Brian Paul [Fri, 26 Jun 2009 22:46:21 +0000 (16:46 -0600)]
docs: bring over news updates from 7.4 branch
Brian Paul [Fri, 26 Jun 2009 22:45:54 +0000 (16:45 -0600)]
docs: bring in 7.4.3 and 7.4.4 release notes, news
Brian Paul [Fri, 26 Jun 2009 22:41:54 +0000 (16:41 -0600)]
docs: increase max varying vars to 16
Brian Paul [Fri, 26 Jun 2009 22:33:46 +0000 (16:33 -0600)]
mesa: raise MAX_VARYING (number of shader varying vars) to 16
16 is the limit for now because of various 32-bit bitfields.
(cherry picked from master, commit
4e762395ef7e8c332c16fd0c11025cfa52763a45)
Ian Romanick [Fri, 26 Jun 2009 20:30:16 +0000 (13:30 -0700)]
intel / DRI2: Additional flush of fake front-buffer to real front-buffer
To maintain correctness, the server will copy the real front-buffer to
a newly allocated fake front-buffer in DRI2GetBuffersWithFormat.
However, if the DRI2GetBuffersWithFormat is triggered by glViewport,
this will copy stale data into the new buffer. Fix this by flushing
the current fake front-buffer to the real front-buffer in
intel_viewport.
Fixes bug #22288.
Brian Paul [Fri, 26 Jun 2009 19:16:08 +0000 (13:16 -0600)]
mesa: bump version to 7.5-rc4
Brian Paul [Fri, 26 Jun 2009 19:15:17 +0000 (13:15 -0600)]
docs: document for/continue bug fix
Brian Paul [Fri, 26 Jun 2009 18:51:05 +0000 (12:51 -0600)]
cell: PIPE_CAP_TGSI_CONT_SUPPORTED query
Brian Paul [Fri, 26 Jun 2009 18:50:58 +0000 (12:50 -0600)]
softpipe: PIPE_CAP_TGSI_CONT_SUPPORTED query
Brian Paul [Fri, 26 Jun 2009 18:50:45 +0000 (12:50 -0600)]
gallium: added PIPE_CAP_TGSI_CONT_SUPPORTED
Brian Paul [Fri, 26 Jun 2009 18:34:03 +0000 (12:34 -0600)]
glsl: move/simplify error checking for 'return' statements
Brian Paul [Fri, 26 Jun 2009 17:33:05 +0000 (11:33 -0600)]
glsl: overhaul 'return' statement handling
A new node type (SLANG_OPER_RETURN_INLINED) is used to denote 'return'
statements inside inlined functions which need special handling.
All glean glsl1 tests pass for EmitContReturn=FALSE and TRUE.
Brian Paul [Thu, 25 Jun 2009 20:01:43 +0000 (14:01 -0600)]
glsl: predicate assignments according to __returnFlag
Fixes glean "function with early return (3)" case (when EmitContReturn=FALSE).
Brian Paul [Thu, 25 Jun 2009 20:02:41 +0000 (14:02 -0600)]
glsl: added slang_variable::is_global field
Brian Paul [Thu, 25 Jun 2009 17:52:37 +0000 (11:52 -0600)]
glsl: silence a problem warning
Brian Paul [Thu, 25 Jun 2009 17:32:37 +0000 (11:32 -0600)]
glsl: code refactoring for return statements
Brian Paul [Thu, 25 Jun 2009 17:12:26 +0000 (11:12 -0600)]
glsl: fix assorted regressions related to early-return-removal
Brian Paul [Thu, 25 Jun 2009 16:20:21 +0000 (10:20 -0600)]
glsl: comments, field reordering
Brian Paul [Thu, 25 Jun 2009 15:50:37 +0000 (09:50 -0600)]
glsl: rework loop nesting code
Brian Paul [Thu, 25 Jun 2009 15:30:53 +0000 (09:30 -0600)]
glsl: remove test for loop unrolling when we hit conditional cont/break
This is no longer needed since we added the new
_slang_loop_contains_continue_or_break() function/test.
Brian Paul [Wed, 24 Jun 2009 23:02:34 +0000 (17:02 -0600)]
glsl: checkpoint: predicate __retVal = expr with __returnFlag
The glean "function with early return (1)" test passes now.
Brian Paul [Wed, 24 Jun 2009 22:37:01 +0000 (16:37 -0600)]
glsl: use new helper functions in _slang_gen_logical_and/or()
Brian Paul [Wed, 24 Jun 2009 22:36:24 +0000 (16:36 -0600)]
glsl: add comments
Brian Paul [Wed, 24 Jun 2009 22:28:32 +0000 (16:28 -0600)]
glsl: checkpoint: replace 'return' with __returnFlag=true;
Needed for "remove early returns" transformation.
Brian Paul [Wed, 24 Jun 2009 22:25:00 +0000 (16:25 -0600)]
glsl: fix up scoping for parent/children in slang_operation_copy()
This will need more testing, but no regressions seen so far.
Brian Paul [Wed, 24 Jun 2009 22:12:20 +0000 (16:12 -0600)]
glsl: fix uninitialized var in _slang_gen_for_without_continue()
Brian Paul [Wed, 24 Jun 2009 22:04:33 +0000 (16:04 -0600)]
glsl: added slang_operation_free_children()
Brian Paul [Wed, 24 Jun 2009 19:16:49 +0000 (13:16 -0600)]
glsl: added slang_oper_num_children() helper
Brian Paul [Wed, 24 Jun 2009 14:40:56 +0000 (08:40 -0600)]
glsl: check-point: declare _returnFlag
Brian Paul [Tue, 23 Jun 2009 23:07:12 +0000 (17:07 -0600)]
glsl: added slang_operation_insert_child()
Brian Paul [Tue, 23 Jun 2009 22:55:10 +0000 (16:55 -0600)]
glsl: use slang_generate_declaration() to consolidate some code
Brian Paul [Tue, 23 Jun 2009 22:17:06 +0000 (16:17 -0600)]
glsl: remove obsolete comment
Brian Paul [Tue, 23 Jun 2009 20:28:43 +0000 (14:28 -0600)]
st/mesa: query PIPE_CAP_TGSI_CONT_SUPPORTED
Brian Paul [Tue, 23 Jun 2009 20:05:11 +0000 (14:05 -0600)]
glsl: implement _slang_gen_while_without_continue()
Brian Paul [Tue, 23 Jun 2009 19:33:28 +0000 (13:33 -0600)]
glsl: fix a bug involving 'continue' in 'for' loops
Need to execute the for loop's increment code before we continue.
Add a slang_assemble_ctx::CurLoopOper field to keep track of the containing
loop and avoid the "cont if true" path in this situation.
Plus, assorted clean-ups.
Brian Paul [Tue, 23 Jun 2009 18:46:11 +0000 (12:46 -0600)]
glsl: added slang_oper_child_const()
Brian Paul [Tue, 23 Jun 2009 16:57:59 +0000 (10:57 -0600)]
glsl: use _slang_loop_contains_continue_or_break() to check for unrolling
The previous test failed for nested loops.
Brian Paul [Thu, 18 Jun 2009 23:25:47 +0000 (17:25 -0600)]
glsl: use new _slang_loop_contains_continue() helper function
Brian Paul [Thu, 18 Jun 2009 23:08:10 +0000 (17:08 -0600)]
glsl: implement continue->break translation for do-while loops
Brian Paul [Thu, 18 Jun 2009 22:57:53 +0000 (16:57 -0600)]
glsl: added slang_operation_literal_int/bool() helper functions
Brian Paul [Thu, 18 Jun 2009 22:57:23 +0000 (16:57 -0600)]
glsl: don't allocate 0-length children array in slang_operation_copy()
Brian Paul [Thu, 18 Jun 2009 20:24:30 +0000 (14:24 -0600)]
glsl: remove debug code, misc clean-up
Brian Paul [Thu, 18 Jun 2009 20:14:53 +0000 (14:14 -0600)]
glsl: implement continue->break transformation for for-loops
Brian Paul [Thu, 18 Jun 2009 20:11:59 +0000 (14:11 -0600)]
glsl: added slang_operation_add_children() and slang_oper_child() helpers
Brian Paul [Thu, 18 Jun 2009 20:11:38 +0000 (14:11 -0600)]
glsl: added slang_assemble_ctx::EmitContReturn field, init
Brian Paul [Thu, 18 Jun 2009 18:43:43 +0000 (12:43 -0600)]
demos: improve argv parsing in fslight.c
Brian Paul [Wed, 17 Jun 2009 19:51:35 +0000 (13:51 -0600)]
glsl: don't unroll loops containing continue/break
Just search the AST in _slang_can_unroll_for_loop().
Brian Paul [Wed, 17 Jun 2009 19:48:52 +0000 (13:48 -0600)]
Revert "slang: if/else/break & if/else/continue work for unrolled loops"
We should just check if the loop contains a continue/break in the
_slang_can_unroll_for_loop() test function...
This reverts commit
989856bde47d699d7e18798df4013fbf962e1d4b.
Conflicts:
src/mesa/shader/slang/slang_codegen.h
Brian Paul [Wed, 17 Jun 2009 17:57:39 +0000 (11:57 -0600)]
glsl: Added gl_shader_state::EmitContReturn field
This is the start of a glsl-continue-return feature branch to support
a GLSL code generator option for 'continue' and 'return' statements.
Some targets don't support CONT or RET statements so we'll need to
try to generate code that does not use them...
Brian Paul [Wed, 24 Jun 2009 01:21:04 +0000 (19:21 -0600)]
docs: updated Mesa release instructions
José Fonseca [Fri, 26 Jun 2009 18:51:57 +0000 (19:51 +0100)]
scons: Tool to build with DirectX SDK.
Also works with MinGW, as long as the path to the DirectX SDK top
directory is set in the DXSDK_DIR environment variable.
José Fonseca [Fri, 26 Jun 2009 18:50:12 +0000 (19:50 +0100)]
scons: Don't use C specific options with g++
Keith Whitwell [Fri, 26 Jun 2009 12:43:10 +0000 (13:43 +0100)]
tgsi: correct handling of return value from util_vsnprintf
We were failing to deal with:
- vsnprintf returns negative value on error.
- vsnprintf returns the number of chars that *would* have been
written on truncation.
Keith Whitwell [Fri, 26 Jun 2009 12:44:20 +0000 (13:44 +0100)]
aux/indices: don't use 'prim' value once it is known to be bad
Theoretical bugfix only - no known case where this might happen.
José Fonseca [Wed, 24 Jun 2009 13:46:07 +0000 (14:46 +0100)]
wgl: Handle flush after a window is destroyed.
Fixes assertion failure with conform.
Brian Paul [Tue, 23 Jun 2009 15:31:27 +0000 (09:31 -0600)]
demos: fix issues in glxcontexts.c
The reshape() function was called when there was no GLX context so
the viewport/modelview/projection setup wasn't doing anything. Move
the call to reshape() into draw().
Also, remove -stereo, -fullscreen options and do some general clean-up.
Roland Scheidegger [Tue, 16 Jun 2009 19:38:58 +0000 (21:38 +0200)]
i965: handle OPCODE_SWZ in the glsl path
glsl compiler will not generate OPCODE_SWZ, and as a first step it would
be translated away to a MOV anyway (why?), but later internally this opcode is
generated (for EXT_texture_swizzling).
(cherry picked from commit
4ef1f8e3b52a06fcf58f78c9c36738531b91dbac)
Owen W. Taylor [Sat, 6 Jun 2009 18:46:22 +0000 (14:46 -0400)]
Disable SGI_swap_control extension for DRI2
We currently don't have support for SGI_swap_control for direct
contexts with DRI2, so disable reporting the extension. Reporting
the extension, and then having glXSwapIntervalSGI() "succeed"
but do nothing can confuse applications.
https://bugs.freedesktop.org/show_bug.cgi?id=22123
(cherry picked from commit
279143c6e808b37c333321b696d80df77f709a04)
Michel Dänzer [Thu, 11 Jun 2009 10:09:10 +0000 (12:09 +0200)]
intel: intel_texture_drawpixels() can't handle GL_DEPTH_STENCIL.
Fixes glean depthStencil test.
(cherry picked from commit
3885b708fdbb7bbd5dd3a247c41fb9a75ee7c057)
Brian Paul [Tue, 9 Jun 2009 21:07:39 +0000 (15:07 -0600)]
i965: added intelFlush() call in intel_get_tex_image()
Fixes the render-to-texture test in progs/tests/getteximage.c
(cherry picked from commit
a03b349153660e449daf4f56d750f1caef23b1a5)
Brian Paul [Fri, 22 May 2009 13:19:27 +0000 (07:19 -0600)]
mesa: s/MAX_VERTEX_PROGRAM_ATTRIBS/MAX_NV_VERTEX_PROGRAM_INPUTS/ in NV funcs
(cherry picked from commit
4dc426c01627a240bd5b148c1804c45b0d5ecd6c)
Brian Paul [Fri, 22 May 2009 13:26:08 +0000 (07:26 -0600)]
mesa: rename MAX_VERTEX_ATTRIBS to MAX_VERTEX_GENERIC_ATTRIBS
Be clearer that this is the number of generic vertex program/shader
attributes, not counting the legacy attributes (pos, normal, color, etc).
(cherry picked from commit
4a95185c9f30c2de7a03bb1a0653f51b53b1111d)
Brian Paul [Thu, 21 May 2009 23:03:21 +0000 (17:03 -0600)]
mesa: s/MAX_VERTEX_PROGRAM_ATTRIBS/MAX_NV_VERTEX_PROGRAM_INPUTS
(cherry picked from commit
d2a74d76c96957cf0294dcf40d29526621ada95e)
Brian Paul [Thu, 21 May 2009 16:05:04 +0000 (10:05 -0600)]
mesa: fix some potential state-restore issues in pop_texture_group()
Call the _mesa_set_enable() functions instead of driver functions, etc.
Also, add missing code for 1D/2D texture arrays.
(cherry picked from commit
aac19609bfd7c950b2577489b06886c8a8097bb2)
Brian Paul [Thu, 21 May 2009 15:52:33 +0000 (09:52 -0600)]
vbo: return VP_NONE from get_program_mode() if running fixed-func vertex program
If we're running a vertex program to emulated fixed-function, we still need
to treat vertex arrays/attributes as if we're in fixed-function mode.
This should probably be back-ported to Mesa 7.5 after a bit more testing.
(cherry picked from commit
dda82137d28aba846dda73da230871c115e30aaf)
Brian Paul [Fri, 1 May 2009 15:30:32 +0000 (09:30 -0600)]
mesa: create/use a fallback texture when bound texture is incomplete
When a GLSL sampler reads from an incomplete texture it should
return (0,0,0,1). Instead of jumping through hoops in all the drivers
to make this happen, just create/install a fallback texture with those
texel values.
Fixes piglit/fp-incomplete-tex on i965 and more importantly, fixes some
GPU lockups when trying to sample from missing surfaces. If a binding
table entry is NULL, it seems that sampling sometimes works, but not
always (lockup).
Todo: create a fallback texture for each type of texture target?
(cherry picked from commit
3f25219c7bf0f090502489928f0f018e62c4f6cf)
Eric Anholt [Fri, 19 Jun 2009 22:57:03 +0000 (15:57 -0700)]
intel: Fix other metaops versus GL_COMPILE_AND_EXECUTE dlists.
Fixes oglconform zbfunc.c and pxtrans-cidraw.c, at least.
(cherry picked from commit
405300bb190f516e16b704050abe3389b366ed27)
Eric Anholt [Fri, 19 Jun 2009 22:49:24 +0000 (15:49 -0700)]
intel: Fix glClear behavior versus display lists.
The CALL_DrawArrays was leaking the clear's primitives into the display
list with GL_COMPILE_AND_EXECUTE. Use _mesa_DrawArrays instead, which
doesn't appear to leak. Fixes piglit dlist-clear test.
(cherry picked from commit
64edde1004f7a69e77877bba24d315a92bcd47c8)
Roland Scheidegger [Fri, 19 Jun 2009 18:00:55 +0000 (20:00 +0200)]
radeons: use dp4 for position invariant vertex programs
Fixes #22181. R200 requires this since DP4 is used in hw tnl mode.
R300 prefers it (should be faster due to no instruction dependencies), but
both methods should be correct (when sw tcl is used though, MUL/MAD might
be faster). Probably doesn't make much difference for R100 since vertex progs
are executed in software anyway, but let's just keep it the same there too.
Michel Dänzer [Fri, 19 Jun 2009 09:19:07 +0000 (11:19 +0200)]
Also release direct rendering resources in glXDestroyGLXPixmap.
Fixes leak running compiz with direct rendering.
Michel Dänzer [Fri, 19 Jun 2009 09:19:08 +0000 (11:19 +0200)]
Always free image offsets memory when re-initializing texture image fields.
Fixes leak running compiz with direct rendering.
José Fonseca [Fri, 19 Jun 2009 14:39:59 +0000 (15:39 +0100)]
scons: Output the friendly "Linking ..." message when creating DLLs with MinGW.
Brian Paul [Fri, 19 Jun 2009 14:40:43 +0000 (08:40 -0600)]
softpipe: return alpha=1 in shadow_compare() function
The alpha value wasn't set at all before so we got unpredictable results.
Note that we don't currently obey GL_DEPTH_TEXTURE_MODE in the state
tracker. For now, we return the result in the default mode (r,r,r,1).
José Fonseca [Thu, 18 Jun 2009 13:47:30 +0000 (14:47 +0100)]
util: Add cast.
It is expected to loose precision here.
José Fonseca [Thu, 18 Jun 2009 13:46:45 +0000 (14:46 +0100)]
draw: Replace pointer arithmetic with cast.
Using uintptr_t as intermediate type for pointer -> integer conversions is
easier to understand and does not cause any size mismatch warnings.
uintptr_t is part of C99, and we already provide a suitable replacement
definition for all platforms we care about.
José Fonseca [Thu, 18 Jun 2009 11:52:10 +0000 (12:52 +0100)]
trace: Use size_t consistently.
José Fonseca [Thu, 18 Jun 2009 11:51:00 +0000 (12:51 +0100)]
pipebuffer: Use a type consistently for sizes/offsets.
Avoids warnings on 64bit builds.
Use regular unsigned since that's what gallium expects, but use a
typedef to facilitate possible changes in the future.
José Fonseca [Wed, 17 Jun 2009 20:28:18 +0000 (21:28 +0100)]
wgl: Fix window resizing in multithread applications.
In multithreading stw_call_window_proc can be called by a thread other
than the thread where the context is bound.
José Fonseca [Wed, 17 Jun 2009 18:24:51 +0000 (19:24 +0100)]
wgl: Move all thread related code together.
Not only for cosmetic reasons, but also because we need to set the
SetWindowsHookEx hook for threads created before the DllMain is called
(threads for each we don't get the DLL_THREAD_ATTACH notification).
Keith Whitwell [Thu, 18 Jun 2009 09:13:55 +0000 (10:13 +0100)]
glew: correct misspelling of glFramebufferTextureLayer
This was misspelt as glFramebufferTexturLayer (missing e), causing
conflicts with the correctly spelt version in glext.h and extfuncs.h.
Keith Whitwell [Thu, 11 Jun 2009 09:52:37 +0000 (10:52 +0100)]
mesa: protect Elements against multiple definitions
Mesa and gallium both have a definition of this macro
Brian Paul [Mon, 8 Jun 2009 20:41:54 +0000 (14:41 -0600)]
intel: remove extra \n from warning string
(cherry picked from commit
42e9bde0fa2276b8f5bb434328eea7665794b127)
Robert Ellison [Thu, 14 May 2009 02:40:23 +0000 (20:40 -0600)]
i965: fix 1D texture borders with GL_CLAMP_TO_BORDER
With 1D textures, GL_TEXTURE_WRAP_T should be ignored (only
GL_TEXTURE_WRAP_S should be respected). But the i965 hardware
seems to follow the value of GL_TEXTURE_WRAP_T even when sampling
1D textures.
This fix forces GL_TEXTURE_WRAP_T to be GL_REPEAT whenever 1D
textures are used; this allows the texture to be sampled
correctly, avoiding "imaginary" border elements in the T direction.
This bug was demonstrated in the Piglit tex1d-2dborder test.
With this fix, that test passes.
(cherry picked from commit
ab6c4fa582972e25f8800c77b5dd5b3a83afc996)
Robert Ellison [Thu, 14 May 2009 02:38:33 +0000 (20:38 -0600)]
i965: send all warnings through _mesa_warning()
One warning message:
drm_i915_getparam: -22
was still being sent to fprintf(). This causes all Piglit tests to fail,
even with MESA_DEBUG=0.
Using _mesa_warning() to emit the message allows the general Mesa controls
for messages like this to be applied.
(cherry picked from commit
bc3270e99f5c39544aaf831742db14796ab83a6a)
Robert Ellison [Fri, 8 May 2009 20:40:38 +0000 (14:40 -0600)]
i965: fix segfault on low memory conditions
When out of memory (in at least one case, triggered by a longrunning
memory leak), this code will segfault and crash. By checking for the
out-of-memory condition, the system can continue, and will report
the out-of-memory error later, a much preferable outcome.
(cherry picked from commit
44a4abfd4f8695809eaec07df8eeb191d6e017d7)
Ian Romanick [Wed, 3 Jun 2009 16:49:05 +0000 (17:49 +0100)]
Fast path when rebinding the same texture in single context environment
If there is no shared context, there is no purpose in rebinding the same
texture. In some artificial tests this improves performance 10% - 30%.
(cherry picked from commit
7f8000db8bd45bb95bda4a4f8535c49b8ef74254)
Eric Anholt [Tue, 2 Jun 2009 13:53:40 +0000 (06:53 -0700)]
i915: Don't put VBOs in graphics memory unless required for an operation.
This saves doing swtnl from uncached memory, which is painful. Improves
clutter test-text performance by 10% since it started using VBOs.
(cherry picked from commit
a945e203d4fe254593bc0c5c5d6caca45e65f9f7)
Eric Anholt [Wed, 27 May 2009 03:45:29 +0000 (20:45 -0700)]
i915: Fall back on NPOT textured metaops on 830-class.
(cherry picked from commit
8ec6e036792decf5149a209e51cb5e93ccc5c754)
Eric Anholt [Wed, 27 May 2009 02:48:08 +0000 (19:48 -0700)]
i915: Restore the Viewport and DepthRange functions on 8xx.
Fixes failed viewport updates on glxgears (and other apps) resize since
e41780fedc2c1f22b43118da30a0103fa68b769f.
Bug #20473.
(cherry picked from commit
0e83e8f51af07a3066519f169f07d9afbf23252e)
Eric Anholt [Wed, 20 May 2009 22:59:07 +0000 (15:59 -0700)]
i956: Make state dependency of SF on drawbuffer bounds match Mesa's.
Noticed while debugging a weird 1D FBO testcase that left its existing
viewport and projection matrix in place when switching drawbuffers. Didn't
fix the testcase, though.
(cherry picked from commit
3a521d84ecc646fcc65fa3fe7c5f1fdbdebe8bc2)
Eric Anholt [Sat, 16 May 2009 00:32:21 +0000 (17:32 -0700)]
intel: Don't complain on falling back from PBO fastpaths.
Instead, stash the debug info under the handy debug flag.
Bug #20053
(cherry picked from commit
22690482e692cb5ed2f84d3e69545c09292e3484)
Eric Anholt [Fri, 15 May 2009 23:24:59 +0000 (16:24 -0700)]
mesa: Mark FBOs with compressed color attachments as FBO-incomplete.
Both EXT_fbo and ARB_fbo agree on this. Fixes a segfault in the metaops
mipmap generation in Intel for SGIS_generate_mipmap of S3TC textures in
Regnum Online.
Bug #21654.
(cherry picked from commit
0307e609aa3e707eeb40051bd664d36f2340ba9b)
Eric Anholt [Fri, 15 May 2009 19:32:51 +0000 (12:32 -0700)]
i915: Use Stencil.Enabled instead of Stencil._Enabled in DrawBuffers.
The _Enabled field isn't updated at the point that DrawBuffers is called,
and the Driver.Enable() function does the testing for stencil buffer
presence anyway.
bug #21608 for Radeon
(cherry picked from commit
4c6f82989983eecc0b3b724716cb3bcb675664c5)
Eric Anholt [Fri, 15 May 2009 18:41:42 +0000 (11:41 -0700)]
i915: Only use the new 945 cube layout for compressed textures.
The docs actually explain this, but not in a terribly clear manner.
This nearly fixes the piglit cubemap testcase, except that something's
going wrong with the nearest filtering at 2x2 sizes in the testcase.
Looks good by visual inspection, though.
Bug #21692
(cherry picked from commit
5c5a46884899ea25cdf25545d6ab3d9a74eafa3a)
Eric Anholt [Thu, 14 May 2009 17:56:32 +0000 (10:56 -0700)]
i965: Fix varying payload reg assignment for the non-GLSL-instructions path.
I don't have a testcase for this, but it seems clearly wrong.
(cherry picked from commit
dc657f3929fbe03275b3fae4ef84f02e74b51114)
Eric Anholt [Thu, 14 May 2009 16:49:45 +0000 (09:49 -0700)]
i965: Fix register allocation of GLSL fp inputs.
Before, if the VP output something that is in the attributes coming into
the WM but which isn't used by the WM, then WM would end up reading subsequent
varyings from the wrong places. This was visible with a GLSL demo
using gl_PointSize in the VS and a varying in the WM, as point size is in
the VUE but not used by the WM. There is now a regression test in piglit,
glsl-unused-varying.
(cherry picked from commit
0f5113deed91611ecdda6596542530b1849bb161)
Eric Anholt [Thu, 14 May 2009 02:08:17 +0000 (19:08 -0700)]
intel: Use FRONT_AND_BACK for StencilOp as well.
(cherry picked from commit
64980125c76b05501a6fe7fe20fe52438f459129)
Eric Anholt [Thu, 14 May 2009 01:18:29 +0000 (18:18 -0700)]
intel: Use GL_FRONT_AND_BACK for stencil clearing.
This comes from a radeon-rewrite fallback fix, but may also fix stencil
clear failure when the polygon winding mode is flipped.
(cherry picked from commit
d866abeffc7e4a29736fa35fb8ac09c3a28a44d6)
Eric Anholt [Sun, 10 May 2009 17:08:32 +0000 (10:08 -0700)]
intel: Skip the DRI2 renderbuffer update when doing Viewport on an FBO.
(cherry picked from commit
d4a42b0ce6455d03be70aa56aacd779be193aca4)
Eric Anholt [Sun, 10 May 2009 16:45:43 +0000 (09:45 -0700)]
intel: Map write-only buffer objects through the GTT when possible.
This looks to be a win of a few percent in cairogears with new vbo code,
thanks to not polluting caches.
(cherry picked from commit
aa422b262509bc0763a50f63a51a1730139ea52f)