Christian König [Mon, 10 Jan 2011 21:09:41 +0000 (22:09 +0100)]
r600g: place buffers into different domains, based on usage flags
Christian König [Mon, 10 Jan 2011 18:15:42 +0000 (19:15 +0100)]
[g3dvl] move to integer verticies
Christian König [Sun, 9 Jan 2011 13:19:14 +0000 (14:19 +0100)]
[g3dvl] use a table of empty block mask instead of calculating it
Christian König [Sun, 9 Jan 2011 12:18:48 +0000 (13:18 +0100)]
r600g: check if hardware blits are possible bevore enabling tilling
Christian König [Sun, 9 Jan 2011 12:18:30 +0000 (13:18 +0100)]
r600g: some merge fixes
Christian König [Sat, 8 Jan 2011 12:24:36 +0000 (13:24 +0100)]
Merge remote branch 'origin/master' into pipe-video
Conflicts:
configure.ac
src/gallium/drivers/r600/eg_asm.c
src/gallium/drivers/r600/r600_asm.c
src/gallium/drivers/r600/r600_asm.h
src/gallium/include/pipe/p_format.h
src/gallium/targets/dri-nouveau/Makefile
Vinson Lee [Sat, 8 Jan 2011 11:03:17 +0000 (03:03 -0800)]
mesa: Clean up header file inclusion in cpuinfo.c.
Christian König [Sat, 8 Jan 2011 10:11:48 +0000 (11:11 +0100)]
r600g: join export instructions
Marek Olšák [Sat, 8 Jan 2011 08:42:17 +0000 (09:42 +0100)]
r300g: fix a surface leak when flushing ZMASK
Marek Olšák [Sat, 4 Dec 2010 03:38:15 +0000 (04:38 +0100)]
r300g: rework command submission and resource space checking
The motivation behind this rework is to get some speed by reducing
CPU overhead. The performance increase depends on many factors,
but it's measurable (I think it's about 10% increase in Torcs).
This commit replaces libdrm's radeon_cs_gem with our own implemention.
It's optimized specifically for r300g, but r600g could use it as well.
Reloc writes and space checking are faster and simpler than their
counterparts in libdrm (the time complexity of all the functions
is O(1) in nearly all scenarios, thanks to hashing).
(libdrm's radeon_bo_gem is still being used in the driver.)
It works like this:
cs_add_reloc(cs, buf, read_domain, write_domain) adds a new relocation and
also adds the size of 'buf' to the used_gart and used_vram winsys variables
based on the domains, which are simply or'd for the accounting purposes.
The adding is skipped if the reloc is already present in the list, but it
accounts any newly-referenced domains.
cs_validate is then called, which just checks:
used_vram/gart < vram/gart_size * 0.8
The 0.8 number allows for some memory fragmentation. If the validation
fails, the pipe driver flushes CS and tries do the validation again,
i.e. it validates only that one operation. If it fails again, it drops
the operation on the floor and prints some nasty message to stderr.
cs_write_reloc(cs, buf) just writes a reloc that has been added using
cs_add_reloc. The read_domain and write_domain parameters have been removed,
because we already specify them in cs_add_reloc.
The space checking has been tested by putting small values in vram/gart_size
variables.
Eric Anholt [Sat, 8 Jan 2011 02:18:50 +0000 (18:18 -0800)]
intel: Make renderbuffer tiling choice match texture tiling choice.
There really shouldn't be any difference between the two for us.
Fixes a bug where Z16 renderbuffers would be untiled on gen6, likely
leading to hangs.
Eric Anholt [Sat, 8 Jan 2011 02:14:37 +0000 (18:14 -0800)]
intel: Use the _BaseFormat from MESA_FORMAT_* in renderbuffer setup.
Marek Olšák [Sat, 8 Jan 2011 01:54:51 +0000 (02:54 +0100)]
docs: fix messed up names with special characters in relnotes-7.9.1
(cherry picked from commit
67aeab0b77fb6be864088e69ea74a010b6543fa1)
Marek Olšák [Sat, 8 Jan 2011 02:06:04 +0000 (03:06 +0100)]
docs: fix messed up names with special characters in relnotes-7.10
(cherry picked from commit
36009724fdd652ab29aa928ba78891afd650e768)
Eric Anholt [Sat, 8 Jan 2011 00:37:17 +0000 (16:37 -0800)]
i915: Drop old checks for the settexoffset hack.
Eric Anholt [Sat, 8 Jan 2011 00:13:12 +0000 (16:13 -0800)]
i915: Don't claim to support AL1616 when neither 830 nor 915 does it.
Fixes an abort in fbo-generatemipmap-formats.
Eric Anholt [Fri, 7 Jan 2011 23:45:33 +0000 (15:45 -0800)]
intel: Add a vtbl hook for determining if a format is renderable.
By relying on just intel_span_supports_format, some formats that
aren't supported pre-gen4 were not reporting FBO incomplete. And we
also complained in stderr when it happened on i915 because draw_region
gets called before framebuffer completeness validation.
Eric Anholt [Fri, 7 Jan 2011 23:15:49 +0000 (15:15 -0800)]
intel: expose ARB_framebuffer_object in the i915 driver.
ARB_fbo no longer disallows mismatched width/height on attachments
(shouldn't be any problem), mixed format color attachments (we only
support 1), and L/A/LA/I color attachments (we already reject them on
965 too). It requires Gen'ed names (driver doesn't care), and adds
FramebufferTextureLayer (we don't do texture arrays). So it looks
like we're already in the position we need to be for this extension.
Bug #27468, #32381.
Christoph Bumiller [Fri, 7 Jan 2011 23:29:41 +0000 (00:29 +0100)]
nvc0: fix reloc domain conflict on buffer migration
Occurred because the code assumed that buf->domain would remain
equal to old_domain.
Christoph Bumiller [Fri, 7 Jan 2011 23:49:17 +0000 (00:49 +0100)]
nvc0: upload user buffers only from draw info min to max index
There are actually applications that profit immensely from this.
Christoph Bumiller [Fri, 7 Jan 2011 23:20:53 +0000 (00:20 +0100)]
nvc0: fix emission of first 3 u8 indices to RING_NI
Christoph Bumiller [Thu, 6 Jan 2011 15:11:36 +0000 (16:11 +0100)]
nvc0: reset mt transfer address after read loop over layers
Christoph Bumiller [Wed, 5 Jan 2011 00:41:49 +0000 (01:41 +0100)]
nvc0: tie buffer memory release to the buffer fence
... instead of the next fence to be emitted. This way we have a
chance to reclaim the storage earlier.
Łukasz Krotowski [Fri, 7 Jan 2011 23:46:38 +0000 (00:46 +0100)]
r300g: Remove invalid assertion.
Invalid after
be1af4394e060677b7db6bbb8e3301e38a3363da (user buffer
creation with width0 == ~0).
Signed-off-by: Marek Olšák <maraeo@gmail.com>
Ian Romanick [Fri, 7 Jan 2011 22:37:08 +0000 (14:37 -0800)]
docs: Import 7.10 release notes from 7.10 branch
Eric Anholt [Fri, 7 Jan 2011 22:26:39 +0000 (14:26 -0800)]
i965: Avoid double-negation of immediate values in the VS.
In general, we have to negate in immediate values we pass in because
the src1 negate field in the register description is in the bits3 slot
that the 32-bit value is loaded into, so it's ignored by the hardware.
However, the src0 negate field is in bits1, so after we'd negated the
immediate value loaded in, it would also get negated through the
register description. This broke this VP instruction in the position
calculation in civ4:
MAD TEMP[1], TEMP[1], CONST[256].zzzz, CONST[256].-y-y-y-y;
Bug #30156
Ian Romanick [Fri, 7 Jan 2011 21:38:26 +0000 (13:38 -0800)]
docs: Import 7.9.1 release notes from 7.9 branch
Christian König [Fri, 7 Jan 2011 17:26:51 +0000 (18:26 +0100)]
r600g: fully implement barrier handling
Henri Verbeet [Fri, 7 Jan 2011 16:06:11 +0000 (17:06 +0100)]
r600g: Also set const_offset if the buffer is not a user buffer in r600_upload_const_buffer().
Henri Verbeet [Fri, 7 Jan 2011 16:06:11 +0000 (17:06 +0100)]
r600g: Update some comments for Evergreen.
Henri Verbeet [Fri, 7 Jan 2011 16:06:11 +0000 (17:06 +0100)]
r600g: Split ALU clauses based on used constant cache lines.
Henri Verbeet [Fri, 7 Jan 2011 16:06:11 +0000 (17:06 +0100)]
r600g: Consistently use the copy of the alu instruction in r600_bc_add_alu_type().
Henri Verbeet [Fri, 7 Jan 2011 16:06:11 +0000 (17:06 +0100)]
r600g: Store kcache settings as an array.
Marek Olšák [Sun, 26 Dec 2010 03:29:44 +0000 (04:29 +0100)]
r300g: derive user buffer sizes at draw time
This only uploads the [min_index, max_index] range instead of [0, userbuf size],
which greatly speeds up user buffer uploads.
This is also a prerequisite for atomizing vertex arrays in st/mesa.
Jian Zhao [Wed, 5 Jan 2011 02:41:20 +0000 (10:41 +0800)]
mesa: fix an error in uniform arrays in row calculating.
Fix the error in uniform row calculating, it may alloc one line
more which may cause out of range on memory usage, sometimes program
aborted when free the memory.
NOTE: This is a candidate for 7.9 and 7.10 branches.
Signed-off-by: Brian Paul <brianp@vmware.com>
Vinson Lee [Fri, 7 Jan 2011 08:08:24 +0000 (00:08 -0800)]
mesa: Directly include mfeatures.h in files that perform feature tests.
Alex Deucher [Fri, 7 Jan 2011 08:10:50 +0000 (03:10 -0500)]
r600c: fix up SQ setup in blit code for Ontario/NI
Dave Airlie [Fri, 7 Jan 2011 00:35:41 +0000 (10:35 +1000)]
r600g: allow constant buffers to be user buffers.
This provides an upload facility for the constant buffers since Marek's
constants in user buffers changes.
gears at least work on my evergreen now.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Alex Deucher [Thu, 6 Jan 2011 23:40:17 +0000 (18:40 -0500)]
r600c: add support for NI asics
Alex Deucher [Thu, 6 Jan 2011 23:05:16 +0000 (18:05 -0500)]
r600g: add support for NI (Northern Islands) GPUs
This adds support for Barts, Turks, and Caicos asics.
Kenneth Graunke [Tue, 4 Jan 2011 22:34:55 +0000 (14:34 -0800)]
i965: Rename various gen6 #defines to match the documentation.
This should make it easier to cross-reference the code and hardware
documentation, as well as clear up any confusion on whether constants
like CMD_3D_WM_STATE mean WM_STATE (pre-gen6) or 3DSTATE_WM (gen6+).
This does not rename any pre-gen6 defines.
Jakob Bornecrantz [Thu, 16 Dec 2010 03:13:21 +0000 (04:13 +0100)]
svga: Ensure that the wrong vdecls don't get used in swtnl path
The draw module set new state that didn't require swtnl which caused need_swtnl to
be unset. This caused the call from to svga_update_state(svga, SVGA_STATE_SWTNL_DRAW)
from the vbuf backend to overwrite the vdecls we setup there to be overwritten with
the real buffers vdecls.
Ian Romanick [Thu, 6 Jan 2011 18:53:38 +0000 (10:53 -0800)]
glsl: Refresh autogenerated lexer and parser files.
For the previous commit.
Ian Romanick [Thu, 6 Jan 2011 18:49:56 +0000 (10:49 -0800)]
glsl: Support the 'invariant(all)' pragma
Previously the 'STDGL invariant(all)' pragma added in GLSL 1.20 was
simply ignored by the compiler. This adds support for setting all
variable invariant.
In GLSL 1.10 and GLSL ES 1.00 the pragma is ignored, per the specs,
but a warning is generated.
Fixes piglit test glsl-invariant-pragma and bugzilla #31925.
NOTE: This is a candidate for the 7.9 and 7.10 branches.
Ian Romanick [Wed, 5 Jan 2011 00:09:00 +0000 (16:09 -0800)]
glsl: Allow less restrictive uses of sampler array indexing in GLSL <= 1.20
GLSL 1.10 and 1.20 allow any sort of sampler array indexing.
Restrictions were added in GLSL 1.30. Commit
f0f2ec4d added support
for the 1.30 restrictions, but it broke some valid 1.10/1.20 shaders.
This changes the error to a warning in GLSL 1.10, GLSL 1.20, and GLSL
ES 1.00.
There are some spurious whitespace changes in this commit. I changed
the layout (and wording) of the error message so that all three cases
would be similar. The 1.10/1.20 and 1.30 text is the same. The only
difference is that one is an error, and the other is a warning. The
GLSL ES 1.00 wording is similar but not quite the same.
Fixes piglit test
spec/glsl-1.10/compiler/constant-expressions/sampler-array-index-02.frag
and bugzilla #32374.
Marek Olšák [Thu, 6 Jan 2011 18:02:12 +0000 (19:02 +0100)]
r300g: fix corruption when nr_cbufs==0 and multiwrites enabled
https://bugs.freedesktop.org/show_bug.cgi?id=32634
Marek Olšák [Sat, 25 Dec 2010 13:49:17 +0000 (14:49 +0100)]
r300g: remove the buffer range checking
It's no longer needed because the upload buffer remains mapped while the CS
is being filled (openarena, ut2004 and others that this code was for do not
use VBOs by default).
Marek Olšák [Mon, 27 Dec 2010 21:20:58 +0000 (22:20 +0100)]
r300g: skip buffer validation of upload buffers when appropriate
because the upload buffers are reused for subsequent draw operations.
Marek Olšák [Wed, 5 Jan 2011 23:26:02 +0000 (00:26 +0100)]
util: add comments to u_upload_mgr and u_inlines
Marek Olšák [Sun, 26 Dec 2010 00:31:42 +0000 (01:31 +0100)]
vbo: remove a redundant call to _ae_invalidate_state
It's called in vbo_exec_invalidate_state too.
Marek Olšák [Wed, 29 Dec 2010 02:35:45 +0000 (03:35 +0100)]
st/mesa: remove unused members in st_context
What were these for?
Marek Olšák [Sun, 26 Dec 2010 21:52:25 +0000 (22:52 +0100)]
tgsi: remove redundant name tables from tgsi_text, use those from tgsi_dump
I also specified the array sizes in the header so that one can use
the Elements macro on it.
Marek Olšák [Sun, 26 Dec 2010 17:43:39 +0000 (18:43 +0100)]
gallium: drivers should reference vertex buffers
So that a state tracker can unreference them after set_vertex_buffers.
Marek Olšák [Wed, 29 Dec 2010 02:17:43 +0000 (03:17 +0100)]
st/mesa: optimize constant buffer uploads
The overhead of resource_create, transfer_inline_write, and resource_destroy
to upload constant data is very visible with some apps in sysprof, and
as such should be eliminated.
My approach uses a user buffer to pass a pointer to a driver. This gives
the driver the freedom it needs to take the fast path, which may differ
for each driver.
This commit addresses the same issue as Jakob's one that suballocates out
of a big constant buffer, but it also eliminates the copy to the buffer.
Marek Olšák [Mon, 27 Dec 2010 17:22:05 +0000 (18:22 +0100)]
st/mesa: do sanity checks on states only in debug builds
Marek Olšák [Mon, 27 Dec 2010 21:32:31 +0000 (22:32 +0100)]
u_upload_mgr: new features
- Added a parameter to specify a minimum offset that should be returned.
r300g needs this to better implement user buffer uploads. This weird
requirement comes from the fact that the Radeon DRM doesn't support negative
offsets.
- Added a parameter to notify a driver that the upload flush occured.
A driver may skip buffer validation if there was no flush, resulting
in a better performance.
- Added a new upload function that returns a pointer to the upload buffer
directly, so that the buffer can be filled e.g. by the translate module.
Marek Olšák [Fri, 24 Dec 2010 12:47:32 +0000 (13:47 +0100)]
u_upload_mgr: keep the upload buffer mapped until it is flushed
The map/unmap overhead can be significant even though there is no waiting on busy
buffers. There is simply a huge number of uploads.
This is a performance optimization for Torcs, a car racing game.
Pierre Allegraud [Thu, 6 Jan 2011 14:58:57 +0000 (07:58 -0700)]
mesa: fix build for NetBSD
See http://bugs.freedesktop.org/show_bug.cgi?id=32859
NOTE: This is a candidate for the 7.9 and 7.10 branches.
Signed-off-by: Brian Paul <brianp@vmware.com>
Brian Paul [Tue, 4 Jan 2011 23:34:12 +0000 (16:34 -0700)]
glext: upgrade to version 67
Vinson Lee [Thu, 6 Jan 2011 08:45:08 +0000 (00:45 -0800)]
mesa: Clean up header file inclusion in version.c.
Include imports.h directly instead of indirectly through context.h.
version.c does use any symbols that are added by context.h.
Vinson Lee [Thu, 6 Jan 2011 08:06:38 +0000 (00:06 -0800)]
nvc0: Fix typo of nvc0_mm.c in SConscript.
Vinson Lee [Thu, 6 Jan 2011 07:27:30 +0000 (23:27 -0800)]
mesa: bump version to 7.11
Vinson Lee [Thu, 6 Jan 2011 07:11:54 +0000 (23:11 -0800)]
mesa: Include mtypes.h in files that use gl_context struct.
Directly include mtypes.h if a file uses a gl_context struct. This
allows future removal of headers that are not strictly necessary but
indirectly include mtypes.h for a file.
Zou Nan hai [Thu, 6 Jan 2011 03:36:23 +0000 (11:36 +0800)]
i965: skip too small size mipmap
this fixes doom3 crash.
Eric Anholt [Thu, 6 Jan 2011 02:27:13 +0000 (18:27 -0800)]
i915: Fix build for previous commit.
Eric Anholt [Thu, 6 Jan 2011 01:02:08 +0000 (17:02 -0800)]
intel: Always allocate miptrees from level 0, not tObj->BaseLevel.
BaseLevel/MaxLevel are mostly used for two things: clamping texture
access for FBO rendering, and limiting the used mipmap levels when
incrementally loading textures. By restricting our mipmap trees to
just the current BaseLevel/MaxLevel, we caused reallocation thrashing
in the common case, for a theoretical win if someone really did want
just levels 2..4 or whatever of their texture object.
Bug #30366
Eric Anholt [Thu, 6 Jan 2011 00:40:56 +0000 (16:40 -0800)]
intel: Drop unused first/lastlevel args to miptree_create_for_region.
We're always making a single-level, 0-baselevel miptree.
Vinson Lee [Thu, 6 Jan 2011 01:46:39 +0000 (17:46 -0800)]
swrast: Include mtypes.h in s_triangle.c.
Include mtypes.h for gl_context symbol.
Vinson Lee [Thu, 6 Jan 2011 00:34:29 +0000 (16:34 -0800)]
st/mesa: Include mtypes.h in st_cb_drawpixels.c.
Include mtypes.h for gl_context symbol.
Eric Anholt [Thu, 6 Jan 2011 00:02:42 +0000 (16:02 -0800)]
intel: Clarify first_level/last_level vs baselevel/maxlevel by deletion.
This has always been ugly about our texture code -- object base/max
level vs intel object first/last level vs image level vs miptree
first/last level. We now get rid of intelObj->first_level which is
just tObj->BaseLevel, and make intelObj->_MaxLevel clearly based off
of tObj->_MaxLevel instead of duplicating its code (incorrectly, as
image->MaxLog2 only considers width/height and not depth!)
Eric Anholt [Wed, 5 Jan 2011 23:48:31 +0000 (15:48 -0800)]
mesa: Consider textures incomplete when maxlevel < baselevel.
See section 3.8.10 of the GL 2.1 specification. There's no way to do
anything sane with that, and drivers would get all sorts of angry.
Eric Anholt [Wed, 5 Jan 2011 22:48:05 +0000 (14:48 -0800)]
i915: Enable LOD preclamping on 8xx like on 915/965.
Fixes lodclamp-between and lodclamp-between-max.
Eric Anholt [Wed, 5 Jan 2011 22:27:41 +0000 (14:27 -0800)]
i915: Implement min/max lod clamping in hardware on 8xx.
This avoids 8xx-specific texture relayout for min/max lod changes.
One step closer to avoiding relayout for base/maxlevel changes!
Eric Anholt [Wed, 5 Jan 2011 22:12:34 +0000 (14:12 -0800)]
intel: Drop TEXTURE_RECTANGLE check in miptree layout setup.
It's already handled by our non-mipmapped MinFilter, since
TEXTURE_RECTANGLE is always NEAREST or LINEAR.
Eric Anholt [Wed, 5 Jan 2011 22:04:08 +0000 (14:04 -0800)]
intel: Clean up redundant setup of firstLevel.
It's always BaseLevel (since TEXTURE_RECTANGLE's baselevel can't be
changed from 0), except for 8xx minlod hilarity.
Eric Anholt [Wed, 5 Jan 2011 22:00:54 +0000 (14:00 -0800)]
intel: Drop a check for GL_TEXTURE_4D_SGIS.
The SGIS_texture4D extension was thankfully never completed, so we
couldn't implement it if we wanted to.
Vinson Lee [Wed, 5 Jan 2011 21:47:02 +0000 (13:47 -0800)]
swrast: Remove unnecessary headers.
Eric Anholt [Wed, 5 Jan 2011 18:04:34 +0000 (10:04 -0800)]
i965: Simplify the renderbuffer setup code.
It was quite a mess by trying to do NULL renderbuffers and real
renderbuffers in the same function. This clarifies the common case of
real renderbuffers.
Michel Dänzer [Wed, 5 Jan 2011 10:41:56 +0000 (11:41 +0100)]
st/xorg: Flesh out colour map support and support depth 8.
Xiang, Haihao [Wed, 5 Jan 2011 05:32:40 +0000 (13:32 +0800)]
i965: use BLT to clear buffer if possible on Sandybridge
This fixes https://bugs.freedesktop.org/show_bug.cgi?id=32713
Eric Anholt [Wed, 5 Jan 2011 00:29:26 +0000 (16:29 -0800)]
i965: Add support for SRGB DXT1 formats.
This makes
fbo-generatemipmap-formats GL_EXT_texture_sRGB-s3tc
match
fbo-generatemipmap-formats GL_EXT_texture_compression_s3tc
and swrast in bad DXT1_RGBA alpha=0 handling, but it means we won't
unpack and repack someone's textures into uncompressed SARGB8 format.
Vinson Lee [Wed, 5 Jan 2011 00:39:19 +0000 (16:39 -0800)]
glcpp: Add test for recursive #define.
Eric Anholt [Tue, 4 Jan 2011 23:25:35 +0000 (15:25 -0800)]
mesa: Fix the baseFormat for GL_COMPRESSED_SLUMINANCE_EXT.
It's just LUMINANCE, not LUMINANCE_ALPHA. Fixes
fbo-generatemipmap-formats GL_EXT_texture_sRGB-s3tc assertion failure
when it tries to pack the L8 channels into LUMINANCE_ALPHA and wonders
why it's trying to do that.
Eric Anholt [Tue, 4 Jan 2011 21:44:34 +0000 (13:44 -0800)]
intel: Merge our choosetexformat fallbacks into core.
We now share the type/format -> MESA_FORMAT_* mappings with software
mesa, and the core supports most of the fallbacks hardware drivers
will want.
Eric Anholt [Fri, 24 Sep 2010 00:17:08 +0000 (17:17 -0700)]
mesa: Make _mesa_choose_tex_format() choose formats out of a supported table.
Right now this is just tweaking the current code to look at the table.
Choosing actually supported formats will come later.
Vinson Lee [Tue, 4 Jan 2011 21:18:19 +0000 (13:18 -0800)]
glcpp: Add division by zero test cases.
Marek Olšák [Wed, 22 Dec 2010 18:29:56 +0000 (19:29 +0100)]
mesa: preserve 10 bits of precision in the texstore general path for ARGB2101010
Use make_temp_float_image instead of _make_temp_chan_image.
The latter converts the texture to 8 bits/component, losing 2 bits.
Marek Olšák [Sat, 18 Dec 2010 09:31:33 +0000 (10:31 +0100)]
st/mesa: advertise GL_ARB_half_float_pixel
This extension doesn't appear to need any driver-specific parts.
Marek Olšák [Tue, 4 Jan 2011 20:54:23 +0000 (21:54 +0100)]
r300/compiler: disable the rename_regs pass for loops
This workaround fixes rendering of kwin thumbnails.
NOTE: This is a candidate for the 7.9 and 7.10 branches.
Alex Deucher [Tue, 4 Jan 2011 20:37:12 +0000 (15:37 -0500)]
r600g: support up to 64 shader constants
From the r600 ISA:
Each ALU clause can lock up to four sets of constants
into the constant cache. Each set (one cache line) is
16 128-bit constants. These are split into two groups.
Each group can be from a different constant buffer
(out of 16 buffers). Each group of two constants consists
of either [Line] and [Line+1] or [line + loop_ctr]
and [line + loop_ctr +1].
For supporting more than 64 constants, we need to
break the code into multiple ALU clauses based
on what sets of constants are needed in that clause.
Note: This is a candidate for the 7.10 branch.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Tom Stellard [Tue, 4 Jan 2011 18:53:49 +0000 (10:53 -0800)]
r300/compiler: Fix black terrain in Civ4
rc_inst_can_use_presub() wasn't checking for too many RGB sources in
Alpha instructions or too many Alpha sources in RGB instructions.
Note: This is a candidate for the 7.10 branch.
Chad Versace [Thu, 16 Dec 2010 19:06:19 +0000 (11:06 -0800)]
glsl: Check that integer vertex outputs are qualified with flat
Perform this check in ast_declarator_list::hir().
From section 4.3.6 of the GLSL 1.30 spec:
"If a vertex output is a signed or unsigned integer or integer
vector, then it must be qualified with the interpolation
qualifier
flat."
Chad Versace [Thu, 16 Dec 2010 00:32:47 +0000 (16:32 -0800)]
glsl: Allow redeclaration of gl_Color and its variants in GLSL 1.30
Allow redeclaration of the following built-in variables with an
interpolation qualifier in language versions >= 1.30:
* gl_FrontColor
* gl_BackColor
* gl_FrontSecondaryColor
* gl_BackSecondaryColor
* gl_Color
* gl_SecondaryColor
See section 4.3.7 of the GLSL 1.30 spec.
Chad Versace [Wed, 15 Dec 2010 23:58:49 +0000 (15:58 -0800)]
glsl: Comment ast_type_qualifier.flags
Eric Anholt [Tue, 4 Jan 2011 18:00:51 +0000 (10:00 -0800)]
intel: When validating an FBO's combined depth/stencil, use the given FBO.
We were looking at the current draw buffer instead to see whether the
depth/stencil combination matched. So you'd get told your framebuffer
was complete, until you bound it and went to draw and we decided that
it was incomplete.
Eric Anholt [Tue, 4 Jan 2011 17:57:21 +0000 (09:57 -0800)]
intel: Fix segfaults from trying to use _ColorDrawBuffers in FBO validation.
The _ColorDrawBuffers is a piece of computed state that gets for the
current draw/read buffers at _mesa_update_state time. However, this
function actually gets used for non-current draw/read buffers when
checking if an FBO is complete from the driver's perspective. So,
instead of trying to just look at the attachment points that are
currently referenced by glDrawBuffers, look at all attachment points
to see if they're driver-supported formats. This appears to actually
be more in line with the intent of the spec, too.
Fixes a segfault in my upcoming fbo-clear-formats piglit test, and
hopefully bug #30278
Christoph Bumiller [Tue, 4 Jan 2011 17:20:05 +0000 (18:20 +0100)]
Merge remote branch 'origin/nvc0'
Brian Paul [Tue, 4 Jan 2011 15:28:16 +0000 (08:28 -0700)]
st/mesa: skip glDrawPixels/glBitmap-related code for ES build
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=32560
Christoph Bumiller [Tue, 4 Jan 2011 15:16:52 +0000 (16:16 +0100)]
nvc0: fix index size method value for u8 indices
Christoph Bumiller [Tue, 4 Jan 2011 00:52:12 +0000 (01:52 +0100)]
nvc0: set the correct FP header bit for multiple colour outputs