profile/extras/intel-gpu-tools.git
10 years agotestdisplay: Respect specified_disp_id in the stereo loop
Damien Lespiau [Mon, 30 Sep 2013 16:02:21 +0000 (17:02 +0100)]
testdisplay: Respect specified_disp_id in the stereo loop

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Remove printing out the eyes geometry
Damien Lespiau [Mon, 30 Sep 2013 15:56:29 +0000 (16:56 +0100)]
testdisplay: Remove printing out the eyes geometry

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agolib: Adjust dump_modes() to the new stereo encoding
Damien Lespiau [Thu, 26 Sep 2013 17:34:06 +0000 (18:34 +0100)]
lib: Adjust dump_modes() to the new stereo encoding

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Use DRM_MODE_FLAG_3D_MASK
Damien Lespiau [Thu, 26 Sep 2013 16:56:01 +0000 (17:56 +0100)]
testdisplay: Use DRM_MODE_FLAG_3D_MASK

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Remove the timing adjustements now that the kernel does it
Damien Lespiau [Tue, 17 Sep 2013 17:59:52 +0000 (18:59 +0100)]
testdisplay: Remove the timing adjustements now that the kernel does it

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agointel_infoframe: Display the VIC in decimal
Damien Lespiau [Mon, 16 Sep 2013 16:52:41 +0000 (17:52 +0100)]
intel_infoframe: Display the VIC in decimal

It's not customary to display the VIC in hexadecimal and lead me to
scratch my head for a couple of seconds. Print it in decimal instead.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Test the stereo 3D modes
Damien Lespiau [Mon, 10 Sep 2012 12:33:26 +0000 (13:33 +0100)]
testdisplay: Test the stereo 3D modes

Now that modes have flags to describe which 3d formats the sink
supports, it's time to test them.

The new test cycles through the supported 3D formats and paint 3D
stereoscopic images taken from publicly available samples:
  http://www.quantumdata.com/apps/3D/sample_BMP.asp

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Provide a full path when opening pngs
Damien Lespiau [Fri, 6 Sep 2013 10:44:41 +0000 (11:44 +0100)]
testdisplay: Provide a full path when opening pngs

This way one doesn't have to be in tests/ for testsdisplay to be able to
open pass.png.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Free the array of connectors
Damien Lespiau [Wed, 4 Sep 2013 13:21:55 +0000 (14:21 +0100)]
testdisplay: Free the array of connectors

That's an array we allocated earlier in this function. Let's be symetric
and free it once done.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Untangle dump_info() from the main testing loop
Damien Lespiau [Thu, 5 Sep 2013 15:49:11 +0000 (16:49 +0100)]
testdisplay: Untangle dump_info() from the main testing loop

-i is just supposed to show some information about the DRM resources.
Right now it works in a quite convoluted way. Untangle this to call
dump_info() when -i is given, exit the program and be done with it.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Fix CRTS typo
Damien Lespiau [Wed, 4 Sep 2013 11:31:18 +0000 (12:31 +0100)]
testdisplay: Fix CRTS typo

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Properly handle the life cycle of framebuffers
Damien Lespiau [Tue, 3 Sep 2013 18:46:19 +0000 (19:46 +0100)]
testdisplay: Properly handle the life cycle of framebuffers

When cycling through the modes, let's make sure to free the previous
framebuffers. This is the perfect occasion to use kmstest_remove_fb().

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Map the fb inside paint_color_key()
Damien Lespiau [Wed, 4 Sep 2013 11:12:37 +0000 (12:12 +0100)]
testdisplay: Map the fb inside paint_color_key()

So the code for this is self-contained. This goes along the way of
reducing the number of global variables in testdisplay.

Take the opportunity to unmap the fb after use as well.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Move the code sanitizing depth into main()
Damien Lespiau [Tue, 3 Sep 2013 13:57:31 +0000 (14:57 +0100)]
testdisplay: Move the code sanitizing depth into main()

It'll be shared by the set_mode() and set_3d_mode() functions.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agolib: Add a helper to write a png from a struct kmstest_fb
Damien Lespiau [Fri, 6 Sep 2013 17:22:53 +0000 (18:22 +0100)]
lib: Add a helper to write a png from a struct kmstest_fb

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agolib: Split create_image_surface() out of create_cairo_ctx()
Damien Lespiau [Fri, 6 Sep 2013 16:26:32 +0000 (17:26 +0100)]
lib: Split create_image_surface() out of create_cairo_ctx()

So we can use it in the next commit.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agolib: Add a helper to paint a PNG using cairo
Damien Lespiau [Fri, 23 Aug 2013 15:46:45 +0000 (16:46 +0100)]
lib: Add a helper to paint a PNG using cairo

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agolib: Dump information about the supported 3D stereo formats
Damien Lespiau [Wed, 12 Sep 2012 13:21:29 +0000 (14:21 +0100)]
lib: Dump information about the supported 3D stereo formats

When dumping the details of a mode, let's add the 3D formats the mode
supports.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Add left/right images of a lovely scene
Damien Lespiau [Fri, 6 Sep 2013 14:17:48 +0000 (15:17 +0100)]
testdisplay: Add left/right images of a lovely scene

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agotestdisplay: Distribute pass.png
Damien Lespiau [Fri, 6 Sep 2013 14:20:35 +0000 (15:20 +0100)]
testdisplay: Distribute pass.png

10 years agotests/gem_evict_everything: tune down forked subtests
Daniel Vetter [Mon, 30 Sep 2013 14:04:38 +0000 (16:04 +0200)]
tests/gem_evict_everything: tune down forked subtests

On new machines with gobloads of memory and cpu cores, but slow swap
on spinning rust we need to limit the runtime a bit for sanity.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests: use igt_assert/igt_require more
Daniel Vetter [Wed, 25 Sep 2013 12:36:59 +0000 (14:36 +0200)]
tests: use igt_assert/igt_require more

With the new _f variants we can replace almost all of them.

Also remove a ton of checks for argc != 1, they're a bit useless ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agointel_l3_parity: Make compilation possible without udev
Ben Widawsky [Mon, 23 Sep 2013 20:46:06 +0000 (13:46 -0700)]
intel_l3_parity: Make compilation possible without udev

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agotests/pc8: fix supports_pc8_plus_residencies
Paulo Zanoni [Mon, 23 Sep 2013 19:05:21 +0000 (16:05 -0300)]
tests/pc8: fix supports_pc8_plus_residencies

Bug caused by bad copy+paste+replace.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
10 years agooverlay: Some very raw usage information
Chris Wilson [Sun, 22 Sep 2013 18:00:59 +0000 (19:00 +0100)]
overlay: Some very raw usage information

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10 years agoNEWS: Start with release notes for 1.5
Daniel Vetter [Sun, 22 Sep 2013 11:15:43 +0000 (13:15 +0200)]
NEWS: Start with release notes for 1.5

Also extend the underlying for 1.4 to the full lenght ;-)

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agointel_l3_parity: Support a daemonic mode
Ben Widawsky [Tue, 10 Sep 2013 21:21:23 +0000 (14:21 -0700)]
intel_l3_parity: Support a daemonic mode

v2: Add a comment explaining the dangers of directly accessing the DFT
register (Daniel)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: Support error injection
Ben Widawsky [Tue, 10 Sep 2013 17:40:54 +0000 (10:40 -0700)]
intel_l3_parity: Support error injection

Haswell added the ability to inject errors which is extremely useful for
testing. Add two arguments to the tool to inject, and uninject.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: Actually support multiple slices
Ben Widawsky [Sun, 8 Sep 2013 03:38:18 +0000 (20:38 -0700)]
intel_l3_parity: Actually support multiple slices

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: slice support
Ben Widawsky [Sun, 8 Sep 2013 03:10:27 +0000 (20:10 -0700)]
intel_l3_parity: slice support

Haswell GT3 adds a new slice which is kept distinct from the old
register interface. Plumb it into the code, though it's only 1 slice
still.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: Hardware info argument
Ben Widawsky [Sun, 8 Sep 2013 23:50:23 +0000 (16:50 -0700)]
intel_l3_parity: Hardware info argument

Add a new command line argument to the tool which will spit out various
parameters for the giving hardware. As a result of this, some new
defines are added to help with the various info.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: Use getopt for the l3 parity tool
Ben Widawsky [Sat, 7 Sep 2013 04:10:50 +0000 (21:10 -0700)]
intel_l3_parity: Use getopt for the l3 parity tool

Add new command line arguments in addition to supporting the old
features. This patch only introduces one feature, the -e argument to
enable a specific row/bank/subbank. Previously you could only enable
all. Otherwise, it has what you expect (we prefer -r -b -s for
specifying the row/bank/subbank).

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: Assert all GEN7+ support
Ben Widawsky [Fri, 6 Sep 2013 18:40:03 +0000 (11:40 -0700)]
intel_l3_parity: Assert all GEN7+ support

v2: Don't assert for Valleyview (Bryan)
Rework code to be a bit more readable.

CC: "Bell, Bryan J" <bryan.j.bell@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_l3_parity: Fix indentation
Ben Widawsky [Tue, 10 Sep 2013 22:34:54 +0000 (15:34 -0700)]
intel_l3_parity: Fix indentation

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agolib/drmtest: Improve printf-like igt_skip_on/require
Daniel Vetter [Thu, 19 Sep 2013 17:56:03 +0000 (19:56 +0200)]
lib/drmtest: Improve printf-like igt_skip_on/require

Ben Widawsky suggested to use vasprintf, which perfectly fits the bill.

Also fix the logic conversion bug in tests/gem_storedw_batches_loop that
crept in again :(

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: igt_assert|require with format strings
Daniel Vetter [Thu, 19 Sep 2013 14:37:07 +0000 (16:37 +0200)]
lib/drmtest: igt_assert|require with format strings

v2: Add a comment about the pitfalls around va_list handling.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_stored_batches_loop: use igt_assert more
Daniel Vetter [Thu, 19 Sep 2013 13:51:33 +0000 (15:51 +0200)]
tests/gem_stored_batches_loop: use igt_assert more

Terser code ftw!

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_overflow: New subtest for overflowing buffer_count
Daniel Vetter [Thu, 19 Sep 2013 09:01:14 +0000 (11:01 +0200)]
tests/gem_reloc_overflow: New subtest for overflowing buffer_count

Luckily everything seems to be fine.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_overflow: Extract reloc_tests
Daniel Vetter [Thu, 19 Sep 2013 08:20:46 +0000 (10:20 +0200)]
tests/gem_reloc_overflow: Extract reloc_tests

I'll be adding more stuff soon ;-)

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_storedw_batches_loop: Add testcase to check secure dispatch
Daniel Vetter [Wed, 18 Sep 2013 16:28:18 +0000 (18:28 +0200)]
tests/gem_storedw_batches_loop: Add testcase to check secure dispatch

v2: Use the mrb_exec function since otherwise we can't pass flags.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_persistent_relocs: Tune down the test a bit
Daniel Vetter [Mon, 16 Sep 2013 20:45:14 +0000 (22:45 +0200)]
tests/gem_persistent_relocs: Tune down the test a bit

Takes too long.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_fenced_exec_thrash: Add interruptible and busy-load subtests
Daniel Vetter [Thu, 12 Sep 2013 14:18:10 +0000 (16:18 +0200)]
tests/gem_fenced_exec_thrash: Add interruptible and busy-load subtests

Usual progression for gem tests up the nasty-scale.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: check that igt_exit is called for subtest tests
Daniel Vetter [Thu, 12 Sep 2013 14:17:33 +0000 (16:17 +0200)]
lib/drmtest: check that igt_exit is called for subtest tests

I get this wrong for almost every conversion to subtests ...

v2: Don't install the check when just listing subtest names.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_fenced_exec_thrash: New subtests that uses all fences
Daniel Vetter [Thu, 12 Sep 2013 12:38:13 +0000 (14:38 +0200)]
tests/gem_fenced_exec_thrash: New subtests that uses all fences

igt runs under the assumption that nothing else (besides maybe fbcon)
is using the gpu. So we should be able to use all fences.

Keep the conservative testcase around though in case someone has a
broken setup.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/*: scrap get_num_fences
Daniel Vetter [Thu, 12 Sep 2013 12:21:56 +0000 (14:21 +0200)]
tests/*: scrap get_num_fences

Just use gem_available_fences directly.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agointel_audio_dump/hsw: rename some audio configuration registers for Haswell
Mengdong Lin [Mon, 9 Sep 2013 19:38:40 +0000 (15:38 -0400)]
intel_audio_dump/hsw: rename some audio configuration registers for Haswell

For Haswell, some audio configuration registers have changed their name and
some bit definitions.

This patch applies the changes, and uses subfunctions to parse registers for
code reuse.

Here is the name change list:
Audio configuration: AUD_CONFIG_x to AUD_TCx_CONFIG
Audio Misc Control: AUD_MISC_CTRL_x to AUD_Cn_MISC_CTRL
Audio M & CTS programming enable: AUD_CTS_ENABLE_x to AUD_TCx_M_CTS_ENABLE
Audio EDID data block: AUD_HDMIW_HDMIEDID_x to AUD_TCx_EDID_DATA
Audio Widget Data Island Packet: AUD_HDMIW_INFOFR_x to AUD_TCx_AUD_INFOFR
Audio Pipe and Converter Configs: AUD_PORT_EN_HD_CFG to AUD_PIPE_CONV_CFG
Audio Digital Converter: AUD_OUT_DIG_CNVT_x to AUD_Cn_DIG_CNVT
Audio Stream Descriptor Format: AUD_OUT_STR_DESC_x to AUD_Cn_STR_DESC
Audio Connect List Entry & Length:  AUD_PINW_CONNLNG_LIST_x to
                                        AUD_TCx_PIN_PIPE_CONN_ENTRY_LNGTH
Audio Connection Select Control: AUD_PINW_CONNLNG_SEL to AUD_PIPE_CONN_SEL_CTRL
Audio DIP & ELD Control State: AUD_DIP_ELD_CTRL_ST_x to AUD_TCx_DIP_ELD_CTRL_ST
Audio HDMI FIFO status: AUD_HDMIW_STATUS to AUD_HDMI_FIFO_STATUS

NOTE:
For Tx, x = A/B/C, meaning Transcoder A/B/C.
For Cn, n = 1/2/3, meaning audio converter 1/2/3.

Signed-off-by: Mengdong Lin <mengdong.lin@intel.com>
Reviewed-by: Haihao Xiang <haihao.xiang@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
10 years agointel_audio_dump/hsw: align code with tab
Mengdong Lin [Mon, 9 Sep 2013 19:38:32 +0000 (15:38 -0400)]
intel_audio_dump/hsw: align code with tab

This patch makes the file to follow kernel coding style:
- replace leading spaces with tabs for alignment
- fix some minor format issues

But the max length of a line is set to 120 characters for readability
on high resolution displays.

Signed-off-by: Mengdong Lin <mengdong.lin@intel.com>
Reviewed-by: Haihao Xiang <haihao.xiang@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
10 years agointel_audio_dump/hsw: remove misuse of PCH transcoder configuration register
Mengdong Lin [Mon, 9 Sep 2013 19:38:20 +0000 (15:38 -0400)]
intel_audio_dump/hsw: remove misuse of PCH transcoder configuration register

The PCH transcoder config register (PCH_TRANS_CONF, 0xf0008) is not the
correct config register for transcoder A, B or C. This register is in
PCH and for CRT display, nothing to do with display audio.

So This patch removes misuse of it as config register for transcoder A/B/C.

Signed-off-by: Mengdong Lin <mengdong.lin@intel.com>
Reviewed-by: Haihao Xiang <haihao.xiang@intel.com>
Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
10 years agobump version to 1.4 and add the release date intel-gpu-tools-1.4
Rodrigo Vivi [Mon, 16 Sep 2013 19:37:08 +0000 (16:37 -0300)]
bump version to 1.4 and add the release date

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
10 years agoNEWS: Fix dates of previous releases
Rodrigo Vivi [Fri, 13 Sep 2013 19:54:56 +0000 (16:54 -0300)]
NEWS: Fix dates of previous releases

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
10 years agooverlay: fix link error due to missing -lrt
Rodrigo Vivi [Mon, 16 Sep 2013 18:34:27 +0000 (15:34 -0300)]
overlay: fix link error due to missing -lrt

CC: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
10 years agolib/drmtest: ducttape over fork race
Daniel Vetter [Fri, 13 Sep 2013 14:43:22 +0000 (16:43 +0200)]
lib/drmtest: ducttape over fork race

Whatever the reason (and I've thought there isn't one) if we fork and
kill right away the child seems to not reliably die. We can work
around this little race by forcing the default SIGQUIT handler. This
should break anything since we reset our atexit handling anyway, so if
the helper needs any atexit handling the special signal helpers will
be reinstated.

Note that inserting sufficient amounts of printf between the fork and
kill makes this unnecessary.

While add it also add the retry loop for the waitpid call, in case
there's another guy constantly interrupting us.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib: add test for igt_fork_signal_helper
Daniel Vetter [Fri, 13 Sep 2013 14:38:59 +0000 (16:38 +0200)]
lib: add test for igt_fork_signal_helper

If we're really fast we've trying to stop the signal helper again
we somehow race somewhere and it'll never happen. So add a testcase for
this. Since I expect more to come for testsuite tests add a separate
make target for them. Run tests with

$ make check

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agoNEWS: smelling fixes
Chris Wilson [Fri, 13 Sep 2013 15:28:15 +0000 (16:28 +0100)]
NEWS: smelling fixes

10 years agoinitial release notes for the next release
Daniel Vetter [Fri, 13 Sep 2013 15:02:07 +0000 (17:02 +0200)]
initial release notes for the next release

Please amend/improve.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agoreconstruct NEWS file for the releases thus far
Daniel Vetter [Fri, 13 Sep 2013 15:01:38 +0000 (17:01 +0200)]
reconstruct NEWS file for the releases thus far

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agooverlay: Fix stale mention to x11-position.c
Damien Lespiau [Fri, 13 Sep 2013 15:00:05 +0000 (16:00 +0100)]
overlay: Fix stale mention to x11-position.c

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agoman: Fixes typo in intel_lid.man
Eero Tamminen [Fri, 13 Sep 2013 14:32:43 +0000 (15:32 +0100)]
man: Fixes typo in intel_lid.man

Signed-off-by: Eero Tamminen <eero.t.tamminen@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
10 years agolib/drmtest: skip suspend tests in simulation
Daniel Vetter [Thu, 12 Sep 2013 12:00:09 +0000 (14:00 +0200)]
lib/drmtest: skip suspend tests in simulation

The simulator doesn't like this nor really support it :(

v2: We've tried to a more gentle resume testing using the pm_test
infrastructure, but that alos failed. So add a FIXME comment that we
need to improve things a bit here.

Cc: Ben Widawsky <benjamin.widawsky@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: drop unused oldsig variabel
Daniel Vetter [Thu, 12 Sep 2013 10:56:20 +0000 (12:56 +0200)]
lib/drmtest: drop unused oldsig variabel

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: Restore default sighandlers
Daniel Vetter [Wed, 11 Sep 2013 13:46:43 +0000 (15:46 +0200)]
lib/drmtest: Restore default sighandlers

Forked tests ended up restoring the sighandlers already inherited
from the parent, resulting in endless signal loops through our
atexit handler.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: don't frob signals in __igt_fork_helper
Daniel Vetter [Wed, 11 Sep 2013 13:05:20 +0000 (15:05 +0200)]
lib/drmtest: don't frob signals in __igt_fork_helper

We shut up the exit handlers already by clearing the array.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: clean up children in an exit handler
Daniel Vetter [Wed, 11 Sep 2013 12:29:00 +0000 (14:29 +0200)]
lib/drmtest: clean up children in an exit handler

Also be a bit more careful with killing them in general.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_persisten_relocs: Update like gem_reloc_vs_pug
Daniel Vetter [Wed, 11 Sep 2013 12:03:47 +0000 (14:03 +0200)]
tests/gem_persisten_relocs: Update like gem_reloc_vs_pug

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_vs_gpu: some fixes
Daniel Vetter [Wed, 11 Sep 2013 11:59:18 +0000 (13:59 +0200)]
tests/gem_reloc_vs_gpu: some fixes

- Clear the igt_helper_process struct to avoid hitting asserts.
- Fix up the logic for enumerating forked processes.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: handle SIGBUS in the exit handlers
Daniel Vetter [Wed, 11 Sep 2013 10:11:54 +0000 (12:11 +0200)]
lib/drmtest: handle SIGBUS in the exit handlers

Our kernel likes to occasionally kill process with a SIGBUS when this
shouldn't ever happen. Hence also handle this signal in the exit
handler infrastructure.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: Reject igt_fork from within igt_fork earlier
Daniel Vetter [Wed, 11 Sep 2013 09:51:40 +0000 (11:51 +0200)]
lib/drmtest: Reject igt_fork from within igt_fork earlier

We reject it in igt_waitchildren already, but earlier is better.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: Fix igt_stop_signal_helper for subtest listing
Daniel Vetter [Wed, 11 Sep 2013 09:09:08 +0000 (11:09 +0200)]
lib/drmtest: Fix igt_stop_signal_helper for subtest listing

We need to bail out early for otherwise we'll hit the
!signal_helper->running assert.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: skip when prefault control isn't available
Daniel Vetter [Wed, 11 Sep 2013 08:47:07 +0000 (10:47 +0200)]
lib/drmtest: skip when prefault control isn't available

Instead of crashing with an igt_assert. Fixes the only crashing test
when running igt on non-intel systems.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: Use ARRAY_SIZE
Daniel Vetter [Wed, 11 Sep 2013 08:40:25 +0000 (10:40 +0200)]
lib/drmtest: Use ARRAY_SIZE

Less fragile ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: double-check that we clean up helper processes
Daniel Vetter [Tue, 10 Sep 2013 15:41:49 +0000 (17:41 +0200)]
lib/drmtest: double-check that we clean up helper processes

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: consolidate the helper process killing in one exit handler
Daniel Vetter [Tue, 10 Sep 2013 13:44:37 +0000 (15:44 +0200)]
lib/drmtest: consolidate the helper process killing in one exit handler

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_vs_gpu: use igt_fork_helper
Daniel Vetter [Tue, 10 Sep 2013 13:29:17 +0000 (15:29 +0200)]
tests/gem_reloc_vs_gpu: use igt_fork_helper

Now we do a waitpid instead of a simple wait which could eat
the "wrong" child ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: create helpers for forking helper threads
Daniel Vetter [Tue, 10 Sep 2013 13:46:08 +0000 (15:46 +0200)]
lib/drmtest: create helpers for forking helper threads

The upshot is that we can share the logic to make sure the helpers
are all properly stoved again in a 2nd step.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: fixup for the prefault rework
Daniel Vetter [Sun, 8 Sep 2013 12:03:20 +0000 (14:03 +0200)]
lib/drmtest: fixup for the prefault rework

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: Avoid calling exit handlers multiple times
Daniel Vetter [Sun, 8 Sep 2013 12:03:03 +0000 (14:03 +0200)]
lib/drmtest: Avoid calling exit handlers multiple times

- reset the count when forking
- don't add the same handler multiple times
- don't restore the exit signal handlers in the forked helper
  process
- reset the exit handler count once called to make sure we don't call
  it multiple times when dying
- don't wait for the signal helper if it's gone already

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_overflow: Add more checks
Daniel Vetter [Wed, 4 Sep 2013 13:25:03 +0000 (15:25 +0200)]
tests/gem_reloc_overflow: Add more checks

For reloc offsets and batch start/len. Doesn't quite fit into the test
subject at hand here, but meh.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_overflow: convert to subtests
Daniel Vetter [Wed, 4 Sep 2013 12:43:06 +0000 (14:43 +0200)]
tests/gem_reloc_overflow: convert to subtests

More will come!

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_exec_bad_domains: enable conflicting write domains test
Daniel Vetter [Wed, 4 Sep 2013 12:27:17 +0000 (14:27 +0200)]
tests/gem_exec_bad_domains: enable conflicting write domains test

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib: fix the fix for gen5 workaround emmision
Imre Deak [Wed, 4 Sep 2013 14:02:19 +0000 (17:02 +0300)]
lib: fix the fix for gen5 workaround emmision

Fix the regression introduced in

commit bfbe813f8fb587017c4e1d73c51395c2837eb395
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Tue May 29 22:14:06 2012 +0200

    lib: fix gen5 workaround emission

Signed-off-by: Imre Deak <imre.deak@intel.com>
10 years agolib/intel_batchbuffer: remove code w/o effect
Imre Deak [Wed, 4 Sep 2013 14:06:00 +0000 (17:06 +0300)]
lib/intel_batchbuffer: remove code w/o effect

Introduced when refactoring the patch in

commit c1ee0bb53269ded7b79966d081518d689639bac7
Author: Imre Deak <imre.deak@intel.com>
Date:   Mon Jul 29 16:43:31 2013 +0300

    intel_batchbuffer: add support for non-32bit blt copies

No functional change.

Signed-off-by: Imre Deak <imre.deak@intel.com>
10 years agotests/gem_pipe_control_store_loop: Add subtest for reused buffers
Daniel Vetter [Wed, 4 Sep 2013 12:08:10 +0000 (14:08 +0200)]
tests/gem_pipe_control_store_loop: Add subtest for reused buffers

This exercises the slightly faulty kernel w/a that Eric fixed in

commit e844b990b1df9242bb91b7d490552f3198946838
Author: Eric Anholt <eric@anholt.net>
Date:   Tue Jul 31 15:35:01 2012 -0700

    drm/i915: Don't forget to apply SNB PIPE_CONTROL GTT workaround.

    If a buffer that was the target of a PIPE_CONTROL from userland was a
    reused one that hadn't been evicted which had not previously had this
    workaround applied, then the early return for a correct
    presumed_offset in this function meant we would not bind it into the
    GTT and the write would land somewhere else.

    Fixes reproducible failures with GL_EXT_timer_query usage in apitrace,
    and I also expect it to fix the intermittent OQ issues on snb that
    danvet's been working on.

    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48019
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52932
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Carl Worth <cworth@cworth.org>
Tested-by: Carl Worth <cworth@cworth.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests: add missing igt_exit() calls
Daniel Vetter [Wed, 4 Sep 2013 11:25:28 +0000 (13:25 +0200)]
tests: add missing igt_exit() calls

Forgotten while converting to subtests, then copy&pasted.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: add igt_skip_on macro
Daniel Vetter [Wed, 4 Sep 2013 11:19:42 +0000 (13:19 +0200)]
lib/drmtest: add igt_skip_on macro

I've fumbled the logic inversion when converting to igt_require way
too often, so lets add something for dummies like me ;-)

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests: Update .gitignore
Daniel Vetter [Wed, 4 Sep 2013 08:36:28 +0000 (10:36 +0200)]
tests: Update .gitignore

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests: add gem_persisten_relocs
Daniel Vetter [Tue, 3 Sep 2013 13:30:02 +0000 (15:30 +0200)]
tests: add gem_persisten_relocs

This reproduces the 3.7 relocation regression ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agointel_gtt: Raw PTE dumper mode
Ben Widawsky [Sun, 1 Sep 2013 19:34:36 +0000 (12:34 -0700)]
intel_gtt: Raw PTE dumper mode

./tools/intel_gtt -d | head
GTT offset |                 PTEs
--------------------------------------------------------
  0x000000 | 0xe4005015 0xe2854015 0xe283e015 0xe283f015
  0x004000 | 0xe28ba015 0xe28bb015 0xe28b6015 0xe28b7015
  0x008000 | 0xe2828015 0xe2829015 0xe282a015 0xe282b015
  0x00c000 | 0xe2928015 0xe2929015 0xe292a015 0xe292b015
  0x010000 | 0xe2918015 0xe2919015 0xe291a015 0xe291b015
  0x014000 | 0xe291c015 0xe291d015 0xe291e015 0xe291f015
  0x018000 | 0xe2920015 0xe2921015 0xe2922015 0xe2923015
  0x01c000 | 0xe2924015 0xe2925015 0xe2926015 0xe2927015

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_gtt: Properly support gen6+ GTT PTEs
Ben Widawsky [Sun, 1 Sep 2013 19:17:23 +0000 (12:17 -0700)]
intel_gtt: Properly support gen6+ GTT PTEs

This finishes the objective in the last patch which was to actually deal
with physical addresses, and not the PTEs.

GEN6+ Provided support for physical addresses above 4GB. I'm not
actually sure what Ironlake supported, and don't feel like firing up the
timemachine.

v2: Add support for gen4, gen5, and haswell.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_gtt: Use function to get the physical address
Ben Widawsky [Sun, 1 Sep 2013 18:49:12 +0000 (11:49 -0700)]
intel_gtt: Use function to get the physical address

The GTT PTEs that the tool is trying to compare is really about
addresses, and not the PTE itself. To accomplish this, make which
calculates the physical address we actually want.

This commit itself doesn't change any functionality; just the wording in
the code.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agogem_vmap_blits: Demote warning to note
Ben Widawsky [Sun, 1 Sep 2013 18:47:57 +0000 (11:47 -0700)]
gem_vmap_blits: Demote warning to note

The warning that vmap isn't supported is useful, but it shouldn't get in
the way of developers (or distros) being able to use -Werror.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agointel_reg_dumper: Silence GCC for uninitialized clock
Ben Widawsky [Sun, 1 Sep 2013 18:47:11 +0000 (11:47 -0700)]
intel_reg_dumper: Silence GCC for uninitialized clock

GCC 4.8.1 seems to think clock may be uninitialized.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
10 years agotests/gem_reloc_vs_gpu: add thrashing tests
Daniel Vetter [Tue, 3 Sep 2013 15:49:02 +0000 (17:49 +0200)]
tests/gem_reloc_vs_gpu: add thrashing tests

Using the i915_gem_drop_caches debugfs interface to thrash without
really thrashing.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_vs_gpu: add forked versions
Daniel Vetter [Tue, 3 Sep 2013 09:57:56 +0000 (11:57 +0200)]
tests/gem_reloc_vs_gpu: add forked versions

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: use igt_require/assert in the prefault helpers
Daniel Vetter [Tue, 3 Sep 2013 10:22:09 +0000 (12:22 +0200)]
lib/drmtest: use igt_require/assert in the prefault helpers

10 years agotests/gem_reloc_vs_gpu: Add faulting reloc tests
Daniel Vetter [Tue, 3 Sep 2013 08:59:13 +0000 (10:59 +0200)]
tests/gem_reloc_vs_gpu: Add faulting reloc tests

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: include sys/mman.h from drmtest.h
Daniel Vetter [Tue, 3 Sep 2013 08:38:29 +0000 (10:38 +0200)]
lib/drmtest: include sys/mman.h from drmtest.h

We need it for mmapping to get at PROT_READ|WRITE anyway.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agolib/drmtest: extract gem_execbuf helper
Daniel Vetter [Tue, 3 Sep 2013 08:37:14 +0000 (10:37 +0200)]
lib/drmtest: extract gem_execbuf helper

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agotests/gem_reloc_vs_gpu: add interruptible version
Daniel Vetter [Tue, 3 Sep 2013 06:57:42 +0000 (08:57 +0200)]
tests/gem_reloc_vs_gpu: add interruptible version

Exercise a bug where we've failed to propagate the error code
correctly.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
10 years agokms_flip: Fix use of fb_width for PAN subtests
Chris Wilson [Mon, 2 Sep 2013 12:50:56 +0000 (13:50 +0100)]
kms_flip: Fix use of fb_width for PAN subtests

We need to be careful to remember that fb-width is not always the same
as hdisplay, since for panning we allocate a larger framebuffer. So fix
up the printfs to use hdisplay/vsisplay since that should be uniform
across the array.

Regression from
commit 919d68901187fa797a9b648fcf87c838fae22fa3
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Aug 29 15:33:53 2013 +0100

    kms_flips: Operate on an array of crtc

Buzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68832
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
10 years agolib/drmtest: Improve output when igt_waitchildren fails
Daniel Vetter [Mon, 2 Sep 2013 06:08:29 +0000 (08:08 +0200)]
lib/drmtest: Improve output when igt_waitchildren fails

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>