Samuel Pitoiset
e3cd101c17
radv/meta: use BDA for clear HTILE mask
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33493 >
2025-02-12 08:02:30 +00:00
Samuel Pitoiset
13e987669c
radv/video: fix missing image offset when computing VA
...
Found by inspection.
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33496 >
2025-02-12 07:38:29 +00:00
Eric Engestrom
b08f9a2dbd
ci: run containers builds on staging branches
...
Fixes: 7152f343d6 ("ci: only trigger the CI for release managers when pushing to staging branch")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33468 >
2025-02-12 07:06:02 +00:00
Job Noorman
c6a932d4b3
ir3/ra: handle phis with preferred regs first
...
Handle phis in two groups: first those which already have a preferred
reg set and then those without. The second group should be rare but by
handling them last, they don't accidentally occupy a preferred reg of
another phi, preventing excessive copying in some cases.
Signed-off-by: Job Noorman <jnoorman@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33489 >
2025-02-12 06:38:30 +00:00
Job Noorman
922bfe4b6e
ir3: add braces around complex if/else block
...
Will need to add code to one of the blocks in the next commit.
Signed-off-by: Job Noorman <jnoorman@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33489 >
2025-02-12 06:38:30 +00:00
Sagar Ghuge
2e0d5ccd91
intel/compiler: Drop primitive leaf desc load code
...
Looks like we are not using the primitive leaf desc loading code part at
all. Let's just drop it.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33497 >
2025-02-12 05:23:05 +00:00
Mike Blumenkrantz
52dfe1e955
zink: never try to oom flush during unsync texture upload
...
this is very broken
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33484 >
2025-02-12 02:35:45 +00:00
Mike Blumenkrantz
2304078261
zink: only enable unsynchronized_texture_subdata with HIC
...
this is otherwise useless
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33484 >
2025-02-12 02:35:45 +00:00
Michael Cheng
873ad6b6d5
anv: Fix missing Perfetto trace for as build
...
The as_build and related functions only appear when MESA_GPU_TRACES=
perfetto is set. By default, when running an RT workload for profiling,
these traces should be recorded alongside other trace points. This
commit ensures that acceleration structure build events are properly
captured when running an RT workload.
v2(Michael Cheng): Move this logic up to anv_device_init_accel_struct_build_state
v3(Michael Cheng): Set emit_markers = true and let the generated
functions handle the check for u_trace_enable and intel_gpu_tracepoint
Signed-off-by: Michael Cheng <michael.cheng@intel.com >
Reviewed-by: Casey Bowman <casey.g.bowman@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33461 >
2025-02-12 00:13:39 +00:00
David Rosca
441252e9e1
radeonsi/uvd: Set correct chroma format for H264 decode
...
Fixes decoding monochrome (chroma_format_idc = 0).
Cc: mesa-stable
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33396 >
2025-02-11 23:38:26 +00:00
David Rosca
110d406302
radeonsi/vcn: Set correct chroma format for H264 decode
...
Fixes decoding monochrome (chroma_format_idc = 0).
Cc: mesa-stable
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33396 >
2025-02-11 23:38:26 +00:00
David Rosca
c28702c35a
frontends/vdpau: Set H264 chroma_format_idc
...
We don't get the actual value from VDPAU, so hardcode to 4:2:0.
Cc: mesa-stable
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33396 >
2025-02-11 23:38:26 +00:00
Rhys Perry
d946a753e3
aco/ra: unconditionally call undo_renames
...
There's no real reason to not do this more.
fossil-db (navi21):
Totals from 2615 (3.29% of 79377) affected shaders:
Instrs: 4729505 -> 4729484 (-0.00%); split: -0.00%, +0.00%
CodeSize: 25210992 -> 25210036 (-0.00%); split: -0.00%, +0.00%
Latency: 31572966 -> 31572435 (-0.00%); split: -0.00%, +0.00%
InvThroughput: 6918552 -> 6918560 (+0.00%); split: -0.00%, +0.00%
VClause: 132152 -> 132116 (-0.03%); split: -0.03%, +0.00%
SClause: 98595 -> 98575 (-0.02%); split: -0.04%, +0.02%
fossil-db (polaris10):
Totals from 1039 (1.68% of 61794) affected shaders:
Instrs: 708761 -> 708766 (+0.00%); split: -0.00%, +0.00%
CodeSize: 3588772 -> 3588792 (+0.00%); split: -0.00%, +0.00%
Latency: 7458892 -> 7459513 (+0.01%); split: -0.00%, +0.01%
InvThroughput: 3494669 -> 3494722 (+0.00%); split: -0.00%, +0.00%
VClause: 16754 -> 16737 (-0.10%)
SClause: 18190 -> 18156 (-0.19%); split: -0.49%, +0.31%
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com >
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33444 >
2025-02-11 23:10:37 +00:00
Rhys Perry
b94b4188a6
aco/ra: reverse renaming of operands outside update_renames
...
This lets us remove some special casing from update_renames and make it
simpler. Doing this in update_renames was also fragile, since the
parallelcopies vector is not final when update_renames is called, so it
might not have been safe to do so in the end.
No fossil-db changes.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com >
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33444 >
2025-02-11 23:10:37 +00:00
Rhys Perry
ebc7355962
aco: validate operands fixed to definitions
...
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com >
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33444 >
2025-02-11 23:10:37 +00:00
Lionel Landwerlin
4864c0a5fc
anv,driconf: Add sampler coordinate precision workaround for Dynasty Warriors
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12584
Cc: mesa-stable
Reviewed-by: José Roberto de Souza <jose.souza@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33488 >
2025-02-11 22:33:24 +00:00
Lionel Landwerlin
57efd752fb
anv: support protected surfaces with display platform
...
Because our buffer are flagged as protected at the GEM level, we can
just passed them to the display driver and it'll do the right thing.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: José Roberto de Souza <jose.souza@intel.com >
Reviewed-by: José Roberto de Souza <None>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26182 >
2025-02-11 22:03:09 +00:00
Lionel Landwerlin
53762e75e8
vulkan: allow support for protected surfaces
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: José Roberto de Souza <jose.souza@intel.com >
Reviewed-by: José Roberto de Souza <None>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26182 >
2025-02-11 22:03:09 +00:00
Lionel Landwerlin
e722665d65
vulkan/wsi: propagate protected swapchain to images
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: José Roberto de Souza <jose.souza@intel.com >
Reviewed-by: José Roberto de Souza <None>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26182 >
2025-02-11 22:03:09 +00:00
Christian Gmeiner
513ba040fc
docs: Add perfetto driver specifics for V3D
...
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com >
Reviewed-by: Maíra Canal <mcanal@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33348 >
2025-02-11 21:01:40 +00:00
Christian Gmeiner
9b6525809e
docs: Update the list of drivers with CPU tracepoints
...
Turnip has supported this feature for some time, and v3dv recently added support.
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com >
Reviewed-by: Maíra Canal <mcanal@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33348 >
2025-02-11 21:01:40 +00:00
Christian Gmeiner
5880a4bb55
docs: Update perfetto with the latest status
...
v3dv supports PPS counters.
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com >
Reviewed-by: Maíra Canal <mcanal@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33348 >
2025-02-11 21:01:40 +00:00
Juan A. Suarez Romero
68db5481f4
v3dv/ci: disable rpi5 job
...
It seems we have some issue with the driver, due the high number of
(random) flakes that appear in different jobs, and that eventually are
causing issues with Marge.
While we don't identify where is the problem, let's disable the job to
avoid interferences with Marge.
Acked-by: Valentine Burley <valentine.burley@collabora.com >
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33490 >
2025-02-11 18:54:05 +00:00
Boris Brezillon
4ae12cc6ff
panvk: Initialize device virtual address space after the VM creation
...
Make sure we're not lacking a lock/heap destroy when we fail to
create the VM.
Fixes: 53fb1d99ca ("panvk: Transition to explicit VA assignment on v10+")
Reported-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33471 >
2025-02-11 18:26:31 +00:00
Boris Brezillon
5f3c6a0f27
panvk/csf: Don't free the resources twice when init_render_desc_ringbuf() fails
...
init_queue() calls cleanup_queue() if anything fails in the middle, which
means finish_render_desc_ringbuf() will be automatically called if
init_render_desc_ringbuf() failed. Get rid of the the error path and
return directly instead. The one exception we have is the dev_addr
allocation, which needs to be explicitly freed if an error occurs between
util_vma_heap_alloc() and pan_kmod_vm_bind().
Reported-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Fixes: 5544d39f44 ("panvk: Add a CSF backend for panvk_queue/cmd_buffer")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33471 >
2025-02-11 18:26:31 +00:00
Faith Ekstrand
a4f3ec207d
nak/repair_ssa: Use a worklist for get_ssa_or_phi()
...
Between Rust's love of the stack, the size of Rust objects, and the
number of parameters we have to pass in order to sort out lifetime
issues, Rust recursion can be quite expensive. Combine that witn
Windows' tiny stack sizes and this call is blowing out the stack on
games running on DXVK and VKD3D-Proton. This gets rid of this bit of
recursion and replaces it with a loop and a worklist.
Reviewed-by: Mel Henning <drawoc@darkrefraction.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33472 >
2025-02-11 18:06:03 +00:00
Faith Ekstrand
dc82d7edff
nvk: Fix a typo in a comment
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33478 >
2025-02-11 17:35:29 +00:00
Faith Ekstrand
6f64962f27
nvk: Fix scissor bounds
...
This code is old, copied from the old nouveau GL driver. As of Pascal,
we have have 32k images so we need 32k scissors as well. Use the
max_image_dimension() helper instead of hard-coding it.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33478 >
2025-02-11 17:35:29 +00:00
David Rosca
18f0807408
radeonsi/video: Fix creating video buffers with AMD_DEBUG=tmz
...
Reviewed-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33136 >
2025-02-11 16:01:21 +00:00
David Rosca
b241a24724
radeonsi/vcn: Use texture instead of video buffer for DPB buffers
...
Video buffer would be allocated encrypted with AMD_DEBUG=tmz.
Reviewed-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33136 >
2025-02-11 16:01:21 +00:00
David Rosca
fcec81363e
frontends/va: Switch to graphics context when creating protected surface
...
Reviewed-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33136 >
2025-02-11 16:01:20 +00:00
David Rosca
62919ef9d0
frontends/va: Require protected context for VAProtectedSliceDataBuffer
...
Instead of switching the protected playback flag on when processing
first VAProtectedSliceDataBuffer, require the context to be created as
protected.
Reviewed-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33136 >
2025-02-11 16:01:20 +00:00
David Rosca
33c47da3fe
frontends/va: Add CreateContext flag to enable protected context
...
Reviewed-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33136 >
2025-02-11 16:01:19 +00:00
David Rosca
f83faf83e7
frontends/va: Don't try to switch to protected buffer in EndPicture
...
It doesn't work because the buffer was already allocated in BeginPicture.
Reviewed-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33136 >
2025-02-11 16:01:19 +00:00
David Rosca
fdf747af3a
radeonsi/video: Avoid stream handle duplicates in PID namespace
...
Add current time when generating the stream handle initial value.
When running inside PID namespace there can be multiple processes
in the system that will share the same PID and with current code
this could result in the same stream handle being used at the same
time from different processes.
This can easily happen with Flatpak when running two instances of the
same application - both processes will have the same PID and we
will use the same stream handles.
For older UVDs kernel will reject the CS if we use duplicated handles.
Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12575
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33374 >
2025-02-11 15:34:33 +00:00
Samuel Pitoiset
8df1ffaa78
radv: use radv_buffer::addr more
...
And remove radv_buffer:offset.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:34 +00:00
Samuel Pitoiset
d92153e998
radv: compute radv_buffer::addr at bind time
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:34 +00:00
Samuel Pitoiset
e7e43f1437
radv: rename radv_buffer::bo_va to addr
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:34 +00:00
Samuel Pitoiset
f70af40c5d
radv: pass addr to radv_copy_buffer()
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:34 +00:00
Samuel Pitoiset
228903aeaf
radv/rmv: pass addr to log_resource_bind_locked()
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:34 +00:00
Samuel Pitoiset
1d58343b43
radv/video: pass addr to send_cmd()
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:34 +00:00
Samuel Pitoiset
4987926e61
radv: remove unused device memory init/finish helpers
...
Also zero-allocate the vulkan object.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:33 +00:00
Samuel Pitoiset
06ac711b06
radv/meta: simplify creating buffers for R32G32B32 operations
...
Not necessary to allocate things.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:33 +00:00
Samuel Pitoiset
1130478e5d
radv/meta: compute the destination addr earlier for query resolves
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:33 +00:00
Samuel Pitoiset
230affd52b
radv/meta: use BDA for query resolves
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33475 >
2025-02-11 15:12:33 +00:00
Dmitry Baryshkov
6f3062dffa
meson: add freedreno (turnip) Vulkan to arm64 defaults
...
On Aarch64 enable freedreno Vulkan driver (turnip) by default. It is
stable enough to be enabled by default on the relevant platform.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33462 >
2025-02-11 14:35:16 +00:00
Georg Lehmann
f921b42c8c
nir/print: print large floats as mantissa + exponent
...
This is silly:
con 64 %18698 = load_const (0xf0f0f0f0f0f0f0f0 = -107730874267432137203343331820822035577514310242782965586097631855966576162301880634213986293205127792322062538351156704152182839736964151026851280176102232488321269248467172131803507875122376996725092200401040958124190100858265776685056.000000 = -1085102592571150096 = 17361641481138401520)
This is better:
con 64 %18698 = load_const (0xf0f0f0f0f0f0f0f0 = -1.077309e+236 = -1085102592571150096 = 17361641481138401520)
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com >
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33456 >
2025-02-11 13:49:05 +00:00
Valentine Burley
6b0ba1109a
zink/ci: Rename a618 suite and expectation files
...
Match the other zink files, including zink-tu-a630 and zink-tu-a750.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33466 >
2025-02-11 13:12:33 +00:00
Valentine Burley
b90f5f305a
freedreno/ci: Unify naming for a306 jobs
...
The a306 jobs were using files named a307 due to Qualcomm's confusing
naming scheme. While the GPUId for this GPU is 307, its marketing name
is Adreno 306.
To avoid confusion, align CI job naming with the marketing name rather
than the GPUId.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/11951
Signed-off-by: Valentine Burley <valentine.burley@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33466 >
2025-02-11 13:12:33 +00:00
Mary Guillemard
98ea66eca8
ci: Enable mesa-clc and precomp-compiler on debian-arm32
...
panfrost requires those options.
Because we have LLVM on the host, we can build their requirement there
and source them from the system in the cross compilation part.
Signed-off-by: Mary Guillemard <mary.guillemard@collabora.com >
Acked-by: Boris Brezillon <boris.brezillon@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32720 >
2025-02-11 12:33:23 +00:00