Dave Airlie [Tue, 5 Oct 2010 05:51:38 +0000 (15:51 +1000)]
r600g: add bo busy backoff.
When we go to do a lot of bos in one draw like constant bufs we need
to avoid bouncing off the busy ioctl, this mitigates by backing off
on busy bos for a short amount of times.
Dave Airlie [Tue, 5 Oct 2010 05:50:58 +0000 (15:50 +1000)]
pb: don't keep checking buffers after first busy
If we assume busy buffers are added to the list in order its unlikely
we'd fine one after the first busy one that isn't busy.
Dave Airlie [Tue, 5 Oct 2010 05:35:52 +0000 (15:35 +1000)]
r600g: add bo fenced list.
this just keeps a list of bos submitted together, and uses them to decide
bo busy state for the whole group.
Brian Paul [Tue, 5 Oct 2010 01:59:23 +0000 (19:59 -0600)]
swrast: fix choose_depth_texture_level() to respect mipmap filtering state
NOTE: this is a candidate for the 7.9 branch.
Marek Olšák [Mon, 4 Oct 2010 19:19:27 +0000 (21:19 +0200)]
r300g: fix microtiling for 16-bits-per-channel formats
These texture formats (like R16G16B16A16_UNORM) were untested until now
because st/mesa doesn't use them. I am testing this with a hacked st/mesa
here.
Marek Olšák [Tue, 5 Oct 2010 00:56:14 +0000 (02:56 +0200)]
update release notes for Gallium
I am trying to be exhaustive, but still I might have missed tons of other
changes to Gallium.
(cherry picked from commit
968a9ec76eadf55e8b58171884e1175d7b8cf59a)
Conflicts:
docs/relnotes-7.9.html
Ian Romanick [Mon, 4 Oct 2010 23:35:09 +0000 (16:35 -0700)]
docs: Add list of bugs fixed in 7.9
Eric Anholt [Mon, 4 Oct 2010 22:07:17 +0000 (15:07 -0700)]
i965: Add support for gen6 FB writes to the new FS.
This uses message headers for now, since we'll need it for MRT. We
can cut out the header later.
Eric Anholt [Mon, 4 Oct 2010 22:03:32 +0000 (15:03 -0700)]
i965: In disasm, gen6 fb writes don't put msg reg # in destreg_conditionalmod.
It instead sensibly appears in the src0 slot.
Eric Anholt [Mon, 4 Oct 2010 18:48:04 +0000 (11:48 -0700)]
i965: Add initial folding of constants into operand immediate slots.
We could try to detect this in expression handling and do it
proactively there, but it seems like less logic to do it in one
optional pass at the end.
Eric Anholt [Sun, 3 Oct 2010 22:15:18 +0000 (15:15 -0700)]
i965: Add trivial dead code elimination in the new FS backend.
The glsl core should be handling most dead code issues for us, but we
generate some things in codegen that may not get used, like the 1/w
value or pixel deltas. It seems a lot easier this way than trying to
work out up front whether we're going to use those values or not.
Eric Anholt [Sun, 3 Oct 2010 22:01:20 +0000 (15:01 -0700)]
i965: Be more conservative on live interval calculation.
This also means that our intervals now highlight dead code.
Vinson Lee [Mon, 4 Oct 2010 22:56:55 +0000 (15:56 -0700)]
r600g: Fix SCons build.
Jerome Glisse [Mon, 4 Oct 2010 14:40:07 +0000 (10:40 -0400)]
r600g: remove dead label & fix indentation
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Jerome Glisse [Mon, 4 Oct 2010 14:38:50 +0000 (10:38 -0400)]
r600g: rename radeon_ws_bo to r600_bo
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Jerome Glisse [Mon, 4 Oct 2010 14:37:32 +0000 (10:37 -0400)]
r600g: use r600_bo for relocation argument, simplify code
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Jerome Glisse [Mon, 4 Oct 2010 14:25:23 +0000 (10:25 -0400)]
r600g: allow r600_bo to be a sub allocation of a big bo
Add bo offset everywhere needed if r600_bo is ever a sub bo
of a bigger bo.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Jerome Glisse [Mon, 4 Oct 2010 14:06:13 +0000 (10:06 -0400)]
r600g: rename radeon_ws_bo to r600_bo
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Krzysztof Smiechowicz [Mon, 4 Oct 2010 18:43:29 +0000 (11:43 -0700)]
nvfx: Pair os_malloc_aligned() with os_free_aligned().
From AROS.
Dave Airlie [Mon, 4 Oct 2010 06:41:49 +0000 (16:41 +1000)]
r600g: TODO domain management
no wonder it was slow, the code is deliberately forcing stuff into GTT,
we used to have domain management but it seems to have disappeared.
Dave Airlie [Mon, 4 Oct 2010 06:26:46 +0000 (16:26 +1000)]
r600g: fix wwarning in bo_map function
Dave Airlie [Mon, 4 Oct 2010 06:24:59 +0000 (16:24 +1000)]
r600g: the code to check whether a new vertex shader is needed was wrong
this code was memcmp'ing two structs, but refcounting one of them afterwards,
so any subsequent memcmp was never going to work.
again this stops unnecessary uploads of vertex program,
Dave Airlie [Mon, 4 Oct 2010 05:58:39 +0000 (15:58 +1000)]
r600g: break out of search for reloc bo after finding it.
this function was taking quite a lot of pointless CPU.
Eric Anholt [Sun, 3 Oct 2010 07:24:09 +0000 (00:24 -0700)]
i965: Fix glean/texSwizzle regression in previous commit.
Easy enough patch, who needs a full test run. Oh, that's right. Me.
Eric Anholt [Sun, 3 Oct 2010 06:27:31 +0000 (23:27 -0700)]
i965: Set up swizzling of shadow compare results for GL_DEPTH_TEXTURE_MODE.
The brw_wm_surface_state.c handling of GL_DEPTH_TEXTURE_MODE doesn't
apply to shadow compares, which always return an intensity value. The
texture swizzles can do the job for us.
Fixes:
glsl1-shadow2D(): 1
glsl1-shadow2D(): 3
Eric Anholt [Sun, 3 Oct 2010 06:44:29 +0000 (23:44 -0700)]
i965: Add support for EXT_texture_swizzle to the new FS backend.
Marek Olšák [Sat, 2 Oct 2010 21:13:12 +0000 (23:13 +0200)]
r300g: add support for L8A8 colorbuffers
Blending with DST_ALPHA is undefined. SRC_ALPHA works, though.
I bet some other formats have similar limitations too.
Marek Olšák [Sat, 2 Oct 2010 19:42:22 +0000 (21:42 +0200)]
r300g: add support for R8G8 colorbuffers
The hw swizzles have been obtained by a brute force approach,
and only C0 and C2 are stored in UV88, the other channels are
ignored.
R16G16 is going to be a lot trickier.
Dave Airlie [Wed, 11 Aug 2010 09:04:05 +0000 (19:04 +1000)]
mesa/st: initial attempt at RG support for gallium drivers
passes all piglit RG tests with softpipe.
Kenneth Graunke [Sat, 2 Oct 2010 02:53:24 +0000 (19:53 -0700)]
i965: Fix incorrect batchbuffer size in gen6 clip state command.
FORCE_ZERO_RTAINDEX should be in the fourth (and final) dword.
Eric Anholt [Sat, 2 Oct 2010 00:18:07 +0000 (17:18 -0700)]
i965: Don't try to emit code if we failed register allocation.
Eric Anholt [Fri, 1 Oct 2010 23:36:17 +0000 (16:36 -0700)]
i965: Fix off-by-ones in handling the last members of register classes.
Luckily, one of them would result in failing out register allocation
when the other bugs were encountered. Applies to
glsl-fs-vec4-indexing-temp-dst-in-nested-loop-combined, which still
fails register allocation, but now legitimately.
Eric Anholt [Fri, 1 Oct 2010 23:35:55 +0000 (16:35 -0700)]
i965: Add a sanity check for register allocation sizes.
Eric Anholt [Fri, 1 Oct 2010 22:42:39 +0000 (15:42 -0700)]
i965: When producing a single channel swizzle, don't make a temporary.
This quickly cuts 8% of the instructions in my glsl demo.
Eric Anholt [Fri, 1 Oct 2010 22:11:09 +0000 (15:11 -0700)]
i965: Restore the forcing of aligned pairs for delta_xy on chips with PLN.
By doing so using the register allocator now, we avoid wasting a
register to make the alignment happen.
Alex Deucher [Sat, 2 Oct 2010 00:13:36 +0000 (20:13 -0400)]
r600c: fix segfault in evergreen stencil code
Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=30551
Vinson Lee [Sat, 2 Oct 2010 00:06:33 +0000 (17:06 -0700)]
r600g: Remove unnecessary headers.
Vinson Lee [Sat, 2 Oct 2010 00:00:43 +0000 (17:00 -0700)]
r600g: Remove unused variable.
Fixes this GCC warning.
r600_shader.c: In function 'tgsi_split_literal_constant':
r600_shader.c:818: warning: unused variable 'index'
Ian Romanick [Fri, 1 Oct 2010 23:40:14 +0000 (16:40 -0700)]
rgtc: Detect RGTC formats as color formats and as compressed formats
Ian Romanick [Fri, 1 Oct 2010 23:35:25 +0000 (16:35 -0700)]
mesa: Trivial correction to comment
Ian Romanick [Fri, 1 Oct 2010 23:23:06 +0000 (16:23 -0700)]
mesa: Fix misplaced #endif
If FEATURE_texture_s3tc is not defined, FXT1 formats would erroneously
fall through to the MESA_FORMAT_RGBA_FLOAT32 case.
Ian Romanick [Fri, 1 Oct 2010 23:21:12 +0000 (16:21 -0700)]
ARB_texture_rg: Add GL_COMPRESSED_{RED,RG} cases in _mesa_is_color_format
Ian Romanick [Fri, 1 Oct 2010 23:07:28 +0000 (16:07 -0700)]
mesa: Add ARB_texture_compression_rgtc as an alias for EXT_texture_compression_rgtc
Change the name in the extension tracking structure to ARB (from EXT).
Vinson Lee [Fri, 1 Oct 2010 23:57:19 +0000 (16:57 -0700)]
savage: Remove unnecessary header.
Vinson Lee [Fri, 1 Oct 2010 23:27:58 +0000 (16:27 -0700)]
glsl: Remove unnecessary header.
Ian Romanick [Fri, 20 Nov 2009 18:45:36 +0000 (10:45 -0800)]
i965: Enable GL_ARB_texture_rg
Ian Romanick [Thu, 30 Sep 2010 23:29:46 +0000 (16:29 -0700)]
mesa: Enable GL_ARB_texture_rg in software paths
Ian Romanick [Thu, 30 Sep 2010 23:30:09 +0000 (16:30 -0700)]
ARB_texture_rg: Allow RED and RG textures as FBO color buffer attachments
Ian Romanick [Wed, 29 Sep 2010 00:24:39 +0000 (17:24 -0700)]
ARB_texture_rg: Add R8, R16, RG88, and RG1616 internal formats
Ian Romanick [Tue, 28 Sep 2010 20:58:28 +0000 (13:58 -0700)]
ARB_texture_rg: Handle RED and RG the same as RGB for tex env
Ian Romanick [Tue, 28 Sep 2010 20:50:34 +0000 (13:50 -0700)]
ARB_texture_rg: Add GL_RED as a valid GL_DEPTH_TEXTURE_MODE
Ian Romanick [Tue, 28 Sep 2010 18:17:27 +0000 (11:17 -0700)]
ARB_texture_rg: Add GL_TEXTURE_{RED,GREEN}_SIZE query support
Ian Romanick [Tue, 28 Sep 2010 18:11:43 +0000 (11:11 -0700)]
ARB_texture_rg: Correct some errors in RED / RG internal format handling
Fixes several problems:
The half-float, float, and integer internal formats depend on
ARB_texture_rg and other extensions.
RG_INTEGER is not a valid internal format.
Generic compressed formats depend on ARB_texture_rg, not
EXT_texture_compression_rgtc.
Use GL_RED instead of GL_R.
Ian Romanick [Fri, 20 Nov 2009 19:03:31 +0000 (11:03 -0800)]
ARB_texture_rg: Add GLX protocol support
Nicolas Kaiser [Fri, 1 Oct 2010 20:39:20 +0000 (13:39 -0700)]
i965g: use Elements macro instead of manual sizeofs
Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Brian Paul <brianp@vmware.com>
Eric Anholt [Fri, 1 Oct 2010 21:09:00 +0000 (14:09 -0700)]
i965: Fix up copy'n'pasteo from moving coordinate setup around for gen4.
Eric Anholt [Fri, 1 Oct 2010 20:37:40 +0000 (13:37 -0700)]
i965: Add real support for pre-gen5 texture sampling to the new FS.
Fixes 36 testcases, including glsl-fs-shadow2d*-bias which fail on the
Mesa IR backend.
richard [Fri, 1 Oct 2010 20:10:02 +0000 (16:10 -0400)]
evergreen : fix z format setting, enable stencil.
Eric Anholt [Fri, 1 Oct 2010 19:15:48 +0000 (12:15 -0700)]
i965: Pre-gen6, map VS outputs (not FS inputs) to URB setup in the new FS.
We should fix the SF to actually give us just the data we need, but
this fixes regressions in the new FS until then.
Fixes:
glsl-kwin-blur
glsl-routing
Eric Anholt [Fri, 1 Oct 2010 18:44:27 +0000 (11:44 -0700)]
i965: Also increment attribute location when skipping unused slots.
Fixes glsl1-texcoord varying.
Eric Anholt [Fri, 1 Oct 2010 17:49:01 +0000 (10:49 -0700)]
i965: Fix the gen6 jump size for BREAK/CONT in new FS.
Since gen5, jumps are in increments of 64 bits instead of increments
of 128-bit instructions.
Eric Anholt [Fri, 1 Oct 2010 17:45:26 +0000 (10:45 -0700)]
i965: Add gen6 attribute interpolation to new FS backend.
Untested, since my hardware is not booting at the moment.
Jerome Glisse [Fri, 1 Oct 2010 14:26:58 +0000 (10:26 -0400)]
r600g: indentation fixes
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Dave Airlie [Fri, 1 Oct 2010 05:56:05 +0000 (15:56 +1000)]
r600g: setup basic loop consts on r600 + evergreen.
this sets up a single loop constant like r600c does.
Dave Airlie [Fri, 1 Oct 2010 05:16:47 +0000 (15:16 +1000)]
r600g: only set the Z export if shader exports it.
Alex Deucher [Fri, 1 Oct 2010 03:13:50 +0000 (23:13 -0400)]
r600c: pull over 6xx/7xx vertex fixes for evergreen
Dave Airlie [Fri, 1 Oct 2010 04:43:02 +0000 (14:43 +1000)]
r600g: flush SH cache on constant change on evergreen
Dave Airlie [Fri, 1 Oct 2010 04:24:14 +0000 (14:24 +1000)]
r600g: fix evergreen draw-buffers
just a typo in the register headers.
Dave Airlie [Fri, 1 Oct 2010 04:00:27 +0000 (14:00 +1000)]
r600g: add cb flushing for extra buffers + depth buffer on r600/evergreen
Dave Airlie [Fri, 1 Oct 2010 03:48:10 +0000 (13:48 +1000)]
r600g: sync vertex/texture cache on resources on evergreen
this gets rid of lots of the instability on evergreen,
which isn't surprising since it really broken not to flush caches.
Dave Airlie [Fri, 1 Oct 2010 03:09:09 +0000 (13:09 +1000)]
r600g: fixup vertex format picking.
there are some vertex formats defined in r600c not in the docs.
Dave Airlie [Fri, 1 Oct 2010 01:38:40 +0000 (11:38 +1000)]
r600g: add assembler support for other vtx fetch fields.
this shouldn't change behaviour, just push the choice of what
to do out to the shader.
Eric Anholt [Fri, 1 Oct 2010 02:18:25 +0000 (19:18 -0700)]
i965: Split the gen4 and gen5 sampler handling apart.
Trying to track the insanity of the different argument layouts for
normal/shadow crossed with normal/lod/bias one generation at a time is
enough.
Fixes: glsl1-texture2D() with bias.
(first test passing in this code that doesn't pass without it!)
Eric Anholt [Fri, 1 Oct 2010 03:09:53 +0000 (20:09 -0700)]
i965: Use the lowering pass for texture projection.
We should end up with the same code, but anyone else with this issue
could share the handling (which I got wrong for shadow comparisons in
the driver before).
Eric Anholt [Fri, 1 Oct 2010 03:07:27 +0000 (20:07 -0700)]
glsl: Add a lowering pass for texture projection.
Dave Airlie [Fri, 1 Oct 2010 01:13:02 +0000 (11:13 +1000)]
r600g: realign evergreen code with r600 code.
fixes segfault in depth-tex-modes-glsl and OA startup.
Alex Deucher [Fri, 1 Oct 2010 00:54:26 +0000 (20:54 -0400)]
r600c: add reloc for CB_COLOR0_ATTRIB
We'll need a reloc for tiling eventually,
so add it now.
Dave Airlie [Fri, 1 Oct 2010 00:52:09 +0000 (10:52 +1000)]
r600g: add reloc for evergreen color attrib
we'll need this for color tiling on evergreen.
Dave Airlie [Fri, 1 Oct 2010 00:19:39 +0000 (10:19 +1000)]
r600g: drop depth quirk on evergreen
none of the EG cards need the quirk.
Dave Airlie [Thu, 30 Sep 2010 23:43:14 +0000 (09:43 +1000)]
r600g: add winsys support for CTL constants.
These need to be emitted, we also need them to do proper vtx start,
instead of abusing index offset.
Dave Airlie [Fri, 1 Oct 2010 00:13:04 +0000 (10:13 +1000)]
r600g: fix evergreen depth flushing.
although evergreen can apparantly sample direct from 24-bit,
just make it work with the current method for now.
Dave Airlie [Thu, 30 Sep 2010 23:39:35 +0000 (09:39 +1000)]
r600g: use Elements macro instead of manual sizeofs
Brian Paul [Thu, 30 Sep 2010 22:41:31 +0000 (16:41 -0600)]
draw: check for null sampler pointers
http://bugs.freedesktop.org/show_bug.cgi?id=30516
Brian Paul [Thu, 30 Sep 2010 17:01:09 +0000 (11:01 -0600)]
gallivm: added some comments
John Doe [Thu, 30 Sep 2010 21:53:36 +0000 (17:53 -0400)]
r600g: keep a mapping around for each bo
Save a lot of call into the kernel and thus improve performances.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
John Doe [Thu, 30 Sep 2010 21:30:25 +0000 (17:30 -0400)]
r600g: don't double count dirty block
This avoid to overcount the number of dwords we need and
thus avoid maximazation of cs buffer use.
Signed-off-by: Jerome Glisse <jglisse@redhat.com
Jerome Glisse [Thu, 30 Sep 2010 21:06:29 +0000 (17:06 -0400)]
evergreeng: avoid overlapping border color btw VS & PS
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Eric Anholt [Thu, 30 Sep 2010 20:26:38 +0000 (13:26 -0700)]
i965: Fix new FS handling of builtin uniforms with packed scalars in structs.
We were pointing each element at the .x channel of the
ParameterValues.
Fixes glsl1-linear fog.
Eric Anholt [Thu, 30 Sep 2010 20:40:22 +0000 (13:40 -0700)]
mesa: Don't reference a W component in setting up a vec3 uniform component.
The 965 driver would try to set up storage for the W component, and
the offsets would get mixed up.
Eric Anholt [Thu, 30 Sep 2010 18:46:24 +0000 (11:46 -0700)]
i965: Fix whole-structure/array assignment in new FS.
We need to walk the type tree to get the right register types for
structure components. Fixes glsl-fs-statevar-call.
Tom Fogal [Thu, 30 Sep 2010 20:39:14 +0000 (14:39 -0600)]
Revert "Prefer intrinsics to handrolled atomic ops."
This reverts commit
5f66b340aa49c6bc8d0acb2d1a6f8e9a7ef2cb2e, quickly
fixing 30514.
Jerome Glisse [Thu, 30 Sep 2010 20:26:33 +0000 (16:26 -0400)]
r600g: fix constant & literal src splitting, also fix mplayer gl2 shader
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Tom Fogal [Mon, 27 Sep 2010 04:32:15 +0000 (22:32 -0600)]
Prefer intrinsics to handrolled atomic ops.
Tom Fogal [Mon, 27 Sep 2010 00:57:59 +0000 (18:57 -0600)]
Implement x86_64 atomics for compilers w/o intrinsics.
Really old gcc's (3.3, at least) don't have support for the
intrinsics we need. This implements a fallback for that case.
Adam Jackson [Thu, 30 Sep 2010 18:08:35 +0000 (14:08 -0400)]
i965: Update renderer strings for sandybridge
Signed-off-by: Adam Jackson <ajax@redhat.com>
Jerome Glisse [Thu, 30 Sep 2010 14:43:26 +0000 (10:43 -0400)]
r600g: use constant buffer instead of register for constant
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Brian Paul [Thu, 30 Sep 2010 16:52:26 +0000 (10:52 -0600)]
gallivm: check for level=0 case in lp_build_minify()
This lets us avoid the shift and max() operations.
José Fonseca [Thu, 30 Sep 2010 16:39:17 +0000 (17:39 +0100)]
gallivm: More comprehensive border usage logic.
Chia-I Wu [Thu, 30 Sep 2010 15:43:58 +0000 (23:43 +0800)]
st/egl: Drop context argument from egl_g3d_get_egl_image.
Fix a regression since
17eace581d25a626a7d75d9d1205d012cbb14a6e.
Nicolas Kaiser [Thu, 30 Sep 2010 14:53:13 +0000 (07:53 -0700)]
st: remove duplicated include
Remove duplicated include.
Signed-off-by: Brian Paul <brianp@vmware.com>