Marek Olšák [Wed, 29 Dec 2021 00:10:07 +0000 (19:10 -0500)]
radeonsi: add into the disk cache key whether cached shaders contain LLVM IR
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Thu, 23 Dec 2021 00:32:06 +0000 (19:32 -0500)]
radeonsi: move si_nir_scan_shader into si_shader_info.c
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Mon, 13 Dec 2021 06:14:44 +0000 (01:14 -0500)]
radeonsi: change si_shader_output_values::vertex_stream to a bitmask
to match si_shader_info.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Sun, 12 Dec 2021 04:56:20 +0000 (23:56 -0500)]
radeonsi: use nir->scratch_size instead of ac_count_scratch_private_memory
It's the same.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Sun, 12 Dec 2021 04:51:03 +0000 (23:51 -0500)]
radeonsi: do opt_large_constants & lower_indirect_derefs after uniform inlining
because loop unrolling caused by uniform inlining can eliminate large constants
and indirect derefs.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Mon, 13 Dec 2021 19:36:37 +0000 (14:36 -0500)]
radeonsi: move smoothing to the main shader part to remove 1 live VGPR
The samplemask VGPR that we had to pass to the epilog increased VGPR usage
by 1 for all shaders. Do it in the main function by using the mono key
structure, which causes on-demand compilation and stall, but we'll save
the VGPR.
57794 shaders in 35145 tests
Totals:
SGPRS: 2715856 -> 2716272 (0.02 %)
VGPRS: 1776168 -> 1718432 (-3.25 %)
Spilled SGPRs: 3704 -> 3630 (-2.00 %)
Spilled VGPRs: 1727 -> 1733 (0.35 %)
Private memory VGPRs: 256 -> 256 (0.00 %)
Scratch size: 2008 -> 2016 (0.40 %) dwords per thread
Code Size:
61429584 ->
61393288 (-0.06 %) bytes
Max Waves: 838645 -> 840484 (0.22 %)
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Tue, 14 Dec 2021 03:37:41 +0000 (22:37 -0500)]
radeonsi: pass sample_coverage VGPR index to the PS prolog instead of guessing
The code was correct, but little confusing. This is cleaner.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Mon, 20 Dec 2021 01:48:45 +0000 (20:48 -0500)]
radeonsi: remove unused si_shader::prolog2
This became unused when the GS prolog was removed.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Sat, 18 Dec 2021 14:22:57 +0000 (09:22 -0500)]
radeonsi: don't bind the ESGS ring twice, handle the difference in the shader
The other shader is changed to modify the descriptor to get the required
values.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Sat, 18 Dec 2021 13:41:37 +0000 (08:41 -0500)]
radeonsi: reorder slots for internal buffers, reuse a slot for GS_QUERY_BUF
GFX10_GS_QUERY_BUF is also renamed.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Sat, 18 Dec 2021 08:57:14 +0000 (03:57 -0500)]
radeonsi: simplify compacted_mrt_index in si_export_mrt_color
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Fri, 17 Dec 2021 20:18:54 +0000 (15:18 -0500)]
radeonsi: export mrtz before color exports
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Fri, 17 Dec 2021 20:15:49 +0000 (15:15 -0500)]
radeonsi: remove unnecessary code that was used to find the last export
We can do it trivially like this.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Fri, 17 Dec 2021 19:21:19 +0000 (14:21 -0500)]
radeonsi: set done=1 for PS exports at the end of si_llvm_build_ps_epilog
so that we can reorder the exports
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Fri, 17 Dec 2021 18:45:56 +0000 (13:45 -0500)]
radeonsi: clean up si_export_mrt_color
Remove the intermediate args variable.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Thu, 23 Sep 2021 16:13:08 +0000 (12:13 -0400)]
radeonsi: make get_thread_id_in_tg non-static
for future work
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Sat, 18 Dec 2021 22:04:32 +0000 (17:04 -0500)]
radeonsi: modifiers can't disable DCC
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Mon, 3 Jan 2022 19:04:57 +0000 (14:04 -0500)]
ac/llvm: remove the num_channels parameter from ac_build_buffer_store_dword
It was used when LLVM didn't support vec3 and we had to pass vec4
with num_channels=3. We no longer need to do that.
This also removes the vec3 splitting or conversion to vec4 in callers.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Thu, 23 Sep 2021 15:11:07 +0000 (11:11 -0400)]
ac/llvm: add vindex into ac_build_buffer_store_dword
for future work
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Marek Olšák [Thu, 9 Dec 2021 04:31:39 +0000 (23:31 -0500)]
amd/registers: work around an assertion in parse_kernel_headers.py
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14266>
Iago Toral Quiroga [Tue, 4 Jan 2022 11:56:42 +0000 (12:56 +0100)]
v3dv: add a refcount mechanism to BOs
Until now we have lived without a refcount mechanism in the driver
because in Vulkan the user is responsible for handling the life
span of memory allocations for all Vulkan objects, however,
imported BOs are tricky because the kernel doesn't refcount
so user-space needs to make sure that:
1. When importing a BO into the same device used to create it
(self-importing) it does not double free the same BO.
2. Frees imported BOs that were not allocated through the same
device.
Our initial implementation always freed BOs when requested,
so we handled 2) correctly but not 1) and we would double-free
self-imported BOs. We tried to fix that in commit
d809d9f3
but that broke 2) and we started to leak BOs for some imports.
This fixes the problem for good by adding refcounts to BOs
so that self-imported BOs have a refcnt > 1 and are only freed
when all references are freed.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5769
Tested-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14392>
Marek Olšák [Thu, 9 Dec 2021 00:00:08 +0000 (19:00 -0500)]
radeonsi: decrease the size of si_pm4_state::pm4 except for cs_preamble_state
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 23:40:58 +0000 (18:40 -0500)]
radeonsi: replace SI_PM4_MAX_DW with a max_dw field
it will vary
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 23:32:33 +0000 (18:32 -0500)]
radeonsi: pack si_pm4_state
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 08:10:24 +0000 (03:10 -0500)]
radv: apply spi_cu_en to CU_EN
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 05:56:03 +0000 (00:56 -0500)]
radv: set COMPUTE_DESTINATION_EN_SEn to spi_cu_en
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 07:15:50 +0000 (02:15 -0500)]
radeonsi: apply spi_cu_en to CU_EN
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 06:11:21 +0000 (01:11 -0500)]
radeonsi: program COMPUTE_STATIC_THREAD_MGMT_SE4..7 on Arcturus
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 05:56:03 +0000 (00:56 -0500)]
radeonsi: set COMPUTE_DESTINATION_EN_SEn to spi_cu_en
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 05:33:10 +0000 (00:33 -0500)]
ac/gpu_info: add AMD_CU_MASK environment variable to set CU_EN
requested internally
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Marek Olšák [Wed, 8 Dec 2021 04:23:21 +0000 (23:23 -0500)]
ac/gpu_info: set cu_mask correctly for Arcturus
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14122>
Emma Anholt [Mon, 20 Dec 2021 23:15:38 +0000 (15:15 -0800)]
i915g: Turn off FP16 in the vertex shaders.
This ended up being turned on in gallivm, but since we use nir_to_tgsi on
the VS and TGSI doesn't have FP16, we can't let that happen.
Fixes:
f814a2449e2c ("llvmpipe: enable FP16 and update CL + traces piglit results.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14403>
satmandu [Fri, 10 Sep 2021 20:58:09 +0000 (20:58 +0000)]
Fix compilation on armv7l with gcc 11.2.0
Cc: mesa-stable
Reviewed-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12810>
Timothy Arceri [Tue, 7 Dec 2021 04:35:48 +0000 (15:35 +1100)]
glsl/glcpp: make sure to expand new token after concatenation
Previously the code was using a hack to change the token type from
INDETIFIER -> OTHER in order to avoid getting in an infinite loop
expanding the tokens. This worked ok until we got to a paste where
the replacement parameters had already had their type changed
to OTHER because the newly created paste token would then
inherit the OTHER type and never get expanded inself.
For example with the follow code:
#define STEP_ONE() \
out_Color = vec4(0.0,1.0,0.0,1.0)
#define GLUE(x,y) x ## _ ## y
#define EVALUATE(x,y) GLUE(x,y)
#define STEP(stepname) EVALUATE(STEP, stepname)()
#define PERFORM_RAYCASTING_STEP STEP(ONE)
This would get all the way to expanding PERFORM_RAYCASTING_STEP to
STEP_ONE() but because it was created via the paste `x ## _ ## y`
it would never get any further.
To fix this we remove the OTHER hack and instead just track if the
token has already been handled via a bool value `explanding`.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5724
Fixes:
28842c2331e6 ("glcpp: Implement token pasting for non-function-like macros")
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14101>
Emma Anholt [Fri, 31 Dec 2021 00:09:50 +0000 (16:09 -0800)]
tgsi/exec: Simplify indirects now that they always use the ADDR file.
This was a lot of extra code in the hot path of getting though
fetch_src_file_channel(). No significant perf difference in softpipe,
though.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14360>
Emma Anholt [Thu, 30 Dec 2021 23:59:15 +0000 (15:59 -0800)]
gallium: Delete PIPE_CAP_TGSI_ANY_REG_AS_ADDRESS
Softpipe was the only driver still using this feature. I had enabled it
in
ba22f014f9db ("softpipe: Enable PIPE_CAP_TGSI_ANY_REG_AS_ADDRESS;") for
an instr count win, but it's really not important to that driver and it's
not worth keeping the knob around just for that.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14360>
Emma Anholt [Thu, 30 Dec 2021 03:23:14 +0000 (19:23 -0800)]
nir_to_tgsi: Use the same address reg mappings as GLSL-to-TGSI did.
It turns out r600 has a bunch of expectations about the Dimension being in
ADDR[1].x, and sampler or atomic indirects being in ADDR[2].x. It's
simpler to just use this static assignment than our dynamic one, anyway.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14360>
Biju Das [Thu, 16 Dec 2021 16:52:57 +0000 (16:52 +0000)]
kmsro: Add 'rcar-du' driver support
rcar-du is used by Renesas R-Car and RZ/G SOCs.
Driver is available in the mainline kernel [1]
[1] https://elixir.bootlin.com/linux/latest/source/drivers/gpu/drm/rcar-du
Suggested-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Show Liu <show.liu.yj@renesas.com>
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14393>
Adam Jackson [Thu, 16 Dec 2021 02:32:58 +0000 (21:32 -0500)]
mesa: Remove unused _mesa_get_linear_format_srgb
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14220>
Adam Jackson [Thu, 16 Dec 2021 02:32:19 +0000 (21:32 -0500)]
mesa: Remove unused _mesa_format_fallback_rgbx_to_rgba
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14220>
Adam Jackson [Thu, 16 Dec 2021 02:21:55 +0000 (21:21 -0500)]
mesa: Remove unused _mesa_bind_texture
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14220>
Adam Jackson [Thu, 16 Dec 2021 01:56:52 +0000 (20:56 -0500)]
mesa: Remove unused _mesa_AllocTextureStorage_sw
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14220>
Adam Jackson [Thu, 16 Dec 2021 01:02:14 +0000 (20:02 -0500)]
mesa: Remove unused _mesa_allow_light_in_model
ctx::_ForceEyeCoords is now never not false, so remove and simplify to
match.
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14220>
Timur Kristóf [Fri, 17 Dec 2021 16:55:36 +0000 (17:55 +0100)]
radv: Support VRS for mesh shaders.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14193>
Timur Kristóf [Fri, 17 Dec 2021 16:54:12 +0000 (17:54 +0100)]
aco: Emit VRS rate when it's per-primitive.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14193>
Timur Kristóf [Fri, 17 Dec 2021 16:51:18 +0000 (17:51 +0100)]
radv: Lower primitive shading rate for mesh shaders.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14193>
Timur Kristóf [Fri, 17 Dec 2021 16:50:35 +0000 (17:50 +0100)]
radv: Note when a mesh shader writes the primitive shading rate.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14193>
Rob Clark [Sun, 26 Dec 2021 20:45:49 +0000 (12:45 -0800)]
clover: Move min image support check
If the gallium driver supports images, but not the minimum image
requirements for CL, then simply don't claim to support images. This
is better than not claiming to support CL at all.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14310>
Samuel Pitoiset [Fri, 15 Oct 2021 11:30:16 +0000 (13:30 +0200)]
radv: add UMR markers for the vertex prolog
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13376>
Samuel Pitoiset [Fri, 15 Oct 2021 09:03:04 +0000 (11:03 +0200)]
radv: dump the VS prolog disassembly to the hang report
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13376>
Samuel Pitoiset [Fri, 15 Oct 2021 08:52:09 +0000 (10:52 +0200)]
radv: save the vertex prolog to the trace BO for debugging
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13376>
Samuel Pitoiset [Fri, 15 Oct 2021 08:03:21 +0000 (10:03 +0200)]
radv,aco: keep track of the prolog disassembly if necessary
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13376>
Samuel Pitoiset [Fri, 15 Oct 2021 07:56:56 +0000 (09:56 +0200)]
aco: do not print prologs disassembly if no disassembler
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13376>
Samuel Pitoiset [Wed, 29 Dec 2021 07:28:22 +0000 (08:28 +0100)]
aco: fix a dynamic-stack-buffer-overflow when printing instructions
Detected by ASAN.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14329>
Dave Airlie [Wed, 29 Dec 2021 18:57:07 +0000 (04:57 +1000)]
mapi: generate correct dispatch for EXT_draw_instanced
These APIs can be exposed in GLES2.0 via EXT_draw_instanced,
they were incorrectly being stuck on GLES 3.0 only.
Fixes piglit ext_draw_instanced-drawarrays
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14338>
Pavel Ondračka [Mon, 3 Jan 2022 14:47:43 +0000 (15:47 +0100)]
r300: Remove broken optimization in rc_transform_KILL
The logic was reversed so this was not only not working but it was
also removing random instructions around. The special IF-KILP-ENDIF
case this optimization was targeting is already transformed to
KILL_IF in the TGSI, so just remove this altogether.
This fixes piglit glsl-fs-discard-04
v2: Update the comment as well
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/343
Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip.gawin@zoho.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14378>
Thomas H.P. Andersen [Sat, 25 Dec 2021 17:26:34 +0000 (18:26 +0100)]
ci: debian-clang: drop -Wno-error=absolute-value
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14302>
Thomas H.P. Andersen [Sat, 25 Dec 2021 17:25:34 +0000 (18:25 +0100)]
broadcom/compiler: fix compile warning -Wabsolute-value
fixes a compile warning with clang
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14302>
Thomas H.P. Andersen [Sat, 25 Dec 2021 17:24:05 +0000 (18:24 +0100)]
xa: fix compile warning for -Wabsolute-value
fixes a compile warning with clang
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14302>
Lionel Landwerlin [Sat, 18 Dec 2021 14:48:00 +0000 (16:48 +0200)]
util/u_trace: protect against reentrant calls
With the Iris driver, the first tracepoint triggers the tracepoint for
"begin of command buffer" which leads to 2 tracepoints using the same
idx.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14254>
Marek Olšák [Sat, 1 Jan 2022 06:36:24 +0000 (01:36 -0500)]
driconf: enable glthread for Minecraft-FTB, Stellaris, Battletech
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14368>
M Henning [Thu, 30 Dec 2021 01:06:26 +0000 (20:06 -0500)]
nouveau/nir: Lower 64-bit phis
Fixes arb_gpu_shader_fp64-fs-non-uniform-control-flow-ubo
on kepler with NV50_PROG_USE_NIR=1
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14343>
Marek Olšák [Tue, 21 Dec 2021 17:56:38 +0000 (12:56 -0500)]
radeonsi: add a debug option that disables DCC for all exported buffers
requested internally
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14283>
Thomas H.P. Andersen [Wed, 29 Dec 2021 10:19:39 +0000 (11:19 +0100)]
zink: malloc/sizeof mismatch
buffer_infos is VkBufferView* but malloc is using the
size of VkImageView. The size is the same but this
fixes a warning
Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14335>
Thomas H.P. Andersen [Thu, 23 Dec 2021 01:07:40 +0000 (02:07 +0100)]
ci: debian-clang: -Wno-error for sometimes-uninitialized
The last of this kind of warning is gone for CPP
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14296>
Thomas H.P. Andersen [Thu, 23 Dec 2021 01:05:47 +0000 (02:05 +0100)]
r600/sb: silence a sometimes-uninitialized warning
Fixes a warning from clang
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14296>
Mike Blumenkrantz [Tue, 16 Nov 2021 15:57:24 +0000 (10:57 -0500)]
mesa/vbo: be more comprehensive for degenerate primitive conversion in dlists
these shouldn't result in any sort of draw, and passing the unsupported primtype
to the driver is invalid
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13845>
Alyssa Rosenzweig [Sun, 2 Jan 2022 01:38:46 +0000 (20:38 -0500)]
pan/bi: Fix load_const of 1-bit booleans
For historical reasons, we ingest 1-bit booleans in NIR but expand them
to 16/32-bit booleans in the backend IR. We need to handle this case
when loading boolean constants, extending from 1-bit to 16/32-bit as
required. This issue is masked by effective constant folding for
booleans, but is visible in a shader from Firefox WebRender.
Fixes:
646e03c4519 ("pan/bi: Temporarily switch back to 0/~0 bools")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reported-by: Icecream95
Closes: #5797
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14371>
Uday Kiran Pichika [Tue, 28 Dec 2021 08:39:46 +0000 (10:39 +0200)]
anv: enable adaptive sync for ANV
Signed-off-by: Uday Kiran Pichika <pichika.uday.kiran@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6420>
Uday Kiran Pichika [Tue, 28 Dec 2021 08:39:03 +0000 (10:39 +0200)]
iris: enable adaptive sync for IRIS
Signed-off-by: Uday Kiran Pichika <pichika.uday.kiran@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6420>
Alyssa Rosenzweig [Mon, 8 Nov 2021 14:52:25 +0000 (09:52 -0500)]
pan/bi: Use fused dual source blending
Instead of emitting a pile of moves to fixed registers at codegen time
and hoping everything works out, add a second staging source to the
BLEND instruction in the intermediate representation containing the dual
source colour, and modify register allocation appropriately. This better
models the operation of blending render target #0 with two sources.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Mon, 8 Nov 2021 14:50:43 +0000 (09:50 -0500)]
pan/bi: Allow an extra staging source
Useful for dual source blending.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Mon, 8 Nov 2021 14:49:19 +0000 (09:49 -0500)]
pan/bi: Use is_staging_src helper
This will allow us to have a separate set of staging sources, which will
be useful for dual-source blending.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Fri, 29 Oct 2021 14:01:22 +0000 (10:01 -0400)]
panfrost: Remove pan_nir_reorder_writeout
Now that dual source stores are fused, there's no ordering hazard so we
can remove this tricky code.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Fri, 29 Oct 2021 13:41:05 +0000 (09:41 -0400)]
panfrost: Combine dual source blends
Reuse the store_combined_output_pan infrastructure for dual source
blending as well. This simplifies ordering.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Fri, 29 Oct 2021 12:48:46 +0000 (08:48 -0400)]
nir: Extend store_combined_output_pan
Extend store_combined_output_pan to take a dual source blend input in
addition to colour, depth, and stencil inputs. Use the last source for
this, and represent the type with the DEST_TYPE index. This is a hack
but there is no SRC2_TYPE and NIR doesn't seem to mind as long as we
know what we mean. This allows the backend to emit a combined "blend
render target #0" instruction taking two sources.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Fri, 29 Oct 2021 13:02:37 +0000 (09:02 -0400)]
panfrost: Simplify blend lowering pass
Combine depth/stencil writeout redundancy. This simplifies the code and
will allow us to add dual source blend handling without duplicating even
more code.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Sat, 1 Jan 2022 15:11:33 +0000 (10:11 -0500)]
nir: Eliminate store_combined_output_pan BASE
It's meaningless for this intrinsic and is just adding noise to the
lowering pass.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Sat, 1 Jan 2022 15:41:58 +0000 (10:41 -0500)]
pan/mdg: Don't read base for combined stores
`base` is meaningless for combined stores, so don't read it. This allows
us to remove the base index from the intrinsic and simplify the lowering
code generating the combined store intrinsic.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Alyssa Rosenzweig [Sat, 1 Jan 2022 15:09:21 +0000 (10:09 -0500)]
pan/bi: Don't read base for combined stores
`base` is meaningless for combined stores, so don't read it. This allows
us to remove the base index from the intrinsic and simplify the lowering
code generating the combined store intrinsic.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13714>
Tatsuyuki Ishi [Sun, 28 Nov 2021 15:12:18 +0000 (00:12 +0900)]
aco: lower masked swizzle to DPP8
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13971>
Tatsuyuki Ishi [Sun, 28 Nov 2021 15:12:04 +0000 (00:12 +0900)]
aco: support DPP8
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13971>
Jesse Natalie [Thu, 30 Dec 2021 00:56:53 +0000 (16:56 -0800)]
gallium/aux: Move index offsetting from prim restart to primconvert
Fixes:
b34fed64 ("u_prim_restart: Fix index scanning with start offset")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5799
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14346>
Bas Nieuwenhuizen [Fri, 31 Dec 2021 16:27:18 +0000 (17:27 +0100)]
radv: Remove dependencies on vk_common entrypoints.
Fixes:
91fe0b5629d ("radv: Delete lots of sync code.")
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14366>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 22:06:48 +0000 (23:06 +0100)]
radv: Rename submit2->submit.
Cleanup after the big switcheroo.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 20:02:30 +0000 (21:02 +0100)]
radv: Remove syncobj reset mechanism.
Now done with the vulkan runtime.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 18:38:08 +0000 (19:38 +0100)]
radv: Delete lots of sync code.
This make things all fall back to the common synchronization functions
which will switch things to the new submission path and to use
vk_fence and vk_semaphore.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
radv: Use common AcquireNextImage2KHR.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
radv: Use common functions for Metro Exodus layer.
Needs to be squashed with the big "switch the world" deletion patch
but kept it separate for review.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 17:55:47 +0000 (18:55 +0100)]
radv: Add new submission path for use by the common sync framework.
Basically just a copy of radv's internal deferred submission handler.
This uses the new winsys submit2 to later be able to move the sem_info
to the winsys.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 16:51:23 +0000 (17:51 +0100)]
radv: Add new cs_submit2 winsys call.
A bit more runtime dependence so we can remove all the syncobj stuff
in radv itself.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 15:51:39 +0000 (16:51 +0100)]
radv: Use vk_command_buffer for preambles.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 15:38:06 +0000 (16:38 +0100)]
radv: Set horizontal sync types.
We basically use 2 types for amdgpu:
1) syncobj. This supports both binary & timeline and autodetects
all the features.
2) emulated timelines if (1) doesn't detect timeline syncobj support.
Note that one has to put these in order of preference so that the
common CreateFence/CreateSemaphore functions decide on the right
type.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Thu, 30 Dec 2021 15:03:37 +0000 (16:03 +0100)]
radv: Add function to allow WSI signalling fences/semaphores.
Added but not enabled yet. Will enable in the big switchover.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 12:39:21 +0000 (13:39 +0100)]
radv: Initialize vk device drm fd.
For common sync.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Thu, 30 Dec 2021 15:21:06 +0000 (16:21 +0100)]
meson: Bump libdrm_amdgpu version req to 2.4.109.
For getting the fd from the device.
This uses the new function introduced with libdrm 2.4.109. This has
already been updated in CI and is already required for mesa common
bits so requiring it for libdrm_amdgpu is no big deal.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 16:59:26 +0000 (17:59 +0100)]
radv: Use vulkan runtime for device lost.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 19:13:39 +0000 (20:13 +0100)]
radv: Use dispatch table for wsi_display.c
Pretty sure this could be moved into WSI.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 21:25:01 +0000 (22:25 +0100)]
radv: Use dispatch table for QueueWaitIdle in the SQTT layer.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 17:52:16 +0000 (18:52 +0100)]
vulkan/runtime: Add sparse bind support.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Bas Nieuwenhuizen [Sun, 28 Nov 2021 17:24:16 +0000 (18:24 +0100)]
vulkan/runtime: Refactor queue submit to take an argument struct.
For merging with the sparse binding structs.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13974>
Daniel Schürmann [Mon, 4 Oct 2021 10:29:04 +0000 (11:29 +0100)]
aco: don't split VOP3P definitions
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13576>