cf0f98458b57418dae25cf53ca180
56308 Author: Eric Anholt <anholt@FreeBSD.org>
56309 Date: Wed Apr 12 11:11:14 2006 -0700
56311 Automatically enable clone mode if we detect two active outputs.
56313 commit a797a6626a8117cd16456a206ed96300e4a0ceeb
56314 Author: Wang Zhenyu <zhenyu.z.wang@intel.com>
56315 Date: Tue Apr 11 16:44:42 2006 -0700
56317 Fix a broken test that resulted in page flipping always being enabled, not just
56318 when the user enabled it.
56320 commit 66d216da3bfe7c143d681b610130e6daf6dfa52f
56321 Author: Eric Anholt <anholt@leguin.anholt.net>
56322 Date: Tue Apr 11 16:37:10 2006 -0700
56324 Replace .cvsignore with .gitignore.
56326 commit 185b5251419724fa9377421d67981daa674908c8
56327 Author: Eric Anholt <anholt@leguin.anholt.net>
56328 Date: Mon Apr 10 20:25:13 2006 -0700
56330 Add missing MIT copyright licenses, which should have been on all versions.
56332 commit c0a2dc608d95f92b0f5a151c623745f09df3afc5
56333 Author: Keith Packard <keithp@neko.keithp.com>
56334 Date: Mon Apr 10 20:09:37 2006 -0700
56336 Save/restore PFIT_CONTROL. Eliminate compiler warnings
56338 commit 498abf474b90b04bc4085397a209cc493d98d15c
56339 Merge: c26f073 3de82ff
56340 Author: Keith Packard <keithp@neko.keithp.com>
56341 Date: Mon Apr 10 19:41:50 2006 -0700
56343 Merge branch 'modesetting-origin' into modesetting
56345 commit c26f0737e2c0f000f67f21ef6007b351160036a4
56346 Merge: 11ad8a5 692a4a4
56347 Author: Keith Packard <keithp@neko.keithp.com>
56348 Date: Mon Apr 10 19:41:41 2006 -0700
56350 Merge in master to modesetting
56352 commit 3de82ff3938c6559c90079be0c28dc507d62f79e
56353 Author: Eric Anholt <anholt@FreeBSD.org>
56354 Date: Mon Apr 10 15:10:15 2006 -0700
56356 Improve LVDS modes when outputting a CRT+LVDS combo at larger than the LVDS's
56357 panel size. This is a hack until we get better clone mode, but it correctly
56358 displays a subset of the root on the LVDS by using a correct pixel clock and
56361 commit 692a4a4bafd4cab197b761cda22368581b78f996
56362 Author: Adam Jackson <ajax@freedesktop.org>
56363 Date: Mon Apr 10 11:27:25 2006 -0700
56367 commit 43c93057c044f33c20b1dd639eaeabefc0901589
56368 Author: Aaron Plattner <aplattner@nvidia.com>
56369 Date: Mon Apr 10 11:26:33 2006 -0700
56371 Add a DrawablePtr argument to the XV functions to pave the way for redirected
56374 commit 39bdfbb4c23c9ba2a44888f22330a892c4becedc
56375 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56376 Date: Mon Apr 10 11:26:06 2006 -0700
56378 x86_64 warning fixes
56380 commit 11ad8a590d65849be00e1be4e9dd52c1159a4f24
56381 Author: Eric Anholt <anholt@FreeBSD.org>
56382 Date: Mon Apr 10 09:39:24 2006 -0700
56384 Remove the int10 POST on EnterVT. It has been reported to cause crashes now,
56385 and I believe it is the responsibility of the kernel to bring the device back to
56386 a mostly-sane state on resume anyway.
56388 commit 4217ce18cec257cad435adf9ddc9258a3c8164ec
56389 Author: Eric Anholt <anholt@FreeBSD.org>
56390 Date: Mon Apr 10 09:37:02 2006 -0700
56394 commit 334512e0604c208ffec914374a76d85720c1dcf9
56395 Author: Eric Anholt <anholt@FreeBSD.org>
56396 Date: Mon Apr 10 09:32:45 2006 -0700
56398 Don't try to use LVDS by default on chipsets that won't have an LVDS attached.
56399 The BIOS tables may still exist, so we can't rely on their presence to indicate
56402 commit 69083a2fc3ca4a3d06c1985c8a630d5628c1110c
56403 Author: Eric Anholt <anholt@FreeBSD.org>
56404 Date: Mon Apr 3 16:29:14 2006 -0700
56406 Remove BIOS save/restore code, fixing VT switching. Removes some other dead
56409 commit 2e5d85fb83def483ab1fd96877aae4a89a962b7f
56410 Author: Eric Anholt <anholt@FreeBSD.org>
56411 Date: Mon Apr 3 16:27:24 2006 -0700
56413 Turn off the VGA plane when we're setting our native modes.
56415 commit ceb08d28f4a4e9f42c9417938b8541bf1b4e8245
56416 Author: Eric Anholt <anholt@FreeBSD.org>
56417 Date: Mon Apr 3 15:32:41 2006 -0700
56419 Remove the hacky-looking Set640x480. I'll be removing things it depends on
56420 (setpipe), and I suspect we'll end up with different hacks for resume, anyway.
56422 commit 1f0ba458d02f7d4777c1669aae02138f3a6628c1
56423 Author: Eric Anholt <anholt@FreeBSD.org>
56424 Date: Mon Apr 3 15:28:47 2006 -0700
56426 Remove more BIOS stuff, particularly Get/SetDisplayDevices. Now, if there's no
56427 hardcoding in the config file, we default to LFP if we detect it from BIOS, and
56428 LFP or CRT if we can get EDID out of them.
56430 commit b20b466aaed54708ae9e4676623c8c394a6f00d5
56431 Author: Eric Anholt <anholt@FreeBSD.org>
56432 Date: Mon Apr 3 14:28:55 2006 -0700
56434 Remove the now-unused displaySize and pipeDisplaySize.
56436 commit d960c3ca1512a58a53b5c24702cb5c97124817ee
56437 Author: Eric Anholt <anholt@FreeBSD.org>
56438 Date: Mon Apr 3 14:12:29 2006 -0700
56440 Remove more BIOS modesetting stuff.
56442 commit 4e3a4827007d624aa3da1a9f5a299837bd601a33
56443 Author: Keith Packard <keithp@neko.keithp.com>
56444 Date: Fri Mar 31 14:05:46 2006 -0800
56446 autodetect LVDS dither. Fix 16bpp depth selection
56448 commit 8d27f8246ed5a73f7a78043e128b47be784186c0
56449 Author: Eric Anholt <anholt@FreeBSD.org>
56450 Date: Tue Mar 28 13:49:41 2006 -0800
56452 Clean up (and I believe fix a couple of bugs in) the divisor selection code.
56454 commit e45581cd073b23a62719f20783d617bd35316fec
56455 Author: Eric Anholt <anholt@FreeBSD.org>
56456 Date: Tue Mar 28 13:40:32 2006 -0800
56458 Fill in clock information from VBT table, so that a good clock value gets
56459 requested for choosing divisors.
56461 commit 33977d23830b5f9bb7d9e2e9c141f91cb127b7de
56462 Author: Eric Anholt <anholt@FreeBSD.org>
56463 Date: Mon Mar 27 18:50:53 2006 -0800
56465 Port code from radeon driver for panel mode validation, which will hopefully get
56466 the right mode chosen on the VAIO. Untested.
56468 commit febdfa967dbe1df487db71ba5f677ef37450bf7a
56469 Author: Eric Anholt <anholt@FreeBSD.org>
56470 Date: Mon Mar 27 15:19:52 2006 -0800
56472 Remove some VBE DDC code that I believe is OBE.
56474 commit f86892c9163473a683ab591adeb9fb1c0be2ef99
56475 Author: Eric Anholt <anholt@FreeBSD.org>
56476 Date: Fri Mar 24 16:00:30 2006 -0800
56478 Make auto panel fitting work by pulling LVDS timing EDID info out of the static
56479 BIOS table, and always using that.
56481 commit f6a3243f40074aee471ec1cda2cc8e5f05116284
56482 Author: Eric Anholt <anholt@FreeBSD.org>
56483 Date: Fri Mar 24 13:49:10 2006 -0800
56485 Add more VBT reading, so we find the panel size.
56487 commit 64756e215016730b5cc2e174f08d47c0288c0ba4
56488 Author: Eric Anholt <anholt@FreeBSD.org>
56489 Date: Fri Mar 24 11:55:01 2006 -0800
56491 Start interpreting VBT information with the goal of pulling out LVDS timings
56494 commit c2d554be671f5368bd98867b8ecf9428e4560a81
56495 Author: Eric Anholt <anholt@FreeBSD.org>
56496 Date: Thu Mar 23 14:33:04 2006 -0800
56498 Save/restore a couple of important VGA plane registers, and have vgahw
56499 save/restore everything it knows about. Also moves the save to just after the
56500 other reg saves, above the VBE save stuff which may have side effects.
56502 commit 8411c126ae66239f8b3a2261e338a723c36aa44f
56503 Author: Eric Anholt <anholt@FreeBSD.org>
56504 Date: Thu Mar 23 11:31:24 2006 -0800
56506 Update to newer airlied DDC code, and do some cleanups as well. Now tries DDC
56507 on LVDS, though my current LVDS gives no results.
56509 commit 6414ad89b9d368a032adf2358a65404f5443ef35
56510 Author: Eric Anholt <anholt@FreeBSD.org>
56511 Date: Thu Mar 23 09:48:18 2006 -0800
56513 Add enough of airlied's DDC/I2C work to pull EDID info from my CRT.
56515 commit 5215e19af58f5c2746c8d281e33ecee86e55f0a5
56516 Author: Eric Anholt <anholt@FreeBSD.org>
56517 Date: Mon Mar 13 10:29:51 2006 -0800
56519 Add some register information for VGACNTRL, and do the reg-debug thing on it.
56521 commit b77bdc2c15640a3b15233f2190179d66bc8b2a4b
56522 Author: Eric Anholt <anholt@debian-sony.jf.intel.com>
56523 Date: Wed Mar 8 15:08:06 2006 -0800
56525 Move the PFIT enabling before the enabling of the pipe, as the specs say we
56526 should. This doesn't work yet, as we apparently need to adjust the pipe to
56527 output at the full resolution of the panel, not the displayed image size,
56528 because PFIT controls scaling between the display planes and the pipe.
56530 commit de4a9e4a7891daa1488d17bf4c22283759f97373
56531 Author: Eric Anholt <anholt@debian-sony.jf.intel.com>
56532 Date: Wed Mar 8 12:54:14 2006 -0800
56534 Add a couple more bits to panel auto-fitting, though it isn't working yet.
56536 commit 02341aa68cf2171e037196019f2553f0d1b1bd0d
56537 Author: Eric Anholt <anholt@FreeBSD.org>
56538 Date: Mon Mar 6 13:23:29 2006 -0800
56540 Make the reg-snapshot code work, add a couple more registers, and use it on VT
56541 switching to hopefully help point out where we aren't restoring enough.
56542 Currently doesn't reveal anything.
56544 commit 2b6fc7ddb020a5511f6f599a2f5c37db27eddbcc
56545 Author: Eric Anholt <anholt@FreeBSD.org>
56546 Date: Fri Mar 3 15:42:15 2006 -0800
56548 Add missed save/restore of DSP*SIZE registers.
56550 commit 12ce799818722473dde2f82739d50ba4ec7f6ecd
56551 Author: Eric Anholt <anholt@FreeBSD.org>
56552 Date: Fri Mar 3 15:38:31 2006 -0800
56554 Add some untested debugging code to for helping figure out VT switch issues.
56556 commit 585cc5f256b8e91460414a26409d0e484a86718c
56557 Author: Eric Anholt <anholt@FreeBSD.org>
56558 Date: Fri Mar 3 10:37:26 2006 -0800
56560 Add more integrated LVDS support code.
56562 commit 33c62e8e654fa2228bc3fd31ec0fe010703f363a
56563 Author: Eric Anholt <anholt@FreeBSD.org>
56564 Date: Thu Mar 2 17:53:38 2006 -0800
56566 Change the order of reg writes for restore, possibly increasing chances of
56567 success. Also save and restore ADPA.
56569 commit 7f4dc09dcafc7f0423b7f3e970a0b19fbbf515dd
56570 Author: Eric Anholt <anholt@FreeBSD.org>
56571 Date: Thu Mar 2 15:55:18 2006 -0800
56573 Remove some dead VBE code.
56575 commit f3ec8653ab4c9b26d249bcf4393326de37ac8321
56576 Author: Eric Anholt <anholt@FreeBSD.org>
56577 Date: Thu Mar 2 14:31:39 2006 -0800
56579 Include config.h, fixing many issues with reading from pI830.
56581 commit 84fc3971d76f78546a0bb7dd57dba52dd893a921
56582 Author: Eric Anholt <anholt@FreeBSD.org>
56583 Date: Thu Mar 2 13:35:40 2006 -0800
56585 Turn off panel power before poking pipe B, and turn it back on if we're doing
56586 LVDS. Also, refuse to do LVDS on pipe A, as the docs say you can't.
56588 commit aca0f942f377d997c5c0b5be790ee7f255d6e046
56589 Author: Eric Anholt <anholt@FreeBSD.org>
56590 Date: Thu Mar 2 13:31:51 2006 -0800
56592 Fix prototype for i830DetectCRT.
56594 commit df6347139bc9de22af981f776d711ac9d44fad1c
56595 Author: Eric Anholt <anholt@FreeBSD.org>
56596 Date: Thu Mar 2 12:52:19 2006 -0800
56598 Use DISPPLANE_32BPP_NO_ALPHA for depth 24, else the empty alpha channel results
56599 in a black display plane.
56601 commit 8e819ced8397f0330fd44efa60fef89bd10ff70e
56602 Author: Eric Anholt <anholt@FreeBSD.org>
56603 Date: Thu Mar 2 11:55:00 2006 -0800
56605 It appears from the specs that you have to wait for vblank after disabling the
56606 plane, not the pipe.
56608 commit ee2410b965321e3d9d64fc500fdcfa6d6dc3a039
56609 Author: Eric Anholt <anholt@FreeBSD.org>
56610 Date: Thu Mar 2 11:43:41 2006 -0800
56612 Move display base setting to a separate function and use it from both mode setup
56613 and AdjustFrame. Rename AdjustFrame to reflect the fact that it doesn't touch
56616 commit 980fb78e759b752a6cae3c87d188c50ab131ba80
56617 Author: Eric Anholt <anholt@FreeBSD.org>
56618 Date: Thu Mar 2 10:49:31 2006 -0800
56620 Correct capitalization of i830SetMode.
56622 commit 2b842c7ee3f9c33cabf16561f11bf1c3a2502d4a
56623 Author: Eric Anholt <anholt@FreeBSD.org>
56624 Date: Thu Mar 2 10:08:20 2006 -0800
56626 Start implementing support for setting modes for multiple output devices,
56627 particularly LVDS. Untested.
56629 commit a09d9cf27af7016bbd9e562c7b1c338154fd3d40
56630 Author: Eric Anholt <anholt@FreeBSD.org>
56631 Date: Thu Mar 2 09:05:13 2006 -0800
56633 Replace I830VESASetMode with simpler i830SetMode, which is all native and
56634 doesn't have BIOS-setup workarounds. Multihead behavior may be wrong, and it is
56637 commit 7edb25789bd13f26134e8f5bf493897c8df322e0
56638 Author: Eric Anholt <anholt@FreeBSD.org>
56639 Date: Wed Mar 1 14:02:51 2006 -0800
56641 Start making i830SetMode aware of multiple pipes.
56643 commit b5915ac77a3887cd3aa4ce0bb77b8a36e1aa5a1b
56644 Author: Eric Anholt <anholt@FreeBSD.org>
56645 Date: Wed Mar 1 13:12:25 2006 -0800
56647 Clean up a couple of warnings.
56649 commit c20e15fbe4daeb7288f5c56cf5467eed13686080
56650 Author: Eric Anholt <anholt@FreeBSD.org>
56651 Date: Wed Mar 1 13:08:38 2006 -0800
56653 Disable I830VESASetVBEMode on restore, since I've removed that code. Note that
56654 this means that VT switching is broken. We're fine with that for now, but at
56655 least now we get a bad display rather than X crashing about an undefined symbol.
56657 commit 7d37c5036956609d0c7ae874d0dc1c616f76f849
56658 Author: Eric Anholt <anholt@FreeBSD.org>
56659 Date: Wed Mar 1 13:06:04 2006 -0800
56661 Re-disable chatty debug code.
56663 commit 2538cbeb5014e40280a3ae782a755512b1dc85fa
56664 Author: Eric Anholt <anholt@FreeBSD.org>
56665 Date: Wed Mar 1 13:01:52 2006 -0800
56667 Enable -Wall on GCC, and do a cleanup of existing warnings.
56669 commit 7746da3b346968ab8d2534fc158d026da67cc7b5
56670 Author: Eric Anholt <anholt@FreeBSD.org>
56671 Date: Wed Mar 1 12:49:06 2006 -0800
56673 Move LVDS panel power state setting to a separate function, and reduce magic
56676 commit 142bc4f91a5d776e7ab44cc5fb2328e3f7267557
56677 Author: Eric Anholt <anholt@FreeBSD.org>
56678 Date: Tue Feb 28 18:50:57 2006 -0800
56680 Write the ADPA (CRTC) register on mode setup. Obtained from airlied.
56682 commit 6877f532c5f41a445d41eb6a9982bf6bcf691bf2
56683 Author: Eric Anholt <anholt@FreeBSD.org>
56684 Date: Tue Feb 28 15:56:06 2006 -0800
56686 Add untested CRT detection code.
56688 commit a085813714818a23aa8d326804f642cd699e0f17
56689 Author: Eric Anholt <anholt@FreeBSD.org>
56690 Date: Tue Feb 28 15:16:59 2006 -0800
56692 Add adjustments of PLL divisor limits for "Almador".
56694 commit 2d4415ff1729554537b20be2b6c878444485e406
56695 Author: Eric Anholt <anholt@FreeBSD.org>
56696 Date: Mon Feb 27 16:57:43 2006 -0800
56698 Remove the hacked local xf86SetModeCRTC now that xf86SetCrtcForModes is taking
56699 effect on all of our modes.
56701 commit 611264ab9b82bca6648a9b27e5ba4b7457c46aa6
56702 Author: Eric Anholt <anholt@FreeBSD.org>
56703 Date: Mon Feb 27 15:44:11 2006 -0800
56705 Remove more VBE stuff and start using xf86ValidateModes. Guessed some params to
56708 commit 9838f639911e8c165ab615fdd9a00e02fa4b3f8a
56709 Author: Eric Anholt <anholt@FreeBSD.org>
56710 Date: Mon Feb 27 14:12:50 2006 -0800
56712 Always ErrorF output debugging info when doing BIOS calls, including file/line.
56714 commit 767944e3782f9941e9fc72a6705cc3115a6e24ac
56715 Author: Eric Anholt <anholt@FreeBSD.org>
56716 Date: Mon Feb 27 13:08:46 2006 -0800
56718 Replace a few magic numbers with symbolic names. Reviewed by md5.
56720 commit 4955cd267e7f8ed70e90b2a3de6f93de2ef859c8
56721 Author: Eric Anholt <anholt@FreeBSD.org>
56722 Date: Mon Feb 27 09:37:47 2006 -0800
56724 Add untested save/restore code, and starting on not using VBEValidateModes
56726 commit 73496b765c9783a8a271b4774a44fa263dd47684
56727 Author: Eric Anholt <anholt@FreeBSD.org>
56728 Date: Mon Feb 27 09:26:51 2006 -0800
56730 Change an if statement to a more obvious but equivalent case statement.
56732 commit 9fe316da7390918dd88940087567314b6f253981
56733 Author: Eric Anholt <anholt@FreeBSD.org>
56734 Date: Thu Feb 23 13:49:31 2006 -0800
56736 Add the manpage's possible names when generated.
56738 commit 1555229f29fa7479d6a7a51f451d04a5ef3460bd
56739 Author: Eric Anholt <anholt@FreeBSD.org>
56740 Date: Thu Feb 23 13:48:26 2006 -0800
56742 Move .cvsignore to .gitignore
56744 commit 139d33ac807fb0fc35c37c3689a6e80238199442
56745 Author: Eric Anholt <anholt@FreeBSD.org>
56746 Date: Thu Feb 23 13:46:30 2006 -0800
56748 Initial add of native CRT modesetting code.
56750 commit 35e4756645891bfc85f25f1ab60c287eba9ef749
56751 Author: Dave Airlie <airlied@linux.ie>
56752 Date: Wed Apr 5 07:12:00 2006 +0000
56754 Fix some argument inversions in xf86DrvMsg function calls
56756 commit 6e5f9b026372eea295c390b18fc41f7018e0e797
56757 Author: Kristian Høgsberg <krh@redhat.com>
56758 Date: Tue Apr 4 18:22:32 2006 +0000
56760 Add more missing #include's, in particular assert.h.
56762 commit fdfce49c54dabf0f6b0d20aa7e5eeb62b5b57eda
56763 Author: Kristian Høgsberg <krh@redhat.com>
56764 Date: Wed Mar 22 23:24:03 2006 +0000
56766 Drop libc wrapper; don't include xf86_ansic.h and add includes now missing.
56768 commit 58817c9909ba659d0070afab0ff7f4e26f722403
56769 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56770 Date: Fri Mar 10 13:07:00 2006 +0000
56772 Enable gamma for hw cursor when ARGB cursor in use (Lukas Hejtmanek)
56774 commit ab4c2f3bf4b26fdd597ea7b7a726ae77778617b2
56775 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56776 Date: Fri Mar 3 16:27:37 2006 +0000
56778 Check the version of shadow being used and turn off rotation if it isn't the
56780 Bump the Xvideo resolution support from 1920x1080 to 1920x1088.
56782 commit be6216cd8c6a47dd053240f7a5d1445b61f79038
56783 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56784 Date: Mon Feb 20 03:25:50 2006 +0000
56786 Fix a rotation problem when DRI is disabled
56788 commit 073c4a6a18e98e2dde114756b5d787b13bb2b7a5
56789 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56790 Date: Mon Feb 20 03:22:26 2006 +0000
56792 Fix a rotation problem when DRI is disabled
56794 commit 7cca2ae73e0114a9df2152ed5ed5b66fc44f65dd
56795 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56796 Date: Sun Feb 19 00:04:49 2006 +0000
56798 fix another build problem
56800 commit 2e01277da0be55687ca901efadcaf5c919826824
56801 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56802 Date: Fri Feb 17 16:16:10 2006 +0000
56806 commit 86fa25369353471dc9a1041df74e66e9029863a7
56807 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56808 Date: Wed Feb 15 18:31:22 2006 +0000
56810 a little more of a workaround until we unplug the frontbuffer from libdri
56812 commit 32f97e02ac3611fd5e3870826dedf701f3b69a51
56813 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56814 Date: Fri Feb 10 21:59:37 2006 +0000
56816 temporary workaround until DRI frontbuffer mapping code is removed from
56819 commit 50e5c2696389cb4d5907a822aa79d5e0e03cee8c
56820 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56821 Date: Thu Jan 26 09:55:15 2006 +0000
56825 commit 91d2ba5743041077754de3bd3565df930b036c7d
56826 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56827 Date: Thu Jan 26 09:05:50 2006 +0000
56831 commit 076971b95ed3b89e1a333adbffceab1c2a5b7743
56832 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56833 Date: Wed Jan 25 22:22:42 2006 +0000
56835 When going dual head only take a portion of memory for the second head
56836 instead of doubling the requirements.
56838 commit 77d1c579f3031789031dd203175d3c9555226f79
56839 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56840 Date: Wed Jan 25 22:17:00 2006 +0000
56842 Fix a glitch in a dual head fix from #3105
56844 commit d8d749080923e88376db1aa18ae2ad722d5113d9
56845 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
56846 Date: Tue Jan 24 15:20:34 2006 +0000
56848 Bump version to 1.5.0.0
56849 Add Intel 945GM support
56850 Add RandR rotation support (full 3D acceleration, HWcursor & Xvideo rotated too)
56851 Remove shadow framebuffer rotation code
56852 Add a new LinearAlloc option to allow more offscreen memory to be allocated
56853 for XVideo applications. This allows HDTV movies to be played via Xvideo.
56855 commit 919d36b6717e9a7d930c5120309778da7b0fe16c
56856 Author: Kevin E Martin <kem@redhat.com>
56857 Date: Wed Dec 21 02:30:00 2005 +0000
56859 Update package version for X11R7 release.
56861 commit 92962f32fccac6df751268190928d8becb558b64
56862 Author: Adam Jackson <ajax@nwnk.net>
56863 Date: Mon Dec 19 16:25:53 2005 +0000
56867 commit d2ef609211306b6bd3861c0e5f6d3a9af3e145fb
56868 Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
56869 Date: Mon Dec 19 09:07:50 2005 +0000
56871 Typo fixes, mailing list & url updates, and other changes to prepare for
56872 X11R6.9 & 7.0 releases.
56874 commit 1c51ff5c652690d66d1f7cba803c0bb26208b509
56875 Author: Kevin E Martin <kem@redhat.com>
56876 Date: Thu Dec 15 00:24:17 2005 +0000
56878 Update package version number for final X11R7 release candidate.
56880 commit c5007a4b46ffb66f5d98abe8bcae6df8c6eb1b31
56881 Author: Kevin E Martin <kem@redhat.com>
56882 Date: Tue Dec 6 22:48:35 2005 +0000
56884 Change *man_SOURCES ==> *man_PRE to fix autotools warnings.
56886 commit 15df4215347854604fd3f6795ac5f971444cf39d
56887 Author: Kevin E Martin <kem@redhat.com>
56888 Date: Sat Dec 3 05:49:33 2005 +0000
56890 Update package version number for X11R7 RC3 release.
56892 commit c50cd5a7f5af7d45550f25f173b512eb89bf41c5
56893 Author: Kevin E Martin <kem@redhat.com>
56894 Date: Fri Dec 2 02:16:08 2005 +0000
56896 Remove extraneous AC_MSG_RESULT.
56898 commit da2c00c0e85a284e3fbd237f3a77212f9647da37
56899 Author: Adam Jackson <ajax@nwnk.net>
56900 Date: Wed Nov 30 18:47:54 2005 +0000
56902 Bump libdrm dep to 2.0.
56904 commit ca36f1da4e254c779b38df30fd655533e548a294
56905 Author: Adam Jackson <ajax@nwnk.net>
56906 Date: Tue Nov 29 23:29:59 2005 +0000
56908 Only build dlloader modules by default.
56910 commit 0b337390d6919e9a3d31ca654c8bc44b30756862
56911 Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
56912 Date: Mon Nov 28 22:04:09 2005 +0000
56914 Change *mandir targets to use new *_MAN_DIR variables set by xorg-macros.m4
56915 update to fix bug #5167 (Linux prefers *.1x man pages in man1 subdir)
56917 commit 159bea027893ce5996a452a6f493372792e95122
56918 Author: Eric Anholt <anholt@freebsd.org>
56919 Date: Mon Nov 21 10:49:14 2005 +0000
56921 Add .cvsignores for drivers.
56923 commit e62939e30e664592085ce394680fd8e7f3d31339
56924 Author: Kevin E Martin <kem@redhat.com>
56925 Date: Sat Nov 19 07:15:38 2005 +0000
56927 Update pkgconfig files to separate library build-time dependencies
56928 from application build-time dependencies, and update package deps
56929 to work with separate build roots.
56931 commit 08f03e4237ce9a54f0dffa37307af402b0bd2e55
56932 Author: Kevin E Martin <kem@redhat.com>
56933 Date: Wed Nov 9 21:31:20 2005 +0000
56935 Bump version numbers for X11R6.9/X11R7 RC2.
56937 commit a600e89a1547908f2018de96b5ecb814e2427b34
56938 Author: Kevin E Martin <kem@redhat.com>
56939 Date: Wed Nov 9 21:15:15 2005 +0000
56941 Update package version number for X11R7 RC2 release.
56943 commit f73c23b4c588c2f5dbc3c36a2f43b8bbb1e0f36a
56944 Author: Kevin E Martin <kem@redhat.com>
56945 Date: Tue Nov 1 15:08:53 2005 +0000
56947 Update pkgcheck depedencies to work with separate build roots.
56949 commit 25a63bfefda420f85b3cba89bdae4b35ef80c84f
56950 Author: Adam Jackson <ajax@nwnk.net>
56951 Date: Fri Oct 21 18:40:19 2005 +0000
56953 Bug #4829: Remove some useless DRI symbol checks.
56955 commit 53c122fa162709f618be26d5940e3488c0cbbafa
56956 Author: Kevin E Martin <kem@redhat.com>
56957 Date: Wed Oct 19 02:48:03 2005 +0000
56959 Update package version number for RC1 release.
56961 commit 4b97a544cac9fb507b4a2a7c8017607ea2cac1bb
56962 Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
56963 Date: Tue Oct 18 00:01:53 2005 +0000
56965 Use @DRIVER_MAN_SUFFIX@ instead of $(DRIVER_MAN_SUFFIX) in macro substitutions
56966 to work better with BSD make
56968 commit c7b7ece6330561b3745f0e1136b57575eadec5ad
56969 Author: Adam Jackson <ajax@nwnk.net>
56970 Date: Mon Oct 17 22:57:33 2005 +0000
56972 More 1.7 braindamage: define EXTRA_DIST in terms of @DRIVER_NAME@ instead of indirectly
56974 commit 0586dbe1c5d0432e8d8da2f3576ac2ccbe139919
56975 Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
56976 Date: Mon Oct 17 00:09:03 2005 +0000
56978 Use sed & cpp to substitute variables in driver man pages
56980 commit 1317f2d1f35241969dcf376abd106ffe3f918432
56981 Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk>
56982 Date: Tue Oct 4 20:01:03 2005 +0000
56984 Check in generated README files
56986 commit bd48e0531a16bdf27fc8549826cc1123ba12d5bd
56987 Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk>
56988 Date: Mon Oct 3 21:22:07 2005 +0000
56990 - For all drivers that have a <driver>.sgml file, add code in their build
56991 system to build the README file at make dist time
56993 - in util/macros/xorg-macros.m4, add a new XORG_CHECK_LINUXDOC macro that
56994 will check if the required tools and files exist, and if so set a
56997 - util/modular/symlink.sh
56998 - Link all the <driver>.sgml to xf86-video-<driver>/README.sgml
56999 - Add all the README.<driver> to the list of excluded files
57001 - xc/programs/Xserver/hw/xfree86/doc/sgml/SiS.sgml: Various changes
57002 to make it spew less warnings when the text file is built.
57004 commit 6fda4e886c30669e346f7f30a38b93a77004a7f6
57005 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57006 Date: Mon Oct 3 09:39:17 2005 +0000
57008 Bug #4353: Fix lockup when suspend and resume when playing video.
57009 Shutdown the overlay on LeaveVT.
57011 commit 154724cc4be1e951649c29e4abdbde964a8c112e
57012 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57013 Date: Tue Sep 27 08:58:23 2005 +0000
57015 Put back the print code to display the DDC information, and remove some bogus
57018 commit 2174848266940e1896970c12717e106a4742dcbc
57019 Author: Adam Jackson <ajax@nwnk.net>
57020 Date: Sat Sep 24 18:45:45 2005 +0000
57022 Disable the {Open,Close}FullScreen DRI protocol.
57023 Remove empty FullScreen stubs from drivers, comment the non-empty ones.
57025 commit b6443515e9d485b1ab8ee266c5b21750d2a5a551
57026 Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
57027 Date: Thu Sep 22 01:58:17 2005 +0000
57029 Don't build xvmc module if DRI support is not present
57031 commit 76033c2c885524c2116ccf0bdd30c89fa2636150
57032 Author: Eric Anholt <anholt@freebsd.org>
57033 Date: Thu Sep 15 06:04:59 2005 +0000
57035 Bug #4458: Check for the xvmc proto package, since it's required to build these.
57038 commit 50f0bf51a67c89bf445a682f8fec34c52ff909a8
57039 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57040 Date: Thu Sep 8 14:51:40 2005 +0000
57042 fix another possible crash
57044 commit 521e357c1846960b48e83fb97f77025f6f91e638
57045 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57046 Date: Thu Sep 8 09:31:04 2005 +0000
57048 Fix a crash situation
57050 commit e8c46c897564f038744b9274949482568d654ecc
57051 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57052 Date: Tue Sep 6 10:31:33 2005 +0000
57054 Fix DirectColor visual colormap issues at 16bpp in the i830 driver.
57056 commit bf6fef41935ebf624cc1ae6faa084dfa0a7eee72
57057 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57058 Date: Tue Sep 6 10:23:57 2005 +0000
57060 Sync the accelerator before calling AdjustFrame to avoid lockups.
57062 commit 5fc5f02d76e077b09bf3ab1c7deeb1263033d57a
57063 Author: Dave Airlie <airlied@linux.ie>
57064 Date: Mon Aug 29 12:06:13 2005 +0000
57066 Add i810 XvMC support, it passes distcheck for me.
57068 commit 86e6935bd50b129519b24ee3d694a3f75b8d4f1c
57069 Author: Dave Airlie <airlied@linux.ie>
57070 Date: Mon Aug 29 12:02:29 2005 +0000
57072 Fix include paths for i810 XvMC add config.h to driDrawable.c
57074 commit 6f1406753da102ccce1898b3f160d6eaaccc9b7c
57075 Author: Kristian Høgsberg <krh@redhat.com>
57076 Date: Fri Aug 26 15:32:40 2005 +0000
57078 Added i830_io.c, which just includes i810_io.c with BUILD_FOR_I830 set.
57080 commit 219beaca91b88dabfe05b8e756d558269f7a5148
57081 Author: Daniel Stone <daniel@fooishbar.org>
57082 Date: Sun Aug 21 06:26:33 2005 +0000
57084 Fix typo in debug ErrorF.
57086 commit fdc6997f2edf4dfdcfd877152028bff2a0306999
57087 Author: Daniel Stone <daniel@fooishbar.org>
57088 Date: Thu Aug 18 09:03:47 2005 +0000
57090 Update autogen.sh to one that does objdir != srcdir.
57092 commit 461d43be78f49e6cb965b7a206c77f6fe1622601
57093 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57094 Date: Mon Aug 15 07:30:05 2005 +0000
57096 Egbert's 64bit fixes for mixed 32/64bit clients
57098 commit 95bd8dbee69b4b35e4c37e19d15e0449c38726b6
57099 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57100 Date: Fri Aug 12 21:23:59 2005 +0000
57102 add missing i830_shadow.c
57104 commit 7476d0413f161ddb3f3cc4a77c3f2a2912d17e60
57105 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57106 Date: Mon Aug 8 09:17:06 2005 +0000
57108 Enable gamma correction bit even for 8bit modes.
57110 The pipeconf register handles the real enablement.
57112 commit a89f89a2423aecb199d5f4af2e9430e180aa1131
57113 Author: Ian Romanick <idr@us.ibm.com>
57114 Date: Fri Aug 5 23:04:33 2005 +0000
57116 I have resolved this by making all the "Direct rendering disabled" messages
57117 X_WARNING and all the "Direct rendering enabled" messages X_INFO. Ideally,
57118 we need to factor some of this stuff out to common code. There is a *LOT* of
57119 duplication among the *_dri.c files under programs/Xserver/hw/xfree86/drivers.
57121 I'm trying to clean up as much low hanging fruit as I can and got tired of
57122 waiting for ajax. :P
57124 commit fd6d7d7f7ea03961c5ecf7db029f12ece93d414a
57125 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57126 Date: Tue Aug 2 16:22:42 2005 +0000
57128 Adjust newPipe code for BIOS's starting at 1200
57130 commit da7b85176fc061995e1b04fa22dc736b6c9353e6
57131 Author: Adam Jackson <ajax@nwnk.net>
57132 Date: Sun Jul 31 17:19:29 2005 +0000
57135 - Disable building the gamma driver
57136 - Fix numerous Imakefiles to get header paths and subdirs right
57137 - Symlink client-side GLX lib source from Mesa
57138 - Add new shader subdirs to the Mesa core build
57139 - Tweak the DDX drivers to not ask for "xf86dri.h"; they don't need
57140 it, since they don't generate protocol, and it was breaking the
57143 commit 1561ec3ca4f0d774f2664e9354d537061a0102d1
57144 Author: Kevin E Martin <kem@redhat.com>
57145 Date: Fri Jul 29 21:22:43 2005 +0000
57147 Various changes preparing packages for RC0:
57148 - Verify and update package version numbers as needed
57149 - Implement versioning scheme
57150 - Change bug address to point to bugzilla bug entry form
57151 - Disable loadable i18n in libX11 by default (use --enable-loadable-i18n
57153 - Fix makedepend to use pkgconfig and pass distcheck
57154 - Update build script to build macros first
57155 - Update modular Xorg version
57157 commit d165b7dc4ec3793e921363cdeb1ad41c0a3653d6
57158 Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk>
57159 Date: Wed Jul 27 21:23:29 2005 +0000
57161 Add xf86driproto to dependencies for drivers with DRI support. Patch from
57164 commit 0cd371c6e57f899c34c4d2e2a8531ef7ce3967a4
57165 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57166 Date: Tue Jul 26 21:55:42 2005 +0000
57168 Bump the DDX version to 1.4.0 to indicate the removal of drmAddress.
57169 Thus allowing 32/64bit clients to cohabit.
57171 commit adef005c2f96395d396834f2173125f0c06d40cb
57172 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57173 Date: Tue Jul 26 08:07:37 2005 +0000
57175 Bug #3850: <https://bugs.freedesktop.org/show_bug.cgi?id=3850>
57176 Fix non-DRI builds.
57178 commit 09e871b1b4a67cfe818613bc012d72965b449f1c
57179 Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk>
57180 Date: Tue Jul 26 00:41:42 2005 +0000
57182 Build system for i810
57184 commit ac920dae7c22ed3956e4f31a0d3baf1909762b76
57185 Author: Adam Jackson <ajax@nwnk.net>
57186 Date: Mon Jul 11 02:29:51 2005 +0000
57188 Prep for modular builds by adding guarded #include "config.h" everywhere.
57190 commit f90f661652e5fe0986eb804f8dff3eeffe0af820
57191 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57192 Date: Sat Jul 9 21:19:11 2005 +0000
57194 Copy mode strings upon mode copy
57196 commit 8fced5b648b0666ca88221e37bc1131ab5c2072d
57197 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57198 Date: Tue Jul 5 09:25:03 2005 +0000
57200 Fix Xvideo when operating in one line mode and video is pushed off the top of
57201 the screen. Also some fixes for Xvideo skew in certain resolutions.
57203 commit 68bc259fa1082f7780e524fc703e926a27fc7f9a
57204 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57205 Date: Mon Jul 4 20:45:39 2005 +0000
57209 commit 83c3e3ee49a0aa3ceaf52e2408cc253fc5dc8af4
57210 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57211 Date: Tue Jun 28 16:13:02 2005 +0000
57213 Add support to check the in-built VESA mode list.
57215 commit 0d64e9e1f5494cff633a1499545d55a0edd0d92e
57216 Author: Adam Jackson <ajax@nwnk.net>
57217 Date: Sat Jun 25 21:16:55 2005 +0000
57219 Bug #3626: _X_EXPORT tags for video and input drivers.
57221 commit 67174d67e740b9082ed0a65a79f08b5df3b4fcfb
57222 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57223 Date: Mon Jun 13 12:15:25 2005 +0000
57225 commit missing piece from resume patch
57227 commit ddab0558d584fb906cc0dfb383cc0e7a122d5336
57228 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57229 Date: Thu Jun 9 09:26:03 2005 +0000
57231 Bug #3496 <https://bugs.freedesktop.org/show_bug.cgi?id=3496>
57232 Remove a bogus check for DRM 1.2 for resume functionality as it's
57233 been present since 1.1
57235 commit 7f49513afe9d4490bdfdac182d50383c089875e0
57236 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57237 Date: Wed Jun 8 17:50:32 2005 +0000
57239 Fix the mode validation such that using function 5f05 to set the refresh rate
57242 commit f8de9b4b9a5ce2756652c6b9ec1e84aedbc3519a
57243 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57244 Date: Tue Jun 7 18:54:10 2005 +0000
57246 Fix text restoration when display switching has been activated and we could be
57247 potentially on a different display than the initial one.
57249 commit 1f538007e9da7d4ad30ec4d4d281fb07ef300ddb
57250 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57251 Date: Mon Jun 6 09:47:29 2005 +0000
57255 commit d8a32bf4274ad91cc2b9ca425f5b0a017c2dd293
57256 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57257 Date: Mon Jun 6 09:40:23 2005 +0000
57259 Bug #3436 <https://bugs.freedesktop.org/show_bug.cgi?id=3054>
57260 Fix some offset, pitch and overlay scaler size problems with
57261 the video overlay. Solves various spontaneous lockups.
57263 commit 005d121d280ff4502f4d6e3d3d585ff0175367bb
57264 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57265 Date: Tue May 31 20:29:01 2005 +0000
57267 and another. sheesh.
57269 commit 27f86d693c32e550d9ebf4d6ff05c2d827754c8b
57270 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57271 Date: Tue May 31 20:27:24 2005 +0000
57275 commit 27feb0f3474c1e21557225e10b13558e002c0ebf
57276 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57277 Date: Tue May 31 17:58:26 2005 +0000
57279 Add Intel(R) 945G support
57281 commit 2d10e8dc8184b0514e8ac9f835a06b85d609bb1b
57282 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57283 Date: Wed May 25 09:10:44 2005 +0000
57285 Change the sequence on how refresh rates are set in the i830 driver. This works
57286 much better with custom modelines.
57288 commit 89ec4ac201c09329d5d732323bdce6250b8592fe
57289 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57290 Date: Fri May 20 15:24:02 2005 +0000
57292 Bugzilla #3292: https://bugs.freedesktop.org/show_bug.cgi?id=3292
57293 Add shadowFB support to the Intel driver (Dima Dorfman)
57295 commit df0964ea8f92c6b278dd87b95380dbc3ada625e5
57296 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57297 Date: Fri May 20 13:24:02 2005 +0000
57299 Bugzilla #2339: https://bugs.freedesktop.org/show_bug.cgi?id=2339
57300 Bugzilla #1064: https://bugs.freedesktop.org/show_bug.cgi?id=1064
57301 Bugzilla #3055: https://bugs.freedesktop.org/show_bug.cgi?id=3055
57302 Much better display switching support for mobile chipsets using
57303 the Fn+F? combination keys.
57304 Add DirectColor support.
57305 Add gamma correction support.
57307 commit b99ef9cd59fce0e02e8829210c7cb8aadb6a3499
57308 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57309 Date: Fri May 20 11:56:45 2005 +0000
57311 Bugzilla #2585: https://bugs.freedesktop.org/show_bug.cgi?id=2585
57312 Support the Intel E7221 server chipset.
57314 commit 3b87fac327b5549678f91a9c961b9fe5c0bf8a37
57315 Author: Adam Jackson <ajax@nwnk.net>
57316 Date: Wed May 18 17:47:35 2005 +0000
57318 Bug #3163: Create new DRIINFO_*_VERSION macros to indicate the
57319 version number of the DRIInfoRec, disambiguating it from the XF86DRI
57320 protocol version number. Modify DRIQueryVersion to return the libdri
57321 version number, which all DDXes were requesting implicitly. Fix the
57322 DDXes to check for the DRIINFO version they were compiled against.
57324 commit 72100ac7cff7485db3e869f1331caddb0bc21a24
57325 Author: Daniel Stone <daniel@fooishbar.org>
57326 Date: Wed Apr 20 12:25:28 2005 +0000
57328 Fix includes right throughout the Xserver tree:
57329 * change "foo.h" to <X11/foo.h> for core headers, e.g. X.h, Xpoll.h;
57330 * change "foo.h", "extensions/foo.h" and "X11/foo.h" to
57331 <X11/extensions/foo.h> for extension headers, e.g. Xv.h;
57332 * change "foo.[ch]" to <X11/Xtrans/foo.[ch]> for Xtrans files.
57334 commit c813bf655908cd72049f346919abbdff8365b944
57335 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57336 Date: Thu Apr 14 22:55:01 2005 +0000
57338 Fix initial DDC parameters when validating modes.
57340 commit e637630908f40f42367c94344d44918b6cdb0eb2
57341 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57342 Date: Fri Apr 1 14:17:24 2005 +0000
57344 Check the VESA BIOS scanline requirements and adjust the displayWidth
57345 accordingly. Fixes an assumption that the screen width was always used as the
57348 commit 5090c8ba6778091aa4ce24b16b058d534ba78dc7
57349 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57350 Date: Mon Mar 21 21:20:39 2005 +0000
57352 a one line fixup to the UseDDC code
57354 commit 8bc9b0d7da5de9a3294687f450afb50452898c7d
57355 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57356 Date: Mon Mar 21 18:57:39 2005 +0000
57358 Fix some bad alignment problems when allocating memory from
57359 the end of AGP space.
57360 Bugzilla #2764 (https://bugs.freedesktop.org/show_bug.cgi?id=2764)
57362 commit 50a105e88cfdfc22fef8cfd2397fc5e9fae17663
57363 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57364 Date: Tue Mar 15 13:10:23 2005 +0000
57366 LFP's may exist on Pipe A in older i830's. Check that case so that Panel ID can
57367 be obtained and Xvideo is correctly setup for panel fitting.
57369 commit eb8b5df739e7be4509423c2bf757a7df35ce02f6
57370 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57371 Date: Tue Mar 15 11:32:10 2005 +0000
57373 Enforce DDC monitor ranges usage if we have them and reject bad
57374 modes. Use NoDDC option to override DDC timings. We currently
57375 only use DS_RANGES, but could use other DDC information, as does
57376 the common layer, to deduce the h/v ranges.
57378 commit 49dcbacf7fa7f29d18c18eec7e2f99e859bf4c7e
57379 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57380 Date: Thu Mar 3 18:01:51 2005 +0000
57382 Limit the maximum refresh rate to 85Hz and remove the bogus
57383 100Hz and 120Hz settings as they don't match close enough and
57384 are really unsupported.
57386 commit 7eaf88d00d7539fde68422ae6566f23993aa2633
57387 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57388 Date: Tue Mar 1 19:53:28 2005 +0000
57390 Fix an off by one error when scaling which caused occasional
57391 overlay color lines at the far right and bottom of the window.
57393 commit 28d65bceaec3ef636f0f5bdd03b38fd76f87c8c6
57394 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57395 Date: Thu Feb 10 13:16:08 2005 +0000
57397 Fix unaligned YV12 videos that caused skew when playing.
57399 commit 6352e65b9cd635e9f266e0e9ed64c5beae9e5c44
57400 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57401 Date: Thu Feb 10 10:41:20 2005 +0000
57403 Fix pixmap cache corruption when Xv was in use due to a memory allocator
57404 calculation problem.
57406 commit f691f959213d3eb5c365edaf5a5a53fa96300491
57407 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57408 Date: Tue Feb 8 20:13:33 2005 +0000
57410 Setup the Xv atoms correctly.
57412 commit 31a4191a97f538b2c6d4b8036796f06443f783c1
57413 Author: Egbert Eich <eich@suse.de>
57414 Date: Mon Feb 7 11:56:31 2005 +0000
57416 Check the pixel clock choosen for a specific refresh rate
57417 against the maximally allowed pixel clock when choosing
57418 mode lines for VBESetVBEMode() and VESA BIOS version >= 3.0
57421 commit a103496e08f699b609be875827ecf1f3afa45e68
57422 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57423 Date: Tue Feb 1 19:47:32 2005 +0000
57425 Check return value of i830InitDma() and fail to initialize the DRI
57427 Also, check the DRM module name that it matches i915 and not the
57428 older i830 driver which some people were still trying to use
57429 and was causing Dma Initialization failures and lockups.
57431 commit 7ff7cec9220815eb5f74cfd95860aaa5f1f263bd
57432 Author: Egbert Eich <eich@suse.de>
57433 Date: Mon Jan 24 15:02:17 2005 +0000
57435 Unmap XCursor AGP memory on VT switch to prevent system from lockup
57438 commit d4c87adf6792ae9af641b9824c547d18cdd1e521
57439 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57440 Date: Mon Jan 10 11:29:02 2005 +0000
57442 Revert some changes to TweakMemorySize
57444 commit 076c5116ae5eda1822da70184cbcc8942b470541
57445 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57446 Date: Thu Jan 6 14:43:14 2005 +0000
57449 - Intel i915GM support to 2D DDX and 3D Mesa drivers.
57450 - PanelID identification
57451 - DRI suspend/resume support
57452 - Detection of monitor changes on VT switches
57453 - Support custom video modes if available in the Video BIOS
57455 * GL_ARB_texture_cube_map
57456 * GL_EXT_blend_equation_separate
57457 * GL_ATI_blend_equation_separate
57458 * GL_ARB_point_parameters
57459 * GL_NV_blend_square
57460 * GL_EXT_cull_vertex
57461 * GL_ARB_depth_texture
57462 * GL_SGIX_depth_texture
57464 * GL_EXT_shadow_funcs
57465 * GL_3DFX_texture_compression_FXT1
57466 By Tungsten Graphics, Keith Whitwell & Alan Hourihane.
57468 commit edb9b6fde0948bb1121a8456500ef9f55f734376
57469 Author: Adam Jackson <ajax@nwnk.net>
57470 Date: Mon Jan 3 16:03:56 2005 +0000
57472 Bug #1488: Fix crash on i830 during startup. (Eric Anholt)
57474 commit 59cecb714b78f7f384c24bf159bd39c81e9c7ee1
57475 Author: Markus Kuhn <Markus.Kuhn@cl.cam.ac.uk>
57476 Date: Sat Dec 4 00:43:09 2004 +0000
57478 Encoding of numerous files changed to UTF-8
57480 commit c24b0e04f3a1b064412e0bea44c4af546ec9fb77
57481 Author: Egbert Eich <eich@suse.de>
57482 Date: Tue Nov 2 08:54:53 2004 +0000
57484 Fixed KGA handling for i810. KGA handling for chips derived from C&T chips is
57485 slightly different. The changes make the code consistent with the C&T (chips)
57488 commit 7ce4898ee319d149cf86e938e7257ea7490d327f
57489 Author: Egbert Eich <eich@suse.de>
57490 Date: Wed Oct 27 09:52:15 2004 +0000
57492 Removing vtSema shilding of accel sync - should have been '&&' instead of '||'
57495 commit 88714c3b2274fcfe50837631b1d38a7cd96f692a
57496 Author: Egbert Eich <eich@suse.de>
57497 Date: Tue Oct 12 19:13:43 2004 +0000
57499 Change order of SetDisplayDevice(), HWRestore(), UnbindGART() and
57500 RestoreBIOSMemSize() to be exactly opposite to the Save procedure in EnterVT()
57501 (Matthias Hopf, Alan Hourihane).
57502 Fixed server crash on reset when a structure allocated in PreInit() was freed on
57504 Fixed ring buffer lock ups that happened because the structure that contained
57505 ringbuffer data was not zeroed after allocation.
57506 Fixed numerous warnings due to signed unsigned comparisons.
57508 commit 687298daed40aaade44e61342a9abe984088541b
57509 Author: Egbert Eich <eich@suse.de>
57510 Date: Mon Oct 11 09:58:04 2004 +0000
57512 Checking if server isn't switched away before calling sync.
57513 Sanity check for possible bugs in aother areas of the code.
57514 Fixing default amount of of allocated video memory from AGP for i810: Use 16MB
57515 if less than 192MB are installed else use 24MB (Matthias Hopf).
57517 commit d2866271b8d42a0d44b65e30667adab96ac8fd54
57518 Author: Adam Jackson <ajax@nwnk.net>
57519 Date: Wed Sep 22 04:38:03 2004 +0000
57521 Bug #1252, #1253, #1255, #1256: Various typo fixes from Dave Jones.
57523 commit 88fd270d23ed1a1761cfdb98367c89ca89735b12
57524 Author: Adam Jackson <ajax@nwnk.net>
57525 Date: Fri Sep 17 03:04:51 2004 +0000
57527 Bug #1192: Remove cfb support from drivers where its use is an option.
57528 Delete xf24_32bpp, as s3virge was the last user. Fix up some comments
57529 to refer to fb rather than cfb.
57531 commit 9e8e19c6fde15fe5b9bb7532316dedeeca54b8d9
57532 Author: Kevin E Martin <kem@redhat.com>
57533 Date: Mon Aug 30 03:19:08 2004 +0000
57535 Fix ARGB cursor problem on i810 (Bug #1195, Alan Hourihane).
57537 commit 15c61546ea1a64d6c7dd60764164143ea7174123
57538 Author: Kevin E Martin <kem@redhat.com>
57539 Date: Thu Aug 26 05:10:56 2004 +0000
57541 Fix i810 compilation without Xinerama (Bug #1180, Simon Schubert).
57543 commit 828c0fb8050dcbed8026365cbcb17d4d4330c268
57544 Author: Kevin E Martin <kem@redhat.com>
57545 Date: Wed Aug 25 00:30:41 2004 +0000
57547 Fix i810 driver startup on Sony VAIO laptop (Bug #1084, Dave Airlie).
57549 commit c4626af83d760c2a6aae90ad478430eac1c4f4ea
57550 Author: Egbert Eich <eich@suse.de>
57551 Date: Fri Jul 30 21:36:40 2004 +0000
57553 Do I810EmitFlush() during screen init to make sure 2D engine is in a sane state
57556 commit e0e01b7d6188a83c7539256ac1710ace04ce132a
57557 Author: Egbert Eich <eich@suse.de>
57558 Date: Fri Jul 30 20:38:27 2004 +0000
57560 add i845 to the list of chips that allow memory size tweaking.
57562 commit 199620cdc924e6aa98683e305843ea91c7e1771d
57563 Author: Adam Jackson <ajax@nwnk.net>
57564 Date: Fri Jul 30 20:30:52 2004 +0000
57566 Bug #400 (partial): Driver fixes for the dlloader. When using
57567 dlloader, all framebuffer formats except cfb and the overlay modes
57568 should work, and r128 and radeon need to be loaded from the ati
57569 driver (both issues to be fixed soon). Tested on i740, s3virge,
57570 mach64, tdfx, vesa, and vga drivers. elfloader users shouldn't be
57573 commit 54e644ffbb63290b7e2ec51b041f5b2df26e164b
57574 Author: Egbert Eich <eich@suse.de>
57575 Date: Fri Jul 30 19:04:14 2004 +0000
57577 Removed bugs on TweakMemorySize() which prevented it from working at all.
57578 Initialized last element of BIOS version number string to 0 to avoid random
57581 commit 3a3b00f18ac5edd7dfef06a96342c95f3cb97f79
57582 Author: Kevin E Martin <kem@redhat.com>
57583 Date: Fri Jul 30 04:44:13 2004 +0000
57585 Fix "DoLoadableServer NO" build
57587 commit 986bc24892e7ce3784fc723fda54b1f89dd75b66
57588 Author: Egbert Eich <eich@suse.de>
57589 Date: Tue Jul 6 14:37:47 2004 +0000
57591 Increase default video RAM size to 16MB when DRI is enabled and more than 128MB
57592 are available (Egbert Eich).
57593 Fixed lockups during mode switch. Problem was introduced when attempting to copy
57594 the behavior during LeaveVT()/EnterVT() but but forgetting to call
57595 I810DRILeave() before I810DRIEnter(). The entire DRILeave()/Enter() scenario has
57596 been commented out as it didn't seem to be necessary (Egbert Eich).
57597 Fix TweakMemorySize() (tested with i855/i865) (Egbert Eich).
57599 commit b7310758e60956a7de14dcf968bb713dfc56738f
57600 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57601 Date: Tue Jun 22 15:12:42 2004 +0000
57605 commit 5ed7e6f8424236dd57d83f6ab1282693260969cb
57606 Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
57607 Date: Tue Jun 22 14:28:46 2004 +0000
57610 Dual Head support for i830, i855, i915.
57611 ARGB cursor support (including i810) for all chipsets.
57613 commit a46e95722d3c2b9dfb9eacf9a390a7c47d427773
57614 Author: Eric Anholt <anholt@freebsd.org>
57615 Date: Wed Jun 16 19:16:54 2004 +0000
57617 Mop-up from the DRI 20040613 import, taking care of some drm.h
57618 header location and type changes, and update the ChangeLog.
57620 commit 35a5cf103e9e41d178f980f685307f8c8063a006
57621 Author: Eric Anholt <anholt@freebsd.org>
57622 Date: Wed Jun 16 09:44:00 2004 +0000
57624 Merge DRI-trunk-20040613 changes in programs/Xserver/hw/xfree86/drivers, with
57625 the following notes:
57627 - Savage and Mach64 (= ati/ati*.[ch] changes) DRI not merged due to insecurity.
57628 - VIA driver converted to new drmContext and drmHandle names.
57629 - Radeon driver merge conflicted in many places, and MergedFB at least could
57630 probably use some checking at this point.
57632 commit bec47f9369b6f02ce4537ab1ac547ba933609edc
57633 Author: Eric Anholt <anholt@freebsd.org>
57634 Date: Wed Jun 16 09:26:18 2004 +0000
57636 DRI trunk-20040613 import
57638 commit c5b1972b50a6dbe57c67d3cc88ae010386b908a6
57639 Author: Eric Anholt <anholt@freebsd.org>
57640 Date: Wed Jun 16 09:23:18 2004 +0000
57642 DRI XFree86-4_3_99_12-merge import
57644 commit 41eba03d0d62fe03a7b437c9c4914016c3125141
57645 Author: Egbert Eich <eich@suse.de>
57646 Date: Fri Apr 23 19:35:55 2004 +0000
57648 Merging XORG-CURRENT into trunk
57650 commit e27494f2a9262b6310c2f1e0a40c406f578fc693
57651 Author: Egbert Eich <eich@suse.de>
57652 Date: Sun Mar 14 08:33:22 2004 +0000
57654 Importing vendor version xf86-4_4_99_1 on Sun Mar 14 00:26:39 PST 2004
57656 commit 5ece535dc7f77185cc8c327a57ae3c02c8b9f918
57657 Author: Egbert Eich <eich@suse.de>
57658 Date: Wed Mar 3 12:12:19 2004 +0000
57660 Importing vendor version xf86-4_4_0 on Wed Mar 3 04:09:24 PST 2004
57662 commit ba904078dd22e013f4112479e7c2669c33e708a8
57663 Author: Egbert Eich <eich@suse.de>
57664 Date: Thu Feb 26 13:35:53 2004 +0000
57666 readding XFree86's cvs IDs
57668 commit 13ea363db30dc0010cd09df29df90074db77f646
57669 Author: Egbert Eich <eich@suse.de>
57670 Date: Thu Feb 26 09:23:19 2004 +0000
57672 Importing vendor version xf86-4_3_99_903 on Wed Feb 26 01:21:00 PST 2004
57674 commit 64b129e800d915f0d82a67ce0f8b1850e67012fc
57675 Author: Egbert Eich <eich@suse.de>
57676 Date: Thu Jan 29 08:08:35 2004 +0000
57678 Importing vendor version xf86-012804-2330 on Thu Jan 29 00:06:33 PST 2004
57680 commit 64d7bbf0ac060a57761aa2ecf21ddbe8169674c0
57681 Author: Kaleb Keithley <kaleb@freedesktop.org>
57682 Date: Fri Dec 19 20:55:10 2003 +0000
57684 XFree86 4.3.99.902 (RC 2)
57686 commit 418eaaee109acd3ea178b147dd29f90318d159f9
57687 Author: Kaleb Keithley <kaleb@freedesktop.org>
57688 Date: Thu Dec 4 22:03:17 2003 +0000
57690 XFree86 4.3.99.901 (RC 1)
57692 commit 72063c52b82faf80a0183f66d2fd5789def76d8f
57693 Author: Kaleb Keithley <kaleb@freedesktop.org>
57694 Date: Tue Nov 25 19:28:37 2003 +0000
57697 Bring the tree up to date for the Cygwin folks
57699 commit e04a1a8f862eb63b546fb74e73ee602fefb82e00
57700 Author: Kaleb Keithley <kaleb@freedesktop.org>
57701 Date: Fri Nov 14 16:48:55 2003 +0000
57705 commit 3ad0d9a73bc0a4f3edb858d5f3a7c36827b7a429
57706 Author: Kaleb Keithley <kaleb@freedesktop.org>
57707 Date: Fri Nov 14 16:48:55 2003 +0000