mesa.git
8 years agoi965: Set fs_inst::base_mrf = -1 by default.
Kenneth Graunke [Thu, 23 Jun 2016 00:01:12 +0000 (17:01 -0700)]
i965: Set fs_inst::base_mrf = -1 by default.

On MRF platforms, we need to set base_mrf to the first MRF value we'd
like to use for the message.  On send-from-GRF platforms, we set it to
-1 to indicate that the operation doesn't use MRFs.

As MRF platforms are becoming increasingly a thing of the past, we've
forgotten to bother with this.  It makes more sense to set it to -1 by
default, so we don't have to think about it for new code.

I searched the code for every instance of 'mlen =' in brw_fs*cpp, and
it appears that all MRF-based messages correctly program a base_mrf.

Forgetting to set base_mrf = -1 can confuse the register allocator,
causing it to think we have a large fake-MRF region.  This ends up
moving the send-with-EOT registers earlier, sometimes even out of
the g112-g127 range, which is illegal.  For example, this fixes
illegal sends in Piglit's arb_gpu_shader_fp64-layout-std430-fp64-shader,
which had SSBO messages with mlen > 0 but base_mrf == 0.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
8 years agoi965: Drop unused return value from intel_finalize_mipmap_tree().
Kenneth Graunke [Sun, 19 Jun 2016 06:40:21 +0000 (23:40 -0700)]
i965: Drop unused return value from intel_finalize_mipmap_tree().

The old return type of GLuint was wonky - it should have been bool.
But nothing actually uses the return value anyway, so we can just drop
that and make it a void function.

In theory, it might make sense to ask whether the texture validated
successfully, but just checking intel_obj->mt != NULL works for that.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
8 years agoi965: Move contents of brw_tex.c into intel_tex_validate.c.
Kenneth Graunke [Sun, 19 Jun 2016 06:36:59 +0000 (23:36 -0700)]
i965: Move contents of brw_tex.c into intel_tex_validate.c.

brw_tex.c is a tiny file containing a single function.  It's closely
tied to the validation logic in intel_tex_validate.c, so it makes sense
to put both in the same file.

While we're at it, update the function to our modern style.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
8 years agoradeonsi: fix fractional odd tessellation spacing for Polaris
Marek Olšák [Fri, 24 Jun 2016 00:17:38 +0000 (02:17 +0200)]
radeonsi: fix fractional odd tessellation spacing for Polaris

ported from Vulkan (and no source explains why this is needed)

Cc: 12.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agoradeonsi: set some VGT context registers on SI-CI
Marek Olšák [Thu, 23 Jun 2016 22:03:26 +0000 (00:03 +0200)]
radeonsi: set some VGT context registers on SI-CI

the kernel sets them, but other UMDs can change them

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agoradeonsi: optimize rendering to linear color buffers
Marek Olšák [Thu, 23 Jun 2016 21:55:24 +0000 (23:55 +0200)]
radeonsi: optimize rendering to linear color buffers

loosely ported from Vulkan

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agoradeonsi: set almost optimal settings in SC_MODE_CNTL_1
Marek Olšák [Thu, 23 Jun 2016 21:26:25 +0000 (23:26 +0200)]
radeonsi: set almost optimal settings in SC_MODE_CNTL_1

ported from Vulkan

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/radeon: let drivers specify SC_MODE_CNTL_1 fields
Marek Olšák [Thu, 23 Jun 2016 20:24:07 +0000 (22:24 +0200)]
gallium/radeon: let drivers specify SC_MODE_CNTL_1 fields

radeonsi will set more fields

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/radeon: disable complicated point clipping against user clip planes
Marek Olšák [Thu, 23 Jun 2016 19:42:39 +0000 (21:42 +0200)]
gallium/radeon: disable complicated point clipping against user clip planes

Nothing in the GL spec says that we should expand points to triangles.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agoradeonsi: fix a compute shader hang with big threadgroups on SI & CI
Marek Olšák [Fri, 24 Jun 2016 00:22:42 +0000 (02:22 +0200)]
radeonsi: fix a compute shader hang with big threadgroups on SI & CI

ported from Vulkan

Cc: 12.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agonvc0: when mapping directly, provide accurate xfer info + start
Ilia Mirkin [Fri, 24 Jun 2016 02:33:59 +0000 (22:33 -0400)]
nvc0: when mapping directly, provide accurate xfer info + start

We were ignoring the incoming box parameters, and were providing totally
bogus stride/layer stride, and other bits, for when a non-full-surface
map was requested.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Tested-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: <mesa-stable@lists.freedesktop.org>
8 years agost/mesa: don't assume that the whole surface gets mapped
Ilia Mirkin [Fri, 24 Jun 2016 02:27:40 +0000 (22:27 -0400)]
st/mesa: don't assume that the whole surface gets mapped

Under some circumstances, the driver may choose to return a temporary
surface instead of a pointer to the original. Make sure to pass the
actual view volume to be mapped to the transfer function rather than
adjusting the map pointer after-the-fact.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agoradeonsi: drop the DRAW_PREAMBLE packet on Polaris
Nicolai Hähnle [Thu, 23 Jun 2016 23:11:09 +0000 (01:11 +0200)]
radeonsi: drop the DRAW_PREAMBLE packet on Polaris

It will be removed from the firmware for the Polaris.

Cc: 12.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoradeonsi: use DRAW_(INDEX_)INDIRECT_MULTI on Polaris
Nicolai Hähnle [Thu, 23 Jun 2016 13:00:53 +0000 (15:00 +0200)]
radeonsi: use DRAW_(INDEX_)INDIRECT_MULTI on Polaris

The non-MULTI variants will be removed in Polaris firmware.

Cc: 12.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agost/mesa: handle negative _ColorDrawBufferIndexes values correctly
Francesco Ansanelli [Thu, 23 Jun 2016 17:56:09 +0000 (19:56 +0200)]
st/mesa: handle negative _ColorDrawBufferIndexes values correctly

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
8 years agowinsys/radeon: add guard pages when R600_DEBUG=check_vm is enabled
Nicolai Hähnle [Wed, 22 Jun 2016 09:22:33 +0000 (11:22 +0200)]
winsys/radeon: add guard pages when R600_DEBUG=check_vm is enabled

This should help flush out GPU VM faults.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agowinsys/amdgpu: add guard pages when R600_DEBUG=check_vm is enabled
Nicolai Hähnle [Wed, 22 Jun 2016 09:22:33 +0000 (11:22 +0200)]
winsys/amdgpu: add guard pages when R600_DEBUG=check_vm is enabled

This should help flush out GPU VM faults.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoradeonsi: report a failure to parse dmesg instead of asserting
Nicolai Hähnle [Mon, 20 Jun 2016 18:24:03 +0000 (20:24 +0200)]
radeonsi: report a failure to parse dmesg instead of asserting

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoradeon: check VM faults from DMA flush
Nicolai Hähnle [Mon, 20 Jun 2016 15:19:28 +0000 (17:19 +0200)]
radeon: check VM faults from DMA flush

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoradeonsi: move gfx fence wait out of si_check_vm_faults
Nicolai Hähnle [Mon, 20 Jun 2016 14:47:06 +0000 (16:47 +0200)]
radeonsi: move gfx fence wait out of si_check_vm_faults

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoradeonsi: extract IB and bo list saving into separate functions
Nicolai Hähnle [Mon, 20 Jun 2016 14:30:29 +0000 (16:30 +0200)]
radeonsi: extract IB and bo list saving into separate functions

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agost/mesa: fix readpixels regression with MESA_pack_invert
Nicolai Hähnle [Wed, 22 Jun 2016 15:36:18 +0000 (17:36 +0200)]
st/mesa: fix readpixels regression with MESA_pack_invert

Fixes an error introduced in commit 3948cd37973696dc319170877382676809659465.

Reported-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoradeonsi: set LLVM denormal flags
Marek Olšák [Tue, 21 Jun 2016 21:28:46 +0000 (23:28 +0200)]
radeonsi: set LLVM denormal flags

- make sure FP32 denormals will stay disabled in LLVM in the future
  (the current default is disabled)
- tell LLVM that FP64 denormals are enabled

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
8 years agoradeonsi: emit 1/sqrt for RSQ
Marek Olšák [Tue, 21 Jun 2016 20:45:29 +0000 (22:45 +0200)]
radeonsi: emit 1/sqrt for RSQ

We don't need the clamped version and we don't have to use any intrinsic.

Stats on Tonga:

15382 shaders in 9128 tests
Totals:
SGPRS: 1230560 -> 1230560 (0.00 %)
VGPRS: 469577 -> 462504 (-1.51 %)
Code Size: 22089908 -> 21730052 (-1.63 %) bytes
LDS: 598 -> 598 (0.00 %) blocks
Scratch: 283648 -> 281600 (-0.72 %) bytes per wave
Max Waves: 125664 -> 126969 (1.04 %)
Wait states: 0 -> 0 (0.00 %)

Totals from affected shaders:
SGPRS: 547280 -> 547280 (0.00 %)
VGPRS: 269132 -> 262059 (-2.63 %)
Code Size: 15709604 -> 15349748 (-2.29 %) bytes
LDS: 198 -> 198 (0.00 %) blocks
Scratch: 74752 -> 72704 (-2.74 %) bytes per wave
Max Waves: 47840 -> 49145 (2.73 %)
Wait states: 0 -> 0 (0.00 %)

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
8 years agor600g: Enable FMA on chips that support it
Jan Vesely [Wed, 15 Jun 2016 23:41:22 +0000 (19:41 -0400)]
r600g: Enable FMA on chips that support it

v2: Merge with PIPE_SHADER_CAP_DOUBLES
    Add CHIP_HEMLOCK

v3: only set the instruction on EG and CM

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
8 years agogallium/u_queue: allow the execute function to differ per job
Marek Olšák [Sat, 11 Jun 2016 15:28:52 +0000 (17:28 +0200)]
gallium/u_queue: allow the execute function to differ per job

so that independent types of jobs can use the same queue.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/u_queue: reduce the number of mutexes by 2
Marek Olšák [Thu, 16 Jun 2016 23:33:12 +0000 (01:33 +0200)]
gallium/u_queue: reduce the number of mutexes by 2

by converting semaphores to condvars and using the main mutex

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/u_queue: add an option to name threads
Marek Olšák [Sun, 12 Jun 2016 11:36:39 +0000 (13:36 +0200)]
gallium/u_queue: add an option to name threads

for debugging

v2: correct the snprintf use

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/u_queue: add an option to have multiple worker threads
Marek Olšák [Sat, 11 Jun 2016 13:40:28 +0000 (15:40 +0200)]
gallium/u_queue: add an option to have multiple worker threads

independent jobs don't have to be stuck on only one thread

v2: use CALLOC & FREE

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/u_queue: rewrite util_queue_fence to allow multiple waiters
Marek Olšák [Sat, 11 Jun 2016 15:51:22 +0000 (17:51 +0200)]
gallium/u_queue: rewrite util_queue_fence to allow multiple waiters

Checking "signalled" is first done without a mutex, then with a mutex.
Also, checking without waiting doesn't lock the mutex. This is racy, but
should be safe.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agogallium/u_queue: use a ring instead of a stack
Marek Olšák [Sun, 12 Jun 2016 10:54:42 +0000 (12:54 +0200)]
gallium/u_queue: use a ring instead of a stack

and allow specifying its size in util_queue_init.

v2: use CALLOC & FREE

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
8 years agoi965: Preserve the internal format of the dri image
Jordan Justen [Wed, 22 Jun 2016 17:34:08 +0000 (17:34 +0000)]
i965: Preserve the internal format of the dri image

Since the OpenGLES API is strict about the internal format matching
the for many operations, we need to preserve it.

See _mesa_es3_error_check_format_and_type in
src/mesa/main/glformats.c.

Fixes ES2-CTS.gtf.GL2ExtensionTests.egl_image.egl_image

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96351
Reported-by: Mark Janes <mark.a.janes@intel.com>
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: Kristian Høgsberg <krh@bitplanet.net>
Cc: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Chad Versace <chad.versace@intel.com>
8 years agoanv: Add anv_render_pass_attachment::store_op
Chad Versace [Wed, 22 Jun 2016 16:17:41 +0000 (09:17 -0700)]
anv: Add anv_render_pass_attachment::store_op

Will be needed for resolving auxiliary surfaces.

I didn't add anv_render_pass_attachment::stencil_store_op, as the driver
would likely never use it, as stencil surfaces never have auxiliary
surfaces.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
8 years agogbm: Fix comments
Gurkirpal Singh [Thu, 23 Jun 2016 10:24:11 +0000 (15:54 +0530)]
gbm: Fix comments

Reviewed-by: Chad Versace <chad.versace@intel.com>
8 years agogbm: doc fixes
Eric Engestrom [Thu, 23 Jun 2016 11:00:44 +0000 (12:00 +0100)]
gbm: doc fixes

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Chad Versace <chad.versace@intel.com>
8 years agoRemove wrongly repeated words in comments
Giuseppe Bilotta [Thu, 23 Jun 2016 17:20:18 +0000 (19:20 +0200)]
Remove wrongly repeated words in comments

Clean up misrepetitions ('if if', 'the the' etc) found throughout the
comments. This has been done manually, after grepping
case-insensitively for duplicate if, is, the, then, do, for, an,
plus a few other typos corrected in fly-by

v2:
    * proper commit message and non-joke title;
    * replace two 'as is' followed by 'is' to 'as-is'.
v3:
    * 'a integer' => 'an integer' and similar (originally spotted by
      Jason Ekstrand, I fixed a few other similar ones while at it)

Signed-off-by: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Reviewed-by: Chad Versace <chad.versace@intel.com>
8 years agosvga: update some comments in svga_buffer_handle()
Brian Paul [Wed, 22 Jun 2016 23:24:12 +0000 (17:24 -0600)]
svga: update some comments in svga_buffer_handle()

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agosvga: add a const qualifier in svga_buffer_upload_piecewise()
Brian Paul [Wed, 22 Jun 2016 23:19:58 +0000 (17:19 -0600)]
svga: add a const qualifier in svga_buffer_upload_piecewise()

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agosvga: minor code refactor for svga_buffer_upload_command()
Brian Paul [Wed, 22 Jun 2016 23:18:17 +0000 (17:18 -0600)]
svga: minor code refactor for svga_buffer_upload_command()

Put the HBS code into a separate function.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agosvga: minor code simplification in svga_context_finish()
Brian Paul [Wed, 22 Jun 2016 22:51:15 +0000 (16:51 -0600)]
svga: minor code simplification in svga_context_finish()

Signed-off-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agoi965: Implement rasterizer discard via SOL unless required for queries.
Kenneth Graunke [Wed, 22 Jun 2016 18:25:26 +0000 (11:25 -0700)]
i965: Implement rasterizer discard via SOL unless required for queries.

We currently use CL_INVOCATION_COUNT for the GL_PRIMITIVES_GENERATED
query, which involves passing all primitives to the clipper.  When
rasterizer discard is enabled, we program the clipper in REJECT_ALL
mode, rather than using the SOL stage's "Rendering Disable" feature.

See commit f09b91f78247409f54c975f56cb10d5f350fe64e for an explanation
of why we implement GL_PRIMITIVES_GENERATED this way.

Apparently the SOL stage's "Rendering Disable" feature is a lot faster
than having the clipper reject all primitives.  It's safe to use when
no GL_PRIMITIVES_GENERATED query is active, as we don't care about
CL_INVOCATION_COUNT incrementing.

This patch makes us use SO_RENDERING_DISABLE when no query is active,
but continues falling back to the clipper in REJECT_ALL mode when the
queries are enabled.  It brings back the perf_debug for the clipper
case (which I removed in commit 1f9445ff57b, thinking it wasn't useful).

Improves performance in Gl32GSCloth by 84.8303% +/- 2.07132% (n = 10)
on my Broadwell GT2 laptop.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
8 years agoi965: Combine 3DSTATE_STREAMOUT emitters and genX_sol_state atoms.
Kenneth Graunke [Wed, 22 Jun 2016 07:33:46 +0000 (00:33 -0700)]
i965: Combine 3DSTATE_STREAMOUT emitters and genX_sol_state atoms.

They're basically the same.  Let's avoid the code duplication.

v2: Fix SO_BUFFER_ENABLE stuff to only happen on Gen < 8 (caught
    by Jason Ekstrand).

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
8 years agoglsl: Don't constant propagate arrays.
Kenneth Graunke [Sat, 30 Apr 2016 00:57:46 +0000 (17:57 -0700)]
glsl: Don't constant propagate arrays.

Constant propagation on arrays doesn't make a lot of sense.  If the
array is only accessed with constant indexes, then opt_array_splitting
would split it up.  Otherwise, we have variable indexing.  If there's
multiple accesses, then constant propagation would end up replicating
the data.

The lower_const_arrays_to_uniforms pass creates uniforms for each
ir_constant with array type that it encounters.  This means that it
creates redundant uniforms for each copy of the constant, which means
uploading too much data.  It can even mean exceeding the maximum number
of uniform components, causing link failures.

We could try and teach the pass to de-duplicate the data by hashing
constants, but it makes more sense to avoid duplicating it in the first
place.  We should promote constant arrays to uniforms, then propagate
the uniform access.

Fixes the TressFX shaders from Tomb Raider, which exceeded the maximum
number of uniform components by a huge margin and failed to link.

On Broadwell:

total instructions in shared programs: 9067702 -> 9068202 (0.01%)
instructions in affected programs: 10335 -> 10835 (4.84%)
helped: 10 (Hoard, Shadow of Mordor, Amnesia: The Dark Descent)
HURT: 20 (Natural Selection 2)

loops in affected programs: 4 -> 0

The hurt programs appear to no longer have a constarray uniform, as
all constants were successfully propagated.  Apparently before this
patch, we successfully unrolled a loop containing array access, but
only after promoting constant arrays to uniforms.  With this patch,
we unroll it first, so all array access is direct, and the array
is split up, and individual constants are propagated.  This seems
better.

Cc: mesa-stable@lists.freedesktop.org
Reported-by: Karol Herbst <nouveau@karolherbst.de>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
8 years agoglsl: Make lower_const_arrays_to_uniforms work directly on constants.
Kenneth Graunke [Sat, 30 Apr 2016 01:05:26 +0000 (18:05 -0700)]
glsl: Make lower_const_arrays_to_uniforms work directly on constants.

There's really no point in looking at ir_dereference_array of a
constant.  It also misses cases like:

  (assign () (var_ref tmp) (constant (array ...) ...))

No changes in shader-db, but keeps it working after the next commit.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
8 years agoi965: Copy propagate before doing variable index lowering.
Kenneth Graunke [Sat, 30 Apr 2016 04:12:15 +0000 (21:12 -0700)]
i965: Copy propagate before doing variable index lowering.

The scalar backend currently doesn't support variable indexing on
temporary arrays, but it does support it on uniform arrays, and
some stages support it for input arrays.  Make sure these are
propagated through before exploding indirects into piles of
if-ladders unnecessarily.

On Broadwell, no instruction count change in shader-db.

total cycles in shared programs: 80675652 -> 80674928 (-0.00%)
cycles in affected programs: 649972 -> 649248 (-0.11%)
helped: 386
HURT: 165

This will help avoid code quality regressions in a future commit.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
8 years agoglsl: Propagate invariant/precise after lowering const arrays.
Kenneth Graunke [Wed, 22 Jun 2016 00:42:59 +0000 (17:42 -0700)]
glsl: Propagate invariant/precise after lowering const arrays.

The new uniform may need precise as well.

Fixes copy propagation of constant array uniforms in Tomb Raider shaders.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
8 years agoglsl: Split arrays even in the presence of whole-array copies.
Kenneth Graunke [Mon, 20 Jun 2016 18:20:51 +0000 (11:20 -0700)]
glsl: Split arrays even in the presence of whole-array copies.

Previously, we failed to split constant arrays.  Code such as

   int[2] numbers = int[](1, 2);

would generates a whole-array assignment:

  (assign () (var_ref numbers)
             (constant (array int 4) (constant int 1) (constant int 2)))

opt_array_splitting generally tried to visit ir_dereference_array nodes,
and avoid recursing into the inner ir_dereference_variable.  So if it
ever saw a ir_dereference_variable, it assumed this was a whole-array
read and bailed.  However, in the above case, there's no array deref,
and we can totally handle it - we just have to "unroll" the assignment,
creating assignments for each element.

This was mitigated by the fact that we constant propagate whole arrays,
so a dereference of a single component would usually get the desired
single value anyway.  However, I plan to stop doing that shortly;
early experiments with disabling constant propagation of arrays
revealed this shortcoming.

This patch causes some arrays in Gl32GSCloth's geometry shaders to be
split, which allows other optimizations to eliminate unused GS inputs.
The VS then doesn't have to write them, which eliminates the entire VS
(5 -> 2 instructions).  It still renders correctly.

No other change in shader-db.

v2: Drop !AOA check and improve a comment (feedback from Tim Arceri).

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
8 years agoglsl: Make constant propagation's folder not propagate into an LHS.
Kenneth Graunke [Mon, 20 Jun 2016 23:48:02 +0000 (16:48 -0700)]
glsl: Make constant propagation's folder not propagate into an LHS.

opt_constant_propagation.cpp contains constant folding code which can
actually do constant propagation in some cases.  It was happily
propagating constants into the left-hand-side of assignments.

For example,

   (assign () (var_ref temp) (constant ...))

would brilliantly be turned into:

   (assign () (constant ...) (constant ....))

This is a bigger hammer than necessary - it prevents propagation
into the left-hand-side altogether.  We could certainly do better
someday.  Notably, the constant propagation pass itself already
takes this approach - it's just the constant propagation pass's
built-in constant folding code (which actually propagates, too)
that was broken.

No change in shader-db, but prevents regressions after future commits.
It seems plausible that this could be hit today, but I haven't seen it
happen.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
8 years agoi965/blorp: Disable vertex element swizzling
Topi Pohjolainen [Wed, 18 May 2016 13:18:59 +0000 (16:18 +0300)]
i965/blorp: Disable vertex element swizzling

Without vertex elements originating directly from vertex fetcher
are not passed to wm-state correctly.

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
8 years agoi965/blorp: Let program data tell if push constants are needed
Topi Pohjolainen [Mon, 16 May 2016 07:18:53 +0000 (10:18 +0300)]
i965/blorp: Let program data tell if push constants are needed

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
8 years agoi965/blorp: Use prog data counters to guide wm/ps setup
Topi Pohjolainen [Wed, 18 May 2016 13:09:49 +0000 (16:09 +0300)]
i965/blorp: Use prog data counters to guide wm/ps setup

just as core upload logic does.

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
8 years agoi965/blorp: Use prog data counters to guide sf/sbe setup
Topi Pohjolainen [Sun, 15 May 2016 08:34:37 +0000 (11:34 +0300)]
i965/blorp: Use prog data counters to guide sf/sbe setup

just as core upload logic does.

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
8 years agoi965: Avoid division by zero.
Ardinartsev Nikita [Thu, 23 Jun 2016 01:28:11 +0000 (18:28 -0700)]
i965: Avoid division by zero.

Fixes regression introduced by af5ca43f2676bff7499f93277f908b681cb821d0

Cc: "12.0 11.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=95419

8 years agoswr: [rasterizer core] fix dependency bug
Tim Rowley [Tue, 14 Jun 2016 23:54:34 +0000 (17:54 -0600)]
swr: [rasterizer core] fix dependency bug

Never be dependent on "draw 0", instead have a bool that makes the draw
dependent on the previous draw or not dependent at all.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer core] use wrap-around safe compares for dependency checking
Tim Rowley [Tue, 14 Jun 2016 23:02:11 +0000 (17:02 -0600)]
swr: [rasterizer core] use wrap-around safe compares for dependency checking

Move drawIDs from 64-bit to 32-bit to increase perf.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer jitter] add support for component packing for 'odd' formats
Tim Rowley [Tue, 14 Jun 2016 18:57:31 +0000 (12:57 -0600)]
swr: [rasterizer jitter] add support for component packing for 'odd' formats

Add early-out if no components are enabled. Add asserts.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer core] track whether GS outputs viewport array index
Tim Rowley [Tue, 14 Jun 2016 16:25:02 +0000 (10:25 -0600)]
swr: [rasterizer core] track whether GS outputs viewport array index

So we can skip the index gather in PA.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer core] GS viewport array index attribute
Tim Rowley [Mon, 13 Jun 2016 15:19:38 +0000 (09:19 -0600)]
swr: [rasterizer core] GS viewport array index attribute

Only adds the attribute mapping to the jitter; no implementation yet.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer core] conservative rasterization frontend support
Tim Rowley [Fri, 10 Jun 2016 17:31:16 +0000 (11:31 -0600)]
swr: [rasterizer core] conservative rasterization frontend support

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer core] stop single threaded crash exit crash
Tim Rowley [Fri, 10 Jun 2016 16:18:45 +0000 (10:18 -0600)]
swr: [rasterizer core] stop single threaded crash exit crash

Function static destructors were getting called by exit
handlers before context teardown.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer jitter] small fetch jit cleanup
Tim Rowley [Thu, 9 Jun 2016 21:17:49 +0000 (15:17 -0600)]
swr: [rasterizer jitter] small fetch jit cleanup

Handle SGV stores separate from the stream fetch code.

Because of this change, there is a potential to jit an extra unused store.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer core] remove old comment
Tim Rowley [Tue, 7 Jun 2016 19:03:27 +0000 (13:03 -0600)]
swr: [rasterizer core] remove old comment

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer jitter] cleanup supporting different llvm versions
Tim Rowley [Tue, 7 Jun 2016 01:13:22 +0000 (19:13 -0600)]
swr: [rasterizer jitter] cleanup supporting different llvm versions

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer jitter] unitialized component fix in fetch jit
Tim Rowley [Thu, 2 Jun 2016 21:24:34 +0000 (15:24 -0600)]
swr: [rasterizer jitter] unitialized component fix in fetch jit

Was trying to store an extra uninitialized component.
Only affects component packing, which isn't enabled (yet).

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer] add support for building avx512 version
Tim Rowley [Wed, 1 Jun 2016 02:01:40 +0000 (20:01 -0600)]
swr: [rasterizer] add support for building avx512 version

Currently, most code paths between AVX2 and AVX512 are identical
(see changes to knobs.h).

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer common] fix include for Intel compiler
Tim Rowley [Fri, 27 May 2016 22:17:47 +0000 (16:17 -0600)]
swr: [rasterizer common] fix include for Intel compiler

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: [rasterizer common] workaround clang for windows __cpuid() bug
Tim Rowley [Thu, 26 May 2016 16:51:48 +0000 (10:51 -0600)]
swr: [rasterizer common] workaround clang for windows __cpuid() bug

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoswr: push/pop DEBUG macro around llvm includes
Tim Rowley [Fri, 17 Jun 2016 02:09:35 +0000 (21:09 -0500)]
swr: push/pop DEBUG macro around llvm includes

llvm redefines DEBUG; adding push/pop prevents a undefined reference
to debug_refcnt_state in llvm-3.7+.

v2: add undef DEBUG

Cc: "12.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
8 years agoinclude: Require MSVC 2013 Update 4.
Jose Fonseca [Thu, 23 Jun 2016 10:18:21 +0000 (11:18 +0100)]
include: Require MSVC 2013 Update 4.

Earlier MSVC 2013 releases have troubles compiling some of our C99 code,
so make sure we have Update 4 to avoid confusion.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
8 years agosvga: rename svga_surface_copy() to svga_resource_copy_region()
Brian Paul [Mon, 20 Jun 2016 18:41:23 +0000 (12:41 -0600)]
svga: rename svga_surface_copy() to svga_resource_copy_region()

To be consistent with the pipe_context function name.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agosvga: don't copy blit_info into local var
Brian Paul [Mon, 20 Jun 2016 18:39:55 +0000 (12:39 -0600)]
svga: don't copy blit_info into local var

There's no reason for doing so.

Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agogallium/util: fix some 4-space indentation in blitter code
Brian Paul [Mon, 20 Jun 2016 19:00:28 +0000 (13:00 -0600)]
gallium/util: fix some 4-space indentation in blitter code

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
8 years agosvga: fix texture array update regression
Charmaine Lee [Tue, 21 Jun 2016 17:12:22 +0000 (10:12 -0700)]
svga: fix texture array update regression

With commit fb9fe35, we start using transfer_inline_write
for memcpy TexSubImage path, but that triggers a regression with
texture array in the svga driver.

With this patch, the direct map code will update the texture array
correctly.

Fixes VMware bug 1679293.

Tested with MTT piglit, glretrace, conform.

Reviewed-by: Brian Paul <brianp@vmware.com>
8 years agosvga: fix index/vertex buffer surface reference at draw
Charmaine Lee [Mon, 16 May 2016 22:12:57 +0000 (15:12 -0700)]
svga: fix index/vertex buffer surface reference at draw

Currently with the SetVertexBuffers optimization, we avoid emitting
redundant DXSetVertexBuffers commands. However, these buffers surfaces
will still need to be referenced, otherwise, in the case of linux,
the subsequent surface discard map will map to the existing mob instead
of a new one, causing rendering artifacts.

With this patch, we'll call resource_rebind() to reference the resources
even if we are avoiding the actual set command. This fixes the
rendering artifacts in the window title area running with unity in
Ubuntu 14.04

Tested with piglit, glretrace.

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
8 years agosvga: fix vertex buffer references in the hw state
Charmaine Lee [Tue, 3 May 2016 01:17:48 +0000 (18:17 -0700)]
svga: fix vertex buffer references in the hw state

This patch fixes three issues with vertex buffer references:
(1) Instead of copy the vertex buffer resource handles to the hw state
    in the context structure, use pipe_resource_reference to properly
    reference the vertex buffer resources in the context.

(2) Make sure to unbind those unused vertex buffer resources.

(3) Force to rebind the vertex buffer resources at the first draw of each
    command buffer to make sure the vertex buffer resources are paged in.

Reviewed-by: Brian Paul <brianp@vmware.com>
8 years agosvga: fix index buffer reference in the hw state
Charmaine Lee [Tue, 3 May 2016 01:12:24 +0000 (18:12 -0700)]
svga: fix index buffer reference in the hw state

Instead of copy the index buffer resource handle to the hw state in
the context structure, use pipe_resource_reference to properly reference
the index buffer resource in the context.

Reviewed-by: Brian Paul <brianp@vmware.com>
8 years agoglsl/mesa: stop duplicating geom and tcs layout values
Timothy Arceri [Wed, 22 Jun 2016 02:41:28 +0000 (12:41 +1000)]
glsl/mesa: stop duplicating geom and tcs layout values

We already store these in gl_shader and gl_program here we
remove it from gl_shader_program and just use the values
from gl_shader.

This will allow us to keep the shader cache restore code as
simple as it can be while making it somewhat clearer where these
values originate from.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
8 years agoglsl/mesa: stop duplicating tes layout values
Timothy Arceri [Wed, 22 Jun 2016 02:41:27 +0000 (12:41 +1000)]
glsl/mesa: stop duplicating tes layout values

We already store this in gl_shader and gl_program here we
remove it from gl_shader_program and just use the values
from gl_shader.

This will allow us to keep the shader cache restore code as
simple as it can be while making it somewhat clearer where these
values originate from.

V2: remove unnecessary NULL check

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Iago Toral <itoral@igalia.com>
8 years ago.mailmap: Fixup my email address
Edward O'Callaghan [Wed, 22 Jun 2016 10:42:30 +0000 (20:42 +1000)]
.mailmap: Fixup my email address

Signed-off-by: Edward O'Callaghan <funfunctor@folklore1984.net>
8 years agost/mesa: expose EXT_vertex_array_bgra when supported by backend
Christian Gmeiner [Mon, 20 Jun 2016 07:44:22 +0000 (09:44 +0200)]
st/mesa: expose EXT_vertex_array_bgra when supported by backend

Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
8 years agoanv: Use different BOs for different scratch sizes and stages
Jason Ekstrand [Thu, 16 Jun 2016 22:26:54 +0000 (15:26 -0700)]
anv: Use different BOs for different scratch sizes and stages

This solves a race condition where we can end up having different stages
stomp on each other because they're all trying to scratch in the same BO
but they have different views of its layout.

Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agogenxml: Make ScratchSpaceBasePointer an address instead of an offset
Jason Ekstrand [Thu, 16 Jun 2016 21:58:25 +0000 (14:58 -0700)]
genxml: Make ScratchSpaceBasePointer an address instead of an offset

While we're here, we also fixup MEDIA_VFE_STATE and rename the field in
3DSTATE_VS on gen6-7.5 to be consistent with the others.

Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoanv: Add an allocator for scratch buffers
Jason Ekstrand [Thu, 16 Jun 2016 21:43:41 +0000 (14:43 -0700)]
anv: Add an allocator for scratch buffers

Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agogenxml: Put append counter fields before MCS in RENDER_SURFACE_STATE on gen7
Jason Ekstrand [Tue, 7 Jun 2016 18:37:46 +0000 (11:37 -0700)]
genxml: Put append counter fields before MCS in RENDER_SURFACE_STATE on gen7

The pack header generation scripts can't handle the case where you have
two addresses in the same dword; they just take whatever is the last one.
This meant that the MCS address wasn't properly getting handled.  Since we
don't care about append counters, we can just re-arrange the XML for now.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoanv,isl: Lower storage image formats in anv
Jason Ekstrand [Thu, 9 Jun 2016 20:56:16 +0000 (13:56 -0700)]
anv,isl: Lower storage image formats in anv

ISL was being a bit too clever for its own good and lowering the format for
us.  This is all well and good *if* we always want to lower it.  However,
the GL driver selectively lowers the format depending on whether the
surface is write-only or not.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Allow for full 31-bit buffer texture sizes
Jason Ekstrand [Fri, 10 Jun 2016 17:45:43 +0000 (10:45 -0700)]
isl/state: Allow for full 31-bit buffer texture sizes

Ivy Bridge and above can handle up to 2^31 elements for RAW buffer
surfaces.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Don't use designated initializers for buffer surface state
Jason Ekstrand [Sat, 11 Jun 2016 01:28:49 +0000 (18:28 -0700)]
isl/state: Don't use designated initializers for buffer surface state

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Add assertions for buffer surface restrictions
Jason Ekstrand [Fri, 10 Jun 2016 01:48:07 +0000 (18:48 -0700)]
isl/state: Add assertions for buffer surface restrictions

Acked-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Don't set SurfacePitch for gen9 1-D textures
Jason Ekstrand [Thu, 9 Jun 2016 22:08:13 +0000 (15:08 -0700)]
isl/state: Don't set SurfacePitch for gen9 1-D textures

This field is ignored by the hardware in this case and, on very large 1-D
textures, it can end up being larger than the maximum allowed value.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Use TILEWALK_XMAJOR for linear surfaces on gen7
Jason Ekstrand [Tue, 7 Jun 2016 21:33:40 +0000 (14:33 -0700)]
isl/state: Use TILEWALK_XMAJOR for linear surfaces on gen7

This matches better what happens on gen8 where the "Tiled Surface" and
"Tile Walke" bits are combined into a single two-bit value.  This is also
more consistent with what the GL driver does.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Emit no-op mip tail setup on SKL
Jason Ekstrand [Tue, 7 Jun 2016 01:22:21 +0000 (18:22 -0700)]
isl/state: Emit no-op mip tail setup on SKL

This hasn't ever been a problem in the past but it is recommended by the
hardware docs.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Only set cube face enables if usage includes CUBE_BIT
Jason Ekstrand [Tue, 7 Jun 2016 01:21:17 +0000 (18:21 -0700)]
isl/state: Only set cube face enables if usage includes CUBE_BIT

It seems safe to set it all the time, but this reduces the diff between
the way i965 does it and what ISL does.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Use the layout for computing qpitch rather than dimensions
Jason Ekstrand [Sun, 5 Jun 2016 04:22:21 +0000 (21:22 -0700)]
isl/state: Use the layout for computing qpitch rather than dimensions

For depth/stencil 1-D textures on SKL, we want them layed out in the old
format that has been used since gen4.  In order for the surface state
fill-out code to handle, this it needs to distinguish based on layout
rather than just dimensionality.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Set the IntegerSurfaceFormat bit on Haswell
Jason Ekstrand [Tue, 7 Jun 2016 15:05:45 +0000 (08:05 -0700)]
isl/state: Set the IntegerSurfaceFormat bit on Haswell

This fixes 688 Vulkan CTS tests on Haswell.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/format: Mark R9G9B9E5 as containing 9-bit unsigned float channels
Jason Ekstrand [Tue, 7 Jun 2016 21:40:47 +0000 (14:40 -0700)]
isl/format: Mark R9G9B9E5 as containing 9-bit unsigned float channels

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Don't set RenderTargetViewExtent for texture surfaces
Jason Ekstrand [Fri, 17 Jun 2016 23:40:24 +0000 (16:40 -0700)]
isl/state: Don't set RenderTargetViewExtent for texture surfaces

The docs specify that this only matters for render targets and surfaces
used with typed dataport messages.  On some platforms (gen4-6) the Depth
field has more bits than RenderTargetViewExtent so we can have textures
with more levels than we can render to.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Set SurfaceArray based on the surface dimension
Jason Ekstrand [Tue, 7 Jun 2016 22:30:00 +0000 (15:30 -0700)]
isl/state: Set SurfaceArray based on the surface dimension

According to the PRM, you can't set SurfaceArray for 3D or buffer textures.
There doesn't seem to be a good reason not to set it when we can.  On the
other hand, if we don't set it we can end up getting strange results for
1-layer array textures such as textureSize() returning the wrong results.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Don't force-disable L2 bypass for everything
Jason Ekstrand [Sun, 5 Jun 2016 03:48:55 +0000 (20:48 -0700)]
isl/state: Don't force-disable L2 bypass for everything

We already set the bit in the few cases where it's required by the docs so
there's no need to set it all the time.  This has no noticable perf impact
for Dota 2 on Vulkan with the time demo I have.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Refactor the setup of clear colors
Jason Ekstrand [Fri, 3 Jun 2016 02:02:23 +0000 (19:02 -0700)]
isl/state: Refactor the setup of clear colors

This commit switches clear colors to use #if's instead of a C if.  This
lets us properly handle SNB where the clear color field doesn't exist.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
8 years agoisl/state: Refactor the per-gen isl_to_gen_h/valign tables
Jason Ekstrand [Fri, 3 Jun 2016 02:00:10 +0000 (19:00 -0700)]
isl/state: Refactor the per-gen isl_to_gen_h/valign tables

This moves the #if's around so that halign and valign have different sets
of #if conditions.  This also prepares us for SNB because isl_to_gen_halign
is not defined at all on gen6.

Reviewed-by: Chad Versace <chad.versace@intel.com>
Cc: "12.0" <mesa-stable@lists.freedesktop.org>