Eric Anholt [Tue, 16 Mar 2021 21:09:04 +0000 (14:09 -0700)]
ci/panfrost: Disable t860/radeonsi testing while the runners are struggling.
These jobs keep getting stuck, so disable for now. Info on what is going
on is at
https://gitlab.freedesktop.org/mesa/mesa/-/issues/3437#note_841920 (the
grunt and stoney runners are maybe still operational at the moment)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9636>
Mike Blumenkrantz [Tue, 16 Mar 2021 19:57:52 +0000 (15:57 -0400)]
zink: unset last_vertex_stage_dirty after applying it
turns out that having extra flags for stuff isn't that useful if you never reset them
Fixes:
9b8c121917d ("zink: flag shaders as needing update when clip_halfz changes")
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9633>
Erik Faye-Lund [Tue, 16 Mar 2021 15:00:33 +0000 (16:00 +0100)]
lavapipe: fix primitive-restart for uint8 indices
I forgot to add an implicit primitive-restart index for uint8 while
adding support for this initially. But we also have a nice little
helper to simplify this, so let's use that instead of open-coding.
Fixes:
6526e6d11b8 ("lavapipe: implement VK_EXT_index_type_uint8")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9622>
Mike Blumenkrantz [Tue, 16 Mar 2021 15:10:09 +0000 (11:10 -0400)]
zink: add missing null check
mismerged from an earlier patch...
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9623>
Daniel Stone [Thu, 11 Mar 2021 11:48:59 +0000 (11:48 +0000)]
CI: Trigger Windows builds for llvmpipe & Vulkan changes
As Windows now builds llvmpipe and lavapipe, we should be sure to run
their builds when these change.
Acked-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9518>
Rob Clark [Tue, 16 Mar 2021 15:49:09 +0000 (08:49 -0700)]
freedreno: Fix fd_fence_finish()
fd_fence_create_unflushed() is called with the driver context, but
fd_fence_finish() is called with the threaded/frontend context, which
is what threaded_context_flush() expects to be passed.
Fixes:
c4e5beef076 ("freedreno: threaded_context async flush support")
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9627>
Anuj Phogat [Mon, 15 Mar 2021 18:58:07 +0000 (11:58 -0700)]
intel: Simplify few version checks involving G4X
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9608>
Anuj Phogat [Thu, 4 Mar 2021 23:30:41 +0000 (15:30 -0800)]
intel: Remove GEN_IS_G4X macro
GEN_GEN and GEN_VERSIONx10 macros provide a consistent way to do platform
version checks. We can avoid platform specific macros.
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9608>
Anuj Phogat [Tue, 9 Mar 2021 22:27:06 +0000 (14:27 -0800)]
intel: Simplify version checks involving haswell
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9608>
Anuj Phogat [Thu, 4 Mar 2021 23:09:42 +0000 (15:09 -0800)]
intel: Remove GEN_IS_HASWELL macro
Use GEN_VERSIONx10 == 75 check in place of GEN_IS_HASWELL macro.
GEN_GEN and GEN_VERSIONx10 macros provide a consistent way to do platform
version checks. We can avoid platform specific macros.
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9608>
Eric Anholt [Tue, 16 Mar 2021 03:28:11 +0000 (20:28 -0700)]
freedreno/a5xx: Fix the max texture buffer size.
The GLES minmax is 65536. The blob vulkan exposes 65536 on both a5xx and
a6xx, but try just doing the same as we do for a6xx.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9617>
Eric Anholt [Tue, 16 Mar 2021 00:26:30 +0000 (17:26 -0700)]
freedreno/a5xx: Fix the texel buffer alignment requirement.
Info comes from the a540 vulkan blob driver minTexelBufferOffsetAlignment.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9617>
Michel Dänzer [Thu, 4 Mar 2021 11:58:56 +0000 (12:58 +0100)]
ci: Run meson tests in strace if it's available and can be used
Keep the strace logs in job artifacts for tests which timed out.
This can be useful for figuring out why a timeout occurred.
strace cannot be used in jobs where ASAN is enabled, because ASAN's
leak checker also uses ptrace(), which isn't possible within strace.
Suggested-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9472>
Michel Dänzer [Thu, 4 Mar 2021 11:37:56 +0000 (12:37 +0100)]
ci: Add strace to the x86_build docker image
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9472>
Michel Dänzer [Tue, 9 Mar 2021 10:53:09 +0000 (11:53 +0100)]
aco/tests: Use _exit in child process
Since the child process doesn't call exec(), exit() attempted to run
atexit handlers registered by the parent process. This could result in
the child process hanging in exit() if there were still disk cache
threads alive when the parent process called fork(). (The CI runners
hit this multiple times when running tests in strace)
Fixes:
6a246f5c6d51 "aco/tests: Fix deadlock for too large test lists"
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9472>
Michel Dänzer [Mon, 8 Mar 2021 17:20:06 +0000 (18:20 +0100)]
ci: Move /usr/bin/time check from meson test wrapper to build script
In jobs where the wrapper isn't used, this leaves the
<build directory>/meson-logs/testlog.txt filename unchanged.
Also prepares for using different wrapper scripts in different jobs.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9472>
Michel Dänzer [Thu, 4 Mar 2021 17:14:03 +0000 (18:14 +0100)]
ci: Drop SIGINT handling from meson test wrapper script
Not needed, I had misread meson code.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9472>
Michel Dänzer [Mon, 8 Mar 2021 17:11:27 +0000 (18:11 +0100)]
ci: Move meson-build.sh to meson/build.sh
A little less clutter in the top .gitlab-ci directory.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9472>
Jesse Natalie [Mon, 1 Mar 2021 21:46:55 +0000 (13:46 -0800)]
d3d12: Really handle null constant buffers
Previous attempt was checking the wrong pointer...
Fixes:
eb27db86 ("d3d12: Handle null constant buffers")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4362
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9349>
Jesse Natalie [Mon, 1 Mar 2021 21:42:21 +0000 (13:42 -0800)]
d3d12: Handle is_new_style_shadow comparison filtering
Note: the tex instruction is modified in place, so the correct
number of expected result components needs to be cached before
the shadow state is modified
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9349>
Jesse Natalie [Mon, 1 Mar 2021 21:31:26 +0000 (13:31 -0800)]
d3d12: Move descriptor pools to screen, and add lock
Surfaces can be shared across contexts, and can even outlive the
original context that created them, so move the pools to the screen.
Since they no longer belong to a single context, they need a lock now.
v2: Samplers moved back to the context
v3: Fix Linux build
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3812
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9349>
Samuel Pitoiset [Tue, 16 Mar 2021 13:57:10 +0000 (14:57 +0100)]
vulkan: add missing vk_shader_module.c/h includes to Makefile
Fixes:
76078ed9fe3 ("vk/util: add unified shader module struct/functions")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9620>
Mike Blumenkrantz [Tue, 27 Oct 2020 14:40:27 +0000 (10:40 -0400)]
zink: cache bufferviews
we can cache bufferviews just like we do surfaces, so let's do that
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:44:03 +0000 (10:44 -0500)]
zink: properly handle null bufferview descriptor states
now that we have this state properly stored we can use it for more accurate
hashing and caching
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:41:21 +0000 (10:41 -0500)]
zink: improve debug asserts for samplers/images during descriptor updates
these are now more accurate for null bufferviews/imageviews
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:39:19 +0000 (10:39 -0500)]
zink: update null descriptor hashes to reflect current descriptor states
also add a bufferview hash since we use these now
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Tue, 27 Oct 2020 17:32:34 +0000 (13:32 -0400)]
zink: add helper function for getting a resource for a descriptor
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Wed, 28 Oct 2020 00:41:15 +0000 (20:41 -0400)]
zink: massively beef up batch tracking for shader images
the struct for these isn't allocated, so we need to ensure that the objects
in it are explicitly tracked on batches when they're used
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Wed, 28 Oct 2020 00:25:06 +0000 (20:25 -0400)]
zink: add extra batch tracking for sampler views
we're going to start invalidating and rebinding resources, which means
we need to be extra sure that we have our lifetimes in order here
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:28:32 +0000 (10:28 -0500)]
zink: simplify bufferview and imageview descriptor state hashing
now that we have the hash for the internal objects available at all times, we
can directly reuse this instead of computing it each time, giving us more
accurate hashes as well as saving us cpu cycles
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:27:08 +0000 (10:27 -0500)]
zink: store bufferview hash to bufferview struct
this lets us do a lot of things more precisely for bufferviews
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:24:53 +0000 (10:24 -0500)]
zink: explicitly use zink_surface objects for sampler/image view objects
we can reuse this codepath instead of NIHing it
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Thu, 5 Nov 2020 15:21:05 +0000 (10:21 -0500)]
zink: use an explicit zink_buffer_view struct
this gives us an object to work with similar to zink_surface
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Fri, 30 Oct 2020 12:18:09 +0000 (08:18 -0400)]
zink: add util function for checking whether shader descriptor is buffer from program
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Fri, 30 Oct 2020 12:17:07 +0000 (08:17 -0400)]
zink: add util function for checking whether a shader descriptor is a buffer
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Tue, 13 Oct 2020 14:12:40 +0000 (10:12 -0400)]
zink: add some asserts for pipeline barriers to check renderpass state
it's illegal to emit barriers during a renderpass
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Mike Blumenkrantz [Tue, 16 Mar 2021 04:34:31 +0000 (00:34 -0400)]
zink: handle cached descriptor set punting
if an active set gets invalidated, it needs to be punted from the active
hash table so it doesn't get reused
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9543>
Daniel Stone [Mon, 15 Mar 2021 20:41:30 +0000 (20:41 +0000)]
CI: Try really hard to get updated Windows TLS certs
Windows doesn't actually distribute a full TLS CA certificate store, but
pulls them in over time with Windows Update. Try to prime it by manually
pulling the certificates and installing them.
This bumps the Windows tag to force a rebuild.
Acked-by: Michel Dänzer <mdaenzer@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9618>
Mike Blumenkrantz [Tue, 27 Oct 2020 16:07:36 +0000 (12:07 -0400)]
zink: track resource_object usage instead of resource usage
the resource object is now a state tracker for the inner resource_object,
so it can safely be destroyed even if the backing resource is still in use
this also requires updating various things to keep descriptor states/caches working
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9544>
Mike Blumenkrantz [Tue, 27 Oct 2020 15:59:33 +0000 (11:59 -0400)]
zink: split out backing resource object create/destroy
moving towards these being able to exist semi-independently...
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9544>
Mike Blumenkrantz [Tue, 27 Oct 2020 15:38:16 +0000 (11:38 -0400)]
zink: move resource internals to a separate struct
this starts the path towards being able to replace the backing objects for
a resource
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9544>
Mike Blumenkrantz [Thu, 8 Oct 2020 18:16:40 +0000 (14:16 -0400)]
zink: break out batch id finding for resource usage into util function
this is generally helpful for figuring out the last batch a resource was
used for
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9544>
Mike Blumenkrantz [Tue, 13 Oct 2020 17:45:50 +0000 (13:45 -0400)]
zink: avoid looping for non-ubo descriptor updates based on set usage
if we know that the descriptor set is cached and already in use by a given
batch, then we also know that all the resources in the set are tracked, which
means that we can skip over some looping during descriptor updates which would
otherwise be used to add tracking for sampler/image views/states
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9567>
Mike Blumenkrantz [Tue, 13 Oct 2020 12:12:59 +0000 (08:12 -0400)]
zink: add batch usage flags for sampler views/states and desc sets
this massively cuts down cpu time for hashing and lookups, resulting in
a noticeable performance increase
the sampler_state batch usage for now is just being set for future use when
it will be leveraged to permit immediate deletion
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9567>
Bas Nieuwenhuizen [Mon, 15 Mar 2021 21:15:39 +0000 (22:15 +0100)]
radv: Use correct DCC compressed block size for sampling.
Don't need to change the pre-GFX9 samplers because this option doesn't
vary there.
Fixes:
f848f2adfae ("radv: Use ac_surface DCC settings for shareable images.")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4455
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4425
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9611>
Jason Ekstrand [Mon, 15 Mar 2021 23:39:19 +0000 (18:39 -0500)]
genxml: Make 1-bit L3$ config register fields bool on Gen7
Otherwise, they look like booleans but, if you put a value other than
0/1 in them, the GenXML generator code will explode.
Fixes:
b6875b0094c "anv: Drop has_slm in emit_l3_config for gen11+"
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9614>
Danylo Piliaiev [Wed, 10 Mar 2021 12:29:31 +0000 (14:29 +0200)]
turnip: implement intrinsic_vulkan_resource_reindex
Descriptor arrays are continuous, so it's just an addition of offset.
Fixes test:
dEQP-VK.spirv_assembly.instruction.compute.variable_pointers.dynamic_offset.select_descriptor_array
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9495>
Eric Anholt [Thu, 11 Mar 2021 19:47:54 +0000 (11:47 -0800)]
ci/freedreno: Add three more a5xx flakes from the last day.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9575>
Mauro Rossi [Fri, 12 Mar 2021 23:08:37 +0000 (00:08 +0100)]
android: i965: Rename files with "intel_" prefix to "brw_"
The necessary changes for Android build were missing in Makefile.sources
Fixes:
9d95e1bd79a8 (i965: Rename files with "intel_" prefix to "brw_")
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9576>
Mike Blumenkrantz [Wed, 10 Mar 2021 22:50:15 +0000 (17:50 -0500)]
anv: use common interfaces for shader modules
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Mike Blumenkrantz [Wed, 10 Mar 2021 22:50:15 +0000 (17:50 -0500)]
tu: use common interfaces for shader modules
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Mike Blumenkrantz [Wed, 10 Mar 2021 22:50:00 +0000 (17:50 -0500)]
v3dv: use common interfaces for shader modules
squashed changes from Alejandro Piñeiro <apinheiro@igalia.com>:
Add call to vk_object_base_init on internal shader_module: we have
some cases where internally we have some shader modules that we don't
create through CreateShaderModule, so in this case we need to manually
call base_init. Not sure why this wasn't needed before.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Mike Blumenkrantz [Wed, 10 Mar 2021 22:49:46 +0000 (17:49 -0500)]
radv: use common interfaces for shader modules
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Mike Blumenkrantz [Wed, 10 Mar 2021 22:49:28 +0000 (17:49 -0500)]
lavapipe: use common interfaces for shader modules
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Mike Blumenkrantz [Fri, 12 Mar 2021 12:30:39 +0000 (07:30 -0500)]
vk/util: add a util macro for initializing stack vk_shader_module structs
radv does a lot of this, so having a central dispatch point will be useful in
case changes are made to this struct in the future
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Mike Blumenkrantz [Wed, 10 Mar 2021 22:48:30 +0000 (17:48 -0500)]
vk/util: add unified shader module struct/functions
there's some extra logging stuff dumped into here to match functionality,
eventually that should also be consolidated into vk_util.c
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9508>
Dave Airlie [Sun, 14 Mar 2021 22:10:36 +0000 (08:10 +1000)]
lavapipe: add host query reset
This just resets the queries on the host size.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9594>
Dave Airlie [Mon, 15 Mar 2021 06:56:48 +0000 (16:56 +1000)]
lavapipe: fix writing availability for queries.
If the availability has to be written it needs to be written
to the correct place.
The host query reset tests fall over this.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9594>
Dave Airlie [Mon, 15 Mar 2021 01:49:33 +0000 (11:49 +1000)]
lavapipe: add missing break
No effect as-is but annoying later.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9594>
Sagar Ghuge [Fri, 12 Mar 2021 03:03:17 +0000 (19:03 -0800)]
anv: Set correct binding table entry count
We can use surface_count as it is to set binding table entry count since
it's already in units.
On Felix's Tigerlake with the GPU at fixed frequency, this patch
improves performance of several games:
- Shadow of the Tomb Raider: +1.5%
- Dota2vk: +1%
- Dark Souls: +1%
v2: (Ken)
- Remove get_binding_table_entry_count() and use surface_count directly.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9548>
Jordan Justen [Sun, 14 Mar 2021 19:54:26 +0000 (12:54 -0700)]
anv: Drop has_slm in emit_l3_config for gen11+
For some gen12+ platforms, L3 config (cfg) can be NULL leading to a
seg-fault in emit_l3_config. But, we don't use has_slm for gen11+, so
we can just avoid declaring the variable.
Reworks:
* Drop has_slm variable for all gens (suggested-by Jason)
Ref:
633dec7163e ("anv: Set L3 full way allocation at context init if L3 cfg is NULL")
Ref: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9534
Fixes:
581e68bc99b ("anv: move L3 config emission to genX_state.c")
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9589>
Anuj Phogat [Thu, 11 Mar 2021 20:30:41 +0000 (12:30 -0800)]
anv: Remove redundant #if checks
Patch removes redundant #if checks and unused code inside another #if
block.
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9604>
Caio Marcelo de Oliveira Filho [Fri, 12 Mar 2021 23:13:30 +0000 (15:13 -0800)]
anv: Lower ViewIndex to zero when multiview is disabled
Vulkan spec says
If multiview is enabled in the render pass, this value will be one
of the bits set in the view mask of the subpass the pipeline is
compiled against. If multiview is not enabled in the render pass,
this value will be zero.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4446
Fixes:
0db70703300 ("anv/pipeline: Add shader lowering for multiview")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9574>
Mike Blumenkrantz [Thu, 5 Nov 2020 01:50:08 +0000 (20:50 -0500)]
zink: stop leaking programs
gfx and compute programs both end up with an extra ref from creation,
so they both need to lose a ref during creation after they get their
shader refs
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9470>
Mike Blumenkrantz [Sun, 13 Dec 2020 22:47:26 +0000 (17:47 -0500)]
zink: unset ctx->program pointers when an unref destroys the object
if we destroy a program object which is currently the "active" program then
we need to unset the pointer to avoid invalid access
also unset injected tcs pointers where appropriate
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9470>
Mike Blumenkrantz [Mon, 14 Dec 2020 18:41:23 +0000 (13:41 -0500)]
zink: return true from program ref functions upon free
this lets us unset the context pointer accurately
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9470>
Mike Blumenkrantz [Tue, 2 Mar 2021 14:30:02 +0000 (09:30 -0500)]
zink: use internal api for first-frame fence
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9290>
Mike Blumenkrantz [Thu, 8 Oct 2020 18:56:37 +0000 (14:56 -0400)]
zink: force explicit fence only on first frame flush
we have implicit sync hooked up for drivers now so we don't need to worry
about drawing over our frontbuffer unexpectedly
still a weird issue remaining where we miss the first frame without an explicit
fence, but I expect that will get resolved eventually by wsi
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9290>
Mike Blumenkrantz [Wed, 23 Dec 2020 16:17:21 +0000 (11:17 -0500)]
zink: force mesa flush implicit fencing on ANV
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9290>
Mike Blumenkrantz [Thu, 25 Feb 2021 15:22:09 +0000 (10:22 -0500)]
zink: determine whether the vulkan driver requires mesa flush wsi
this is used for ANV's implicit fencing (until we get wsi)
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9290>
Mike Blumenkrantz [Thu, 29 Oct 2020 21:19:12 +0000 (17:19 -0400)]
zink: use a safer iteration for fb surfaces during rp init
doing a pointer iteration looks cool, but if the array is full, then it
goes out of bounds and we crash
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9542>
Mike Blumenkrantz [Tue, 9 Feb 2021 20:40:38 +0000 (15:40 -0500)]
zink: implement a surface cache
this is a global cache for all surface objects, enabling some memory
optimizations as well as improved reuse of cached descriptors
loosely based on patches from Antonio Caggiano
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9541>
Mike Blumenkrantz [Tue, 9 Feb 2021 20:49:21 +0000 (15:49 -0500)]
zink: ralloc screen objects
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9541>
Michel Zou [Thu, 11 Mar 2021 20:32:40 +0000 (21:32 +0100)]
swr: fix array-bounds warning
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9502>
Michel Zou [Wed, 10 Mar 2021 21:12:43 +0000 (22:12 +0100)]
swr: fix deprecated llvm 11 declaration warning
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9502>
Michel Zou [Wed, 10 Mar 2021 20:42:25 +0000 (21:42 +0100)]
swr: fix unused SplitString warning
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9502>
Michel Zou [Wed, 10 Mar 2021 20:34:14 +0000 (21:34 +0100)]
swr: Fix SWR_CONTEXT pre-declaration
Silents a warning on msvc
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9502>
Michel Zou [Wed, 10 Mar 2021 20:25:05 +0000 (21:25 +0100)]
swr: fix win32 intrinsics
The doc doesnt mention Index is altered when mask is null.
This is consistent with both llvm & migw implementations.
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9502>
Michel Zou [Wed, 10 Mar 2021 20:19:20 +0000 (21:19 +0100)]
swr: extern declaration for win32 intrinsics
This fixes compilation on msvc because llvm redefines these too
Closes #4417
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9502>
Samuel Pitoiset [Thu, 11 Mar 2021 16:14:42 +0000 (17:14 +0100)]
radv: rework radv_use_dcc_for_image() a bit
To make it clear that only GFX8-9 have missing DCC features.
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/9526>
Samuel Pitoiset [Thu, 11 Mar 2021 15:59:11 +0000 (16:59 +0100)]
radv: remove useless DCC disable check for 3D images on GFX10+
addrlib uses the S swizzle mode which disables DCC completely.
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/9526>
Samuel Pitoiset [Fri, 12 Mar 2021 15:06:19 +0000 (16:06 +0100)]
radv: add missing SQTT events for copy_commands2/create_renderpass2
A bunch of entrypoints were missing.
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/9560>
Marcin Ślusarz [Mon, 15 Mar 2021 09:35:06 +0000 (10:35 +0100)]
anv: fix memory allocation error handling
Reported by Coverity.
Fixes:
0a7224f3ff7 ("anv: group as many command buffers into a single execbuf")
Signed-off-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9596>
Lionel Landwerlin [Mon, 15 Mar 2021 09:09:00 +0000 (11:09 +0200)]
intel: install intel_device_info
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Suggested-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9595>
Pierre-Eric Pelloux-Prayer [Fri, 12 Mar 2021 08:19:57 +0000 (09:19 +0100)]
frontends/va: fix protected slice data buffer read size
Read vlVaBuffer->size bytes instead of an hardcoded value.
Fixes:
deb7dc82f62 ("frontends/va: handle protected slice data buffer")
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Boyuan Zhang <Boyuan.Zhang@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9553>
Pierre-Eric Pelloux-Prayer [Thu, 11 Mar 2021 20:54:36 +0000 (21:54 +0100)]
radeonsi/sqtt: fix user event max size
Larger strings can corrupt rgp files.
Fixes:
5dc823304b1 ("radeonsi/sqtt: forward string markers to sqtt")
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Boyuan Zhang <Boyuan.Zhang@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9553>
Iago Toral Quiroga [Fri, 12 Mar 2021 08:54:49 +0000 (09:54 +0100)]
broadcom/compiler: be more flexible scheduling TMU writes
V3D 4.x allows more flexibility, so take advantage of that. Generally,
we can reorder any writes in the same sequence, so long as they are
not the sequence terminator (which must always be last, since it is
the one triggering the operation), and TMUD writes, since these must
be ordered with respect to each other.
total instructions in shared programs:
13735183 ->
13731927 (-0.02%)
instructions in affected programs: 903057 -> 899801 (-0.36%)
helped: 2358
HURT: 746
Instructions are helped.
total max-temps in shared programs: 2322020 -> 2322009 (<.01%)
max-temps in affected programs: 619 -> 608 (-1.78%)
helped: 19
HURT: 11
Inconclusive result (value mean confidence interval includes 0).
total sfu-stalls in shared programs: 31494 -> 31489 (-0.02%)
sfu-stalls in affected programs: 182 -> 177 (-2.75%)
helped: 40
HURT: 40
Inconclusive result (value mean confidence interval includes 0).
total inst-and-stalls in shared programs:
13766677 ->
13763416 (-0.02%)
inst-and-stalls in affected programs: 901343 -> 898082 (-0.36%)
helped: 2349
HURT: 746
Inst-and-stalls are helped.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9555>
Iago Toral Quiroga [Fri, 12 Mar 2021 08:29:07 +0000 (09:29 +0100)]
broadcom/compiler: flag wrtmuc with a read dependency on last_tmu_config
Instead of using a write depdency. We use last_tmu_config to ensure ordering
of instructions participating in different TMU sequences. To this end,
all sequence terminators flag a write dependency on last_tmu_config, but
wrtmuc is not a sequence terminator, so we can be more flexible by flagging
it as a read depedency. This would prevent it to be moved into a previous
sequence (since it cannot be moved past the previous sequence terminator due
to the read depedency), but it allows it to be reordered with instructions in
the same sequence, which allows us to pair it up more effectively. Particularly,
it allows to pair up a wrtmuc with the sequence terminator of the same sequence,
turning code like this:
nop ; mov tmut, r0 ; thrsw; wrtmuc (tex[0].p0 | 0x3)
nop ; nop ; wrtmuc (tex[0].p1 | 0x0)
nop ; mov tmus, r1
Into this:
nop ; mov tmut, r0 ; thrsw; wrtmuc (tex[0].p0 | 0x3)
nop ; mov tmus, r1 ; wrtmuc (tex[0].p1 | 0x0)
total instructions in shared programs:
13755738 ->
13735183 (-0.15%)
instructions in affected programs: 2510921 -> 2490366 (-0.82%)
helped: 10963
HURT: 485
Instructions are helped.
total max-temps in shared programs: 2322828 -> 2322020 (-0.03%)
max-temps in affected programs: 11303 -> 10495 (-7.15%)
helped: 608
HURT: 19
Max-temps are helped.
total sfu-stalls in shared programs: 31545 -> 31494 (-0.16%)
sfu-stalls in affected programs: 235 -> 184 (-21.70%)
helped: 62
HURT: 11
Sfu-stalls are helped.
total inst-and-stalls in shared programs:
13787283 ->
13766677 (-0.15%)
inst-and-stalls in affected programs: 2525187 -> 2504581 (-0.82%)
helped: 10989
HURT: 477
Inst-and-stalls are helped.
v2: add a comment explaining the read depdency (Piñeiro).
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9555>
Timothy Arceri [Sat, 6 Mar 2021 03:52:46 +0000 (14:52 +1100)]
util/fossilize_db: remove compression from foz db helper
We now handle compression in the shared cache item creation code.
Compressing the cache item header with the already compressed blob
doesn't help much so lets just remove it.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9593>
Timothy Arceri [Thu, 4 Mar 2021 05:39:47 +0000 (16:39 +1100)]
util/disk_cache: make use of the new compression helpers
This makes compression use more consistent between the zstd and
zlib libraries. It also reduces the amount of code required for
zlib use.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9593>
Timothy Arceri [Thu, 4 Mar 2021 05:34:34 +0000 (16:34 +1100)]
util: create some standalone compression helpers
This will be used by the following patch. It allows us to detangle
compression from the disk cache code, and abstract the underlying
compression libraries we use.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9593>
Mike Blumenkrantz [Wed, 7 Oct 2020 18:32:16 +0000 (14:32 -0400)]
zink: use _mesa_set_search_and_add() for set management
this avoids extra hash lookups
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9565>
Mike Blumenkrantz [Tue, 6 Oct 2020 20:59:02 +0000 (16:59 -0400)]
zink: optimize resource usage tracking
we already have the batch usage info here for the resource, so if we know
the resource is already used on the batch then we don't need to also perform
a hash lookup to double check that it's really there
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9565>
Dave Airlie [Sun, 14 Mar 2021 22:01:50 +0000 (08:01 +1000)]
lavapipe: enable EXT_scalar_block_layout
The frontend should handle this.
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9590>
Dave Airlie [Sun, 14 Mar 2021 21:53:24 +0000 (07:53 +1000)]
lavapipe: enable KHR_uniform_buffer_standard_layout
I think the frontend should handle all of this fine.
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9590>
Vinson Lee [Thu, 25 Feb 2021 06:11:12 +0000 (22:11 -0800)]
iris: Fix typos.
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9382>
Vinson Lee [Mon, 1 Mar 2021 02:44:27 +0000 (18:44 -0800)]
nv50/ir: Add constructor for NV50LegalizePostRA.
Fix defect reported by Coverity Scan.
Uninitialized pointer field (UNINIT_CTOR)
member_not_init_in_gen_ctor: The compiler-generated constructor
for this class does not initialize r63.
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9326>
Mike Blumenkrantz [Sun, 14 Mar 2021 23:22:39 +0000 (19:22 -0400)]
zink: ci updates
disabling qbo piglit test for now due to lots of brokenness in zink which
makes them flaky, will be reenabled soon
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9566>
Mike Blumenkrantz [Thu, 8 Oct 2020 19:19:50 +0000 (15:19 -0400)]
zink: remove extraneous flush in transfer_map_region_flush
this was only needed to cover up some other bugs:
* missing barriers for buffer sampler/image descriptors
* weirdness with first frame handling
there's better ways of handling both cases, and now they're handled better
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9566>
Mike Blumenkrantz [Wed, 7 Oct 2020 19:55:16 +0000 (15:55 -0400)]
zink: move buffer<->image copying to pipe_context::resource_copy_region hook
that's a todo item off the list
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9566>