Zack Rusin [Thu, 3 Sep 2009 20:51:31 +0000 (16:51 -0400)]
st/xorg: add linear and radial gradient acceleration
Zack Rusin [Thu, 3 Sep 2009 19:33:31 +0000 (15:33 -0400)]
st/xorg: radial gradient shader
Zack Rusin [Thu, 3 Sep 2009 18:18:05 +0000 (14:18 -0400)]
st/xorg: linear gradient shader
Alex Deucher [Thu, 3 Sep 2009 19:22:16 +0000 (15:22 -0400)]
r600: make sure the active vertex shader bo is re-added to persistent list.
Brian Paul [Thu, 3 Sep 2009 19:02:41 +0000 (13:02 -0600)]
glapi: fix incorrect/missing return types for glFenceSync(), glClientWaitSync()
Brian Paul [Thu, 3 Sep 2009 18:47:25 +0000 (12:47 -0600)]
mesa: rename gl_sync_object::Status to StatusFlag
There's a symbol collision with X11/Xlib.h #define Status int
in the Mesa xlib code. This seems the simpliest way to work around this.
Ian Romanick [Thu, 3 Sep 2009 18:22:05 +0000 (11:22 -0700)]
docs: Document new extension support for 7.6 release.
Eric Anholt [Tue, 1 Sep 2009 21:26:22 +0000 (14:26 -0700)]
Add ARB_sync to the xorg sw dri driver.
Eric Anholt [Tue, 1 Sep 2009 19:37:51 +0000 (12:37 -0700)]
savage: Fix driver build post-ARB_sync.
Like s3v, clean up absurd use of Xlib in the driver, avoiding namespace
pollution.
Eric Anholt [Tue, 1 Sep 2009 19:31:24 +0000 (12:31 -0700)]
s3v: Fix driver build for ARB_sync.
This driver was including Xlibint.h to get the CARD32 typedef to use for
32-bit unsigned integers, which #defined Status to something. CARD32 isn't
actually a 32-bit unsigned integer, so replacing with uint32_t usage should
fix bugs on 64-bit along with the build.
Eric Anholt [Tue, 1 Sep 2009 19:20:10 +0000 (12:20 -0700)]
intel: Add support for ARB_sync.
We currently weasel out of supporting the timeout parameter, but otherwise
this extension looks ready, and should make the common case happy.
Ian Romanick [Tue, 1 Sep 2009 19:15:51 +0000 (12:15 -0700)]
ARB sync / swrast: Use GL_ARB_sync_functions instead of GL_ARB_sync. Oops.
Ian Romanick [Tue, 1 Sep 2009 18:58:36 +0000 (11:58 -0700)]
Don't dereference function pointers in calls.
I'm apparently alone in prefering this calling convention, so I'll be
a team player. :p Based on review comments by Brian Paul and Eric
Anholt.
Ian Romanick [Tue, 1 Sep 2009 18:57:40 +0000 (11:57 -0700)]
Add syncobj.c to SConscript
Ian Romanick [Tue, 1 Sep 2009 18:56:34 +0000 (11:56 -0700)]
Whitespace and include file clean-up.
Based on review comments by Brian Paul.
Ian Romanick [Tue, 1 Sep 2009 18:24:54 +0000 (11:24 -0700)]
Use MIN2 instead of open-coded version
Based on review comments by Brian Paul.
Ian Romanick [Mon, 31 Aug 2009 21:57:50 +0000 (14:57 -0700)]
ARB sync: Fix delete behavior and context destruction behavior
I believe this resolves the outstanding issues WRT sync object
deletetion. I have also added a large comment at the top of syncobj.c
describing the expected memory management behavior. I'm still a
little uncertain about the locking on ctx->Shared.
Ian Romanick [Mon, 31 Aug 2009 21:49:33 +0000 (14:49 -0700)]
Add dummy list node type 'struct simple_node'
Ian Romanick [Sat, 29 Aug 2009 03:10:05 +0000 (20:10 -0700)]
ARB sync: Add support for GL_ARB_sync to swrast
This isn't quite right yet. The delete behavior and the context
clean-up needs some work.
Ian Romanick [Fri, 28 Aug 2009 22:53:00 +0000 (15:53 -0700)]
Eliminate trailing whitespace in extension_helper.c
Ian Romanick [Fri, 28 Aug 2009 22:50:28 +0000 (15:50 -0700)]
ARB sync: Regenerate files from previous commit
Ian Romanick [Fri, 28 Aug 2009 22:49:27 +0000 (15:49 -0700)]
ARB sync: Add base infrastructure for GL_ARB_sync
Ian Romanick [Fri, 28 Aug 2009 22:37:25 +0000 (15:37 -0700)]
ARB sync: Add infrastructure for glGetInteger64v
Brian Paul [Thu, 3 Sep 2009 18:10:53 +0000 (12:10 -0600)]
mesa: fix typos in separable filter functions
Brian Paul [Wed, 2 Sep 2009 21:39:13 +0000 (15:39 -0600)]
mesa: compute tex env program key size based on nr_enabled_units
When no textures are enabled, the key size is 4 bytes. If unit[0] is
enabled, the key size is 16 bytes.
Inspired by a patch from Chris Wilson.
Brian Paul [Thu, 3 Sep 2009 17:46:28 +0000 (11:46 -0600)]
swrast: wrap comment to 80 columns, remove dead code
Brian Paul [Thu, 3 Sep 2009 17:41:29 +0000 (11:41 -0600)]
mesa: use new combined PBO validate/map helpers in convolution/filter funcs
Brian Paul [Thu, 3 Sep 2009 17:23:05 +0000 (11:23 -0600)]
mesa: use new combined PBO validate/map helpers
Brian Paul [Thu, 3 Sep 2009 17:22:27 +0000 (11:22 -0600)]
mesa: combined PBO validate/map helpers
Brian Paul [Thu, 3 Sep 2009 16:41:14 +0000 (10:41 -0600)]
mesa: use new _mesa_map_pbo_source/dest() functions in more places
This trims down the code a bit. The next step would be to combine
the validate and map operations into one helper...
Brian Paul [Thu, 3 Sep 2009 17:29:18 +0000 (11:29 -0600)]
mesa: consolidate PBO map/unmap helpers
Instead of _mesa_map_readpix_pbo() use _mesa_map_pbo_source().
Instead of _mesa_map_drawpix_pbo() and _mesa_map_bitmap_pbo() use
_mesa_map_pbo_dest().
Brian Paul [Thu, 3 Sep 2009 16:36:49 +0000 (10:36 -0600)]
intel: helper to debug bufmgr (disabled)
Jakob Bornecrantz [Thu, 3 Sep 2009 18:30:54 +0000 (19:30 +0100)]
i915g: Up the number of relocs
Jakob Bornecrantz [Thu, 3 Sep 2009 18:30:36 +0000 (19:30 +0100)]
i915g: Don't run in sync with HW
Jakob Bornecrantz [Thu, 3 Sep 2009 18:28:38 +0000 (19:28 +0100)]
i915g: Don't send flush flags to HW
Zack Rusin [Thu, 3 Sep 2009 16:44:15 +0000 (12:44 -0400)]
st/xorg: cleanup and adjust to the changes in the ureg code
Zack Rusin [Thu, 3 Sep 2009 16:38:10 +0000 (12:38 -0400)]
st/xorg: adjust enums in preperation for gradients
Brian Paul [Thu, 3 Sep 2009 15:41:41 +0000 (09:41 -0600)]
mesa: change ctx->Driver.BufferData() to return GLboolean for success/failure
Return GL_FALSE if we failed to allocate the buffer. Then raise
GL_OUT_OF_MEMORY in core Mesa.
Keith Whitwell [Thu, 3 Sep 2009 14:16:25 +0000 (15:16 +0100)]
aux/tgsi: pull back ureg work from 0.1 branch
Manual merge of ureg changes on the branch. Too much unrelated stuff
for a proper merge.
Christoph Bumiller [Thu, 3 Sep 2009 10:54:41 +0000 (12:54 +0200)]
nv50: move centroid, flat bits when making interp long
Before this, just the perspective divide bit was moved in
convert_to_long of the load interpolant instruction.
Christoph Bumiller [Thu, 3 Sep 2009 10:39:46 +0000 (12:39 +0200)]
nv50: fix vbo_static_attrib method size error
The VTX_ATTR_3/2/1F methods also had size 4 ...
a stupid copy/paste error.
Christoph Bumiller [Thu, 3 Sep 2009 10:37:05 +0000 (12:37 +0200)]
nv50: fix draw_elements_inline_u08/16
We've been emitting the same two indices over and over
without incrementing map.
Cooper Yuan [Thu, 3 Sep 2009 09:06:18 +0000 (17:06 +0800)]
r300g: specify point/line/triangle have stuffed texture coord
Dave Airlie [Thu, 3 Sep 2009 04:14:47 +0000 (14:14 +1000)]
r600: visual depth has no meaning here.
fbos get angry when this happens.
Dave Airlie [Thu, 3 Sep 2009 04:14:31 +0000 (14:14 +1000)]
r600: make sure the active shader bo is re-added to persistent list.
Dave Airlie [Thu, 3 Sep 2009 02:01:52 +0000 (12:01 +1000)]
radeon: pass internal format into the miptree.
We need to figure out if the compression format changes.
without this texcmp segfaults if you change format
enough times.
Dave Airlie [Thu, 3 Sep 2009 02:01:27 +0000 (12:01 +1000)]
radeon/dri2: add gl20 bits for r300/r600 just like dri1 does
Dave Airlie [Thu, 3 Sep 2009 01:33:53 +0000 (11:33 +1000)]
tests/texcmp: glewInit in wrong place
Zack Rusin [Wed, 2 Sep 2009 22:39:21 +0000 (18:39 -0400)]
st/xorg: fix warnings
Zack Rusin [Wed, 2 Sep 2009 22:38:23 +0000 (18:38 -0400)]
st/xorg: handle xrender repeat modes
Zack Rusin [Wed, 2 Sep 2009 21:21:28 +0000 (17:21 -0400)]
st/xorg: flip the coordinate system
Zack Rusin [Wed, 2 Sep 2009 21:16:03 +0000 (17:16 -0400)]
st/xorg: handle solid fills in the fragment shader
Eric Anholt [Wed, 2 Sep 2009 19:17:28 +0000 (12:17 -0700)]
Revert "i965: Use VBOs in the VBO module on 965, now that we have ARB_map_buffer_range."
This reverts commit
00413d87426f14df47d90ba3c995e1889e9f88ca. Even with
fixes, using ARB_map_buffer_range in the VBO module isn't showing up as a
significant win, and some cases apparently regressed.
Bug #23624.
Eric Anholt [Wed, 2 Sep 2009 19:07:37 +0000 (12:07 -0700)]
intel: Add support for FlushMappedBufferRange for ARB_map_buffer_range.
This should help for the usage by the VBO module, where we would upload
the whole remaining chunk of the buffer for a series of range maps that should
cover just a segment of it.
Eric Anholt [Wed, 2 Sep 2009 18:33:08 +0000 (11:33 -0700)]
intel: Sync a synchronized READ_BIT map buffer range with GL drawing to it.
It's probably uncommon, but would obviously have gone wrong.
Eric Anholt [Wed, 2 Sep 2009 05:07:37 +0000 (22:07 -0700)]
intel: Move MapBufferRange mesa state setting up to cover the 915 case.
Jakob Bornecrantz [Wed, 2 Sep 2009 17:48:49 +0000 (19:48 +0200)]
st/xorg: Guard against realy old versions of Xorg
Brian Paul [Wed, 2 Sep 2009 16:38:46 +0000 (10:38 -0600)]
Revert "mesa: fix the non-GNU C bit-field case"
This reverts commit
4b08e7498230eac30eea1721f33994b30999acd4.
Don't know what I was thinking there.
Brian Paul [Wed, 2 Sep 2009 16:12:15 +0000 (10:12 -0600)]
mesa: fix the non-GNU C bit-field case
Gary Wong [Wed, 2 Sep 2009 16:11:15 +0000 (10:11 -0600)]
mesa: silence gcc bit-field warning
Brian Paul [Wed, 2 Sep 2009 15:43:00 +0000 (09:43 -0600)]
mesa: replace 8 with NUM_UNITS
Brian Paul [Wed, 2 Sep 2009 15:25:28 +0000 (09:25 -0600)]
mesa: remove accidentally committed printf
Brian Paul [Wed, 2 Sep 2009 14:45:34 +0000 (08:45 -0600)]
mesa: added #ifdef __GNUC__ around GLubyte bitfield usage
It would be nice if there were a #pragma or something to disable the warnings:
main/texenvprogram.c:87: warning: type of bit-field ‘Source’ is a GCC extension
main/texenvprogram.c:88: warning: type of bit-field ‘Operand’ is a GCC extension
but there doesn't appear to be a way to do that.
Chris Wilson [Wed, 2 Sep 2009 12:11:25 +0000 (05:11 -0700)]
mesa: Compact state key for TexEnv program cache
By rearranging the bitfields within the key we can reduce the size
of the key from 644 to 196 bytes, reducing the cost of both the
hashing and equality tests.
Michel Dänzer [Wed, 2 Sep 2009 14:01:21 +0000 (16:01 +0200)]
st/xorg: Remove code no longer support by xserver Git
Shouldn't be necessary with older versions either.
Michel Dänzer [Wed, 2 Sep 2009 14:00:30 +0000 (16:00 +0200)]
st/xorg: Add support for xserver Git master DRI2 interface.
Reorganize to handle both interfaces with common code.
Michel Dänzer [Wed, 2 Sep 2009 13:41:41 +0000 (15:41 +0200)]
st/dri: Slight consolidation of DRI2 buffer handling.
Always use / assume a combined depth/stencil buffer.
Also update drawable formats from surfaces returned by the driver.
Michel Dänzer [Wed, 2 Sep 2009 13:41:40 +0000 (15:41 +0200)]
st/xorg: Add EXA UploadToScreen and DownloadFromScreen hooks.
Xiang, Haihao [Wed, 2 Sep 2009 05:27:25 +0000 (13:27 +0800)]
i965: CS FENCE in URB_FENCE is 11-bits wide
Xiang, Haihao [Thu, 20 Aug 2009 10:19:36 +0000 (18:19 +0800)]
i965: validate sf state
Eric Anholt [Mon, 31 Aug 2009 17:13:22 +0000 (10:13 -0700)]
mesa: Make MultiDrawElements submit multiple primitives at once.
Previously, MultiDrawElements just called DrawElements a bunch of times.
By sending several primitives down the pipeline at once, we avoid a bunch
of validation. On my GL demo, this improves fps by 2.5% (+/- .41%) and
reduces CPU usage by 70.5% (+/- 2.9%) (n=3).
Reviewed by: Ian Romanick <ian.d.romanick@intel.com>
Ben Skeggs [Wed, 2 Sep 2009 02:03:42 +0000 (12:03 +1000)]
nv50: SWZ is the same as MOV from our perspective
Brian Paul [Tue, 1 Sep 2009 23:40:07 +0000 (17:40 -0600)]
mesa: skip bitmap drawing code if width==0 or height==0
Brian Paul [Tue, 1 Sep 2009 23:12:47 +0000 (17:12 -0600)]
intel: use _mesa_expand_bitmap() to skip an intermediate buffer
Brian Paul [Tue, 1 Sep 2009 23:11:59 +0000 (17:11 -0600)]
st/mesa: use new _mesa_expand_bitmap() function
Brian Paul [Tue, 1 Sep 2009 23:11:35 +0000 (17:11 -0600)]
mesa: new _mesa_expand_bitmap() function
Brian Paul [Tue, 1 Sep 2009 22:26:03 +0000 (16:26 -0600)]
mesa: remove redundant assignments
Brian Paul [Tue, 1 Sep 2009 22:22:02 +0000 (16:22 -0600)]
mesa: more clean-ups
Brian Paul [Tue, 1 Sep 2009 22:10:57 +0000 (16:10 -0600)]
mesa: change conditional to match the previous one
Brian Paul [Tue, 1 Sep 2009 22:05:07 +0000 (16:05 -0600)]
mesa: updated #includes
Brian Paul [Tue, 1 Sep 2009 22:03:28 +0000 (16:03 -0600)]
mesa: remove unused texenv_fragment_program::ctx field
Brian Paul [Tue, 1 Sep 2009 22:01:12 +0000 (16:01 -0600)]
mesa: remove unused ureg::abs field
Brian Paul [Tue, 1 Sep 2009 21:58:49 +0000 (15:58 -0600)]
mesa: remove unused ureg:negateabs field
Brian Paul [Tue, 1 Sep 2009 21:57:36 +0000 (15:57 -0600)]
mesa: more comments, clean-ups
Brian Paul [Tue, 1 Sep 2009 21:51:23 +0000 (15:51 -0600)]
mesa: simplify translate_tex_src_bit()
Brian Paul [Tue, 1 Sep 2009 21:49:55 +0000 (15:49 -0600)]
mesa: minor code clean-ups, comments
Brian Paul [Tue, 1 Sep 2009 21:34:16 +0000 (15:34 -0600)]
mesa: replace 8 with MAX_TEXTURE_UNITS
Brian Paul [Tue, 1 Sep 2009 18:43:23 +0000 (12:43 -0600)]
dri: remove unused meta_clear_tris()
Brian Paul [Tue, 1 Sep 2009 18:41:32 +0000 (12:41 -0600)]
intel: use BUFFER_BITS_COLOR
Brian Paul [Tue, 1 Sep 2009 18:39:30 +0000 (12:39 -0600)]
intel: fix incorrect parameter type for intel_bufferobj_map_range()
Brian Paul [Tue, 1 Sep 2009 18:32:01 +0000 (12:32 -0600)]
radeon: trim down #includes
Brian Paul [Tue, 1 Sep 2009 18:28:48 +0000 (12:28 -0600)]
radeon: use _mesa_meta_clear()
Brian Paul [Tue, 1 Sep 2009 18:24:30 +0000 (12:24 -0600)]
mesa: obey stencil write mask in _mesa_meta_draw_pixels()
Brian Paul [Tue, 1 Sep 2009 18:23:52 +0000 (12:23 -0600)]
intel: set Length/Offset fields in intel_bufferobj_map()
Brian Paul [Tue, 1 Sep 2009 18:23:01 +0000 (12:23 -0600)]
intel: use _mesa_meta_copy_pixels() when do_blit_copypixels() fails
Also, trim down #includes.
Brian Paul [Tue, 1 Sep 2009 18:18:55 +0000 (12:18 -0600)]
intel: trim down #includes
Brian Paul [Tue, 1 Sep 2009 18:16:13 +0000 (12:16 -0600)]
intel: use _mesa_meta_draw_pixels()
The textured quad path is slightly faster and will work with POT textures
on i945.
Brian Paul [Tue, 1 Sep 2009 18:07:52 +0000 (12:07 -0600)]
intel: trim down #includes
Brian Paul [Tue, 1 Sep 2009 18:05:41 +0000 (12:05 -0600)]
intel: use _mesa_meta_clear(), it's a bit faster
Pauli Nieminen [Tue, 1 Sep 2009 23:03:13 +0000 (02:03 +0300)]
radeon: Fix OQ to set ful lstate as dirty too.