Kristian Høgsberg [Tue, 23 Mar 2010 01:57:26 +0000 (21:57 -0400)]
glx: Suppress BadDrawable from DRI2CopyRegion
This can happen when an X window is destroyed behind our back. We use
DRI2CopyRegion behind the scenes in many places (like flushing the fake
front to the real front) so we have to ignore X errors triggered in that
case.
The glean test cases trigger this consistently as they don't destroy the
GLX drawable nicely, they just destroy the X window.
Corbin Simpson [Tue, 23 Mar 2010 01:12:16 +0000 (18:12 -0700)]
gallium/docs: Cleanup and clarify point ras info.
...This state's interdependent? Really? Needs moar cleanup.
Corbin Simpson [Tue, 23 Mar 2010 01:11:40 +0000 (18:11 -0700)]
radeong: Always initialize this variable.
May or may not be responsible for slight increases in ipers FPS.
Corbin Simpson [Mon, 22 Mar 2010 10:10:19 +0000 (03:10 -0700)]
r300g: Cleanup a few old warnings.
Those paths aren't important anymore, and a debugging warning won't stop
a hardlock anyway.
Jesse Barnes [Mon, 22 Mar 2010 23:39:11 +0000 (16:39 -0700)]
GLX/OML: honor OML semantics even if target, divisor and remainder are 0
This change passes a remainder of 1 to the server with the
DRI2SwapBuffers request, causing it to honor the OML semantics for the
swap rather than falling through to glXSwapBuffers behavior. The
remainder actually ends up ignored since the divisor is 0, but we need
to differentiate the OML and standard behavior somehow.
Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Eric Anholt [Mon, 22 Mar 2010 17:17:19 +0000 (10:17 -0700)]
i965: Add disasm for SNB MATH opcode.
Eric Anholt [Mon, 22 Mar 2010 16:04:25 +0000 (09:04 -0700)]
i965: Enable normal clipping on SNB.
Rejecting all doesn't seem to be helping get the pipeline lit up.
Eric Anholt [Sat, 20 Mar 2010 00:14:12 +0000 (17:14 -0700)]
i965: Force single program flow in SNB GS, to match gen4 GS.
Eric Anholt [Sat, 20 Mar 2010 00:08:42 +0000 (17:08 -0700)]
i965: Correct copy and wasted field shifts for SNB GS URB.
Eric Anholt [Fri, 19 Mar 2010 23:00:31 +0000 (16:00 -0700)]
i965: Enable VS on SNB.
It appears that the thing that was killing VS threads was the
gratuitous NOP that replaced the gratuitous jump from OPCODE_END to
the nearby OPCODE_END implementation. With that gone, we can move on
to the rest of the pipeline.
Eric Anholt [Fri, 19 Mar 2010 22:57:50 +0000 (15:57 -0700)]
i965: Remove gratuitous jump or nop from OPCODE_END to vertex emit.
Just emit the URB write at END time. Subroutine code that sits after
OPCODE_END won't be executed since we've ended the thread at the point
that the URB write is done.
Eric Anholt [Fri, 19 Mar 2010 22:04:17 +0000 (15:04 -0700)]
i965: Ignore execution mask for the mov(m0, g0) of VS URB write header on SNB.
Otherwise, we may not get the FFTID set up which would break freeing
of resources.
Eric Anholt [Fri, 19 Mar 2010 19:34:53 +0000 (12:34 -0700)]
i965: Allow FS constants to be used as immediates instead of push/pull.
The hope is to later take advantage of the reduced constant usage to
free up regs. This only covers the GLSL path at the moment, because
the brw_wm_emit path doesn't get the information as to whether a float
value is a constant or a uniform.
Eric Anholt [Fri, 19 Mar 2010 19:27:38 +0000 (12:27 -0700)]
i965: Add INTEL_DEBUG=glsl_force to force brw_wm_glsl.c.
I keep finding the desire to force this path to debug it instead of
cooking up goofy-looking testcases to do so.
Eric Anholt [Fri, 19 Mar 2010 18:44:43 +0000 (11:44 -0700)]
i965: Optimize OPCODE_CMP by using BRW_SEL to choose results.
Tested with piglit glsl-fs-sqrt-branch, fp-cmp.vpfp.
Brian Paul [Mon, 22 Mar 2010 21:55:43 +0000 (15:55 -0600)]
st/glx: fix depth/stencil format selection code
This fixes a pretty big performance regression caused by commit
3475e88442c16fb2b50b903fe246b3ebe49da226.
When the user does not request a stencil buffer it's important that we
don't use a depth/stencil format (or at least make it our last choice).
If the user calls glClear(GL_DEPTH_BUFFER_BIT) when we have a combined
depth/stencil buffer, that causes us to hit the clear_with_quad() path
which can be much, much slower than calling pipe_context::clear().
Also, try to use a shallower depth format before a deeper one.
Brian Paul [Mon, 22 Mar 2010 21:49:38 +0000 (15:49 -0600)]
st/mesa: rename st_framebuffer() to st_ws_framebuffer()
Be clear that this function is not just a cast wrapper.
Kristian Høgsberg [Mon, 22 Mar 2010 21:54:35 +0000 (17:54 -0400)]
intel: Call intel_prepare_render() before looking up regions.
Fixes #27213.
Vinson Lee [Mon, 22 Mar 2010 20:02:09 +0000 (13:02 -0700)]
st/mesa: Remove unnecessary headers.
Brian Paul [Mon, 22 Mar 2010 19:36:27 +0000 (13:36 -0600)]
glslcompiler: fix build again
Simply use the libglapi.a archive instead of individual .o files.
Fixes the non-debug build.
Vinson Lee [Mon, 22 Mar 2010 18:34:18 +0000 (11:34 -0700)]
cso: Remove unnecessary header.
Francisco Jerez [Mon, 22 Mar 2010 17:10:25 +0000 (18:10 +0100)]
dri/nouveau: Rectangle texture fixes.
Francisco Jerez [Mon, 22 Mar 2010 17:28:51 +0000 (18:28 +0100)]
dri/nouveau: Some render to texture fixes.
Francisco Jerez [Mon, 22 Mar 2010 17:25:30 +0000 (18:25 +0100)]
dri/nouveau: Fix swrast fallbacks when the read and draw buffers aren't the same.
Francisco Jerez [Mon, 22 Mar 2010 17:21:13 +0000 (18:21 +0100)]
dri/nouveau: Expose EXT_framebuffer_blit.
Francisco Jerez [Mon, 22 Mar 2010 13:23:40 +0000 (14:23 +0100)]
dri/nouveau: Plug in some mesa_meta functions instead of the swrast variants.
Francisco Jerez [Mon, 22 Mar 2010 13:15:52 +0000 (14:15 +0100)]
dri/nouveau: Emit dirty states on nouveau_validate_framebuffer.
Fixes broken intermediate frames when a window is being resized
(regression caused by
878eef8c4).
Luca Barbieri [Mon, 22 Mar 2010 17:36:16 +0000 (18:36 +0100)]
nvfx: fix sampler views support
The code was half converted, resulting in texturing being totally broken.
Brian Paul [Mon, 22 Mar 2010 15:08:09 +0000 (09:08 -0600)]
glslcompiler: fix build breakage
Brian Paul [Mon, 22 Mar 2010 15:07:46 +0000 (09:07 -0600)]
Merge branch '7.8'
Conflicts:
src/gallium/drivers/cell/ppu/cell_screen.c
src/mesa/state_tracker/st_cb_drawpixels.c
Brian Paul [Mon, 22 Mar 2010 15:00:13 +0000 (09:00 -0600)]
glslcompiler: fix build breakage
Brian Paul [Mon, 22 Mar 2010 14:23:14 +0000 (08:23 -0600)]
mesa: remove return, do as the comment says
Pauli Nieminen [Mon, 22 Mar 2010 13:14:16 +0000 (15:14 +0200)]
mesa: Add end of line to the end of a debug output.
Chia-I Wu [Mon, 22 Mar 2010 05:30:54 +0000 (13:30 +0800)]
docs: Update the path to build libgl-xlib.
Vinson Lee [Mon, 22 Mar 2010 02:21:11 +0000 (19:21 -0700)]
progs/tests: Add cva_huge to SCons build.
Vinson Lee [Mon, 22 Mar 2010 02:09:54 +0000 (19:09 -0700)]
progs/tests: Include stddef.h for ptrdiff_t on all platforms.
stddef.h is the standard C header that defines ptrdiff_t.
Fixes build of cva_huge on Mac OS X.
Pauli Nieminen [Sun, 21 Mar 2010 21:23:21 +0000 (23:23 +0200)]
r200: Fix emit size prediction to account elt splitting.
Emit sizes prediction didn't account for render splitting in
hwtnl path.
Pauli Nieminen [Sun, 21 Mar 2010 20:55:13 +0000 (22:55 +0200)]
r200: Don't flush when closing elts in KMS.
Flush in middle of rendering in KMS is not allowed because
buffers are discarded in flush.
Fixes crash when emiting split indices with RADEON_DEBUG=all.
Marek Olšák [Sun, 21 Mar 2010 21:17:14 +0000 (22:17 +0100)]
r300g: disable tiling for YUV formats
Marek Olšák [Sun, 21 Mar 2010 21:05:27 +0000 (22:05 +0100)]
r300g: fix scons build
Marek Olšák [Sun, 21 Mar 2010 20:47:41 +0000 (21:47 +0100)]
r300g: accelerate blitting for all 64-bit texture formats
Marek Olšák [Sun, 14 Feb 2010 16:43:32 +0000 (17:43 +0100)]
r300g: add and enable square microtiling
It requires DRM 2.1.0 (e.g. kernel 2.6.34) and is disabled on older ones.
Finally, the texture tiling implementation is now complete. Uff.
Marek Olšák [Sun, 21 Mar 2010 20:34:57 +0000 (21:34 +0100)]
r300g: cleanup tiling flags propagation
Marek Olšák [Sun, 21 Mar 2010 20:34:18 +0000 (21:34 +0100)]
r300g: put common defines into one file
Marek Olšák [Sun, 21 Mar 2010 20:22:21 +0000 (21:22 +0100)]
r300g: remove unused code
Pauli Nieminen [Sun, 21 Mar 2010 18:16:17 +0000 (20:16 +0200)]
r200: Fix swtnl fallback to flush pending rendering before transition.
Flush after transition would emit wrong state that could cause
wrong state emited for pending rendering operation.
Fixes wan once from extrement tuxracer that is using per vertex
materials.
Pauli Nieminen [Sun, 21 Mar 2010 18:02:10 +0000 (20:02 +0200)]
r200: Fix mixed indetion in r200TclFallback.
George Sapountzis [Sun, 21 Mar 2010 17:36:28 +0000 (19:36 +0200)]
Revert "drm/sw: does not need sw_winsys"
This reverts commit
97a694e2211dc804090f282d8e096c028f29579f.
George Sapountzis [Sun, 21 Mar 2010 17:33:17 +0000 (19:33 +0200)]
Revert "gallium: add soft screen helper"
This reverts commit
f87a5f6499f51f651c2a9f2d4682875b22926905.
George Sapountzis [Sun, 21 Mar 2010 17:33:09 +0000 (19:33 +0200)]
Revert "drm/sw: just s/softpipe_create_screen/gallium_soft_create_screen/"
This reverts commit
5d524cce9c4fcc18ed977801d59ba7bb911020db.
George Sapountzis [Sun, 21 Mar 2010 17:32:40 +0000 (19:32 +0200)]
Revert "fix scons build"
This reverts commit
29ec05164838f13d9535271796a50fa213d07912.
George Sapountzis [Sun, 21 Mar 2010 17:20:15 +0000 (19:20 +0200)]
fix scons build
George Sapountzis [Sun, 21 Mar 2010 13:37:40 +0000 (15:37 +0200)]
Merge branch 'master' of git+ssh://git.freedesktop.org/git/mesa/mesa
George Sapountzis [Sun, 21 Mar 2010 13:32:50 +0000 (15:32 +0200)]
Revert "Makefile.template: respect LIBRARY_DEFINES in mkdep"
This reverts commit
9ec29e31919e85f9230867f43841c0e74be930d3.
Christoph Bumiller [Sun, 21 Mar 2010 13:05:20 +0000 (14:05 +0100)]
nv50: report driver as GLSL capable
There's still no hint of optimization in the shaders, but
we support conditionals, loops and even integer opcodes so
it isn't that big a lie.
George Sapountzis [Sun, 21 Mar 2010 13:05:21 +0000 (15:05 +0200)]
Merge branch 'master' of git://anongit.freedesktop.org/git/mesa/mesa
George Sapountzis [Sun, 21 Mar 2010 13:01:36 +0000 (15:01 +0200)]
drm/sw: just s/softpipe_create_screen/gallium_soft_create_screen/
This partially reverts commit
f87a5f6499f51f651c2a9f2d4682875b22926905.
Christoph Bumiller [Sun, 21 Mar 2010 11:26:08 +0000 (12:26 +0100)]
nv50: don't validate arrays on clear
The vertex elements and buffers might not be valid anymore.
Christoph Bumiller [Sun, 21 Mar 2010 12:17:02 +0000 (13:17 +0100)]
nv50: get rid of the static_init stateobj
Relocations of per-screen buffers are now emitted directly,
and include the necessary method to get changes in constbuf
addresses committed to the hw.
It should also be a bit cheaper than the way stateobjs emit
relocation markers, use a little less pushbuf space.
George Sapountzis [Fri, 19 Mar 2010 00:38:11 +0000 (02:38 +0200)]
gallium: add soft screen helper
George Sapountzis [Fri, 19 Mar 2010 00:38:11 +0000 (02:38 +0200)]
dri1_api: need not include drm.h
George Sapountzis [Fri, 19 Mar 2010 00:38:11 +0000 (02:38 +0200)]
drm/sw: does not need sw_winsys
George Sapountzis [Fri, 19 Mar 2010 00:38:11 +0000 (02:38 +0200)]
drm/sw: llvmpipe texture_from_handle
Not sure, but judging by softpipe, this hook was forgotten.
George Sapountzis [Fri, 19 Mar 2010 00:38:10 +0000 (02:38 +0200)]
Makefile.template: respect LIBRARY_DEFINES in mkdep
George Sapountzis [Fri, 19 Mar 2010 00:38:10 +0000 (02:38 +0200)]
rename dri_sw to drisw_util for consistency
George Sapountzis [Fri, 19 Mar 2010 00:38:09 +0000 (02:38 +0200)]
dri_inteface: add define for checking presence of drm.h
__NOT_HAVE_DRM_H is a like a feature, defined by default on specific platforms
and allows to be defined externally as well.
__NOT_HAVE_DRM_H should only be used by xserver and mesa swrast_dri drivers
George Sapountzis [Fri, 19 Mar 2010 23:51:00 +0000 (01:51 +0200)]
glx: minor cosmetic
George Sapountzis [Fri, 19 Mar 2010 00:38:09 +0000 (02:38 +0200)]
glx: swapBuffers prototype has changed
Pauli Nieminen [Sun, 21 Mar 2010 10:42:10 +0000 (12:42 +0200)]
vbo: Fix vbo_split_copy to pass correct max_index to draw.
vbo_split_copy was passing one past the max_index to draw function
which caused _tnl_draw_prims function to read uninitialized values
from copied array.
Bug was spoted in valgrind report of progs/tests/cva_huge.
Pauli Nieminen [Sun, 21 Mar 2010 11:11:47 +0000 (13:11 +0200)]
tests: Add test for huge client arrays that has to be split.
When running this test case in valgrind report includes read of unitialized
value in _tnl_draw_prims. The bug doesn't cause any vissible errors.
Bug is caused by vbo_split_copy that is calling draw function with max_index
one past the end instead of the end.
Francisco Jerez [Sun, 21 Mar 2010 10:49:46 +0000 (11:49 +0100)]
dri/nouveau: Reemit the light model state when lighting goes on/off.
Francisco Jerez [Sun, 21 Mar 2010 10:49:20 +0000 (11:49 +0100)]
dri/nouveau: Random cleanups.
Dave Airlie [Sun, 21 Mar 2010 09:36:15 +0000 (19:36 +1000)]
r300g: fix occlusion query hangs.
Okay need to revist the whole OQ stuff anyways, glean test asserts
which is never good.
I'm liking the cached bufmgr restrictions less and less, I think I'll
probably play with the fence and/or busy stuff ASAP and try and clean it up.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Sun, 21 Mar 2010 08:25:31 +0000 (18:25 +1000)]
r300g: add bufmgr cached for vertex/index buffers.
This adds the cached bufmgr for vertex/index buffers on top of the kernel
manager.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Sun, 21 Mar 2010 08:23:22 +0000 (18:23 +1000)]
gallium: cached bufmgr add map busy check
Suggested by Jose on the list, probably not perfect but will let me get
past this for now, testing with a fenced bufmgr on top of this, was slower,
Also this doesn't let you do the busy early exit optimisation either from
what I can see.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Chia-I Wu [Sun, 21 Mar 2010 08:32:34 +0000 (16:32 +0800)]
Merge remote branch 'origin/gallium-st-api-dri'
Marek Olšák [Sun, 21 Mar 2010 03:49:35 +0000 (04:49 +0100)]
r300g: align misaligned ushort vertex indices
Marek Olšák [Sun, 21 Mar 2010 03:39:33 +0000 (04:39 +0100)]
r300g: fix misaligned generated offset for ubyte vertex indices
Marek Olšák [Sat, 20 Mar 2010 21:14:59 +0000 (22:14 +0100)]
r300g: split the vertex buffer alignment validation
Marek Olšák [Sat, 20 Mar 2010 20:26:19 +0000 (21:26 +0100)]
r300g: skip null vertex buffers
Marek Olšák [Sat, 20 Mar 2010 17:31:11 +0000 (18:31 +0100)]
r300/compiler: fix assertion failure in the r500-fragprog emission path
Corbin Simpson [Sun, 21 Mar 2010 00:18:22 +0000 (17:18 -0700)]
r300g: Bump immediate limits.
Seems like a decent idea, especially since the big barrier now is
getting the VBOs back from the VRAM boundary.
Corbin Simpson [Sun, 21 Mar 2010 00:16:46 +0000 (17:16 -0700)]
r300g: Correctly hax max_index on pipe_vertex_buffers.
Still not happy with this, but at least things seem to work.
Corbin Simpson [Sat, 20 Mar 2010 23:15:36 +0000 (16:15 -0700)]
radeong: Use TRUE and FALSE.
Corbin Simpson [Sat, 20 Mar 2010 23:14:20 +0000 (16:14 -0700)]
radeong: Test the BO ptr for NULL instead of using a bool.
Slight memory savings, but more importantly also avoids a potential NULL
dereference.
Dave Airlie [Sat, 20 Mar 2010 20:34:11 +0000 (06:34 +1000)]
r300g: winsys fix tiling change flushes.
If we change the tiling on a buffer we need to flush it, the old
radeon_buffer.c code had this but it crossed streams when I ported to
radeon_drm_buffer.c and I missed it. Should fix some piglit regressions.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Sat, 20 Mar 2010 20:32:48 +0000 (06:32 +1000)]
r300g: add buffer/texture referenced checks.
I've no idea about the comments that were in there, just pass
this down to the winsys.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Brian Paul [Sat, 20 Mar 2010 18:05:25 +0000 (12:05 -0600)]
docs: note GL_EXT_packed_float, GL_EXT_shared_exponent extensions
Brian Paul [Sat, 20 Mar 2010 18:04:52 +0000 (12:04 -0600)]
mesa: added GL3 query for GL_TEXTURE_SHARED_SIZE
Brian Paul [Sat, 20 Mar 2010 17:55:24 +0000 (11:55 -0600)]
mesa: added missing glGet query for GL_MAX_ARRAY_TEXTURE_LAYERS_EXT
Brian Paul [Sat, 20 Mar 2010 17:52:12 +0000 (11:52 -0600)]
mesa: added new GL3 buffer object queries
And clean up the error checking code.
Brian Paul [Sat, 20 Mar 2010 17:50:55 +0000 (11:50 -0600)]
mesa: added GL3 buffer attachment aliases
Brian Paul [Sat, 20 Mar 2010 17:50:29 +0000 (11:50 -0600)]
mesa: added GL3 ContextFlags field and query code
Marek Olšák [Sat, 20 Mar 2010 17:31:11 +0000 (18:31 +0100)]
r300/compiler: fix assertion failure in the r500-fragprog emission path
Brian Paul [Sat, 20 Mar 2010 16:33:10 +0000 (10:33 -0600)]
docs: remove the 'Last updated date'
Chia-I Wu [Sat, 20 Mar 2010 15:58:15 +0000 (23:58 +0800)]
glapi: Fix aliases to non-static functions.
The bug is triggered by
41a87a43e11c664935349f938022d58d3e22da4e.
glBlitFramebuffer, for example, is an alias to the non-static
glBlitFramebufferEXT. We should define glBlitFramebuffer as an alias to
_dispatch_stub_NNN.
Chia-I Wu [Thu, 18 Mar 2010 03:36:57 +0000 (11:36 +0800)]
glapi: Correctly generate static disatches for X86.
The entry point names, instead of the function name, should be used to
test if the entry point should be statically dispatched.
George Sapountzis [Fri, 19 Mar 2010 19:16:21 +0000 (19:16 +0000)]
st/dri: fix bug in make_current
George Sapountzis [Fri, 19 Mar 2010 00:38:10 +0000 (02:38 +0200)]
st/dri: fix bug in allocate_textures
Marek Olšák [Tue, 9 Mar 2010 01:10:36 +0000 (02:10 +0100)]
r300g: remove hacks from translate_vertex_data_swizzle
Fixing RGBA 4ub vertex colors.