Commit graph

42315 commits

Author SHA1 Message Date
Kenneth Graunke
e0c1fc3283 glsl/Makefile: glcpp doesn't need libglsl.a.
Also, add a 'glcpp' target so you can type 'make glcpp' instead of
'make glcpp/glcpp'.
2011-02-10 17:37:01 -08:00
Marek Olšák
032b162ce8 r300g: plug a memory leak in winsys 2011-02-11 02:34:15 +01:00
Marek Olšák
862ebb411b r300g: remove unneeded code in winsys
We don't need the read/write flags.
2011-02-11 01:32:44 +01:00
Marek Olšák
98f344c504 r300g: fix warning 2011-02-11 01:18:53 +01:00
Marek Olšák
7da5105fb3 configure.ac: remove libdrm_radeon dependency for r300g and r600g 2011-02-11 01:16:06 +01:00
Marek Olšák
6ccab620a0 r300g: import the last bits of libdrm and cleanup the whole thing
Based on Dave's branch.

The majority of this commit is a cleanup, mainly renaming things.
There wasn't much code to import, just ioctl calls.

Also done:
- implemented unsynchronized bo_map (important optimization!)
- radeon_bo_is_referenced_by_cs is no longer a refcount hack
- dropped the libdrm_radeon dependency

I'm surprised that this has resulted in less code in the end.
2011-02-11 01:07:25 +01:00
Marek Olšák
c0beaf6e6d st/mesa: allow rendering to sRGB textures if EXT_fb_srgb is unsupported
In this case, we always use the corresponding linear format in create_surface,
therefore we should check for linear format support as well.
2011-02-11 01:07:21 +01:00
Ian Romanick
4c1dc1c4d7 i915: Force lowering of all types of indirect array accesses in the FS
NOTE: This is a candidate for the 7.9 and 7.10 branches.
2011-02-10 13:26:49 -08:00
Ian Romanick
53b8b68843 i915: Calculate partial result to temp register first
Previously the SNE and SEQ instructions would calculate the partial
result to the destination register.  This would cause problems if the
destination register was also one of the source registers.

Fixes piglit tests glsl-fs-any, glsl-fs-struct-equal,
glsl-fs-struct-notequal, glsl-fs-vec4-operator-equal,
glsl-fs-vec4-operator-notequal.

NOTE: This is a candidate for the 7.9 and 7.10 branches.
2011-02-10 13:26:49 -08:00
José Fonseca
05a16b8e1c u_upload_mgr: Use PIPE_TRANSFER_FLUSH_EXPLICIT.
This can avoid DMAing the whole buffer when just a fraction was changed.
2011-02-10 20:55:10 +00:00
José Fonseca
d60f07489e scons: Disable -mstackrealign on MinGW again.
It's still broken, however it doesn't show up on debug builds.
2011-02-10 16:29:10 +00:00
Kristian Høgsberg
1a8899d630 configure.ac: Fix typo 2011-02-10 10:45:27 -05:00
Marek Olšák
fea4ad8f66 r300g: implement accelerated copy_region for compressed formats 2011-02-10 11:27:35 +01:00
Marek Olšák
7c24a4c6a8 r300g: add a way to change texture properties arbitrarily
So that we can implement resource_copy on arbitrary data.
2011-02-10 11:27:35 +01:00
Marek Olšák
56ba7e913f r300g: consolidate buffers and textures to r300_resource
Transfers and create/destroy are still handled separately.
2011-02-10 11:27:35 +01:00
Marek Olšák
ce9c0d2801 r300g: simplify WRITE_RELOC API and cleanup 2011-02-10 11:27:35 +01:00
Marek Olšák
ac366af9fd u_blitter: let the driver check whether there's a recursion 2011-02-10 11:27:34 +01:00
Marek Olšák
fc9170d0cf r300g: use format from pipe_surface instead of pipe_resource 2011-02-10 02:11:38 +01:00
Marek Olšák
2314a2f45f Revert "r300g: support sRGB colorbuffers"
This partially reverts commit 91eba2567e.

Conflicts:

	src/gallium/drivers/r300/r300_blit.c
2011-02-10 01:43:27 +01:00
Dave Airlie
21b0996dfc mesa/st: enable GL_EXT_framebuffer_sRGB
If the formats don't match we need to update the surface with the new
format.

if we can render to SRGB formats, enable the extension

Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-10 10:14:27 +10:00
José Fonseca
3b8bb7b449 scons: Restrict whole program optimization to release builds.
It just takes forever, so it is inadequate for continuous testings
(i.e., checked and profile build types).
2011-02-09 18:31:21 +00:00
José Fonseca
2d95885308 svga: Don't advertise pixel shader addr register support.
It's not fully supported.
2011-02-09 18:31:21 +00:00
Kristian Høgsberg
04c5cc5b8b eglplatform.h: Define Wayland native platform types
This is conditional on WL_EGL_PLATFORM being #defined, so application
must make sure to include wayland-egl.h before including any egl header.
2011-02-09 10:58:20 -05:00
Christoph Bumiller
b6e3130a3b nvc0: serialize on PIPE_FLUSH_RENDER_CACHE as well
Effects were easily visible in piglit/fbo-generatemipmap-formats.
2011-02-09 16:05:00 +01:00
Christoph Bumiller
fc798dc37d nvc0: fix stride of NVC0_3D_RT methods 2011-02-09 16:05:00 +01:00
Christoph Bumiller
95f0aa0e52 nvc0: correct storage type for 16 bit surface formats 2011-02-09 16:05:00 +01:00
Christoph Bumiller
0bd04cdd12 nvc0: make CSE work for ops with multiple results 2011-02-09 16:05:00 +01:00
Christoph Bumiller
0691530b7f nvc0: replace branching with predicated insns where feasible 2011-02-09 16:05:00 +01:00
Christoph Bumiller
0f776fea43 nvc0: implement local memory load and store ops 2011-02-09 16:05:00 +01:00
Christoph Bumiller
4124feabcb nvc0: make sure phi-ops really have one source per in-block 2011-02-09 16:05:00 +01:00
Christoph Bumiller
7401590ded nv50,nvc0: do not forget to apply sign mode to saved TGSI inputs 2011-02-09 16:05:00 +01:00
Christoph Bumiller
c485368efe nvc0: do not generate a backwards jump if a loop ends with BRK 2011-02-09 16:05:00 +01:00
Christoph Bumiller
8e240e6153 nvc0: store only one value per basic block for TGSI regs 2011-02-09 16:05:00 +01:00
Christoph Bumiller
d5263e4093 nv50,nvc0: fix condition code change when commuting SET sources 2011-02-09 16:04:59 +01:00
Christoph Bumiller
8f05134580 nvc0: set basic block on manual instruction insertion 2011-02-09 16:04:59 +01:00
Christoph Bumiller
92d8af582d nvc0: try to fix register conflicts for vector instructions
Vector here means using multiple 32 bit regs which are forced to be
consecutive in the register file.

This still isn't quite nice.
2011-02-09 16:04:59 +01:00
Christoph Bumiller
c62fc50c88 nvc0: reset texture base address after read transfer 2011-02-09 16:04:59 +01:00
Christoph Bumiller
d3ea15f5ca nvc0: don't combine memory loads across block boundaries 2011-02-09 16:04:59 +01:00
Christoph Bumiller
f0d7429623 nvc0: detect no-op MIN/MAX, do CSE earlier to succeed more often 2011-02-09 16:04:59 +01:00
Thomas Hellstrom
a7293cbe5c mesa/st: Clean up vertex buffer unreferencing
Avoid accessing draw module internal structures outside of the draw module.
Unreference vertex buffers in error path.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
2011-02-09 09:45:34 +01:00
Marek Olšák
c264786809 gallium/docs: fix typo 2011-02-09 05:21:20 +01:00
Brian Paul
f70c98b6a6 r200: add cast to silence warning 2011-02-08 19:25:04 -07:00
Brian Paul
3b0c2eb07c mesa: remove unused BITFIELD64 macros 2011-02-08 19:25:04 -07:00
Brian Paul
6f2f449414 mesa: remove _mesa_create_context_for_api()
Just add the gl_api parameter to _mesa_create_context().
2011-02-08 19:25:04 -07:00
Brian Paul
5e4ca1ccc9 mesa: remove _mesa_initialize_context_for_api()
Just add the gl_api parameter to _mesa_initialize_context().
2011-02-08 19:25:04 -07:00
Brian Paul
2634e92dc0 mesa: add/update VERBOSE_API logging 2011-02-08 19:20:43 -07:00
Brian Paul
7230e1a228 st/mesa: fix shader deletion regression
Fixes a regression from commit 5cbff0932e.
The problem is *some* glDrawPixels fragment programs need to be deleted,
but not all.  Use an explicit flag to indicate whether or not the program
needs to be deleted.

This should fix http://bugs.freedesktop.org/show_bug.cgi?id=34049
2011-02-08 18:23:59 -07:00
Kenneth Graunke
f6f04ae29c i965: Add missing DEFINE_BITS for brw dirty bits.
These are only used for debugging, but should be there.
Found by inspection.
2011-02-08 16:56:18 -08:00
Kenneth Graunke
58b7e37722 i965: Separate the BRW_NEW_(VS|WM)_CONSTBUF dirty bits.
These were incorrectly defined to the same value - likely due to a cut
and paste error.  Found by inspection.
2011-02-08 16:55:20 -08:00
Kenneth Graunke
71acbb54f4 i965: Rename a few more commands to match the documentation. 2011-02-08 16:06:47 -08:00