Kenneth Graunke [Thu, 30 Jun 2011 07:00:36 +0000 (00:00 -0700)]
i965: Emit 3DSTATE_VF_STATISTICS in OUT_BATCH style.
This is a little different from most because it's a single DWord;
there's no length field.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Thu, 30 Jun 2011 06:50:05 +0000 (23:50 -0700)]
i965: Convert 3DSTATE_GLOBAL_DEPTH_OFFSET_CLAMP to OUT_BATCH style.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Henri Verbeet [Sun, 3 Jul 2011 22:57:43 +0000 (00:57 +0200)]
mesa: Fix the BindSampler unit limit.
I'm not sure about this one. The current code actually follows the spec, but
considering the spec is supposed to be written against GL 3.2 I'd say the spec
is broken. I filled out a spec feedback form over a month ago, but either the
form is broken, or nobody cares.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Henri Verbeet [Sun, 3 Jul 2011 22:57:42 +0000 (00:57 +0200)]
mesa: Fix a couple of TexEnv unit limits.
NOTE: This is a candidate for the 7.11 branch.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Henri Verbeet [Sun, 3 Jul 2011 22:57:42 +0000 (00:57 +0200)]
mesa: Use the Elements macro for the sampler index assert in validate_samplers().
This is probably nicer if the array size ever changes.
NOTE: This is a candidate for the 7.11 branch.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Henri Verbeet [Sun, 3 Jul 2011 22:57:42 +0000 (00:57 +0200)]
mesa: Allow sampling from units >= MAX_TEXTURE_UNITS in shaders.
The total number of units used by a shader is limited to MAX_TEXTURE_UNITS,
but the actual indices are only limited by MAX_COMBINED_TEXTURE_IMAGE_UNITS,
since they're shared between vertex and fragment shaders.
NOTE: This is a candidate for the 7.11 branch.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Henri Verbeet [Sun, 3 Jul 2011 22:57:42 +0000 (00:57 +0200)]
mesa: Check the texture against all units in unbind_texobj_from_texunits().
NOTE: This is a candidate for the 7.11 branch.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Christoph Bumiller [Thu, 7 Jul 2011 12:58:29 +0000 (14:58 +0200)]
nv50,nvc0: use screen instead of context for flush notifier
Context may become NULL and we still have to be able to flush
pending fences.
Christoph Bumiller [Tue, 28 Jun 2011 14:06:56 +0000 (16:06 +0200)]
nv50: more formats for nv50_resource_copy_region conversion path
It's not supposed to do conversion, but st sometimes asks us to.
Sometimes conversion is even wrong (e.g. between UNORM and SRGB).
This should now include all formats the 2D engine supports.
Stéphane Marchesin [Thu, 7 Jul 2011 07:15:05 +0000 (00:15 -0700)]
i915g: Apply optimization to ADDS/MUL and only if we're not saturating.
Thomas Hellstrom [Thu, 7 Jul 2011 08:18:24 +0000 (10:18 +0200)]
st/xa: Fix up error reporting. Disable component alpha.
Component alpha is temporarily disabled since it seems a bit buggy.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Wed, 6 Jul 2011 19:58:33 +0000 (21:58 +0200)]
st/xa: Fix render to xa_format_a8, which is backed by a gallium L8 texture
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Brian Paul [Wed, 6 Jul 2011 23:18:32 +0000 (17:18 -0600)]
i915g: move code after declaration
Ian Romanick [Tue, 28 Jun 2011 00:59:58 +0000 (17:59 -0700)]
linker: Assign locations for fragment shader output
Fixes an assertion failure in the piglib out-01.frag
ARB_explicit_attrib_location test. The locations set via the layout
qualifier in fragment shader were not being applied to the shader
outputs. As a result all of these variables still had a location of
-1 set.
This may need some more work for pre-3.0 contexts. The problem is
dealing with generic outputs that lack a layout qualifier. There is
no way for the application to specify a location
(glBindFragDataLocation is not supported) or query the location
assigned by the linker (glGetFragDataLocation is not supported).
NOTE: This is a candidate for the 7.10 and 7.11 branches.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38624
Reviewed-by: Eric Anholt <eric@anholt.net>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Cc: Vinson Lee <vlee@vmware.com>
Ian Romanick [Sat, 25 Jun 2011 00:30:41 +0000 (17:30 -0700)]
glsl: Don't choke when printing an anonymous function parameter
NOTE: This is a candidate for the 7.10 and 7.11 branches.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38584
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Ian Romanick [Sat, 25 Jun 2011 00:12:31 +0000 (17:12 -0700)]
ir_to_mesa: Allocate temporary instructions on the visitor's ralloc context
And don't delete them. Let ralloc clean them up. Deleting the
temporary IR leaves dangling references in the prog_instruction. That
results in a bad dereference when printing the IR with MESA_GLSL=dump.
NOTE: This is a candidate for the 7.10 and 7.11 branches.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38584
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Ian Romanick [Mon, 27 Jun 2011 23:33:13 +0000 (16:33 -0700)]
glsl: Track initial mask in constant propagation live set
The set of values initially available (before any kills) must be
tracked with each constant in the set. Otherwise the wrong component
can be selected after earlier components have been killed.
NOTE: This is a candidate for the 7.10 and 7.11 branches.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=37383
Reviewed-by: Eric Anholt <eric@anholt.net>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Cc: Matthias Bentrup <matthias.bentrup@googlemail.com>
Stéphane Marchesin [Wed, 6 Jul 2011 21:10:51 +0000 (14:10 -0700)]
i915g: Fix optimize so that it actually gets used.
Brian Paul [Wed, 6 Jul 2011 21:04:17 +0000 (15:04 -0600)]
mesa: fix texstore addressing bugs for depth/stencil formats
Using GLuint pointers worked when the pixel size was four bytes
or the row stride was a multiple of four but was otherwise broken.
Fixes failures found with the piglit fbo-stencil test.
This helps to fix https://bugs.freedesktop.org/show_bug.cgi?id=38729
NOTE: This is a candidate for the 7.11 branch.
Brian Paul [Wed, 6 Jul 2011 21:02:58 +0000 (15:02 -0600)]
softpipe: add missing stencil format case in convert_quad_stencil()
Part of the fix for https://bugs.freedesktop.org/show_bug.cgi?id=38729
NOTE: This is a candidate for the 7.11 branch
Brian Paul [Wed, 6 Jul 2011 21:02:25 +0000 (15:02 -0600)]
tgsi: add some debug/print code in exec_tex(), disabled
Brian Paul [Wed, 6 Jul 2011 18:15:10 +0000 (12:15 -0600)]
softpipe: minor clean-ups in sp_quad_depth_test.c
Stéphane Marchesin [Wed, 6 Jul 2011 18:44:19 +0000 (11:44 -0700)]
i915g: Fix optimization, also make it more generic.
Eric Anholt [Tue, 28 Jun 2011 23:39:15 +0000 (16:39 -0700)]
mesa: Don't error on glFeedbackBuffer(size = 0, buffer = NULL)
The existing error result doesn't appear in the GL 2.1 or 3.2
compatibility specs, and triggers an unexpected GL error in Intel's
oglconform when it tries to reset the feedback state after usage so
that the "diff the state at error time vs. context init time" code
doesn't generate spurious diffs. The unexpected GL error then
translates into testcase failure. Brian wants the safety check on
buffer = NULL, though, so that people can't as easily set up a broken
buffer.
Eric Anholt [Tue, 28 Jun 2011 22:04:22 +0000 (15:04 -0700)]
mesa: Don't skip glGetProgramLocalParam4dvARB if there was already an error.
Like the previous commit, but fixes
ARB_vertex_program/getlocal4d-with-error.
v2: Move the success case line into the conditional, use ASSIGN_4V more.
Eric Anholt [Tue, 28 Jun 2011 20:55:44 +0000 (13:55 -0700)]
mesa: Don't skip glGetProgramEnvParam4dvARB if there was already an error.
Fixes a bug caught by oglconform, and now piglit
ARB_vertex_program/getenv4d-with-error. The wrapping of an existing
GL function made it so that we couldn't distinguish an error in
looking up our arguments from an existing error. Instead, make a
helper function to choose the param, and use it from multiple callers.
v2: Move the success case line into the conditional, use COPY_4V more.
Eric Anholt [Wed, 6 Jul 2011 05:02:33 +0000 (22:02 -0700)]
glsl: Suppress warning from matching_signature change.
gcc isn't smart enough to see that we only look at matched_score after
we've initialized it (because match != NULL happens at the same time)
Eric Anholt [Wed, 6 Jul 2011 04:59:33 +0000 (21:59 -0700)]
glsl: Fix make clean for dricore.
Kenneth Graunke [Wed, 29 Jun 2011 23:47:30 +0000 (16:47 -0700)]
i965/fs: Fix message register allocation in FB writes.
Commit
6750226e6d915742ebf96bae2cfcdd287b85db35 bumped the base MRF to
m2 instead of m0, but failed to adjust inst->mlen, which was being set
to the highest MRF. Subtracting the base MRF solves the issue.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Völgyes Dávid [Wed, 6 Jul 2011 14:10:20 +0000 (10:10 -0400)]
Fixes for leaks reported by cppcheck.
Stéphane Marchesin [Wed, 6 Jul 2011 09:19:48 +0000 (02:19 -0700)]
i915g: Improve flushing using heuristics.
Stéphane Marchesin [Tue, 5 Jul 2011 03:21:37 +0000 (20:21 -0700)]
i915g: introduce the tiny shader optimizer.
Stéphane Marchesin [Mon, 4 Jul 2011 02:44:02 +0000 (19:44 -0700)]
i915g: Add comment about DDX/DDY.
Stéphane Marchesin [Mon, 4 Jul 2011 02:43:19 +0000 (19:43 -0700)]
i915g: Move back to the old method for target format fixup.
Vadim Girlin [Mon, 4 Jul 2011 14:30:42 +0000 (18:30 +0400)]
r600g: fix buffer overflow check in r600_query_begin
Vadim Girlin [Mon, 4 Jul 2011 14:30:41 +0000 (18:30 +0400)]
r600g: fix bo map usage flags in r600_query_begin
Vadim Girlin [Mon, 4 Jul 2011 14:30:40 +0000 (18:30 +0400)]
r600g: reduce flushes for queries
Vadim Girlin [Mon, 4 Jul 2011 14:30:39 +0000 (18:30 +0400)]
r600g: fix buffer offset in r600_query_begin
Henri Verbeet [Mon, 4 Jul 2011 23:58:47 +0000 (01:58 +0200)]
r600g: Get rid of the state_inlines headers.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Henri Verbeet [Mon, 4 Jul 2011 23:58:46 +0000 (01:58 +0200)]
r600g: Get rid of an unused include in r600_texture.c.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Henri Verbeet [Mon, 4 Jul 2011 23:58:46 +0000 (01:58 +0200)]
r600g: Get rid of some unused functions.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Henri Verbeet [Mon, 4 Jul 2011 23:58:46 +0000 (01:58 +0200)]
r600g: Use the actual Evergreen functions to query format support on Evergreen.
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Henri Verbeet [Mon, 4 Jul 2011 23:58:46 +0000 (01:58 +0200)]
r600g: Get rid of the unused "family" parameter to r600_is_vertex_format_supported().
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
José Fonseca [Mon, 4 Jul 2011 11:31:30 +0000 (12:31 +0100)]
draw: Fix fetch_max calculation.
It should be
max_index = start + count - 1
instead of
max_index = count - 1
Thomas Hellstrom [Mon, 4 Jul 2011 08:21:35 +0000 (10:21 +0200)]
Merge branch 'xa_branch'
Conflicts:
configure.ac
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Sat, 2 Jul 2011 08:37:17 +0000 (10:37 +0200)]
st/glx: Fix endless loop in drawable validation
This fixes a regression introduced with commit
"st-api: Rework how drawables are invalidated v3"
where the glx state tracker manager would invalidate a drawable each time it
checks the drawable dimensions, even during a validate call, which
resulted in an endless loop, since the state tracker would immediately
detect the new invalidation and rerun the validate...
This change marks the drawable invalid only if the drawable dimensions actually
changed during the validate, which will result in at most a single
unnecessary validate by the context running a validate during which the
dimensions changed.
To avoid unnecessary validates altogether, we need to implement yet another
st-api change: Returning the current time stamp from the validate function,
as suggested by Chia-I Wu. The glx state tracker manager could then return
the stamp resulting from the last drawable dimension check.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Mon, 4 Jul 2011 07:26:27 +0000 (09:26 +0200)]
st/xa: Fix crosscompile builds with nonstandard ld locations
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Mon, 4 Jul 2011 07:24:19 +0000 (09:24 +0200)]
xa-vmwgfx: Remove unnecessary include
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Henri Verbeet [Sun, 3 Jul 2011 20:29:34 +0000 (22:29 +0200)]
r600g: Support the PIPE_FORMAT_R16_FLOAT colorformat.
NOTE: This is a candidate for the 7.11 branch.
Benjamin Franzke [Sat, 2 Jul 2011 11:46:42 +0000 (13:46 +0200)]
configure: Require libudev for drm & wayland egl platforms
NOTE: This is a candidate for the 7.11 branch.
Benjamin Franzke [Sat, 2 Jul 2011 11:46:09 +0000 (13:46 +0200)]
configure: Fix typo in gbm check for egl drm platform
NOTE: This is a candidate for the 7.11 branch.
Benjamin Franzke [Sat, 2 Jul 2011 11:45:14 +0000 (13:45 +0200)]
configure: Enable st/gbm if st/egl has drm platform
NOTE: This is a candidate for the 7.11 branch.
Benjamin Franzke [Sat, 2 Jul 2011 11:41:35 +0000 (13:41 +0200)]
egl_dri2: Fix compilation if udev devel files are not installed
NOTE: This is a candidate for the 7.11 branch.
Chia-I Wu [Sat, 2 Jul 2011 08:57:30 +0000 (17:57 +0900)]
egl: add copyright notices
The list of copyright holders could be incomplete. Please update
directly or notify me if your name is missing.
Chia-I Wu [Sat, 2 Jul 2011 00:49:17 +0000 (09:49 +0900)]
autoconf: swrast does not require libdrm
This fixes
$ ./configure --disable-driglx-direct \
--with-dri-drivers=swrast \
--with-gallium-drivers=
José Fonseca [Fri, 1 Jul 2011 18:04:57 +0000 (19:04 +0100)]
scons: Disable saving options automatically.
It makes things too random, as settings for temporary trials get stored
permannently, and it make difficult to build several platforms from the
same tree.
So disable it, again.
José Fonseca [Fri, 1 Jul 2011 16:06:58 +0000 (17:06 +0100)]
svga: Use the correct element size.
Instead of always using the first element's size.
This fixes flashing floor on CINEBENCH R10.
José Fonseca [Fri, 1 Jul 2011 16:05:30 +0000 (17:05 +0100)]
u_upload_mgr: Use pipe_buffer_map_range istead of pipe_buffer_map.
pipe_buffer_map_range makes stricter assertions, and would have saved us
grief detecting a bug in svga user buffer uploads.
Micael Dias [Thu, 30 Jun 2011 02:33:47 +0000 (03:33 +0100)]
Gallium: fix buffer overflow
Signed-off-by: José Fonseca <jfonseca@vmware.com>
Kristian Høgsberg [Fri, 1 Jul 2011 13:34:34 +0000 (09:34 -0400)]
driclient: Remove
Not used or maintained. We talked about removing it a few releases ago,
there were no objections but it just never happened. Now it's gone.
José Fonseca [Fri, 1 Jul 2011 12:51:18 +0000 (13:51 +0100)]
scons: Actually add src/glx/SConscript.
Thomas Hellstrom [Thu, 30 Jun 2011 06:52:19 +0000 (08:52 +0200)]
svga: Fix multiple uploads of the same user-buffer.
If a user-buffer was referenced twice by a draw command, the affected ranges
were uploaded separately, with only the last one being referenced by the
hardware. Make sure we upload only a single range.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Wed, 29 Jun 2011 11:17:02 +0000 (13:17 +0200)]
svga: Ignore redefine_user_buffer calls
We currently always treat contents of user-buffers as volatile so
we don't need to take any particular action when the state tracker
announces that the contents has changed.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Brian Paul [Thu, 16 Jun 2011 16:31:53 +0000 (10:31 -0600)]
svga: fix incorrect user buffer size computation for instance divisor case
See preceeding commit for more info.
Brian Paul [Wed, 15 Jun 2011 22:41:13 +0000 (16:41 -0600)]
svga: fix incorrect user buffer size computation
Viewperf uses some unusual vertex arrays where the stride is less
than the element size. In this case, the stride was 4 while the
element size was 12. The difference of 8 bytes causes us to miss
uploading the tail bit of the array data.
Typically the stride is >= the element size so there was no problem
with other apps.
José Fonseca [Tue, 31 May 2011 12:33:34 +0000 (13:33 +0100)]
svga: Handle null buffers in svga_buffer_is_user_buffer().
Thomas Hellstrom [Thu, 17 Mar 2011 20:09:38 +0000 (21:09 +0100)]
gallium/svga: Upload only parts of user-buffers that we actually use
Stream user buffer contents rather than trying to maintain persistent
host / hardware copies.
Resulting negative array offsets are not allowed by the hardware,
(well, at least not according to header files), so adjust index bias
to make all array offsets positive.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Thu, 17 Mar 2011 14:09:48 +0000 (15:09 +0100)]
gallium/svga: Make use of u_upload_flush().
This enables us to pack more data into single upload buffers.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Thomas Hellstrom [Thu, 10 Mar 2011 13:53:46 +0000 (14:53 +0100)]
gallium/util: Upload manager optimizations
Make sure that the upload manager doesn't upload data that's not
dirty. This speeds up the viewperf test proe-04/1 a factor 5 or so on svga.
Also introduce an u_upload_unmap() function that can be used
instead of u_upload_flush() so that we can pack
even more data in upload buffers. With this we can basically reuse the
upload buffer across flushes.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
José Fonseca [Thu, 17 Mar 2011 16:08:21 +0000 (16:08 +0000)]
svga: Flush when switching between HW to SW TNL, after updating need_swtnl.
Also, only flush when going from HW TNL to SW TNL, given it is impossible
for the buffers resulting from SWTNL to be ever referred by HW TNL path.
Chia-I Wu [Fri, 1 Jul 2011 09:12:16 +0000 (18:12 +0900)]
egl: fix a compiler warning
Chia-I Wu [Fri, 1 Jul 2011 09:13:38 +0000 (18:13 +0900)]
st/egl: fix a typo
Wrong goto labels.
Chia-I Wu [Fri, 1 Jul 2011 09:19:31 +0000 (18:19 +0900)]
st/egl: remove unused header
Chia-I Wu [Mon, 27 Jun 2011 00:19:02 +0000 (09:19 +0900)]
autoconf: fix --disable-glx
libdrm is used in multiple places. Always check for it and set
have_libdrm. Each user can then check the variable.
This is useful when only EGL and DRI drivers are needed.
Chia-I Wu [Mon, 27 Jun 2011 04:06:39 +0000 (13:06 +0900)]
autoconf: fix --disable-dri
Define GLX_INDIRECT_RENDERING and GLX_DIRECT_RENDERING when $enable_glx,
not $enable_dri.
Chia-I Wu [Sun, 26 Jun 2011 04:24:32 +0000 (13:24 +0900)]
autoconf: add --enable-{dri,glx,osmesa}
The idea is that DRI driver, libGL and libOSMesa are libraries that can
be independently enabled, yet --with-driver does not allow us to easily
do that, if not impossible. This also matches what
--enable-{egl,xorg,d3d1x} do for the respective libraries.
There are two libGL providers: Xlib-based and DRI-based. They cannot
coexist. To be able to choose between them, --enable-xlib-glx is also
added.
With this commit, --with-driver=dri can be replaced by
$ ./configure --enable-dri --enable-glx --disable-osmesa
--with-driver=xlib can be replaced by
$ ./configure --disable-dri --enable-glx --enable-osmesa \
--enable-xlib-glx
and --with-driver=osmesa can be replaced by
$ ./configure --disable-dri --disable-glx --enable-osmesa
Some combinations that cannot be supported with --with-driver will
produce errors at the moment. But in the future, we would like to
support, for example,
$ ./configure --enable-dri --disable-glx --enable-egl
(build libEGL and DRI drivers, but not libGL)
Note that this commit still keeps --with-driver for transitional
purpose.
Vinson Lee [Thu, 30 Jun 2011 22:48:41 +0000 (15:48 -0700)]
gallivm: Pass in CPU name to createTargetMachine when on llvm-3.0.
llvm-3.0svn revision 134127 changed createTargetMachine to take in
an additional argument of the CPU name.
Vinson Lee [Thu, 30 Jun 2011 22:07:57 +0000 (15:07 -0700)]
gallivm: Rename TargetInstrDesc to MCInstrDesc when using llvm-3.0.
llvm-3.0svn revision 134021 renamed TargetInstrDesc to MCInstrDesc.
Marek Olšák [Thu, 30 Jun 2011 20:53:13 +0000 (22:53 +0200)]
docs: add GL3 GLX todo
Vadim Girlin [Wed, 29 Jun 2011 12:29:18 +0000 (16:29 +0400)]
r600g: fix check for empty cs
Marek Olšák [Thu, 30 Jun 2011 19:27:46 +0000 (21:27 +0200)]
docs: update GL3 status - vertex textures
They work fine on r600g.
José Fonseca [Thu, 30 Jun 2011 16:36:37 +0000 (17:36 +0100)]
scons: Expose pkg-config in a simpler manner.
José Fonseca [Fri, 17 Jun 2011 19:12:18 +0000 (20:12 +0100)]
scons: Buid libGL.so (WIP).
José Fonseca [Thu, 30 Jun 2011 09:43:57 +0000 (10:43 +0100)]
scons: Make declaration-after-statement and pointer-arith just warnings.
Necessary, in order to build the whole tree.
Stéphane Marchesin [Thu, 30 Jun 2011 09:05:50 +0000 (02:05 -0700)]
i915g: Implement surface format fixup without adding an extra instruction.
We also avoid writing output color twice, which might not work when we run out of phases.
Stéphane Marchesin [Thu, 30 Jun 2011 04:59:01 +0000 (21:59 -0700)]
i915g: update TODO.
Chia-I Wu [Thu, 30 Jun 2011 01:23:50 +0000 (10:23 +0900)]
target/egl-static: fix a compiler warning
Stéphane Marchesin [Wed, 29 Jun 2011 23:52:28 +0000 (16:52 -0700)]
i915g: Try to do better in the shader compiler.
- Copy i915c's support for phases, that should allow us to run a coupe more shaders.
- Fix the error messages.
- Still try to proceed when we get a shader that's too long.
Stéphane Marchesin [Wed, 29 Jun 2011 03:45:07 +0000 (20:45 -0700)]
Gallium:draw:aaline and aapoint: Restore the old hooks when we destroy our stage.
Kenneth Graunke [Wed, 15 Jun 2011 06:13:27 +0000 (23:13 -0700)]
glsl: Don't use MOD_TO_FRACT lowering on GLSL 1.30's % operator.
MOD_TO_FRACT was designed to lower the GLSL 1.20 mod() function, which
operates on floating point values. However, we also use ir_binop_mod
for GLSL 1.30's % operator, which operates on integers.
For now, make MOD_TO_FRACT only apply to floating-point mod operations.
In the future, we may want to add a lowering pass for integer-based mod.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Wed, 15 Jun 2011 05:47:04 +0000 (22:47 -0700)]
glsl: Fix DIV_TO_MUL_RCP lowering for uint result types.
f2i results in an int/ivec; we need i2u to get a uint/uvec.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Wed, 15 Jun 2011 05:21:41 +0000 (22:21 -0700)]
glsl: Distinguish "type mismatch" error messages for modulus operator.
Previously, it would simply say "type error" in three different cases:
- The LHS is not an integer
- The RHS is not an integer
- The LHS and RHS have different base types (int vs. uint)
Now the error messages state the specific problem.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Tue, 14 Jun 2011 23:28:32 +0000 (16:28 -0700)]
glsl: Find the "closest" signature when there are multiple matches.
Previously, ir_function::matching_signature had a fatal bug: if a
function had more than one non-exact match, it would simply return NULL.
This occured, for example, when looking for max(uvec3, uvec3):
- max(vec3, vec3) -> score 1 (found first)
- max(ivec3, ivec3) -> score 1 (found second...used to return NULL here)
- max(uvec3, uvec3) -> score 0 (exact match...the right answer)
This did not occur for max(ivec3, ivec3) since the second match found
was an exact match.
The new behavior is to return a match with the lowest score. If there
is an exact match, that will be returned. Otherwise, a match with the
least number of implicit conversions is chosen.
Fixes piglit tests max-uvec3.vert and glsl-inexact-overloads.shader_test.
NOTE: This is a candidate for the 7.10 and 7.11 branches.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Wed, 15 Jun 2011 06:23:49 +0000 (23:23 -0700)]
glsl: Use i2u and u2i to implement constructor conversions.
Inspired by a patch from Bryan Cain <bryancain3@gmail.com>.
Fixes piglit tests:
- ctor-int-uint.vert
- ctor-ivec4-uvec4.vert
- ctor-uint-int.vert
- ctor-uvec4-ivec4.vert
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Tue, 14 Jun 2011 20:04:32 +0000 (13:04 -0700)]
i965/fs: Implement new ir_unop_u2i and ir_unop_i2u opcodes.
No MOV is necessary since signed/unsigned integers share the same
bit-representation; it's simply a question of interpretation. In
particular, the fs_reg::imm union shouldn't need updating.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Wed, 15 Jun 2011 04:41:10 +0000 (21:41 -0700)]
ir_to_mesa: "Support" u2f, i2u, and u2i operations by doing nothing.
Mesa IR actually stores all numbers as floating point, so this is
totally a farce, but we may as well keep it going.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Kenneth Graunke [Tue, 14 Jun 2011 19:40:09 +0000 (12:40 -0700)]
glsl: Revert "fix conversions from uint to bool and from..."
Reverts commit
f41e1db3273a31285360241c4342f0a403ee0b03
"fix conversions from uint to bool and from float/bool to uint"
f2i, b2i, and b2i should not accept uint types. Use i2u and u2i.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Bryan Cain [Wed, 15 Jun 2011 06:34:11 +0000 (23:34 -0700)]
glsl: Add ir_unop_i2u and ir_unop_u2i operations.
These are necessary to handle int/uint constructor conversions. For
example, the following code currently results in a type mismatch:
int x = 7;
uint y = uint(x);
In particular, uint(x) still has type int.
This commit simply adds the new operations; it does not generate them,
nor does it add backend support for them.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Eric Anholt [Fri, 24 Jun 2011 20:21:29 +0000 (13:21 -0700)]
glsl: Use the default values of ir_assignment() in lower_mat_op_to_vec.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Eric Anholt [Fri, 24 Jun 2011 20:17:07 +0000 (13:17 -0700)]
glsl: Allow ir_assignment() constructor to not specify condition.
We almost never want to specify a condition, and when we do we're
already thinking about it (because we're writing a lowering pass
generating the condition), so a default argument should make the code
more pleasant to read.
NOTE: This is a candidate for the 7.11 branch (we want to be able to
cherry-pick future code).
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>