mesa.git
4 years agoac: add more fields to ac_gpu_info
Samuel Pitoiset [Thu, 20 Feb 2020 10:35:23 +0000 (11:35 +0100)]
ac: add more fields to ac_gpu_info

For RGP traces.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3900>

4 years agoci: Enable -Werror on meson-vulkan and meson-testing.
Eric Anholt [Tue, 17 Dec 2019 05:23:02 +0000 (21:23 -0800)]
ci: Enable -Werror on meson-vulkan and meson-testing.

I want to make sure that I don't introduce warnings in turnip where we
have active work going on, and I also want to make sure that the drivers
we care about testing are warnings-clean.

As with the previous -Werror change, this is for CI only and doesn't
affect end-user builds.

Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3607>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3607>

4 years agoaco: Fix signed-vs-unsigned warning.
Eric Anholt [Tue, 4 Feb 2020 23:12:18 +0000 (15:12 -0800)]
aco: Fix signed-vs-unsigned warning.

The previous instance of this comparision was 1u to avoid the warning, fix
this one too.

Fixes: dba71de5c636 ("aco: only create parallelcopy to restore exec at loop exit if needed")
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3607>

4 years agogallium/u_vbuf: silence a warning by using unreachable
Marek Olšák [Thu, 27 Feb 2020 00:16:01 +0000 (19:16 -0500)]
gallium/u_vbuf: silence a warning by using unreachable

Reviewed-by: Eric Anholt <eric@anholt.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3970>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3970>

4 years agomesa: fix 11 warnings
Marek Olšák [Wed, 26 Feb 2020 22:21:43 +0000 (17:21 -0500)]
mesa: fix 11 warnings

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3970>

4 years agonir: fix 5 warnings
Marek Olšák [Wed, 26 Feb 2020 22:20:51 +0000 (17:20 -0500)]
nir: fix 5 warnings

Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3970>

4 years agogallivm: fix 5 warnings
Marek Olšák [Wed, 26 Feb 2020 22:00:19 +0000 (17:00 -0500)]
gallivm: fix 5 warnings

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3970>

4 years agonir: replace GCC unroll with an option that works on GCC < 8.0
Marek Olšák [Wed, 26 Feb 2020 21:57:37 +0000 (16:57 -0500)]
nir: replace GCC unroll with an option that works on GCC < 8.0

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3970>

4 years agomesa: fix incorrect prim.begin/end for glMultiDrawElements
Marek Olšák [Fri, 14 Feb 2020 03:56:54 +0000 (22:56 -0500)]
mesa: fix incorrect prim.begin/end for glMultiDrawElements

This has no effect on Gallium, but it affects tnl.

Cc: 19.3 20.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Mathias Fröhlich <mathias.froehlich@web.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agomesa: optimize glMultiDrawArrays, call Draw only once (v2)
Marek Olšák [Wed, 12 Feb 2020 22:16:45 +0000 (17:16 -0500)]
mesa: optimize glMultiDrawArrays, call Draw only once (v2)

v2: use the macros

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agomesa: don't unroll glMultiDrawElements if one count is 0
Marek Olšák [Wed, 12 Feb 2020 22:21:35 +0000 (17:21 -0500)]
mesa: don't unroll glMultiDrawElements if one count is 0

let the driver skip or submit an empty draw call.

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/3990>

4 years agomesa: clean up glMultiDrawElements code, use alloca for small draw count (v2)
Marek Olšák [Wed, 12 Feb 2020 20:30:30 +0000 (15:30 -0500)]
mesa: clean up glMultiDrawElements code, use alloca for small draw count (v2)

v2: use calloc, add reusable macros

Reviewed-by: Mathias Fröhlich <mathias.froehlich@web.de> (v1)
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> (v1)
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agomesa: move num_instances and base_instance out of _mesa_prim
Marek Olšák [Tue, 11 Feb 2020 02:42:56 +0000 (21:42 -0500)]
mesa: move num_instances and base_instance out of _mesa_prim

They are never used by multi draws and internal draws.

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/3990>

4 years agomesa: remove redundant _mesa_prim::is_indexed
Marek Olšák [Tue, 11 Feb 2020 01:25:52 +0000 (20:25 -0500)]
mesa: remove redundant _mesa_prim::is_indexed

Instead, check (ib != NULL) like all other drivers.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agomesa/i965: remove _mesa_prim::indirect_offset
Marek Olšák [Tue, 11 Feb 2020 01:19:37 +0000 (20:19 -0500)]
mesa/i965: remove _mesa_prim::indirect_offset

Only i965 was using it.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agogallium/u_threaded: convert dividing by index_size to a bit shift
Marek Olšák [Wed, 26 Feb 2020 20:10:47 +0000 (15:10 -0500)]
gallium/u_threaded: convert dividing by index_size to a bit shift

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agogallium/u_threaded: fix uploading user indices with start != 0
Marek Olšák [Fri, 14 Feb 2020 20:28:28 +0000 (15:28 -0500)]
gallium/u_threaded: fix uploading user indices with start != 0

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agogallium: pass cso_velems_state into cso_context instead of pipe_vertex_element
Marek Olšák [Tue, 21 Jan 2020 03:59:49 +0000 (22:59 -0500)]
gallium: pass cso_velems_state into cso_context instead of pipe_vertex_element

This removes one memcpy from the CSO hashing code.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agogallium/cso_hash: inline struct cso_hash_data
Marek Olšák [Tue, 11 Feb 2020 23:15:02 +0000 (18:15 -0500)]
gallium/cso_hash: inline struct cso_hash_data

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agogallium/cso_hash: pack cso_node better
Marek Olšák [Wed, 22 Jan 2020 01:17:32 +0000 (20:17 -0500)]
gallium/cso_hash: pack cso_node better

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agomesa: remove unused "indirect" parameter from Driver.Draw
Marek Olšák [Tue, 11 Feb 2020 01:02:30 +0000 (20:02 -0500)]
mesa: remove unused "indirect" parameter from Driver.Draw

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agoi965: stop using "indirect" parameter from Driver.Draw (non-indirect)
Marek Olšák [Tue, 11 Feb 2020 00:54:44 +0000 (19:54 -0500)]
i965: stop using "indirect" parameter from Driver.Draw (non-indirect)

The parameter will be removed.

v2: added UNUSED, removed "!!"

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3990>

4 years agoanv: Remove unused field `urb.total_size`
Caio Marcelo de Oliveira Filho [Thu, 27 Feb 2020 19:02:17 +0000 (11:02 -0800)]
anv: Remove unused field `urb.total_size`

This was used before the URB calculation functions were shared by GL
and Vulkan.  Also drop the substruct for the remaining, `l3_config` is
a good name on its own.

Also-written-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3981>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3981>

4 years agopan/midgard: Use address analysis for globals, etc
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:41:17 +0000 (09:41 -0500)]
pan/midgard: Use address analysis for globals, etc

..instead of opencoding for constants and doing the rest as ALU.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Add address analysis framework
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:38:21 +0000 (09:38 -0500)]
pan/midgard: Add address analysis framework

Midgard has the ability to calculate addresses as part of the load/store
pipeline. We'd like to make use of this to avoid doing this work on the
ALU pipes. To do so, when emitting globals/SSBOs/shareds, we walk the
tree looking for address arithmetic to try to parse out something the
hardware can work with, letting the original instructions be DCE'd
ideally. This analysis is done at the NIR level to properly account for
some messy details of vectorization which we'd rather not poke at the
backend level. (Originally I wrote this as a MIR pass but I'm fairly
sure it was wrong.)

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Force address alignment
Alyssa Rosenzweig [Thu, 27 Feb 2020 15:07:32 +0000 (10:07 -0500)]
pan/midgard: Force address alignment

I thought we already had this but... maybe not..

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Round up bytemasks when promoting uniforms
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:37:28 +0000 (09:37 -0500)]
pan/midgard: Round up bytemasks when promoting uniforms

Fixes crashes with uniform promotion in certain mixed type
circumstances.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Fix load/store argument sizing
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:36:46 +0000 (09:36 -0500)]
pan/midgard: Fix load/store argument sizing

The swizzles are as-if they were 32-bit regardless of the bitness of the
operation, but the source sizes can and do change depending on the
flags. Account for this in the analysis.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Add LDST_ADDRESS property
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:35:50 +0000 (09:35 -0500)]
pan/midgard: Add LDST_ADDRESS property

Many load/store ops (used for globals, SSBOs, shared memory, etc) have
the ability to compute addresses directly. Mark off which ones behave
like this.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Extract nir_ssa_index helper
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:35:12 +0000 (09:35 -0500)]
pan/midgard: Extract nir_ssa_index helper

In case we don't have a nir_src.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Partially fix 64-bit swizzle alignment
Alyssa Rosenzweig [Thu, 27 Feb 2020 14:15:00 +0000 (09:15 -0500)]
pan/midgard: Partially fix 64-bit swizzle alignment

When mixing 32/64-bit, we need to align the 32-bit registers to get the
required alignment. This isn't quite enough yet, though, since user
swizzles could bypass and will need to be lowered to 32-bit moves
(outstanding todo).

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Allow fusing inverted sources for inverted ops
Alyssa Rosenzweig [Wed, 26 Feb 2020 18:51:07 +0000 (13:51 -0500)]
pan/midgard: Allow fusing inverted sources for inverted ops

It doesn't make a difference to the actual algorithm, so let's get rid
of them.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopan/midgard: Allow inverted inverted ops
Alyssa Rosenzweig [Wed, 26 Feb 2020 18:50:46 +0000 (13:50 -0500)]
pan/midgard: Allow inverted inverted ops

We'd like to transform `inand.not` back to `iand` and so forth.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agopanfrost: Increase SSBO/image limit from 4->8
Alyssa Rosenzweig [Wed, 26 Feb 2020 18:37:10 +0000 (13:37 -0500)]
panfrost: Increase SSBO/image limit from 4->8

Fixes an error compiling some shaders.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3978>

4 years agoetnaviv: disable INT_FILTER for ASTC
Jonathan Marek [Thu, 20 Feb 2020 02:14:35 +0000 (21:14 -0500)]
etnaviv: disable INT_FILTER for ASTC

Tested on GC3000: INT_FILTER bit is incompatible with ASTC formats.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3927>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3927>

4 years agoanv: Remove unused field xfb_used from anv_pipeline
Caio Marcelo de Oliveira Filho [Wed, 26 Feb 2020 23:54:33 +0000 (15:54 -0800)]
anv: Remove unused field xfb_used from anv_pipeline

Since we only use xfb_info for GEN >= 8, make the ifdef cover that
local variable.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3973>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3973>

4 years agoci: Include db410c support in the ARM container.
Eric Anholt [Tue, 28 Jan 2020 20:02:39 +0000 (12:02 -0800)]
ci: Include db410c support in the ARM container.

I'm working on moving the db410c CI from docker to LAVA, which means we
get to boot a custom kernel.  To do that, we need to enable ARCH_QCOM in
the kernel, save the dtb around, and include abootimg in our container so
that we can generate combined kernel/dtb/ramdisk images for fastboot.

LAVA's fastboot support is unable to pack the overlay into an abootimg
image, just a cpio rootfs.  We could flash the cpio rootfs after overlay
addition, but that takes 2 minutes to do, and causes wear on the devices.
Instead, we'll bring up the network at boot and use wget to fetch the
overlay.  We'll want network support anyway, so that we can transfer the
failure xmls back to the gitlab job's artifacts at some point.

Since the msm GPU and realtek network firmware increase our payload by
3MB, add in firmware compression so that it doesn't waste as much RAM on
devices not using it.

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3928>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3928>

4 years agoci: Shrink the arm64 kernel build a bit.
Eric Anholt [Tue, 28 Jan 2020 20:24:21 +0000 (12:24 -0800)]
ci: Shrink the arm64 kernel build a bit.

No sense building some of these subsystems for just testing the GPU.
Saves some container rebuild time and kernel contents to move around.

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3928>

4 years agoci: Stop disabling ACPI in the LAVA arm64 kernel build.
Eric Anholt [Tue, 11 Feb 2020 22:34:24 +0000 (14:34 -0800)]
ci: Stop disabling ACPI in the LAVA arm64 kernel build.

I can't figure out why, but it's necessary to get QCOM_CLK_APCS_MSM8916
for db410c.

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3928>

4 years agoci: Remove LLVM from ARM test drivers.
Eric Anholt [Mon, 24 Feb 2020 18:31:33 +0000 (10:31 -0800)]
ci: Remove LLVM from ARM test drivers.

The LLVM libraries were a significant fraction of the entire payload
(55M/250M uncompressed) into the initramfs of the test boards, but
LLVM is only used for the draw module used in select/feedback (which
isn't even tested in CI on ARM yet).

Assume that llvmpipe draw is safe enough for ARM given the coverage on
x86, and disable LLVM for these jobs.

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3928>

4 years agoci: Split out radv build-testing on arm64
Rohan Garg [Thu, 20 Feb 2020 15:37:48 +0000 (16:37 +0100)]
ci: Split out radv build-testing on arm64

radv needs libllvm which increases our ramdisk size
significantly. Since this driver is only build tested,
we can split it out into a separate job.

Signed-off-by: Rohan Garg <rohan.garg@collabora.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3928>

4 years agogitlab-ci: Skip dEQP-GLES3.functional.shaders.derivate.*
Tomeu Vizoso [Thu, 27 Feb 2020 14:09:30 +0000 (15:09 +0100)]
gitlab-ci: Skip dEQP-GLES3.functional.shaders.derivate.*

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agogitlab-ci: Remove GLES3 test from Panfrost fails list
Tomeu Vizoso [Thu, 27 Feb 2020 13:18:02 +0000 (14:18 +0100)]
gitlab-ci: Remove GLES3 test from Panfrost fails list

Seems to have been fixed recently.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agogitlab-ci: Use PAN_MESA_DEBUG=gles3 for Panfrost
Tomeu Vizoso [Thu, 27 Feb 2020 12:30:29 +0000 (13:30 +0100)]
gitlab-ci: Use PAN_MESA_DEBUG=gles3 for Panfrost

We can drop now the GLES version overrides now that we have a DEBUG flag
that enables all what is expected from a GLES 3.0 implementation.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Add PAN_MESA_DEBUG=gles3 option
Alyssa Rosenzweig [Thu, 20 Feb 2020 14:30:42 +0000 (09:30 -0500)]
panfrost: Add PAN_MESA_DEBUG=gles3 option

This enables experimental GLES3.0 support without ES3.1/2 hacks.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Expose PIPE_CAP_PRIMITIVE_RESTART
Alyssa Rosenzweig [Thu, 20 Feb 2020 14:43:34 +0000 (09:43 -0500)]
panfrost: Expose PIPE_CAP_PRIMITIVE_RESTART

It works just fine, we just forgot to expose the CAP.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Simplify stack shift calculation
Alyssa Rosenzweig [Tue, 25 Feb 2020 20:40:20 +0000 (15:40 -0500)]
panfrost: Simplify stack shift calculation

I'm not sure why I never saw smaller values, but here you go.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Reserve an extra page for spilling
Alyssa Rosenzweig [Tue, 25 Feb 2020 20:36:11 +0000 (15:36 -0500)]
panfrost: Reserve an extra page for spilling

I'm not sure what this is for, but the blob does it and I'd rather not
poke farther than needed into hardware-internal details.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Default to 256 threads for TLS
Alyssa Rosenzweig [Tue, 25 Feb 2020 20:34:51 +0000 (15:34 -0500)]
panfrost: Default to 256 threads for TLS

I'm not sure where I got the impression 1024 was the right number. From
kbase:

   #define THREAD_MT_DEFAULT 256

(where MT = "max threads" and the threads to allocate for TLS is <= max
threads). Let's cut out memory footprint for spilling by 75% :)

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Fix param getting
Alyssa Rosenzweig [Tue, 25 Feb 2020 16:52:52 +0000 (11:52 -0500)]
panfrost: Fix param getting

....Oops....

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Don't set shared->unk0
Alyssa Rosenzweig [Tue, 25 Feb 2020 20:32:23 +0000 (15:32 -0500)]
panfrost: Don't set shared->unk0

This field controls the size of per-thread temporaries (somehow this is
separate from the regular stack for register spilling..), though I'm not
certain on the details. Regardless this value of 0x1e despite being used
in places by the blob seems wrong and is interfering with correct sizing
of the stack.

We don't use non-spilling scratchpad yet, so this is just here to fix
some details of spilling.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Update spilling comment framebuffer->shared
Alyssa Rosenzweig [Tue, 25 Feb 2020 20:34:16 +0000 (15:34 -0500)]
panfrost: Update spilling comment framebuffer->shared

All of this should apply equally with compute shaders, as far as I know.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Fix padded_vertex_count generation
Alyssa Rosenzweig [Tue, 25 Feb 2020 13:16:52 +0000 (08:16 -0500)]
panfrost: Fix padded_vertex_count generation

These two cases were flipped from the notes, leading to underestimates
of the padded vertex count, manifesting as visual corruption (random
geometry messed up). This issue was raised when noticing the corruption
went away when dramaticlaly oversizing max_index on an instanced indexed
draw, and then checking that padded_count >= vertex_count -- which
turned out *not* to be the case on certain inputs, a clear issue. Hence
looking into this routine...

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Fix gl_VertexID/InstanceID
Alyssa Rosenzweig [Thu, 20 Feb 2020 15:03:29 +0000 (10:03 -0500)]
panfrost: Fix gl_VertexID/InstanceID

Fixes a bunch of tests in dEQP-GLES3.functional.instanced.*.

Fixes: 027944c7c8c ("panfrost: Avoid reading GPU memory when packing
vertices")

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopan/midgard: Don't spill near a branch
Alyssa Rosenzweig [Mon, 24 Feb 2020 14:36:18 +0000 (09:36 -0500)]
pan/midgard: Don't spill near a branch

Fixes
dEQP-GLES2.functional.shaders.indexing.varying_array.vec2_dynamic_loop_write_
static_read with register pressure forced down.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopan/decode: Dump scratchpad size if present
Alyssa Rosenzweig [Tue, 25 Feb 2020 20:40:46 +0000 (15:40 -0500)]
pan/decode: Dump scratchpad size if present

This will help us narrow the size required for thread local storage.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3950>

4 years agopanfrost: Implement index buffer cache
Alyssa Rosenzweig [Wed, 19 Feb 2020 15:32:20 +0000 (10:32 -0500)]
panfrost: Implement index buffer cache

For index bufer resources (not user index buffers), we're able to cache
results. In practice, the cache works pretty dang well. It's still
important that the min/max computation is efficient (since when the
cache misses it'll run at draw-time and we don't want jank), but this
can eliminate a lot of computations entirely.

We use a custom data structure for caching. Search is O(N) to the size
but sizes are capped so it's effectively O(1). Insertion is O(1) with
automatic oldest eviction, on the assumption that the oldest results are
the least likely to still be useful. We might also experiment with other
heuristics based on actual usage later.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3880>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3880>

4 years agopanfrost: Combine get_index_buffer with bound computation
Alyssa Rosenzweig [Wed, 19 Feb 2020 15:15:13 +0000 (10:15 -0500)]
panfrost: Combine get_index_buffer with bound computation

These operations are intertwined since there are optimizations that will
want to "double dip". In particular for user index buffers we'd want to
upload simultaneous with index computation. For resources we'd like to
keep resource related code together.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3880>

4 years agoradeon/jpeg: fix the jpeg dt_pitch with YUYV format
Leo Liu [Fri, 7 Feb 2020 00:20:09 +0000 (19:20 -0500)]
radeon/jpeg: fix the jpeg dt_pitch with YUYV format

The dt_pitch should be same as NV12 format from decoder views,
and it finally got corrected with gfx9 surface's fixes from MR
https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3738

Signed-off-by: Leo Liu <leo.liu@amd.com>
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agost/va: add support YUY2
Pierre-Eric Pelloux-Prayer [Thu, 13 Feb 2020 18:24:43 +0000 (19:24 +0100)]
st/va: add support YUY2

YUY2 is a duplicate of YUYV and is used by gstreamer for 4:2:2.

Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agost/va: enable 4:2:2 chroma format
Pierre-Eric Pelloux-Prayer [Wed, 12 Feb 2020 10:22:06 +0000 (11:22 +0100)]
st/va: enable 4:2:2 chroma format

Everything is in place to support them.

Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agoradeonsi: fix surf_pitch for subsampled surface
Pierre-Eric Pelloux-Prayer [Tue, 4 Feb 2020 17:56:59 +0000 (18:56 +0100)]
radeonsi: fix surf_pitch for subsampled surface

gfx9.surf_pitch is supposed to be in blocks (or elements) but addrlib
returns a pitch in pixels.
This cause a mismatch between surface->bpe and surface.u.gfx9.surf_pitch.

For subsampled formats like uyvy (bpe is 2) this breaks in various places:
 - sdma copy
 - video rendering (see issue https://gitlab.freedesktop.org/mesa/mesa/issues/2363)
   when the vl_compositor_gfx_render method is used

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agogallium/vl: add 4:2:2 support
Pierre-Eric Pelloux-Prayer [Tue, 4 Feb 2020 17:57:08 +0000 (18:57 +0100)]
gallium/vl: add 4:2:2 support

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2363
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agogallium/video: remove pipe_video_buffer.chroma_format
Pierre-Eric Pelloux-Prayer [Wed, 5 Feb 2020 09:21:20 +0000 (10:21 +0100)]
gallium/video: remove pipe_video_buffer.chroma_format

chroma_format depends on buffer_format so use the format_to_chroma_format
helper instead of storing it next to buffer_format.
This avoids bugs where one value is changed without updating the other.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agoformat: add format_to_chroma_format
Pierre-Eric Pelloux-Prayer [Wed, 5 Feb 2020 09:20:06 +0000 (10:20 +0100)]
format: add format_to_chroma_format

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agoradeonsi: test subsampled format in testdma
Pierre-Eric Pelloux-Prayer [Thu, 6 Feb 2020 17:18:38 +0000 (18:18 +0100)]
radeonsi: test subsampled format in testdma

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3738>

4 years agoac/llvm: flush denorms for nir_op_fmed3 on GFX8 and older gens
Samuel Pitoiset [Wed, 26 Feb 2020 14:12:55 +0000 (15:12 +0100)]
ac/llvm: flush denorms for nir_op_fmed3 on GFX8 and older gens

The hardware doesn't flush denorms, exactly like fmin/fmax, so
we have to do it manually. This doesn't fix anything known.

Fixes: d6a07732c9c ("ac: use llvm.amdgcn.fmed3 intrinsic for nir_op_fmed3")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3962>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3962>

4 years agoac/llvm: fix 16-bit fmed3 on GFX8 and older gens
Samuel Pitoiset [Wed, 26 Feb 2020 14:09:40 +0000 (15:09 +0100)]
ac/llvm: fix 16-bit fmed3 on GFX8 and older gens

16-bit med3 is only supported on GFX9+.

Fixes dEQP-VK.spirv_assembly.instruction.amd_trinary_minmax.mid3.f16.*.

Fixes: d6a07732c9c ("ac: use llvm.amdgcn.fmed3 intrinsic for nir_op_fmed3")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3962>

4 years agoac/llvm: fix 64-bit fmed3
Samuel Pitoiset [Wed, 26 Feb 2020 14:04:38 +0000 (15:04 +0100)]
ac/llvm: fix 64-bit fmed3

Lower 64-bit fmed3 because LLVM doesn't expose an intrinsic.

Fixes dEQP-VK.spirv_assembly.instruction.amd_trinary_minmax.mid3.f64.*.

Fixes: d6a07732c9c ("ac: use llvm.amdgcn.fmed3 intrinsic for nir_op_fmed3")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3962>

4 years agomesa: Flush vertices before changing the OpenGL state.
Mathias Fröhlich [Wed, 26 Feb 2020 06:49:27 +0000 (07:49 +0100)]
mesa: Flush vertices before changing the OpenGL state.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
CC: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3958>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3958>

4 years agomesa: Check for OpenGL state change before flushing vertices.
Mathias Fröhlich [Wed, 26 Feb 2020 06:35:27 +0000 (07:35 +0100)]
mesa: Check for OpenGL state change before flushing vertices.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3958>

4 years agogallivm/nir: handle mod 0 better.
Dave Airlie [Wed, 26 Feb 2020 02:13:13 +0000 (12:13 +1000)]
gallivm/nir: handle mod 0 better.

I haven't seen this crash but TGSI does it so best align with
it to avoid future issues.

Fixes: 44a6b0107b3J (gallivm: add nir->llvm translation (v2))
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3956>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3956>

4 years agogallivm/nir: fix integer divide SIGFPE
Dave Airlie [Wed, 26 Feb 2020 02:10:54 +0000 (12:10 +1000)]
gallivm/nir: fix integer divide SIGFPE

Blender was crashing with a SIGFPE even though the divide by 0
logic was kicking in. I'm not sure why TGSI doesn't get into this state.

The problem was is the numerator was INT_MIN we'd replace the div by
0 with a divide by -1, which is an exception for INT_MIN as INT_MIN/-1
== INT_MAX + 1 (too large for 32-bits). Instead for integer divides
just replace the mask values with 0x7fffffff. Also fix up the
result handling so it aligns with TGSI usage. (gives 0)

Fixes: c717ac1247c3 ("gallivm/nir: wrap idiv to avoid divide by 0 (v2)")
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3956>

4 years agogallivm/tgsi: fix stream id regression
Dave Airlie [Tue, 25 Feb 2020 21:03:33 +0000 (07:03 +1000)]
gallivm/tgsi: fix stream id regression

This broke TGSI GS shaders with llvmpipe, it wasn't looking at the
right immediates and it should be cast to an integer type.

Fixes: 163d5fde0669 (gallium/swr: Enable GL_ARB_gpu_shader5: multiple streams)
Reviewed-by: Krzysztof Raszkowski <krzysztof.raszkowski@intel.com>
Acked-by: Jan Zielinski <jan.zielinski@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3949>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3949>

4 years agomesa: call FLUSH_VERTICES before updating CoordReplace
Marek Olšák [Thu, 20 Feb 2020 02:43:56 +0000 (21:43 -0500)]
mesa: call FLUSH_VERTICES before updating CoordReplace

Reviewed-by: Mathias Fröhlich <mathias.froehlich@web.de>
Cc: 20.0 <mesa-stable@lists.freedesktop.org>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3947>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3947>

4 years agomesa: remove leftovers from ARB_shadow_ambient
Marek Olšák [Thu, 20 Feb 2020 01:14:59 +0000 (20:14 -0500)]
mesa: remove leftovers from ARB_shadow_ambient

Reviewed-by: Mathias Fröhlich <mathias.froehlich@web.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3947>

4 years agocube_face_index: Use fabsf instead of fabs since we know it's floats
Albert Astals Cid [Wed, 26 Feb 2020 18:33:47 +0000 (19:33 +0100)]
cube_face_index: Use fabsf instead of fabs since we know it's floats

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3933>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3933>

4 years agocube_face_coord: Use fabsf instead of fabs since we know it's floats
Albert Astals Cid [Mon, 24 Feb 2020 23:36:44 +0000 (00:36 +0100)]
cube_face_coord: Use fabsf instead of fabs since we know it's floats

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3933>

4 years agoiris: Apply the flushes when switching pipelines.
Rafael Antognolli [Thu, 20 Feb 2020 19:02:52 +0000 (11:02 -0800)]
iris: Apply the flushes when switching pipelines.

Even though the workaround description says:

   "all the listed commands are non-pipelined and hence flush caused due
   to pipeline mode change must not cause performance issues..."

My understanding is that we still need to have the flushes. Also, the
flushes are required not only to stall the pipeline, but also to clear
caches, so I don't think they can simply be discarded.

Additionally, while doing some testing that increased the number of
surface STATE_BASE_ADDRESS emitted, I got a lot more GPU hangs. Adding
these flushes fixes those hangs.

Fixes: b8fbb39a (iris: Implement Gen12 workaround for non pipelined
                 state)

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3908>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3908>

4 years agogallium/hash_table: remove some function wrappers
Marek Olšák [Wed, 5 Feb 2020 19:52:38 +0000 (14:52 -0500)]
gallium/hash_table: remove some function wrappers

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3722>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3722>

4 years agogallium/hash_table: turn it into a wrapper around util/hash_table
Marek Olšák [Wed, 5 Feb 2020 19:47:36 +0000 (14:47 -0500)]
gallium/hash_table: turn it into a wrapper around util/hash_table

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3722>

4 years agogallium/hash_table: use the same callback signatures as util/hash_table
Marek Olšák [Wed, 5 Feb 2020 19:19:40 +0000 (14:19 -0500)]
gallium/hash_table: use the same callback signatures as util/hash_table

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3722>

4 years agogallium/hash_table: consolidate hash tables with FD keys
Marek Olšák [Wed, 5 Feb 2020 19:27:21 +0000 (14:27 -0500)]
gallium/hash_table: consolidate hash tables with FD keys

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3722>

4 years agogallium/hash_table: consolidate hash tables with pointer keys
Marek Olšák [Wed, 5 Feb 2020 19:10:48 +0000 (14:10 -0500)]
gallium/hash_table: consolidate hash tables with pointer keys

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3722>

4 years agoamd/addrlib: fix build on non-x86 platforms
Greg V [Sun, 26 Jan 2020 11:22:45 +0000 (14:22 +0300)]
amd/addrlib: fix build on non-x86 platforms

regparm(0) attribute does not work on aarch64 (and presumably powerpc64 and others).
Default to not specifying any calling convention on non-amd64/i386 platforms.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3567>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3567>

4 years agotgsi_to_nir: set num_images and num_samplers with holes correctly
Marek Olšák [Mon, 24 Feb 2020 23:16:48 +0000 (18:16 -0500)]
tgsi_to_nir: set num_images and num_samplers with holes correctly

This fixes the copy_uv shader from st/omx, because it uses image 0 and 2
and image 1 isn't declared.

Cc: 20.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3936>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3936>

4 years agonir: Drop nir_tex_instr::texture_array_size
Jason Ekstrand [Tue, 25 Feb 2020 03:16:01 +0000 (21:16 -0600)]
nir: Drop nir_tex_instr::texture_array_size

It's set by lots of things and we spend a lot of time maintaining it but
no one actually uses the value for anything useful.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3940>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3940>

4 years agofreedreno/computerator: Fix defined-but-not-used warnings from lex/yacc.
Eric Anholt [Wed, 26 Feb 2020 00:25:05 +0000 (16:25 -0800)]
freedreno/computerator: Fix defined-but-not-used warnings from lex/yacc.

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3954>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3954>

4 years agoturnip: Fix compiler warning about casting a nondispatchable handle.
Eric Anholt [Sat, 22 Feb 2020 00:38:26 +0000 (16:38 -0800)]
turnip: Fix compiler warning about casting a nondispatchable handle.

Fixes: 1c5d84fcae71 ("turnip: hook up cmdbuffer event set/wait")
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3916>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3916>

4 years agogitlab-ci: Move to 5.5 kernel plus fixes for Panfrost
Tomeu Vizoso [Thu, 20 Feb 2020 10:58:38 +0000 (11:58 +0100)]
gitlab-ci: Move to 5.5 kernel plus fixes for Panfrost

There's two fixes that help with stability when running dEQP on Kevin
Chromebooks.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3876>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3876>

4 years agopanfrost: Remove some more prints to stdout
Tomeu Vizoso [Fri, 21 Feb 2020 12:47:38 +0000 (13:47 +0100)]
panfrost: Remove some more prints to stdout

They can confuse test runners.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3876>

4 years agogitlab-ci: Run GLES3 tests in dEQP on Panfrost
Tomeu Vizoso [Wed, 18 Dec 2019 15:11:33 +0000 (16:11 +0100)]
gitlab-ci: Run GLES3 tests in dEQP on Panfrost

We are able to run only 1/5th of the tests in around the same time that
dEQP-GLES2 takes, so do that for now while more DUTs are installed.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3876>

4 years agomesa/st: toggle EXT_texture_norm16 based on format support
Tapani Pälli [Wed, 26 Feb 2020 08:30:42 +0000 (10:30 +0200)]
mesa/st: toggle EXT_texture_norm16 based on format support

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2556
Fixes: 7f467d4f738 ("mesa: GL_EXT_texture_norm16 extension plumbing")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3941>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3941>

4 years agoi965: toggle on EXT_texture_norm16
Tapani Pälli [Wed, 26 Feb 2020 08:29:49 +0000 (10:29 +0200)]
i965: toggle on EXT_texture_norm16

Fixes: 7f467d4f738 ("mesa: GL_EXT_texture_norm16 extension plumbing")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3941>

4 years agomesa: introduce boolean toggle for EXT_texture_norm16
Tapani Pälli [Wed, 26 Feb 2020 08:27:04 +0000 (10:27 +0200)]
mesa: introduce boolean toggle for EXT_texture_norm16

Fixes: 7f467d4f738 ("mesa: GL_EXT_texture_norm16 extension plumbing")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3941>

4 years agonir/lower_double_ops: add note for lowering mod
Juan A. Suarez Romero [Wed, 8 Jan 2020 10:48:29 +0000 (10:48 +0000)]
nir/lower_double_ops: add note for lowering mod

Add a note to clarify that while Vulkan allows mod(x,y) to be in [0, y]
range, OpenGL does not allow it, so the lowering ensures the result is
always in [0, y) range, as this lowering is shared by the Vulkan and
OpenGL implementation.

Reviewed-by: Elie Tournier <elie.tournier@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3315>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3315>

4 years agoradv: fix creating null devices if KHR_display is enabled
Samuel Pitoiset [Wed, 26 Feb 2020 09:21:24 +0000 (10:21 +0100)]
radv: fix creating null devices if KHR_display is enabled

Found this while replaying pipelines with Fossilize, it worked
fine with vkpipeline-db.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3959>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3959>

4 years agogitlab-ci: Add add a set of lima flakes
Andreas Baierl [Wed, 26 Feb 2020 08:17:57 +0000 (09:17 +0100)]
gitlab-ci: Add add a set of lima flakes

Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Andreas Baierl <ichgeh@imkreisrum.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3957>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3957>

4 years agoradv: make use of ac_gpu_info::max_wave64_per_simd
Samuel Pitoiset [Thu, 20 Feb 2020 08:27:07 +0000 (09:27 +0100)]
radv: make use of ac_gpu_info::max_wave64_per_simd

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3899>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3899>

4 years agoradv/gfx10: adjust the number of VGPRs used to compute waves
Samuel Pitoiset [Thu, 20 Feb 2020 08:26:03 +0000 (09:26 +0100)]
radv/gfx10: adjust the number of VGPRs used to compute waves

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3899>