zink: fix zink_destroy_screen for early screen creation fails

Fixes crashes on at least my system with multiple GPUs.

Fixes: 0c2045553fe ("zink: use screen destructor for creation fails")
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25844>

etnaviv: Skip 'empty' cmd streams

Make use of etna_cmd_stream_mark_end_of_context_init(..) to mark the
end of context init states. If the cmd stream is untouched until
etna_cmd_stream_flush(..) gets called the drm layer will downgrade this
specific cmd stream to a noop.

This safes us from calling into the kernel with an 'empty' cmd stream.

Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25849>

etnaviv: drm: Be able to mark end of context init

Add etna_cmd_stream_mark_end_of_context_init(..) which stores the current
offset of the cmd stream as offset_end_of_context_init. This information
will be used to downgrade a cmd stream flush to a noop.

Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25849>

pvr: Only setup the bgobj to load if we have a load_op

We only need to setup the background object if we have a load_op,
to perform the loads/clears.

This fixes segfaults for cases where we don't have a load_op.

Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25703>

ci/b2c: change artifacts path to match baremetal and LAVA

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25815>

ci/ci_run_n_monitor: keep monitoring if a job is still running

When running multiple targets, the script was returning if any of the
targets failed.
But sometimes some of the targets are still running and we want to
monitor it until it finishes, even if some of the targets failed.

So check if any targeted jobs are in pending or running before deciding
to finish.

Signed-off-by: Helen Koike <helen.koike@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25826>

ci/freedreno: disable Adreno 660 Vulkan pre-merge

No need to move it to manual/nightly - we already run full runs there.

Revert when Adreno 660 gets stable again!

Reported-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25839>

radeonsi: fixes compilaton error when perfetto is disabled

Signed-off-by: Saroj Kumar <saroj.kumar@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25827>

ci/freedreno: fix copy paste causing a618_gl being run only in manual pipeline

Fixes: 9d442b459a43 ("ci/freedreno: handle disabling farm properly for each FD/Collabora farm")
Closes: #9992

Acked-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25816>

svga: Unify gmr and mob surface pool managers

The svga driver contains several buffer managers with gmr* and mob* prefixes,
but they are of similar functionality. The underlying vmwgfx kernel driver
allocates memory of one type - mobs if supported by the virtual hardware,
and gmrs otherwise - and returns an opaque pointer to userspace, so several
managers are reduntant.

This patch reduces the number of buffer managers by unifying similar managers
and performing refactors where necessary.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25832>

svga: Make surfaces shareable at creation.

There is no need to have a separate shareable pool of surfaces, we can make
all surfaces shareable.
This change makes all surfaces shareable when creating them.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25832>

ci/venus: mark more flaky tests after recent cts uprev

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25799>

ci/marge_queue: add missing python-dateutils to requirements.txt

marge_queue.py requires python-dateutils to run, so add it.

Signed-off-by: Helen Koike <helen.koike@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25828>

mesa: add GL_APPLE_sync support

This extension is a backport of GL_ARB_sync to GLES 1.1 and 2.0
and reuses token IDs and entry point prototypes from that extension,
so adding support is pretty trivial.

Signed-off-by: Lucas Stach <dev@lynxeye.de>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25789>

zink: properly alias shared memory

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25763>

zink: wrap shared memory blocks in a struct

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25763>

frontends/va: Fix locking in vlVaBeginPicture

The assert in vlVaSetSurfaceContext would very rarely fail because
the mutex was already unlocked when calling this function from
vlVaBeginPicture.

Keep the lock until returning from vlVaBeginPicture, as that's what
other functions are already doing.

Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25823>

pvr: Implement VK_EXT_host_query_reset

Signed-off-by: Vlad Schiller <vlad-radu.schiller@imgtec.com>
Reviewed-by: Matt Coster <matt.coster@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25783>

pvr: Implement VK_EXT_texel_buffer_alignment

This commit will implement the VK_EXT_texel_buffer_alignment
extension and add the texture_baseaddress_byte_aligned feature.

Signed-off-by: Vlad Schiller <vlad-radu.schiller@imgtec.com>
Reviewed-by: Matt Coster <matt.coster@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25781>

pvr: Enable VK_KHR_bind_memory2 extension

The two functions 'vkBindBufferMemory2' and 'vkBindImageMemory2' are aleady
implemented, and it seems that the flag does not need to be enabled for the
tests to pass.

Signed-off-by: Vlad Schiller <vlad-radu.schiller@imgtec.com>
Reviewed-by: Matt Coster <matt.coster@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25779>

pvr: Implement VK_KHR_external_semaphore

In order for the tests to pass, this commit also enables
the VK_KHR_external_semaphore_fd extension.

Signed-off-by: Vlad Schiller <vlad-radu.schiller@imgtec.com>
Reviewed-by: Matt Coster <matt.coster@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25784>

pvr: Implement VK_KHR_external_fence

In order for the tests to pass, this commit also enables
the VK_KHR_external_fence_fd extension.

Signed-off-by: Vlad Schiller <vlad-radu.schiller@imgtec.com>
Reviewed-by: Matt Coster <matt.coster@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25784>

anv: document the draw indirect optimization ring mode

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: merge gfx9/11 indirect draw generation shaders

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: add ring buffer mode to generated draw optimization

When the number of draw calls is very large, instead of allocating
large amounts of batch buffer space for the draws, use a ring buffer
and process the draw calls by batches.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8645
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: index indirect data buffer with absolute offset

This will help for a follow up change where we will respawn the shader
multiple times in a loop and the base offset will be edited by the
shader itself.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: factor out host/gpu internal shaders interfaces

This will prevent host/gpu structure definitions to go out of sync.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

util/glsl2spirv: add ability to pass defines

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: move generation batch fields to a sub-struct

Just tyding things a bit since we're about to add more.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: avoid MI commands to copy draw indirect count

We can just make the address of the count available to the generation
shader.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: identify internal shader in NIR

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: extract out draw call generation

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: fix generated draws gl_DrawID with more than 8192 indirect draws

This applies only to Gfx9.

We're writting out of bounds to a wrong location.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 1d9cf8f381 ("anv: add gfx9 generated draw support")
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

anv: move generation shader return instruction to last draw lane

If we dispatch exactly a multiple of 8192 items, there is additional
lane left to generate the jump instruction.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: c950fe97a0 ("anv: implement generated (indexed) indirect draws")
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25361>

Revert "etnaviv: use correct blit box sizes when copying resource"

Revert commit b4c24d5978a5, as it causes a regression in
dEQP-GLES2.functional.texture.specification.basic_copytexsubimage2d.2d_rgb
and some other testcases. Needs further investigation.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/10007
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25825>

dri: Drop a duplicate mesa vs pipe format table.

mesa_format is a typedef of enum pipe_format, and the nice mesa format
names are a subset of the available pipe formats.  No need to keep a
duplicate table here.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25650>

ci: drop skip for glx-swap-copy.

This will always be a skip now that we dropped GLX_OML_swap_method.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25650>

glx: Delete support for GLX_OML_swap_method.

This was an ill-advised extension.  While we advertised SWAP_COPY support,
we might implement it with a back-copy from the front buffer.  And we
never advertised EXCHANGE because we couldn't guarantee it.  So, if you
actually used this extension to try to reduce app redraws of the back
buffer, you might actually increase the bandwidth you used.  Whoops.

Instead, GLX_EXT_buffer_age and the similar EGL extension give you
feedback on what's left in your back buffer, letting you do minimum
redraws.

This reduces our GLX visual+fbconfig count from 1410 to 940 on an llvmpipe
X server.  Reducing visual counts will improve test runtime for
visual-iterating tests like piglit's glx-visuals-*.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25650>

ci_run_n_monitor: Poll mesa/mesa and user/mesa for pipelines at the same time.

Now you don't fail if you're trying to test a mesa/mesa MR pipeline and
gitlab takes more than 10s to create it.  And you don't have to wait 10
seconds to get things started (aka see if your regex was right) if you're
testing a user/mesa fork pipeline.

Fixes: 941d92408ee5 ("bin/ci_run_n_monitor: automatically pick MR pipelines when they exist")
Closes: #9894
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25810>

radeonsi/vcn: set jpeg reg version for gfx 1150

select appropriate jpeg register version for gfx1150

Signed-off-by: Saleemkhan Jamadar <saleemkhan.jamadar@amd.com>
Reviewed-by: Leo Liu <Leo.Liu@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25278>

amd/common: update DCC for gfx11.5

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25278>

amd, radeonsi: Add code to enable gfx11.5

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25278>

ac/nir: extract must_wait_attr_ring helper

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25278>

amd/common: add registers for gfx11.5

Built from amd-staging-drm-next c5a7d38c2c7fc + Alex D
"drm/amdgpu: update to the latest GC 11.5 header" patch.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25278>

amd/common: update addrlib for gfx11.5

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25278>

d3d12: Allocate d3d12_video_buffer with higher alignment for compatibility

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: Add support for PIPE_VIDEO_CAP_ENC_MAX_TILE_ROWS/COLS

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

frontend/va: Add support for VAConfigAttribEncMaxTileRows/Cols

Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: AV1 Encode - Fix VAConfigAttribEncMaxRefFrames reporting

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: AV1 encode - add fallback for app passing unsupported pic_params.InterpolationFilter

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: AV1 encode - Add lower resolution fallback check for uniform tile support

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: Do not destroy codec when destroying video buffer

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: Video Decode - Wait for GPU completion before destroying decoder in-flight objects

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: Fixes AV1 tx_mode_support reporting and unsupported tx_mode overriding

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

d3d12: Upgrade to D3D12 Agility SDK 1.611 Video interface

Reviewed-by: Giancarlo Devich <gdevich@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

frontends/va: Add profile param when querying PIPE_VIDEO_CAP_ENC_QUALITY_LEVEL

Pass the specific profile as drivers may only support
quality levels for specific codecs or profiles.

Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25819>

venus: set lvp queries as saturate on overflow

Test case 'dEQP-VK.pipeline.monolithic.timestamp.misc_tests.consistent_results'..
  Fail (Results are inconsistent: B32=0xffffffff B64=0x10103ba2cdd4e G32=0xba2cdd4e G64=0x10103ba2cdd4e)

Fixes: e6cffa1f0e4e ("venus: use feedback for vkGetQueryPoolResults")

Signed-off-by: Juston Li <justonli@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25820>

venus: Fix -Wmaybe-uninitialized

  src/virtio/vulkan/vn_ring.c: In function ‘vn_ring_submit’:
  src/virtio/vulkan/vn_ring.c:259:4: warning: ‘cur_seqno’ may be used uninitialized [-Wmaybe-uninitialized]
    259 |    vn_ring_retire_submits(ring, cur_seqno);
        |    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Signed-off-by: Lina Versace <linyaa@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25797>

svga: ignore sampler view resource if not used by shaders

Currently bound sampler view resources are relevant only if the
currently bound shaders also access them. So when checking for shader
resource collision, we only need to check those shader resources that
are actively used by the shaders.

This fixes a regression with manhattan on SVGA device when only the
active state changes are sent to the driver and a no longer relevant
shader resource is included in the shader resource collision check.
This causes a backing resource to be unnecessarily created and the content
never propagated to the original resource.

Fixes: aaa4b0e6186 ("st/mesa: move check_program_state code into _mesa_update_state")
Reviewed-by: Neha Bhenden <bhenden@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25801>

egl/wayland: only add more registry listeners for hardware devices

these extensions shouldn't be exposed for pure software drivers

fixes #9948

Fixes: 1b4e877def1 ("egl/wayland: use more registry listeners to better handle device init")

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25817>

ci: don't set GALLIUM_DRIVER for zink

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25640>

glx: add automatic zink fallback loading between hw and sw drivers

if loading the default hardware driver fails, implicitly loading zink
should now be preferable to hitting the software fallback now that zink
has all the same capabilities

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25640>

egl: add automatic zink fallback loading between hw and sw drivers

if loading the default hardware driver fails, implicitly loading zink
should now be preferable to hitting the software fallback now that zink
has all the same capabilities

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25640>

egl/glx: don't load non-sw zink without dri3 support

this is going to be broken, so don't bother trying

also add LIBGL_KOPPER_DRI2 so people can continue to footgun if they
really really want to

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25640>

egl: init dri3 version info during screen creation

no functional changes

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25640>

egl: unify dri2_egl_display creation

this is the same for all platforms

no functional changes

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25640>

radeonsi: Add tracepoints in radeonsi driver

Add initialization code for u_trace and tracepoints in the
driver code.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23664>

radeonsi: Add u_trace init code in radeonsi

Add u_trace initialization and callbacks in new files si_utrace.h/c

Change si_context in si_pipe.h to add si_ds_queue and u_trace to
collect traces.

Add si_record_ts function in si_gfx_cs.c to add timestamp in command
stream.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23664>

radeonsi: Add perfetto support in radeonsi

Add perfetto code in new files si_perfetto.h/cc which add tracepoint
begin and end event and calls to the generated code from python
si_tracepoints.py

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23664>

radv: Wait for bottom of pipe in ACE gang wait postamble.

This makes sure that the semaphore is only written when all work
on the compute queue is finished.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Add queue family argument to some functions.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Use new WRITE_DATA helper in more places.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Refactor WRITE_DATA helper function.

Create a version of this function that takes a CS and queue family.
move it to radv_cs.h so it can be called from multiple other files.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Move radv_cp_wait_mem to radv_cs.h and add queue family argument.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Only call si_cp_dma_wait_for_idle on GFX and ACE queues.

This command is only supported on GFX and ACE.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Use const on vi_alpha_is_on_msb arguments.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Use const device argument in radv_sdma_copy_buffer.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

radv: Rename SDMA file to radv_sdma.c

It already contains more than just the buffer image copies.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>

nvk: Implement VK_EXT_dynamic_rendering_unused_attachments

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9632
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25813>

nvk: Enable descriptorIndexing

This is the Vulkan 1.2 equivalent of VK_EXT_descriptor_indexing
extension enable.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25812>

nvk: Advertise shaderOutputLayer and shaderOutputViewportIndex

We already advertise the extension but it's moved into feature bits in
Vulkan 1.2 and we didn't have those turned on.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25812>

nvk: Re-arrange Vulkan 1.2 features to match the header

Also, conditional render ended up in the Vulkan 1.2 section which is
wrong.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25812>

features: Mark VK_EXT_attachment_feedback_loop_layout done for NVK

Fixes: c93b30daa373 ("nvk: Advertise VK_EXT_attachment_feedback_loop_layout")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25812>

nvk: Advertise VK_EXT_attachment_feedback_loop_layout

We don't need to do anything with the layout.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25808>

anv: uninitialize queues before utrace

We need to shut down the runtime queue threads before tearing down
anything else.

Gets rid of helgrind errors like this :

==212772== Possible data race during write of size 4 at 0xADCBFB0 by thread #1
==212772== Locks held: 1, at address 0x6B8F260
==212772==    at 0x8AC3EFF: simple_mtx_destroy (simple_mtx.h:97)
==212772==    by 0x8ACB24D: intel_ds_device_fini (intel_driver_ds.cc:603)
==212772==    by 0x6CBD4D4: anv_device_utrace_finish (anv_utrace.c:471)
==212772==    by 0x6C71577: anv_DestroyDevice (anv_device.c:3679)
==212772==    by 0x6B2F1E2: loader_layer_destroy_device (loader.c:4358)
==212772==    by 0x6B3F10B: vkDestroyDevice (trampoline.c:983)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: cc5843a573 ("anv: implement u_trace support")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/10010
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25805>

ci/freedreno: Generalize the implicit_unmap timeouts.

dedicated_alloc.image caused a job flake today.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25804>

anv: don't uninitialize bvh_bo_pool is not initialized

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 3e8d2617e1 ("anv: use buffer pools for BVH build buffers")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/10009
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25802>

ci/zink: Cut zink-lvp coverage in half.

It's taking 23 minutes of deqp-runner time.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci/freedreno: Move skqp testing to a618.

The limozeen runners are quite under-committed, while we had 10 jobs
running for our 9 cheza runners, causing an unnecessary wait even when the
farms were idle.

I've removed the old 630 skqp bits, so we can just track what's being
tested on 618.  618 had skqp toml and xfails laying around alreaady.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci/freedreno: Fold a630_egl into a630_gl.

The GL job was only taking 5 minutes in deqp, and egl is 2 minutes.  Saves
a boot, and gets us closer to starting all a630 jobs at the same time if
the farm was idle.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci/iris: Drop parallel on kbl piglit to 2.

We have only 6 of these boards since one died in May, and 7 jobs allocated
to them.  So you ended up with a 5 minute delay on each pipeline with an
otherwise-idle farm while you waited for the first batch of jobs to
complete so you could get the last one started.  It turns out that piglit
was taking 3 minutes of runtime each, so we can just shard piglit 2 ways,
stay under runtime, and not over-allocate the farm

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci/bare-metal: Default our boards to a 20-minute timeout for the whole job.

If the test time is 10 minutes, and boot is 1.5 minutes, that gives you
8.5 minutes of slop for intermittent costs like container download.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci/bare-metal: Drop the 2 vs 1 exit code from poe_run.

This exit code was used for the intra-job retries on the other bare-metal
run scripts, but poe_run doesn't do that.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci: Stop doing internal retries in bare-metal.

We have job-level retry on failure now, and will continue to need to in
order to work around fd.o infrastructure flakes.  If we stop doing retry
inside the job, then we can crank down the gitlab-level timeouts on test
jobs to be closer to our CI guidelines and avoid blocking a runner for an
hour when things go wrong (for example, cheza #16 failing to boot in a
recognized way and continuously looping due to the intra-job retry).
Plus, the job logs will be more readable when you don't have two boots in
one job, and we'll get the flakes surfaced in our monitoring dashboards.

If internal retries were really doing useful work we may see an increase
in flakes as a result of this.  I'm committing to turning off boards or
reducing coverage as necessary to handle this.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

ci/zink: Only test half of piglit pre-merge on anv.

Unfortunately we're at 16 minutes of runtime for this job, which is much
too far past our guidelines.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25790>

zink: fix legacy depth texture rewriting for single component reads

if only a single component is read, this instruction can (and must) be
rewritten to use that component since depth sampling in vulkan is single-component

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25795>

zink: update pointer for GPL pipeline cache entry formats

this otherwise points to ctx->gfx_pipeline_state.rendering_formats,
which is broken for background compiles

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25795>

nir/lower_fragcolor: preserve location_frac

this otherwise breaks component-based outputs

cc: mesa-stable

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25794>

venus: Enable VK_EXT_graphics_pipeline_library behind debug flag

Until the performance issues get resolved, enable the extension
if `VN_DEBUG=gpl`.

Also enable the required (and trivial) VK_KHR_pipeline_library.

Signed-off-by: Lina Versace <linyaa@google.com>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22419>

venus: Do pipeline fixes for VK_EXT_graphics_pipeline_library

Signed-off-by: Lina Versace <linyaa@google.com>
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22419>