review.tizen.org Git - platform/upstream/mesa.git/log

turnip/perfetto: handle gpu timestamps being non-monotonic

Perfetto requires time in clock snaphots to be monotonic, otherwise
the clock would be excluded.
GPU timestamps start from zero after every suspend-resume cycle
which makes them non-monotonic.

As a solution on msm we check whether GPU was just resumed and
remember previous highest timestamp to then add it to the next
timestamps.

If the functionality to get whether gpu is resumed is unavailable
or doesn't work - we fallback to a check for a discontinuity
in timestamps. For kgsl we always use fallback.

Fixes renderstage timeline disappearing in AGI.

Or you could avoid the issue altogether by preventing GPU from going to
sleep by increasing auto suspend delay e.g.:

echo 5000 > /sys/devices/platform/soc\@0/3d00000.gpu/power/autosuspend_delay_ms

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14391>

turnip/trace: process u_trace chunks on queue submission

tu_QueuePresentKHR was not the best place since application
isn't required to call it.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14391>

turnip: rename tu_drm_get_timestamp into tu_device_get_gpu_timestamp

It is not drm specific and will be implemented in kgsl.

Change parameter to tu_device along the way.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14391>

turnip/trace: refactor creation and usage of trace flush data

Fixes the case when last cmd buffer in submission doesn't have
tracepoints leading to flush data not being freed.

Added a few comments, renamed things, refactored allocations - now
the data flow should be a bit more clean.

Extracted submission data creation into tu_u_trace_submission_data_create
which would be later used in in tu_kgsl.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14391>

turnip/perfetto: Optimize timestamp synchronization

We shouldn't do ioctl to get timestamp if perfetto isn't connected.
Also it's better to sync timestamps after submission since the
call could block until GPU is resumed.

Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14391>

virgl/ci: make crosvm-runner pass variables in a secure way

crosvm-runner.sh was using `export -p` to create an environment script
for the virtualized system, but this command will dump every declared
environment variable in the system, which includes Gitlab's CI variables
with sensitive data, such as passwords and auth tokens.

Replacing `export -p` to `generate-env.sh`, which only exports the
necessary variables for Mesa CI jobs.

Extra changes:

* Stop changing ${PWD} variable programmatically in scripts. ${PWD} is a
variable used by most prolific coreutils and bash commands, such as `cd`
and `pwd`, besides it is set by subshells [1]; changing this variable
may lead to complex situations.
As drop-in replacement for ${PWD}, use ${DEQP_BIN_DIR} to flag that
there is a special folder where dEQP should be run.
* Double quote path and array variables. See: https://github.com/koalaman/shellcheck/wiki/SC2086
* Do not export variables directly from commands output. See: https://github.com/koalaman/shellcheck/wiki/SC2155

[1]
```
$ cd /tmp
$ export PWD=test; bash -c 'echo $PWD'
/tmp
```

v2:
- Revert $DEQP_BIN_DIR quoting in crosvm-runner.sh and crosvm-init.sh
- Log all the passed variables to stdout, to help with debugging when
new variable are needed to be put in `generate-env.sh`

v3:
- Revert $DEQP_BIN_DIR quoting leftovers

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14626>

ci/freedreno: Update hashes for closed traces.

These two had different pixel results from last time someone updated them,
but things still look fine.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14757>

tu: Report code size in pipeline statistics

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14754>

intel/ci: expected failure for 1.3 with older CTS

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14706>

relnotes/features: updates for Vulkan 1.3

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14706>

anv: Advertise Vulkan 1.3

Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14706>

anv: switch a bunch of struct/enum to 1.3 versions

Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14706>

anv: Implement 1.3 features/properties

Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14706>

anv: SPIR-V 1.6 shaders imply ALLOW_VARYING_SUBGROUP_SIZE

Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14706>

panvk: Fix pointer corruption in panvk_add_wait_event_syncobjs

nr_in_fences was being incremented to point to an
illegal address

Fixes: 1e23004600c3 ("panvk: Add vkEvents support")
Cc: mesa-stable
Signed-off-by: Manas Chaudhary <manas.chaudhary@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14744>

zink: unify some context casts in zink_create_sampler_view

no functional changes

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14746>

anv: silence wsi debug logging

this is triggered by mesa's own wsi handling, so stop printing nonsense

Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14743>

vulkan/wsi/x11: document implementation

To extend the shared understanding of our code base and ease contributing
document purpose and flow for many of our internal functions.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6536>

radv/ci: mark the dEQP fails related to a missing VKCTS 1.3 as expected

Now that RADV is exposing Vulkan 1.3 by default, VKCTS is getting
confused by it and fails 2 tests:

- dEQP-VK.api.version_check.version: This version of CTS does not
support Vulkan device version 1.3.204 (Fail)
- dEQP-VK.info.device_properties: deviceProperties apiVersion not
valid (Fail)

Mark both of these failures as expected, while we wait for VKCTS 1.3
to be released.

Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14729>

v3dv: document why we don't expose VK_EXT_scalar_block_layout

And since this is an optional feature in Vulkan 1.2, fill in the
corresponding feature query while we are at it.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14731>

v3dv: rework Vulkan 1.2 feature queries

Fill them into a VkPhysicalDeviceVulkan12Features struct like we
do for Vulkan 1.1, and then read them from there.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14731>

v3dv: implement VK_KHR_imageless_framebuffer

Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14704>

v3dv: drop signature of undefined function

This is a left over from when we added multi-version support in the
driver, where we turned this helper into a versioned scheme.

Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14704>

ci/traces: Always generate the junit XML.

While it's not the primary interface to interpreting trace job failures,
it was set in all the traces jobs it looks like and it's low cost anyway.

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/traces: Drop PIGLIT_REPLAY_UPLOAD_TO_MINIO.

You have to do this as part of the traces workflow, otherwise there are no
baseline images for your driver to compare to in the HTML summary.

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/traces: Drop the baseline file creation for trace results.

It's always empty for traces. This reduces more noise in the job logs so
people are more likely to see the link to the HTML.

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/traces: Clean up the failure report message.

You really want to be reviewing the HTML summary with image diffs, not the
junit XML (though we do still generate it so you get the results in the
gitlab UI if that's how you like to interact with it).

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/traces: Drop the PIGLIT_PROFILES setting for traces replay.

Now that the script just does traces, no need to conditionalize this
stuff.

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/traces: Rename the piglit/run.sh script to piglit-traces.sh.

That's the only use of this script that's left.

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/llvmpipe,softpipe: Switch piglit testing to piglit-runner.

The new runner reduces the runtime by about 1/3 thanks to using rust
instead of python, and includes automatic flake handling so you don't just
have to skip flaky tests.  The wrapper script also includes IRC flake
reporting (so one can track and update the flakes list to improve CI
reliability), always uploading results to CI for review (so you can
diagnose flakes and look at timings), has a prettier regressions report
and a helpful timing report, and is the same as what's used by all the HW
runners as well.

The downside is that by dropping the massive list of skips, you no longer
get flagged if Mesa refactors end up accidentally disabling extensions and
thus making tests skip.  For that, I've started on
https://gitlab.freedesktop.org/anholt/deqp-runner/-/merge_requests/33 so
that hardware drivers get extension checking coverage too.

Thanks to the perf improvement, we get to drop one of the jobs for
llvmpipe.

xfail lists were mostly sed-jobs from the prior expectations lists.  The
exceptions to that you'll find in the form of whitespace around the
affected test group (usually changes of capitalization or
special-characters), or an explanation for the more interesting changes
(which thankfully we can now record in the xfails lists!).

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

ci/llvmpipe: Drop the skip of piglit edgeflag test.

I think the fail was fixed by 39ea95330fb5 ("mesa: ensure parameter list
capacity before associating uniform storage"), it doesn't reproduce for me
any more.

Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

Revert "ci: freedreno: Update a530 dEQP fail expectation list"

This reverts commit a35c5540e4241cb4473f5cc9cb1a53bdeee85ada. Another
patch doing so had already landed, so 530 was now failing all its (manual)
tests.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14604>

freedreno: Add missing generated header dependency

Fixes:

  In file included from ../mesa-freedreno-22.0.0_pre/src/gallium/drivers/freedreno/ir3/ir3_cmdline.c:45:
  In file included from ../mesa-freedreno-22.0.0_pre/src/gallium/drivers/freedreno/ir3/ir3_gallium.h:34:
  In file included from ../mesa-freedreno-22.0.0_pre/src/gallium/drivers/freedreno/freedreno_util.h:31:
  ../mesa-freedreno-22.0.0_pre/src/freedreno/drm/freedreno_ringbuffer.h:35:10: fatal error: 'adreno_common.xml.h' file not found
  #include "adreno_common.xml.h"
           ^~~~~~~~~~~~~~~~~~~~~
  1 error generated.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14735>

dri_interface: Remove the remaining DRI1 API definitions

None of these are used anymore, and as a bonus we can drop the dance
around the libdrm headers.

Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14738>

anv: Re-enable CCS_E on TGL+

Commit e6147895885 ("anv: Also disallow CCS_E for multi-LOD images")
accidentally disabled CCS_E on TGL+ because it checked for
image->vk.mip_levels > 0 instead of image->vk.mip_levels > 1.

Instead of reverting it, we remove the code which disables CCS_E for
mipmapped or arrayed images now that we've sufficiently handled the
clear color issue in other ways.

Fixes: e6147895885 ("anv: Also disallow CCS_E for multi-LOD images")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14723>

anv: Use ANV_FAST_CLEAR_DEFAULT_VALUE for CCS on TGL+

On TGL, if a block of fragment shader outputs match the surface's clear
color, the HW may convert them to fast-clears (see HSD 14010672564).
This can lead to rendering corruptions if not handled properly. We
restrict the clear color to zero to avoid issues that can occur with:

- Texture view rendering (including blorp_copy calls)
- Images with multiple levels or array layers

Fixes: e6147895885 ("anv: Also disallow CCS_E for multi-LOD images")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14723>

anv: Disable CCS_E for some 8/16bpp copies on TGL+

CCS_E is currently disabled on TGL+, but we'll enable it soon. We choose
to explicitly disable it for certain copy operations to avoid CTS
failures in the following groups:

- dEQP-VK.drm_format_modifiers.export_import.*
- dEQP-VK.synchronization*

Fixes: e6147895885 ("anv: Also disallow CCS_E for multi-LOD images")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14723>

anv: Drop assert against modifier with aux on gfx12

Commit b664349973b ("anv: Enable implicit CCS for external images")
introduced support for implicit CCS with I915_FORMAT_MOD_Y_TILED.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14723>

docs: use http-redirect when possible

GitLab Pages has added a feature to do proper HTTP redirects, which
are genreally better than the HTML redirects we currently use.

Unfortunately, it doesn't support redirecting to other domains, all
paths must start with a slash. So there's sadly *one* redirect this
doesn't work for. So let's leave that one using a HTML redirect, and
use HTTP redirects when we can.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14369>

softpipe: Dispatch 4 CS invocations per tgsi_exec thread.

We were executing 1 non-helper invocation and 3 helpers per CS tgsi_exec
machine, which was a total waste of the CPU when we could trivially have
all 4 invocations do real work (at least in the common case of a
gl_WorkGroupSize.x >= 4).

This didn't have the effect on dEQP that I was hoping for, as it turns out
that its shaders are almost all 1x1x1 workgroups. However, it does reduce
the runtime of piglit arb_compute_shader-local-id from 2:10 to 47 seconds
on my system.

Part of #4097

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14728>

softpipe: Initialize the CS dispatch mask at machine setup time.

It's not modified later, so no need to reset it per barrier or per
workgroup.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14728>

softpipe: Improve some local var naming in compute shaders.

These aren't dimensions, they're gl_LocalInvocationID.xyz.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14728>

tgsi_exec: Fix shared var stores for >1 real invocation, and overflow checks.

The shared var store overflow checks left a lot of overflowing
opportunities available, while the buffer storage path did proper
checking. But, more importantly for this branch, it always used the first
invocation's offset for each invocation in the quad (which only worked so
far because softpipe only dispatched a single non-helper invocation
per quad).

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14728>

turnip: fix leaks of submit requests.

Fixes: 479a1c40 ("turnip: Porting to common vulkan implementation for synchronization.")

Signed-off-by: Hyunjun Ko <zzoon@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14727>

zink: return 256 for PIPE_CAP_MIN_MAP_BUFFER_ALIGNMENT

this isn't the minimum allowed by the driver, but zink doesn't return
the minimum allowed by the driver anyway and hasn't in a very long time

instead, it suballocates using a minimum alignment of 256 bytes, so use
that instead

fixes (in caselists):
KHR-GL46.map_buffer_alignment.functional

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14720>

zink: fix unused variable warning

fixes: 4ed30be3

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14739>

docs/release-calendar: add another 21.3.x since 22.0 has been delayed a bit

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14737>

docs: update calendar and link releases notes for 21.3.5

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14737>

docs: add release notes for 21.3.5

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14737>

.mailmap: Switch Jason Ekstrand to @collabora.com

Jason is starting at Collabora on the 24th. More details at
https://www.jlekstrand.net/jason/blog/2022/01/hello-collabora/

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14586>

tu: VkExternalImageFormatProperties is optional

..even if external image info has valid external handles.

Fixes: 26380b3a9f8 ("turnip: Add driver skeleton (v2)")

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14730>

radeon/vcn: Updating render_pic_list for correction

In order to keep track of reference frame buffer address changing,
using past_ref to compare with render_pic_list, once the one in
past_ref is valid and if render_pic_list has that entry, it will
need to update it to the latest one in ref[i].

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5868

Signed-off-by: Ruijing Dong <ruijing.dong@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14646>

frontend/va: Keep surface buf addr before reallocation

The reference buffer address is used as the indication in h264 DPB
Tier2, when reference buffer was reallocated, h264 DPB would lose
track of that reference picture. Adding a pointer obsolete_buf in
vlVaSurface data structure for tracking this released buffer, also
in h264_picture_desc adding a private field, which contains
past_ref[16] for tracking previously released buffer vs current
buffer for reference frames.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5868

Signed-off-by: Ruijing Dong <ruijing.dong@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14646>

zink: reorder fbfetch flag-setting to avoid null deref

this avoids dereferencing pg->dd which is allocated a few lines later

Fixes: 417477f60ed ("zink: always use lazy (non-push) updating for fbfetch descriptors")

fixes (radv):
dEQP-GLES31.functional.blend_equation_advanced.basic.multiply

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14718>

radv: fix optimized MSAA copies with suballocated images

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Fixes: deb4685df39 ("radv: implement optimized MSAA copies using FMASK")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5829
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14705>

zink: handle bogus xfb draws

drawing unpopulated xfb data is legal(?) and tested in cts, and the correct
operation is to just drop the draw, so do that here

fixes (nvidia):
GTF-GL46.gtf40.GL3Tests.transform_feedback2.transform_feedback2_api

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14694>

anv: Enable VK_KHR_dynamic_rendering

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: implement the meat of VK_KHR_dynamic_rendering

Includes a fake framebuffer allocation that's necessary for the code we
still use from the regular render passes.

v3: (Lionel)
- Reuse the attachment count from the faux render pass, remove now
unused function
- Add a cmd_buffer_end_rendering function to match begin_rendering,
making use of the split stuff from end_subpass

v4: (Lionel)
- Don't bother with mark_images_writen or resolves on suspend case
- Remove flush at the end of end_rendering, it's not needed

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: split end_subpass into more discrete components

v3: Split cmd_buffer_end_subpass instead of doing parts conditionally (Lionel)

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: Split attachment clearing code into their own functions

v3: Avoid recalculating parameters the caller already had (Lionel)

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: allocate fake render pass for continuation command buffers

v4: Assert if there's no VkCommandBufferInheritanceRenderingInfoKHR (Lionel)

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: Split out state attachments allocation

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: allocate fake render pass on pipeline creation

v3: (Lionel)
- Handle VkPipelineRenderingCreateInfoKHR not being present
- Rename dynamic_pass and set it for regular render passes too

v4: C99 is good (Lionel)

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: add functions to set up fake render passes

There's two of them because they can be created from three points in the
code that provide different details and this is the least ugly way I
could think of for now.

v2: Avoid allocations (Lionel)

v3: Move definition closer to its usage (Lionel)

v4: (Lionel)
- Simplify anv_dynamic_pass_init_full
- Zero out pass/subpass to avoid stall pointers

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv: Remove unused struct member

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

anv/pass: Don't set first_subpass_layout for stencil-only attachments

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13980>

panfrost: Remove NO_BLEND_PACKS quirk

Now unused.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Simplify format class selection

This was made way more complicated than it needs to be for a Midgard-only pass.
The only caller doesn't care about the class, only if it's native or not.
Simplify it appropriately.

It really isn't that hard.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Don't set NO_BLEND_PACKS on Bifrost

It doesn't make sense on Bifrost -- the only consumer of the quirk is
pan_lower_framebuffer, a Midgard-only pass.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Remove MIDGARD_{NO_TYPED_BLEND_STORES,MISSING_LOADS}

These "quirks" are common for Midgard, yet are only consumed by
pan_lower_framebuffer -- a Midgard-only pass. So the quirks should be removed
and inlined into their users. Thid removes MIDGARD_QUIRKS altogether.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Remove NO_TILE_ENABLE_MAP quirk

Function of architecture. Add a comment to the sole consumer of the quirk bit
about why it's used.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Remove MIDGARD_BROKEN_FP16 quirk

Unused.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Remove MIDGARD_SFBD quirk

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

panfrost: Remove HAS_SWIZZLES quirk

It's a function of the major arch.

Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14724>

d3d12: Set caps for tesselation

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Handle input clip array size in the shader key

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Update varying creation logic to handle location_frac

When multiple variables are packed into the same location, we need
to re-construct variables that read/write the same components of that
register so that the DXIL signature is correct. We could try to
merge these variables, but getting the types right sounds harder than
just preserving the multiple individual variables.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Add a state variable for patch_vertices_in

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Handle passthrough TCS in the case where eval is bound

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Handle patch_vertices and patch topology

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Link tesselation control and eval shaders

GLSL puts a bunch of tessellation info in the eval shaders, because
passthrough control shaders can exist. D3D12 puts it in the control
(hull) shader instead. So, when specializing, copy info from domain
to hull. For initial compiles (no domain shader), just make something
up.

D3D12 also requires the domain and hull shaders to have identical
patch constant signatures. Use the existing infrastructure and extend
it to also propagate patch constants. Notably, patch constant locations
are outside of the 64-bit range value so they require a separate pass
to avoid shifts larger than 64.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Initial plumbing for tesselation

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

d3d12: Enable PIPE_CAP_TGSI_TEXCOORD

This is required to be able to use the necessary number of varyings,
otherwise we hit asserts because mesa/st starts assigning varyings
locations above 64 due to the +9 reserving these texcoords.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Handle clip/cull distance as an input to tess shaders

In order to get the semantics right, we need to know how many of the clip/
cull fields are designated for which purpose. In the case of a shader that
can receive these fields as both input and output, the shader_info property
is reserved to store the output info. We could add a dedicated input field
to shader_info, but since it'd probably only be useful for us, just send
it through a side channel during shader linking.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Location_frac needs to be included in sort order

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Primitive ID should only be added as a sysval in geometry shaders

Docs say that its presence in signatures as a "shadow" element (meaning it's not
accessed via load/store, but with a dedicated opcode) is legacy. It seems it
wasn't carried forward when HS/DS were added in D3D11.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Emit DS PSV validation and entrypoint metadata

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Handle domain location intrinsic

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Handle load_output in the HS stage as reading a previously written patch constant

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Handle load_per_vertex_output as LoadOutputControlPoint

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: For load_input from DS, use loadPatchConstant

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: For store_output from HS, use storePatchConstant

In HS, store_per_vertex_output maps to storeOutput in DXIL. The data
that isn't per-vertex is patch constants.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Add a pass for hull and domain shaders to shrink tess level vars

DXIL validation will complain if the tess factor signature entries have the
wrong number of components for the shader's domain. Make sure that both
hull and domain shaders have the right number, and drop loads and stores
from the removed components.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Add patch constant signature into PSV and as container blob

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Gather patch const signature and handle tess factor in it

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: When sorting patch varyings, adjust location to be in normal varying range

This way, patch varyings come before the patch sysvals (tess levels).

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Overlap patch and non-patch varyings so both are separately 0-indexed

Also add tess factors to the list of sysvals that can cause vars to be sorted last.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Fix I/O signatures for tess shaders

- Skip patch variables, those go into a separate patch constant signature
- Use nir_is_arrayed_io and only strip one level of array when it's true

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>

microsoft/compiler: Emit HS PSV validation and entrypoint metadata

Note that this requires the shader info "tess" data to be correct.
For GLSL tess control shaders, only the output primitive count is
automatically available. The rest will need to be either guessed
or filled in from a matching tess eval (domain) shader. This is handled
by the d3d12 driver in a later patch.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Bill Kristiansen <billkris@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14399>