docs: add sha sum for 24.1.7

When barriers are used in invalid shaders with non-uniform control flow
we might get a hang. Forcing 32-wide group can help by making it more
probable that barrier instruction is executed by at least one channel
in each thread, and thus hang will be avoided. This shouldn't affect
Xe2+, where active-thread-only barriers are used anyway.

Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/11497
Signed-off-by: Sviatoslav Peleshko <sviatoslav.peleshko@globallogic.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30581>
(cherry picked from commit 7e52b67801)

.gitlab-ci/bare-metal/poe_run.py

@@ -64,7 +64,7 @@ class PoERun:
         if not boot_detected:
             self.print_error(
                 "Something wrong; couldn't detect the boot start up sequence")
-            return 1
+            return 2
 
         self.logger.create_job_phase("test")
         for line in self.ser.lines(timeout=self.test_timeout, phase="test"):

.gitlab-ci/build/gitlab-ci.yml

@@ -190,7 +190,7 @@ debian-build-testing:
     .gitlab-ci/run-shellcheck.sh
     section_switch yamllint "yamllint"
     .gitlab-ci/run-yamllint.sh
-    section_switch meson "meson"
+    section_end yamllint
     .gitlab-ci/meson/build.sh
     .gitlab-ci/prepare-artifacts.sh
   timeout: 15m
@@ -207,7 +207,6 @@ shader-db:
   before_script:
     - !reference [.download_s3, before_script]
   script: |
-    section_switch shader-db "shader-db"
     .gitlab-ci/run-shader-db.sh
   artifacts:
     paths:

.gitlab-ci/common/generate-env.sh

@@ -84,6 +84,7 @@ VARS=(
     MESA_IMAGE_PATH
     MESA_IMAGE_TAG
     MESA_LOADER_DRIVER_OVERRIDE
+    MESA_SPIRV_LOG_LEVEL
     MESA_TEMPLATES_COMMIT
     MESA_VK_ABORT_ON_DEVICE_LOSS
     MESA_VK_IGNORE_CONFORMANCE_WARNING

.gitlab-ci/container/debian/x86_64_build.sh

@@ -85,7 +85,7 @@ python3 -m pip install --break-system-packages -r .gitlab-ci/lava/requirements.t
 
 # install bindgen
 RUSTFLAGS='-L native=/usr/local/lib' cargo install \
-  bindgen-cli --version 0.62.0 \
+  bindgen-cli --version 0.65.1 \
   --locked \
   -j ${FDO_CI_CONCURRENT:-4} \
   --root /usr/local

.gitlab-ci/image-tags.yml

@@ -16,7 +16,7 @@ variables:
    DEBIAN_BASE_TAG: "20240412-pycparser"
 
    DEBIAN_X86_64_BUILD_IMAGE_PATH: "debian/x86_64_build"
-   DEBIAN_BUILD_TAG: "20240408-cbindgen"
+   DEBIAN_BUILD_TAG: "20240723-241cbindgen"
 
    DEBIAN_X86_64_TEST_BASE_IMAGE: "debian/x86_64_test-base"
 

.gitlab-ci/lava/utils/constants.py

@@ -23,3 +23,10 @@ KNOWN_ISSUE_R8152_PATTERNS: tuple[str, ...] = (
 # This is considered noise, since LAVA produces this log after receiving a package of feedback
 # messages.
 LOG_DEBUG_FEEDBACK_NOISE = "Listened to connection for namespace 'dut' done"
+
+A6XX_GPU_RECOVERY_WATCH_PERIOD_MIN = 3
+A6XX_GPU_RECOVERY_FAILURE_MAX_COUNT = 30
+A6XX_GPU_RECOVERY_FAILURE_MESSAGE = (
+    "cx gdsc didn't collapse",
+    "Timeout waiting for GMU OOB",
+)

.gitlab-ci/lava/utils/lava_log_hints.py

@@ -1,8 +1,9 @@
 from __future__ import annotations
 
 import re
+from datetime import datetime, timedelta
 from dataclasses import dataclass, field
-from typing import TYPE_CHECKING, Any, Sequence
+from typing import TYPE_CHECKING, Any, Optional, Sequence
 
 if TYPE_CHECKING:
     from lava.utils import LogFollower
@@ -13,6 +14,9 @@ from lava.utils.constants import (
     KNOWN_ISSUE_R8152_MAX_CONSECUTIVE_COUNTER,
     LOG_DEBUG_FEEDBACK_NOISE,
     KNOWN_ISSUE_R8152_PATTERNS,
+    A6XX_GPU_RECOVERY_WATCH_PERIOD_MIN,
+    A6XX_GPU_RECOVERY_FAILURE_MESSAGE,
+    A6XX_GPU_RECOVERY_FAILURE_MAX_COUNT,
 )
 from lava.utils.log_section import LogSectionType
 
@@ -29,6 +33,8 @@ class LAVALogHints:
     log_follower: LogFollower
     r8152_issue_consecutive_counter: int = field(default=0, init=False)
     reboot_counter: int = field(default=0, init=False)
+    a6xx_gpu_recovery_fail_counter: int = field(default=0, init=False)
+    a6xx_gpu_first_fail_time: Optional[datetime] = field(default=None, init=False)
 
     def raise_known_issue(self, message) -> None:
         raise MesaCIKnownIssueException(
@@ -44,6 +50,7 @@ class LAVALogHints:
                 continue
             self.detect_r8152_issue(line)
             self.detect_forced_reboot(line)
+            self.detect_a6xx_gpu_recovery_failure(line)
 
     def detect_r8152_issue(self, line):
         if self.log_follower.phase in (
@@ -77,3 +84,23 @@ class LAVALogHints:
                     self.raise_known_issue(
                         "Forced reboot detected during test phase, failing the job..."
                     )
+
+    # If the a6xx gpu repeatedly fails to recover over a short period of time,
+    # then successful recovery is unlikely so cancel the job preemptively.
+    def detect_a6xx_gpu_recovery_failure(self, line: dict[str, Any]) -> None:
+        if search_known_issue_patterns(A6XX_GPU_RECOVERY_FAILURE_MESSAGE, line["msg"]):
+            time_of_failure = datetime.fromisoformat(line["dt"])
+            self.a6xx_gpu_recovery_fail_counter += 1
+
+            if self.a6xx_gpu_first_fail_time is None:
+                self.a6xx_gpu_first_fail_time = time_of_failure
+
+            if self.a6xx_gpu_recovery_fail_counter == A6XX_GPU_RECOVERY_FAILURE_MAX_COUNT:
+                time_since_first_fail = time_of_failure - self.a6xx_gpu_first_fail_time
+                if time_since_first_fail <=  timedelta(minutes=A6XX_GPU_RECOVERY_WATCH_PERIOD_MIN):
+                    self.raise_known_issue(
+                        "Repeated GPU recovery failure detected: cancelling the job"
+                    )
+                else:
+                    self.a6xx_gpu_first_fail_time = None
+                    self.a6xx_gpu_recovery_fail_counter = 0

.gitlab-ci/test/gitlab-ci.yml

@@ -33,7 +33,7 @@
   tags:
     - placeholder-job
 
-rustfmt:
+.rustfmt:
   extends:
     - .formatting-check
     - .lint-rustfmt-rules

.gitlab-ci/tests/utils/test_lava_log.py

@@ -16,7 +16,13 @@ from lava.utils import (
     fix_lava_gitlab_section_log,
     hide_sensitive_data,
 )
-from lava.utils.constants import KNOWN_ISSUE_R8152_MAX_CONSECUTIVE_COUNTER
+from lava.utils.constants import (
+    KNOWN_ISSUE_R8152_MAX_CONSECUTIVE_COUNTER,
+    A6XX_GPU_RECOVERY_WATCH_PERIOD_MIN,
+    A6XX_GPU_RECOVERY_FAILURE_MESSAGE,
+    A6XX_GPU_RECOVERY_FAILURE_MAX_COUNT,
+)
+from lava.utils.lava_log_hints import LAVALogHints
 
 from ..lava.helpers import (
     create_lava_yaml_msg,
@@ -390,3 +396,44 @@ def test_detect_failure(messages, expectation):
     lf = LogFollower(starting_section=boot_section)
     with expectation:
         lf.feed(messages)
+
+def test_detect_a6xx_gpu_recovery_failure(frozen_time):
+    log_follower = LogFollower()
+    lava_log_hints = LAVALogHints(log_follower=log_follower)
+    failure_message = {
+        "dt": datetime.now().isoformat(),
+        "msg": A6XX_GPU_RECOVERY_FAILURE_MESSAGE[0],
+        "lvl": "feedback",
+    }
+    with pytest.raises(MesaCIKnownIssueException):
+        for _ in range(A6XX_GPU_RECOVERY_FAILURE_MAX_COUNT):
+            lava_log_hints.detect_a6xx_gpu_recovery_failure(failure_message)
+            # Simulate the passage of time within the watch period
+            frozen_time.tick(1)
+            failure_message["dt"] = datetime.now().isoformat()
+
+def test_detect_a6xx_gpu_recovery_success(frozen_time):
+    log_follower = LogFollower()
+    lava_log_hints = LAVALogHints(log_follower=log_follower)
+    failure_message = {
+        "dt": datetime.now().isoformat(),
+        "msg": A6XX_GPU_RECOVERY_FAILURE_MESSAGE[0],
+        "lvl": "feedback",
+    }
+    # Simulate sending a tolerable number of failure messages
+    for _ in range(A6XX_GPU_RECOVERY_FAILURE_MAX_COUNT - 1):
+        lava_log_hints.detect_a6xx_gpu_recovery_failure(failure_message)
+        frozen_time.tick(1)
+        failure_message["dt"] = datetime.now().isoformat()
+
+    # Simulate the passage of time outside of the watch period
+    frozen_time.tick(60 * A6XX_GPU_RECOVERY_WATCH_PERIOD_MIN + 1)
+    failure_message = {
+        "dt": datetime.now().isoformat(),
+        "msg": A6XX_GPU_RECOVERY_FAILURE_MESSAGE[1],
+        "lvl": "feedback",
+    }
+    with does_not_raise():
+        lava_log_hints.detect_a6xx_gpu_recovery_failure(failure_message)
+    assert lava_log_hints.a6xx_gpu_first_fail_time is None, "a6xx_gpu_first_fail_time is not None"
+    assert lava_log_hints.a6xx_gpu_recovery_fail_counter == 0, "a6xx_gpu_recovery_fail_counter is not 0"

VERSION

@@ -1 +1 @@
-24.1.1
+24.1.7

docs/_exts/bootstrap.py

@@ -96,7 +96,7 @@ class BootstrapHTML5TranslatorMixin:
         self.body.append(tag)
 
 def setup_translators(app):
-    if app.builder.default_translator_class is None:
+    if app.builder.format != "html":
         return
 
     if not app.registry.translators.items():
@@ -111,10 +111,6 @@ def setup_translators(app):
         app.set_translator(app.builder.name, translator, override=True)
     else:
         for name, klass in app.registry.translators.items():
-            if app.builder.format != "html":
-                # Skip translators that are not HTML
-                continue
-
             translator = types.new_class(
                 "BootstrapHTML5Translator",
                 (

docs/conf.py

@@ -111,8 +111,10 @@ html_copy_source = False
 # Add any paths that contain custom static files (such as style sheets) here,
 # relative to this directory. They are copied after the builtin static files,
 # so a file named "default.css" will overwrite the builtin "default.css".
-html_static_path = [
-  '_static/',
+html_static_path = []
+
+html_extra_path = [
+  '_extra/',
   'release-maintainers-keys.asc',
   'features.txt',
   'libGL.txt',
@@ -120,8 +122,6 @@ html_static_path = [
   'README.VCE',
 ]
 
-html_extra_path = []
-
 html_redirects = [
   ('webmaster', 'https://www.mesa3d.org/website/'),
   ('developers', 'https://www.mesa3d.org/developers/'),
@@ -218,7 +218,7 @@ graphviz_output_format = 'svg'
 
 # -- Options for hawkmoth -------------------------------------------------
 
-hawkmoth_root = os.path.abspath('..')
+hawkmoth_root = os.path.abspath(os.pardir)
 hawkmoth_clang = [
   '-Idocs/header-stubs/',
   '-Iinclude/',

docs/envvars.rst

@@ -1443,7 +1443,11 @@ RADV driver environment variables
    ``noopt``
       disable various optimizations
    ``nosched``
-      disable instructions scheduling
+      disable pre-RA, ILP and VOPD instruction scheduling
+   ``nosched-ilp``
+      disable ILP instruction scheduling
+   ``nosched-vopd``
+      disable VOPD instruction scheduling
    ``perfinfo``
       print information used to calculate some pipeline statistics
    ``liveinfo``

docs/features.txt

@@ -860,7 +860,7 @@ Rusticl OpenCL 1.2 -- all DONE:
   Separate compilation and linking of programs          DONE
   Extend cl_mem_flags                                   DONE
   clEnqueueFillBuffer, clEnqueueFillImage               DONE
-  Add CL_MAP_WRITE_INVALIDATE_REGION to cl_map_flags    in progress (flag is ignored)
+  Add CL_MAP_WRITE_INVALIDATE_REGION to cl_map_flags    DONE
   New image types                                       DONE
   clCreateImage                                         DONE
   clEnqueueMigrateMemObjects                            DONE

docs/relnotes.rst

@@ -3,6 +3,12 @@ Release Notes
 
 The release notes summarize what's new or changed in each Mesa release.
 
+-  :doc:`24.1.7 release notes <relnotes/24.1.7>`
+-  :doc:`24.1.6 release notes <relnotes/24.1.6>`
+-  :doc:`24.1.5 release notes <relnotes/24.1.5>`
+-  :doc:`24.1.4 release notes <relnotes/24.1.4>`
+-  :doc:`24.1.3 release notes <relnotes/24.1.3>`
+-  :doc:`24.1.2 release notes <relnotes/24.1.2>`
 -  :doc:`24.1.1 release notes <relnotes/24.1.1>`
 -  :doc:`24.1.0 release notes <relnotes/24.1.0>`
 -  :doc:`24.0.5 release notes <relnotes/24.0.5>`
@@ -418,6 +424,12 @@ The release notes summarize what's new or changed in each Mesa release.
    :maxdepth: 1
    :hidden:
 
+   24.1.7 <relnotes/24.1.7>
+   24.1.6 <relnotes/24.1.6>
+   24.1.5 <relnotes/24.1.5>
+   24.1.4 <relnotes/24.1.4>
+   24.1.3 <relnotes/24.1.3>
+   24.1.2 <relnotes/24.1.2>
    24.1.1 <relnotes/24.1.1>
    24.1.0 <relnotes/24.1.0>
    24.0.5 <relnotes/24.0.5>

docs/relnotes/24.1.1.rst

@@ -19,7 +19,7 @@ SHA256 checksum
 
 ::
 
-    TBD.
+    0038826c6f7e88d90b4ce6f719192fa58ca7dedf4edcaa1174cf7bd920ef89ea  mesa-24.1.1.tar.xz
 
 
 New features

docs/relnotes/24.1.2.rst

@@ -0,0 +1,228 @@
+Mesa 24.1.2 Release Notes / 2024-06-19
+======================================
+
+Mesa 24.1.2 is a bug fix release which fixes bugs found since the 24.1.1 release.
+
+Mesa 24.1.2 implements the OpenGL 4.6 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 4.6. OpenGL
+4.6 is **only** available if requested at context creation.
+Compatibility contexts may report a lower version depending on each driver.
+
+Mesa 24.1.2 implements the Vulkan 1.3 API, but the version reported by
+the apiVersion property of the VkPhysicalDeviceProperties struct
+depends on the particular driver being used.
+
+SHA256 checksum
+---------------
+
+::
+
+    a2c584c8d57d3bd8ba11790a6e9ae3713f8821df96c059b78afb29dd975c9f45  mesa-24.1.2.tar.xz
+
+
+New features
+------------
+
+- None
+
+
+Bug fixes
+---------
+
+- free_zombie_shaders() leave context in a bad state (access violation occurs)
+- [NINE]Far Cry 1 trees flicker regression[bisected][traces]
+- Vulkan: Most sync2 implementations are missing new access flags
+- Incorrect buffer_list advance when writing disjoint image descriptors
+- ANV: Block shadows in Cyberpunk on Intel A770
+- ACO ERROR: Temporary never defined or are defined after use
+- VAAPI ffmpeg encoding breaks with mesa-24.1.0
+- [ANV] Graphics memory allocation in Total War: Warhammer 3
+- NVK: Vulkan apps simply terminated with segfault under wayland and Xwayland
+- NVK: VK_ERROR_OUT_OF_DEVICE_MEMORY on swapchain creation
+- freedreno + perfetto missing dependency on adreno_common.xml.h
+
+
+Changes
+-------
+
+Amol Surati (1):
+
+- nine: avoid using post-compacted indices with state expecting pre-compacted ones
+
+Boris Brezillon (2):
+
+- pan/bi: Fix dynamic indexing of push constants
+- panvk: Fix Cube/2DArray/3D img -> buf copies
+
+Caio Oliveira (1):
+
+- intel/brw: Fix typo in DPAS emission code
+
+Daniel Schürmann (1):
+
+- aco/assembler: fix MTBUF opcode encoding on GFX11
+
+Danylo Piliaiev (1):
+
+- freedreno: Make fd_pps_driver.h usable without including other FD sources
+
+Dave Airlie (4):
+
+- nvk: Only enable WSI modifiers if the extension is supported.
+- draw/texture: handle mip_offset[0] being != 0 for layered textures.
+- nouveau/nvc0: increase overallocation on shader bo to 2K
+- radv/video: fix layered decode h264/5 tests.
+
+David Heidelberg (1):
+
+- rusticl: add -cl-std only when it's not defined
+
+David Rosca (2):
+
+- radeonsi: Fix si_compute_clear_render_target with 422 subsampled formats
+- radv/video: Add missing VCN 3.0.2 to decoder init switch
+
+Eric Engestrom (15):
+
+- docs: add sha256sum for 24.1.1
+- .pick_status.json: Update to 50e5067be77bf8f34de6616e8edca2af2cf8d310
+- v3dv: add missing bounds check in VK_EXT_4444_formats
+- .pick_status.json: Update to cc82f7f8ace50f68b06c53ad347e36d411ae9dab
+- radv/ci: fix manual rules
+- .pick_status.json: Update to 41dd1c52b1d091b36f8931c4a57d3b6dc361bc84
+- v3d/drm-shim: emulate a rpi4 instead of a rpi3
+- .pick_status.json: Update to a80a1c983844bca646d5f07d65c695a84f964bfe
+- egl: fix teardown when using xcb
+- .pick_status.json: Mark f017beb29ce6e3469da33caff2c9a493799faca6 as denominated
+- .pick_status.json: Update to 7dcba7e873c6b753930e2fdc8c714bb4da1a22dd
+- glx: fix build -D glx-direct=false
+- .pick_status.json: Update to 10d21d410068f2ca32fe898f6b4b690993d90daa
+- .pick_status.json: Mark a9fff07c2e2b1e52b00b30dc16781209f7761c04 as denominated
+- .pick_status.json: Update to 887f0e0af664b11c081b4140931e7213240c7b41
+
+Erik Faye-Lund (3):
+
+- mesa/main: remove stale prototype
+- mesa/main: do not allow RGBA_INTEGER et al in gles3
+- panvk: move macro-definition to header
+
+Faith Ekstrand (5):
+
+- nak: Only convert the written portion of the buffer in NirInstrPrinter
+- nak: BMov is always variable-latency
+- nak: Only copy-prop neg into iadd2/3 if no carry is written
+- nak/legalize: Fold immediate sources before instructions
+- nouveau: Fix a race in nouveau_ws_bo_destroy()
+
+Friedrich Vock (2):
+
+- radv/rt: Fix memory leak when compiling libraries
+- aco/spill: Don't spill phis with all-undef operands
+
+Georg Lehmann (1):
+
+- radeonsi: set COMPUTE_STATIC_THREAD_MGMT_SE2-3 correctly on gfx10-11
+
+Iago Toral Quiroga (1):
+
+- broadcom/compiler: initialize payload_conflict for all initial nodes
+
+Iván Briano (1):
+
+- vulkan/runtime: pColorAttachmentInputIndices is allowed to be NULL
+
+Job Noorman (14):
+
+- ir3: fix crash in try_evict_regs with src reg
+- ir3: fix handling of early clobbers in calc_min_limit_pressure
+- ir3: set offset on splits created while spilling
+- ir3: correctly set wrmask for reload.macro
+- ir3: don't remove intervals for non-killed tex prefetch sources
+- ir3: don't remove collects early while spilling
+- ir3: expose instruction indexing helper for merge sets
+- ir3: make indexing instructions optional in ir3_merge_regs
+- ir3: index instructions before fixing up merge sets after spilling
+- ir3: move liveness recalculation inside ir3_ra_shared
+- ir3: restore interval_offset after liveness recalculation in shared RA
+- ir3: add ir3_cursor/ir3_builder helpers
+- ir3: refactor ir3_spill.c to use the ir3_cursor/ir3_builder API
+- ir3: only add live-in phis for top-level intervals while spilling
+
+Karol Herbst (2):
+
+- rusticl/spirv: do not pass a NULL pointer to slice::from_raw_parts
+- rusticl/memory: copies might overlap for host ptrs
+
+Konstantin Seurer (2):
+
+- ac/llvm: Fix DENORM_FLUSH_TO_ZERO with exact instructions
+- ac/llvm: Enable helper invocations for vote_all/any
+
+Lionel Landwerlin (4):
+
+- anv: fix pipeline flag fields
+- anv: limit aux invalidations to primary command buffers
+- anv: ensure completion of surface state copies before secondaries
+- intel/fs: fix lower_simd_width for MOV_INDIRECT
+
+Lucas Fryzek (1):
+
+- llvmpipe: query winsys support for dmabuf mapping
+
+Marek Olšák (1):
+
+- Revert "radeonsi: fix initialization of occlusion query buffers for disabled RBs"
+
+Mary Guillemard (2):
+
+- panvk: Add missing null check in DestroyCommandPool
+- panvk: Check for maxBufferSize in panvk_CreateBuffer
+
+Mike Blumenkrantz (2):
+
+- lavapipe: fix mesh+task binding with shader objects
+- mesa/st: fix zombie shader handling for non-current programs
+
+Patrick Lerda (1):
+
+- radeonsi: fix assert triggered on gfx6 after the tessellation update
+
+Qiang Yu (2):
+
+- glsl: respect GL_EXT_shader_image_load_formatted when image is embedded in a struct
+- radeonsi: add missing nir_intrinsic_bindless_image_descriptor_amd
+
+Rhys Perry (4):
+
+- aco: don't combine vgpr into writelane src0
+- aco/gfx6: set glc for buffer_store_byte/short
+- aco: remove some missing label resets
+- aco: insert s_nop before discard early exit sendmsg(dealloc_vgpr)
+
+Samuel Pitoiset (4):
+
+- radv: fix creating unlinked shaders with ESO when nextStage is 0
+- radv: don't assume that TC_ACTION_ENA invalidates L1 cache on gfx9
+- radv: fix incorrect buffer_list advance for multi-planar descriptors
+- radv: always save/restore all shader objects for internal operations
+
+Sviatoslav Peleshko (3):
+
+- anv,driconf: Add fake non device local memory WA for Total War: Warhammer 3
+- intel/brw: Actually retype integer sources of sampler message payload
+- intel/elk: Actually retype integer sources of sampler message payload
+
+Timur Kristóf (1):
+
+- ac/nir/tess: Fix per-patch output LDS mapping.
+
+Valentine Burley (2):
+
+- tu: Handle the new sync2 flags
+- tu: Remove declaration of unused update_stencil_mask function
+
+Zan Dobersek (1):
+
+- tu: fix ZPASS_DONE interference between occlusion queries and autotuner

docs/relnotes/24.1.3.rst

@@ -0,0 +1,199 @@
+Mesa 24.1.3 Release Notes / 2024-07-03
+======================================
+
+Mesa 24.1.3 is a bug fix release which fixes bugs found since the 24.1.2 release.
+
+Mesa 24.1.3 implements the OpenGL 4.6 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 4.6. OpenGL
+4.6 is **only** available if requested at context creation.
+Compatibility contexts may report a lower version depending on each driver.
+
+Mesa 24.1.3 implements the Vulkan 1.3 API, but the version reported by
+the apiVersion property of the VkPhysicalDeviceProperties struct
+depends on the particular driver being used.
+
+SHA256 checksum
+---------------
+
+::
+
+    63236426b25a745ba6aa2d6daf8cd769d5ea01887b0745ab7124d2ef33a9020d  mesa-24.1.3.tar.xz
+
+
+New features
+------------
+
+- None
+
+
+Bug fixes
+---------
+
+- Incorrect colours on desktop and apps
+- nir: Incorrect nir_opt_algebraic semantics for signed integer constants causing end-to-end miscompiles
+- src/gallium/frontends/clover/meson.build:93:40: ERROR: Unknown variable "idep_mesaclc".
+- [radv] GPU hang in Starfield on RX 5700 XT
+- Indika: flickering black artifacting on the snow
+- MESA 24.1 - broken zink OpenGL under Windows
+- Blue flickering rectangles on AMD RX 7600
+- GPU hangs on AMD Radeon RX 6400 on a fragment shader
+
+
+Changes
+-------
+
+Alyssa Rosenzweig (1):
+
+- nir: fix miscompiles with rules with INT32_MIN
+
+Bas Nieuwenhuizen (1):
+
+- util/disk_cache: Fix cache marker refresh.
+
+Connor Abbott (1):
+
+- ir3: Make sure constlen includes stc/ldc.k/ldg.k instructions
+
+Daniel Schürmann (1):
+
+- aco/spill: Unconditionally add 2 SGPRs to live-in demand
+
+Dylan Baker (2):
+
+- clc: remove check for null pointer that cannot be true in llvm_mod_to_spirv
+- anv/grl: add some validation that we're not going to overflow
+
+Eric Engestrom (14):
+
+- docs: add sha256sum for 24.1.2
+- ci/shader-db: drop extra nesting section
+- ci/debian-build-testing: drop extra nesting section
+- ci: fix section_end in debian-build-testing
+- .pick_status.json: Update to 1ff86021a7a06d2548482c40b1584042e298f58e
+- .pick_status.json: Update to c4a38c658317bc8d17447fd6ee3e717a96ca9948
+- .pick_status.json: Update to dd85b50d182a2bd1c67d9a8f858d93fc4dded91c
+- .pick_status.json: Update to 037eaa962b56ff70ecf889ace05020635964e23c
+- .pick_status.json: Update to 68215332a8cd87d8109ee4c3b50e04df223d9c83
+- .pick_status.json: Update to 6b5a12611bff70ffb3c736de29ff5631efbb8770
+- .pick_status.json: Update to 076cbf605e84ad2f7353099af95969702aac5b77
+- .pick_status.json: Mark 41698eee96b17ab11773ca92bf557d35bc72e207 as denominated
+- .pick_status.json: Mark 7033623acd8b7bae8bc52911d4d1c3223726a8f9 as denominated
+- .pick_status.json: Mark 5ca85d75c05de9df7c3170122dfdb04bc795b43a as denominated
+
+Erik Faye-Lund (4):
+
+- nir: fix utf-8 encoding-issue
+- Revert "docs: use html_static_path for static files"
+- docs: use os.pardir
+- docs: fix bootstrap-extension
+
+Faith Ekstrand (1):
+
+- nir/format_convert: Smash NaN to 0 in pack_r9g9b9e5()
+
+Iago Toral Quiroga (2):
+
+- broadcom/compiler: don't spill in between multop and umul24
+- broadcom/compiler: fix per-quad spilling
+
+Jesse Natalie (2):
+
+- wgl: Delete pixelformat support query
+- wgl: Fix flag check for GDI compat
+
+José Expósito (1):
+
+- llvmpipe: Init eglQueryDmaBufModifiersEXT num_modifiers
+
+José Roberto de Souza (2):
+
+- anv: Remove block promoting non CPU mapped bos to coherent
+- anv: Fix assert in xe_gem_create()
+
+Julian Orth (1):
+
+- egl/wayland: ignore unsupported driver configs
+
+Karol Herbst (7):
+
+- util/u_printf: properly handle %%
+- rusticl/memory: assume minimum image_height of 1
+- rusticl/memory: fix clFillImage for buffer images
+- rusticl: add new CL_INVALID_BUFFER_SIZE condition for clCreateBuffer
+- rusticl: add bsymbolic to linker flags
+- rusticl/queue: gracefully stop the worker thread
+- nir/schedule: add write dep also for shared_atomic
+
+Konstantin Seurer (4):
+
+- llvmpipe: Lock shader access to sample_functions
+- llvmpipe: Stop using a sample_functions pointer as cache key
+- llvmpipe: Only evict cache entries if a fence is available
+- lavapipe: Always call finish_fence after lvp_execute_cmd_buffer
+
+Lionel Landwerlin (6):
+
+- anv: fix vkCmdWaitEvents2 handling
+- anv: add a protected scratch pool
+- anv: prepare 2 variants of all shader instructions
+- anv: allocate compute scratch using the right scratch pool
+- anv: emit the right shader instruction for protected mode
+- anv: workaround flaky xfb query results on Gfx11
+
+Luc Ma (1):
+
+- meson: Build pipe-loader when build-tests is true
+
+Mary Guillemard (1):
+
+- panvk: Report correct min value for discreteQueuePriorities
+
+Michel Dänzer (2):
+
+- egl/dri: Use packed pipe_format
+- dri: Go back to hard-coded list of RGBA formats
+
+Mike Blumenkrantz (2):
+
+- dri: rename 'implicit' param from earlier series
+- zink: null check pipe loader config before use
+
+Neha Bhende (1):
+
+- svga: Retrieve stride info from hwtnl->cmd.vdecl for swtnl draws
+
+Patrick Lerda (1):
+
+- clover: fix meson opencl-spirv option
+
+Paulo Zanoni (2):
+
+- anv/xe: fix declaration of memory flags for integrated non-LLC platforms
+- anv/sparse: fix TR-TT page table bo size and flags
+
+Pierre-Eric Pelloux-Prayer (2):
+
+- ac/llvm: implement WA in nir to llvm
+- ac/surface: reject modifiers with retile_dcc and bpe != 32
+
+Qiang Yu (2):
+
+- nir: fix lower array to vec metadata preserve
+- nir: fix clip cull distance lowering metadata preserve
+
+Rhys Perry (3):
+
+- aco/insert_exec_mask: ensure top mask is not a temporary at loop exits
+- vtn: ensure TCS control barriers have a large enough memory scope
+- aco: skip continue_or_break LCSSA phis when not needed
+
+Samuel Pitoiset (2):
+
+- radv/amdgpu: fix chaining CS with external IBs on compute queue
+- radv: fix incorrect cache flushes before decompressing DCC on compute
+
+Tapani Pälli (1):
+
+- isl: fix condition for enabling sampler route to lsc

docs/relnotes/24.1.4.rst

@@ -0,0 +1,173 @@
+Mesa 24.1.4 Release Notes / 2024-07-17
+======================================
+
+Mesa 24.1.4 is a bug fix release which fixes bugs found since the 24.1.3 release.
+
+Mesa 24.1.4 implements the OpenGL 4.6 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 4.6. OpenGL
+4.6 is **only** available if requested at context creation.
+Compatibility contexts may report a lower version depending on each driver.
+
+Mesa 24.1.4 implements the Vulkan 1.3 API, but the version reported by
+the apiVersion property of the VkPhysicalDeviceProperties struct
+depends on the particular driver being used.
+
+SHA256 checksum
+---------------
+
+::
+
+    7cf7c6f665263ad0122889c1d4b076654c1eedea7a2f38c69c8c51579937ade1  mesa-24.1.4.tar.xz
+
+
+New features
+------------
+
+- None
+
+
+Bug fixes
+---------
+
+- [regression][bisected] VMware Xv video displays as black rectangle
+- Blender 4.2,4.3 crashes when rendering with motion blur on RDNA3 cards (OpenGL/radeonsi)
+- Transparent background in Blender 3D view with nouveau
+- Stuttering and delays with AV1 decoding in Mesa 24 using VAAPI on AMD Phoenix, works fine on Mesa 23
+- 24.0.7 AV1 VA-API dropping frames
+- Gnome shell (wayland) crashes when opening any window
+- DRI Intel drivers fix a problem in Redhat 7 (Mesa 18), but are not included for Redhat 8 (Mesa versions  v23, v24)
+- Vulkan: ../src/nouveau/vulkan/nvk_physical_device.c:1109: VK_ERROR_INCOMPATIBLE_DRIVER
+- RADV: Smooth lines affect triangle rendering
+- [armhf build error][regression] error: ‘StringMapIterator’ was not declared in this scope; did you mean ‘llvm::StringMapIterator’?
+
+
+Changes
+-------
+
+Aleksi Sapon (1):
+
+- lavapipe: build "Windows" check should use the host machine, not the \`platforms` option.
+
+Connor Abbott (3):
+
+- tu: Make cs writeable for GMEM loads when FDM is enabled
+- tu: Fix fdm_apply_load_coords patchpoint size
+- ir3: Fix stg/ldg immediate offset on a7xx
+
+Dave Airlie (2):
+
+- anv/video: use correct offset for MPR row store scratch buffer.
+- radv/video: advertise mutable/extended for dst video images.
+
+David Rosca (1):
+
+- Reapply "radeonsi/vcn: AV1 skip the redundant bs resize"
+
+Doug Brown (1):
+
+- xa: add missing stride setup in renderer_draw_yuv
+
+Eric Engestrom (6):
+
+- docs: add sha256sum for 24.1.3
+- [24.1 only] ci: disable rustfmt
+- .pick_status.json: Update to d9e41e8a8ca3a8a22628513b44764fa7675ec288
+- .pick_status.json: Update to ae3e0ae26a4678b317727dc08ae64aee6577374d
+- .pick_status.json: Update to a04dc1a4517bbe359fb246a79cd38c99f250c826
+- .pick_status.json: Update to 2d260314f101540298edf973f5393e3468ed84ba
+
+Erico Nunes (1):
+
+- lima: fix surface reload flags assignment
+
+Faith Ekstrand (5):
+
+- nvk: Silently fail to enumerate if not on nouveau
+- nvk: Bump the sparse alignment requirement on buffers to 64K
+- nvk: Align sparse-bound images to the sparse binding size
+- zink/kopper: Set VK_COMPOSITE_ALPHA_OPAQUE_BIT when PresentOpaque is set
+- nvk: Drop the sparse alignment back down to 4096
+
+Karol Herbst (8):
+
+- rusticl/program: move binary parsing into its own function
+- rusticl/program: make binary API not crash on errors
+- rusticl/program: use blob.h to parse binaries
+- rusticl/program: update binary format
+- rusticl/buffer: harden bound checks against overflows
+- rusticl/context: move SVM pointer tracking into own type
+- rusticl/ptr: add a few APIs to TrackedPointers
+- rusticl/memory: complete rework on how mapping is implemented
+
+Kenneth Graunke (1):
+
+- intel/nir: Don't needlessly split u2f16 for nir_type_uint32
+
+Konstantin Seurer (1):
+
+- radv: Fix smooth lines with dynamic polygon mode and topology
+
+Marek Olšák (5):
+
+- ac/surface: finish display DCC for gfx11.5
+- radeonsi: replace si_shader::scratch_bo with scratch_va, don't set it on gfx11+
+- radeonsi: don't update compute scratch if the compute shader doesn't use it
+- ac: add radeon_info::has_scratch_base_registers
+- radeonsi: lock a mutex when updating scratch_va for compute shaders
+
+Mary Guillemard (1):
+
+- pan/kmod: Avoid deadlock on VA allocation failure on panthor
+
+MastaG (1):
+
+- gallivm: Call StringMapIterator from llvm:: scope
+
+Mike Blumenkrantz (7):
+
+- st/pbo: fix MESA_COMPUTE_PBO=spec crash on shutdown
+- st/pbo_compute: special case stencil extraction from Z24S8
+- zink: propagate valid buffer range to real buffer when mapping staging
+- zink: track the "real" buffer range from replacement buffers
+- zink: modify some buffer mapping behavior for buffer replacement srcs
+- mesa/st: load state params for feedback draws with allow_st_finalize_nir_twice
+- egl/x11/sw: fix partial image uploads
+
+Patrick Lerda (1):
+
+- st/pbo_compute: fix async->nir memory leak
+
+Paulo Zanoni (2):
+
+- anv: reimplement the anv_fake_nonlocal_memory workaround
+- iris: fix iris_xe_wait_exec_queue_idle() on release builds
+
+Pierre-Eric Pelloux-Prayer (4):
+
+- radeonsi: fix buffer_size in si_compute_shorten_ubyte_buffer
+- Revert "ac, radeonsi: remove has_syncobj, has_fence_to_handle"
+- winsys/radeon: fill lds properties
+- radeonsi: fix crash in si_update_tess_io_layout_state for gfx8 and earlier
+
+Samuel Pitoiset (1):
+
+- radv: disable VK_EXT_sampler_filter_minmax on TAHITI and VERDE
+
+Sviatoslav Peleshko (1):
+
+- mesa: Fix PopAttrib not restoring states that changed on deeper stack level
+
+Tatsuyuki Ishi (1):
+
+- vk_cmd_queue_gen: Exclude CmdDispatchGraphAMDX
+
+Tim Huang (2):
+
+- amd: add GFX v11.5.2 support
+- amd/vpelib: support VPE IP v6.1.3
+
+msizanoen (1):
+
+- egl/wayland: Fix direct scanout with EGL_EXT_present_opaque

docs/relnotes/24.1.5.rst

@@ -0,0 +1,196 @@
+Mesa 24.1.5 Release Notes / 2024-07-31
+======================================
+
+Mesa 24.1.5 is a bug fix release which fixes bugs found since the 24.1.4 release.
+
+Mesa 24.1.5 implements the OpenGL 4.6 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 4.6. OpenGL
+4.6 is **only** available if requested at context creation.
+Compatibility contexts may report a lower version depending on each driver.
+
+Mesa 24.1.5 implements the Vulkan 1.3 API, but the version reported by
+the apiVersion property of the VkPhysicalDeviceProperties struct
+depends on the particular driver being used.
+
+SHA256 checksum
+---------------
+
+::
+
+    02761ffd965dd64b95421ebfca1191d73724aba00f30034009237564f34cf976  mesa-24.1.5.tar.xz
+
+
+New features
+------------
+
+- None
+
+
+Bug fixes
+---------
+
+- piglit: cl-api-build-program crashes
+- i915g: glGenerateMipmap() fails with 2048×2048 textures
+- Build fails without Vulkan
+- [Bisected] Recent compile issue in libnak
+- GPU Hang in Metal Gear Rising Revengeance
+- VK_ERROR_DEVICE_LOST A770 DXVK Fallout 3
+- RustiCL (or maybe not...): radeonsi freezes after 2 hours of simulation, zink works just fine
+
+
+Changes
+-------
+
+Alessandro Astone (1):
+
+- egl/gbm: Walk device list to initialize DRM platform
+
+Dave Airlie (1):
+
+- gallivm/sample: fix sampling indirect from vertex shaders
+
+David Rosca (1):
+
+- Revert "frontends/va: Fix AV1 slice_data_offset with multiple slice data buffers"
+
+Deborah Brouwer (1):
+
+- ci/lava: Detect a6xx gpu recovery failures
+
+Dylan Baker (5):
+
+- compilers/clc: Add missing break statements.
+- mesa: fix memory leak when using shader cache
+- tgsi_to_nir: free disk cache value if the size is wrong
+- crocus: properly free resources on BO allocation failure
+- crocus: check for depth+stencil before creating resource
+
+Emma Anholt (1):
+
+- dri: Fix a pasteo in dri2_from_names()
+
+Eric Engestrom (12):
+
+- docs: add sha256sum for 24.1.4
+- .pick_status.json: Update to 534f0019d714f2331f29d1281a009bb439ebdb17
+- .pick_status.json: Mark 4e9c16b035159c01b10fadeb5930f62efba14058 as denominated
+- .pick_status.json: Update to 0cc23b652401600e57c278d8f6fe6756b13b9f6a
+- .pick_status.json: Update to 0bdc2f180f1fbce16d96d718119b4eed1e5a6731
+- [24.1 only] ci: bump cbindgen-cli to 0.65 to match actual requirements
+- nak: fix meson typo
+- venus: initialize bitset in CreateDescriptorPool()
+- meson: xcb & xcb-randr are needed by the loader whenever x11 is built
+- .pick_status.json: Update to ad90bf0500e07b1bc35f87a406f284c0a7fa7049
+- ci/baremetal: fix logic for retrying boot when it failed
+- .pick_status.json: Update to 235ce3df9baad0d7f0895e58c647914da00d7351
+
+Faith Ekstrand (4):
+
+- nvk: Fix indirect cbuf binds pre-Turing
+- nvk: Don't advertise sparse residency on Maxwell A
+- nvk: Reject sparse images on Maxwell A and earlier
+- nak/spill_values: Don't assume no trivial phis
+
+Francisco Jerez (5):
+
+- intel/brw: Implement null push constant workaround.
+- intel/dev: Add devinfo flag for TBIMR push constant workaround.
+- anv/gfx12.5: Pass non-empty push constant data to PS stage for TBIMR workaround.
+- iris/gfx12.5: Pass non-empty push constant data to PS stage for TBIMR workaround.
+- iris: Pin pixel hashing table BO from iris_batch submission instead of from iris_state.
+
+GKraats (3):
+
+- i915g: fix generation of large mipmaps
+- i915g: fix mipmap-layout for npots
+- i915g: fix max_lod at mipmap-sampling
+
+Georg Lehmann (2):
+
+- aco/optimizer: update temp_rc when converting to uniform bool alu
+- spirv: ignore more function param decorations
+
+Iván Briano (1):
+
+- anv: get scratch surface from the correct pool
+
+Jesse Natalie (1):
+
+- microsoft/clc: Split struct copies before vars_to_ssa in pre-inline optimizations
+
+Jessica Clarke (3):
+
+- Revert "meson: Do not require libdrm for DRI2 on hurd"
+- Revert "meson: fix with_dri2 definition for GNU Hurd"
+- meson: egl: Build egl_dri2 driver even for plain DRI
+
+José Roberto de Souza (2):
+
+- isl: Fix Xe2 protected mask
+- anv: Propagate protected information to blorp_batch_isl_copy_usage()
+
+Karol Herbst (17):
+
+- rusticl/event: make set_status handle error status properly
+- rusticl/queue: do not overwrite event error states
+- rusticl/queue: properly check all dependencies for an error status
+- rusticl/event: properly implement CL_EXEC_STATUS_ERROR_FOR_EVENTS_IN_WAIT_LIST
+- rusticl/queue: properly implement in-order queue error checking
+- rusticl/event: return execution errors when doing a blocking enqueue
+- rusticl/mesa: handle failures with u_upload_data
+- rusticl/mesa: set take_ownership to true in set_constant_buffer_stream
+- nak: allow clippy::not_unsafe_ptr_arg_deref lints
+- clc: force linking of spirvs with mismatching pointer types in signatures
+- spirv: generate info for FunctionParameterAttribute
+- spirv: initial parsing of function parameter decorations
+- spirv: handle function parameters passed by value
+- rusticl: fix clippy lint having bounds defined in multiple places
+- rusticl/program: protect against 0 length in slice::from_raw_parts
+- rusticl/api: protect against 0 length in slice::from_raw_parts
+- rusticl/spirv: protect against 0 length in slice::from_raw_parts
+
+Lionel Landwerlin (4):
+
+- brw: fix uniform rebuild of sources
+- isl: account for protection in base usage checks
+- anv: properly flag image/imageviews for ISL protection
+- anv: propagate protected information for blorp operations
+
+M Henning (1):
+
+- nak: Add minimum bindgen requirement
+
+Matt Turner (4):
+
+- intel/clc: Free parsed_spirv_data
+- intel/clc: Free disk_cache
+- intel/brw: Use REG_CLASS_COUNT
+- intel/elk: Use REG_CLASS_COUNT
+
+Paulo Zanoni (1):
+
+- anv/trtt: fix the process of picking device->trtt.queue
+
+Samuel Pitoiset (1):
+
+- radv: do not expose ImageFloat32AtomicMinMax on GFX11_5
+
+Sushma Venkatesh Reddy (1):
+
+- intel/clflush: Utilize clflushopt in intel_invalidate_range
+
+Vlad Schiller (2):
+
+- pvr: Handle VK_STRUCTURE_TYPE_EXPORT_MEMORY_ALLOCATE_INFO
+- pvr: Handle VK_STRUCTURE_TYPE_IMAGE_FORMAT_LIST_CREATE_INFO
+
+X512 (2):
+
+- egl/haiku: fix double free of BBitmap
+- egl/haiku: fix synchronization problems, add missing header
+
+Yiwei Zhang (1):
+
+- venus: fix a race condition between gem close and gem handle tracking

docs/relnotes/24.1.6.rst

@@ -0,0 +1,184 @@
+Mesa 24.1.6 Release Notes / 2024-08-14
+======================================
+
+Mesa 24.1.6 is a bug fix release which fixes bugs found since the 24.1.5 release.
+
+Mesa 24.1.6 implements the OpenGL 4.6 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 4.6. OpenGL
+4.6 is **only** available if requested at context creation.
+Compatibility contexts may report a lower version depending on each driver.
+
+Mesa 24.1.6 implements the Vulkan 1.3 API, but the version reported by
+the apiVersion property of the VkPhysicalDeviceProperties struct
+depends on the particular driver being used.
+
+SHA256 checksum
+---------------
+
+::
+
+    da94c0908d5662467369b69ed8236da1e1577141a6e7d25171a9bf56383b34e8  mesa-24.1.6.tar.xz
+
+
+New features
+------------
+
+- None
+
+
+Bug fixes
+---------
+
+- u_debug_stack_test.capture_not_overwritten fails on i386
+- turnip: vulkaninfo crashed
+- turnip-freereno Build error (/usr/local/include/xf86drm.h:40:10: fatal error: drm.h: No such file or directory)
+- tu: compilation failing when compiling turnip with only kgsl and X11 enabled
+- samplerCube constructor in fragment shader no longer converting uvec2 bindless texture handles ( segmentation fault , core dumped )
+- anv: gen9.5 flakiness in dEQP-VK.multiview.dynamic_rendering.depth.*
+- interpolateAt precision lowering unhandled by glsl_to_nir()
+- [radeonsi][bisected][regression] glClientWaitSync() quickly times out with INT64_MAX timeout
+- zink/tu: glcts flake on a750
+- [Regression][Vulkan][TGL][Bisected]vkCmdCopyQueryPoolResults failed to write buffer with compute pipeline on Mesa 24.1
+- Worms Revolution: not rendering explosion effects
+- crash on pushbuf_validate nvc0_blit do_blit_framebuffer
+- [radeonsi] Asterix & Obelix XXLRomastered: river misrendered (completely black)
+
+
+Changes
+-------
+
+David Heidelberg (1):
+
+- llvmpipe: Silence "possibly uninitialized value" warning for ssbo_limit (cont)
+
+David Rosca (1):
+
+- radeonsi/vcn: Add decode DPB buffers as CS dependency
+
+Echo J (1):
+
+- util: Fix the integer addition in os_time_get_absolute_timeout()
+
+Eric Engestrom (7):
+
+- docs: add sha256sum for 24.1.5
+- .pick_status.json: Update to c90e2bccf756004e48f9e7e71e555db0d03c1b98
+- .pick_status.json: Update to cc2dbb8ea5329b509d79eedb6c0cbb9a1903b5ad
+- .pick_status.json: Mark 89863a050bea429d9574a307bc28953bb60accaf as denominated
+- .pick_status.json: Mark 93f9afa1e039cbf681adcc6d170aec987d9f0f65 as denominated
+- ci: pass MESA_SPIRV_LOG_LEVEL from job to the test
+- .pick_status.json: Mark 34753cefd828f5db8e9e14cee9c4e45cee5baaaa as denominated
+
+Eric R. Smith (1):
+
+- panfrost: use RGB1 component ordering for R5G6B5 pixel formats
+
+Faith Ekstrand (3):
+
+- nvk: Require color or depth/stencil attachment support for input attachments
+- nvk: Support STORAGE_READ_WITHOUT_FORMAT on buffers
+- zink: Align descriptor buffers to descriptorBufferOffsetAlignment
+
+Georg Lehmann (2):
+
+- nir/lower_int64: replace uadd_sat with ior for find_lsb64 and ufind_msb64
+- aco/gfx10+: set lateKill for sgprs used by wave64 VALU writing a mask
+
+Hans-Kristian Arntzen (2):
+
+- wsi/common: Do not update present mode with MESA_VK_WSI_PRESENT_MODE.
+- wsi/x11: Bump maximum number of outstanding COMPLETE events.
+
+Iván Briano (1):
+
+- intel/rt: fix terminateOnFirstHit handling
+
+Karmjit Mahil (1):
+
+- tu: Set \`TU_ACCESS_CCHE_READ` for transfer ops with read access
+
+Karol Herbst (12):
+
+- Revert "rusticl/queue: gracefully stop the worker thread"
+- nouveau: handle realloc failure inside cli_kref_set
+- mesa: check for enabled extensions for \*UID enums
+- zink: lower 64 bit find_lsb, ufind_msb and bit_count
+- zink: lower 8/16 bit alu ops vk spirv doesn't allow
+- rusticl/kernel: properly respect device thread limits per dimension
+- rusticl/memory: Fix memory unmaps after rework
+- rusticl/image: take pitches into account when allocating memory for maps
+- rusticl/image: properly sync mappings content for 1Dbuffer images
+- rusticl/queue: add clSetCommandQueueProperty
+- util/u_printf: do not double print format string with unused arugments
+- rusticl/memory: fix sampler argument size check
+
+Konstantin Seurer (1):
+
+- aco: print s_delay_alu INSTSKIP>3 correctly
+
+Lionel Landwerlin (6):
+
+- anv: fix check on pipeline mode to track buffer writes
+- vulkan/runtime: allow null/empty debug names
+- anv: reuse object string for RMV token
+- anv: add missing MEDIA_STATE_FLUSH for internal shaders
+- anv/blorp: force CC_VIEWPORT reallocation when programming 3DSTATE_VIEWPORT_STATE_POINTERS_CC
+- brw/rt: fix ray_object_(direction|origin) for closest-hit shaders
+
+Marek Olšák (2):
+
+- nir/opt_algebraic: use fmulz for fpow lowering to fix incorrect rendering
+- radeonsi: fix buffer coherency issues on gfx6-8,12 due to missing PFP->ME sync
+
+Matt Turner (2):
+
+- util: Add ATTRIBUTE_OPTIMIZE(flags)
+- util: Force emission of stack frame in stack unit test
+
+Mike Blumenkrantz (7):
+
+- dri: link with libloader
+- kopper: check swapchain size after possible loader image resize
+- pipe-loader: fix driconf memory management
+- egl: fix zink init
+- dri: fix kms_swrast screen fail
+- egl/wayland: bail on zink init in non-sw mode if extension check fails
+- zink: fix partial update handling
+
+Pavel Ondračka (1):
+
+- r300: bias presubtract fix
+
+Rhys Perry (1):
+
+- docs: update ACO_DEBUG documentation for scheduler options
+
+Rob Clark (2):
+
+- tu: Fix issues with 16k (or larger) page sizes
+- freedreno/drm/virtio: Fix issues with 16k (or larger) page sizes
+
+Sil Vilerino (1):
+
+- Revert "d3d12: Video Encode - Remove PIPE_VIDEO_PROFILE_MPEG4_AVC_BASELINE as not supported" This reverts commit d6bb4ddc638f3ee37fbbe066c631dad80aaeb2d3. Fixes: d6bb4ddc638 ("d3d12: Video Encode - Remove PIPE_VIDEO_PROFILE_MPEG4_AVC_BASELINE as not supported")
+
+Tapani Pälli (1):
+
+- anv: fix a cmd_buffer reference in simple shader
+
+Timothy Arceri (3):
+
+- nir: set disallow_undef_to_nan for legacy ARB asm programs
+- glsl: fix glsl to nir support for lower precision builtins
+- glsl: always copy bindless sampler packing constructors to a temp
+
+Valentine Burley (2):
+
+- vulkan/wsi: Refactor can_present_on_device
+- tu: Always report that we can present on kgsl
+
+WANG Xuerui (1):
+
+- meson: Additionally probe -mtls-dialect=desc for TLSDESC support

docs/relnotes/24.1.7.rst

@@ -0,0 +1,190 @@
+Mesa 24.1.7 Release Notes / 2024-08-29
+======================================
+
+Mesa 24.1.7 is a bug fix release which fixes bugs found since the 24.1.6 release.
+
+Mesa 24.1.7 implements the OpenGL 4.6 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 4.6. OpenGL
+4.6 is **only** available if requested at context creation.
+Compatibility contexts may report a lower version depending on each driver.
+
+Mesa 24.1.7 implements the Vulkan 1.3 API, but the version reported by
+the apiVersion property of the VkPhysicalDeviceProperties struct
+depends on the particular driver being used.
+
+SHA256 checksum
+---------------
+
+::
+
+    SHA256: ecd2e7b1c73998f4103542f39c6b8c968d251637ccc8caa42641aecb86cd2566  mesa-24.1.7.tar.xz
+    SHA512: 890f6387963bbb86a85305ea963cca326a3b3f8f8758ae2082fd62c52df77c2883a359341e91f36614fea59219394ef78f799a334080699a0bb71d984a68cb01  mesa-24.1.7.tar.xz
+
+
+New features
+------------
+
+- None
+
+
+Bug fixes
+---------
+
+- [ANV] Breaking Limit benchmark does not run
+- Possible race condition when generating enums.h
+- [anv] possible regression found on 982106e6768e6f69d71710e314325c813591452d
+- [radeonsi] regression with bindless textures (bisected)
+- Artifacts on Black Myth Wukong
+- DCS world radeon rx 7900 flicker
+- radv/aco: Scheduler bug in ACO somewhere in parallel-gs
+- Once human black artifacts consume the screen
+- xe2: Non-compiler assertion failure replaying cp2077 fossil
+- Mesa doesn't build with the following config options
+
+
+Changes
+-------
+
+Connor Abbott (4):
+
+- tu: Fix off-by-one in UBO CP_LOAD_STATE size
+- ir3, tu: Use a UBO for VS primitive params on a750+
+- tu: Fix passthrough D24S8 attachments
+- tu: Treat partially-bound depth/stencil attachments as passthrough
+
+Dave Airlie (2):
+
+- radv/video: fix reporting video format props for encode.
+- llvmpipe: make sure to duplicate the fd handle before giving out
+
+David Heidelberg (2):
+
+- etnaviv: rename enums_h appropriately
+- etnaviv: build dependency for the etnaviv tests
+
+David Rosca (1):
+
+- frontends/va: Fix leaks with multiple coded buffer segments
+
+Eric Engestrom (8):
+
+- docs: add sha256sum for 24.1.6
+- .pick_status.json: Update to 7b92e11e16ad1e947035c2af40d54076457e81e0
+- .pick_status.json: Update to d7f7aede15d5e169e7f2a3b279adf4b192e6bfe2
+- .pick_status.json: Mark 77f783462a9758b22e211c72a57ec7c36d6c09fd as denominated
+- .pick_status.json: Mark a33ef21106a3b34cb359f0272a103c8b8066fbe6 as denominated
+- .pick_status.json: Update to 25ba90fd888cef431c2098c8afdb0a2bbd34b303
+- .pick_status.json: Mark 8aa52ac6666c835169f490703e557f892648e89a as denominated
+- vc4: Add missing libvc4_neon build dependencies
+
+Faith Ekstrand (3):
+
+- vulkan: Add null descriptor bits to vk_pipeline_robustness_state
+- nouveau/mme: Fix add64 of immediates on Fermi
+- nvk: Disable conditional rendering around CopyQueryPoolResults
+
+Friedrich Vock (1):
+
+- aco: Fix 1D->2D dispatch conversion on <gfx9
+
+GKraats (2):
+
+- i915g: Screen corruption  with ENOBUFS caused by fence register shortage
+- i915g: fix count of buffers at i915_drm_batchbuffer_validate_buffers
+
+Ian Romanick (2):
+
+- anv: Protect against OOB access to anv_state_pool::buckets
+- anv: Larger memory pools for huge shaders
+
+Job Noorman (2):
+
+- ir3: update merge set affinity in shared RA
+- ir3: fix clearing merge sets after shared RA
+
+Karol Herbst (3):
+
+- rusticl/mem: do not check against image base alignment for 1Dbuffer images
+- vtn: ignore volatile on functions for now
+- rusticl/device: limit CL_DEVICE_IMAGE_MAX_BUFFER_SIZE more aggressively
+
+Konstantin (2):
+
+- radv: Handle instruction encodings > 8 bytes when splitting disassembly
+- radv: Handle repeated instructions when splitting disassembly
+
+Lepton Wu (1):
+
+- egl/android: Fix wrong pipe format for RGB_565
+
+Lionel Landwerlin (6):
+
+- vulkan/runtime: fix GetBufferMemoryRequirements2 for maintenance4
+- anv: fix extended buffer flags usages
+- anv: only set 3DSTATE_CLIP::MaximumVPIndex once
+- anv: optimize CLIP::MaximumVPIndex setting
+- anv: always use workaround_address, not workaround_bo
+- nir/divergence: add missing load_constant_base_ptr
+
+Mary Guillemard (1):
+
+- panvk: Fix NULL deref on model name when device isn't supported
+
+Matt Turner (1):
+
+- nir: Skip opt_if_merge when next_if has block ending in a jump
+
+Mike Blumenkrantz (5):
+
+- glx/dri2: strdup driver name
+- zink: bail on choose_pdev immediately if no devices are available
+- st/pbo: reject vs/fs pbo ops if rowstride < width
+- zink: don't skip cbuf store ops if resolve is set
+- tc: set resolve on renderpass info if blit terminates the renderpass
+
+Nanley Chery (2):
+
+- iris: Invalidate state cache for some depth fast clears
+- intel/isl: Fix packing of SINT formats
+
+Pavel Ondračka (1):
+
+- r300: fix RGB10_A2 CONSTANT_COLOR blending
+
+Rhys Perry (8):
+
+- aco: form hard clauses in VS prologs
+- aco: split selection_control_remove into rarely_taken and never_taken
+- aco: only remove branch jumping over SMEM/barrier if it's never taken
+- aco: ignore exec and literals when mitigating VALUMaskWriteHazard
+- aco: also consider VALU reads for VALUMaskWriteHazard
+- aco: don't consider sa_sdst=0 before SALU write to fix VALUMaskWriteHazard
+- aco: check SALU writing lanemask later for VALUMaskWriteHazard
+- aco: preserve bitsets after a lane mask is written
+
+Samuel Pitoiset (1):
+
+- aco: fix bogus assert in RT prolog on GFX11+
+
+Sviatoslav Peleshko (3):
+
+- brw,elk: Fix opening flags on dumping shader binaries
+- anv: Release correct BO in anv_cmd_buffer_set_ray_query_buffer
+- anv: Add full subgroups WA for the shaders with barriers in Breaking Limit
+
+Tapani Pälli (1):
+
+- gbm: depend on libdrm indepedent of dri2 setting
+
+Timothy Arceri (4):
+
+- nir: create validate_tex_src_texture_deref() helper
+- nir: add nir_tex_src_{sampler,texture}_deref_intrinsic
+- glsl: make use of new tex src deref intrinsic
+- nir/glsl: set deref cast mode during function inlining
+
+Valentine Burley (1):
+
+- llvmpipe: Only use udmabuf with libdrm

include/GL/internal/dri_interface.h

@@ -968,7 +968,7 @@ struct __DRIswrastExtensionRec {
    int (*queryBufferAge)(__DRIdrawable *drawable);
 
    /**
-    * createNewScreen() with the driver extensions passed in and implicit load flag.
+    * createNewScreen() with the driver extensions passed in and driver_name_is_inferred load flag.
     *
     * \since version 6
     */
@@ -976,7 +976,7 @@ struct __DRIswrastExtensionRec {
                                     const __DRIextension **loader_extensions,
                                     const __DRIextension **driver_extensions,
                                     const __DRIconfig ***driver_configs,
-                                    bool implicit,
+                                    bool driver_name_is_inferred,
                                     void *loaderPrivate);
 
 };
@@ -995,7 +995,7 @@ typedef __DRIscreen *
                              const __DRIextension **extensions,
                              const __DRIextension **driver_extensions,
                              const __DRIconfig ***driver_configs,
-                             bool implicit,
+                             bool driver_name_is_inferred,
                              void *loaderPrivate);
 
 typedef __DRIdrawable *
@@ -1253,7 +1253,7 @@ struct __DRIdri2ExtensionRec {
    __DRIcreateNewScreen2Func            createNewScreen2;
 
    /**
-    * createNewScreen with the driver's extension list passed in and implicit load flag.
+    * createNewScreen with the driver's extension list passed in and driver_name_is_inferred load flag.
     *
     * \since version 5
     */

include/GL/internal/mesa_interface.h

@@ -60,7 +60,7 @@ struct __DRImesaCoreExtensionRec {
    __DRIcreateContextAttribsFunc createContext;
 
    /* driver function for finishing initialization inside createNewScreen(). */
-   const __DRIconfig **(*initScreen)(struct dri_screen *screen, bool implicit);
+   const __DRIconfig **(*initScreen)(struct dri_screen *screen, bool driver_name_is_inferred);
 
    int (*queryCompatibleRenderOnlyDeviceFd)(int kms_only_fd);
 

include/kopper_interface.h

@@ -97,6 +97,7 @@ struct kopper_loader_info {
    struct kopper_vk_surface_create_storage bos;
    int has_alpha;
    int initial_swap_interval;
+   bool present_opaque;
 };
 
 #define __DRI_KOPPER_LOADER "DRI_KopperLoader"

meson.build

@@ -335,7 +335,6 @@ if with_aco_tests and not with_amd_vk
 endif
 
 with_microsoft_clc = get_option('microsoft-clc').enabled()
-with_clc = with_microsoft_clc or with_intel_clc or with_gallium_asahi
 with_spirv_to_dxil = get_option('spirv-to-dxil')
 
 if host_machine.system() == 'darwin'
@@ -506,22 +505,28 @@ if not have_mtls_dialect
   if meson.is_cross_build() and not meson.can_run_host_binaries()
     warning('cannot auto-detect -mtls-dialect when cross-compiling, using compiler default')
   else
-    # -fpic to force dynamic tls, otherwise TLS relaxation defeats check
-    gnu2_test = cc.run('int __thread x; int main() { return x; }',
-                       args: ['-mtls-dialect=gnu2', '-fpic'],
-                       name: '-mtls-dialect=gnu2')
-    if gnu2_test.returncode() == 0 and (
-          # check for lld 13 bug: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5665
-          host_machine.cpu_family() != 'x86_64' or
-          # get_linker_id misses LDFLAGS=-fuse-ld=lld: https://github.com/mesonbuild/meson/issues/6377
-          #cc.get_linker_id() != 'ld.lld' or
-          cc.links('''int __thread x; int y; int main() { __asm__(
-                "leaq x@TLSDESC(%rip), %rax\n"
-                "movq y@GOTPCREL(%rip), %rdx\n"
-                "call *x@TLSCALL(%rax)\n"); }''', name: 'split TLSDESC')
-          )
-      c_cpp_args += '-mtls-dialect=gnu2'
-    endif
+    # The way to specify the TLSDESC dialect is architecture-specific.
+    # We probe both because there is not a fallback guaranteed to work for all
+    # future architectures.
+    foreach tlsdesc_arg : ['-mtls-dialect=gnu2', '-mtls-dialect=desc']
+      # -fpic to force dynamic tls, otherwise TLS relaxation defeats check
+      tlsdesc_test = cc.run('int __thread x; int main() { return x; }',
+                            args: [tlsdesc_arg, '-fpic'],
+                            name: tlsdesc_arg)
+      if tlsdesc_test.returncode() == 0 and (
+            # check for lld 13 bug: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5665
+            host_machine.cpu_family() != 'x86_64' or
+            # get_linker_id misses LDFLAGS=-fuse-ld=lld: https://github.com/mesonbuild/meson/issues/6377
+            #cc.get_linker_id() != 'ld.lld' or
+            cc.links('''int __thread x; int y; int main() { __asm__(
+                  "leaq x@TLSDESC(%rip), %rax\n"
+                  "movq y@GOTPCREL(%rip), %rdx\n"
+                  "call *x@TLSCALL(%rax)\n"); }''', name: 'split TLSDESC')
+            )
+        c_cpp_args += tlsdesc_arg
+        break
+      endif
+    endforeach
   endif
 endif
 
@@ -563,9 +568,7 @@ if with_vulkan_icd_dir == ''
   with_vulkan_icd_dir = join_paths(get_option('datadir'), 'vulkan/icd.d')
 endif
 
-# GNU/Hurd includes egl_dri2, without drm.
-with_dri2 = (with_dri or with_any_vk) and (with_dri_platform == 'drm' or with_dri_platform == 'apple' or
-  host_machine.system() == 'gnu')
+with_dri2 = (with_dri or with_any_vk) and (with_dri_platform == 'drm' or with_dri_platform == 'apple')
 with_dri3 = get_option('dri3').disable_auto_if(not (system_has_kms_drm and with_dri2)).allowed()
 
 if with_any_vk and (with_platform_x11 and not with_dri3)
@@ -842,10 +845,11 @@ if with_gallium_rusticl
 
   add_languages('rust', required: true)
   rustc = meson.get_compiler('rust')
-
-  with_clc = true
 endif
 
+with_clover_spirv = with_gallium_opencl and get_option('opencl-spirv')
+with_clc = with_microsoft_clc or with_intel_clc or with_gallium_asahi or with_gallium_rusticl or with_clover_spirv
+
 dep_clc = null_dep
 if with_gallium_opencl or with_clc
   dep_clc = dependency('libclc')
@@ -1000,6 +1004,7 @@ endforeach
 _attributes = [
   'const', 'flatten', 'malloc', 'pure', 'unused', 'warn_unused_result',
   'weak', 'format', 'packed', 'returns_nonnull', 'alias', 'noreturn',
+  'optimize',
 ]
 foreach a : cc.get_supported_function_attributes(_attributes)
   pre_args += '-DHAVE_FUNC_ATTRIBUTE_@0@'.format(a.to_upper())
@@ -1696,8 +1701,7 @@ with_gallium_drisw_kms = false
 if system_has_kms_drm
   dep_libdrm = dependency(
     'libdrm', version : '>=' + _drm_ver,
-    # GNU/Hurd includes egl_dri2, without drm.
-    required : (with_dri2 and host_machine.system() != 'gnu') or with_dri3
+    required : with_dri2 or with_dri3
   )
 else
   # We should prevent libdrm from being available when the target doesn't have it to avoid transitive
@@ -1828,8 +1832,7 @@ endif
 pre_args += '-DLLVM_AVAILABLE=' + (with_llvm ? '1' : '0')
 pre_args += '-DDRAW_LLVM_AVAILABLE=' + (with_llvm and draw_with_llvm ? '1' : '0')
 
-with_opencl_spirv = (_opencl != 'disabled' and get_option('opencl-spirv')) or with_clc
-if with_opencl_spirv
+if with_clover_spirv or with_clc
   chosen_llvm_version_array = dep_llvm.version().split('.')
   chosen_llvm_version_major = chosen_llvm_version_array[0].to_int()
   chosen_llvm_version_minor = chosen_llvm_version_array[1].to_int()
@@ -2080,11 +2083,11 @@ with_dri3_modifiers = false
 with_dri3_explicit_sync = false
 with_xcb_keysyms = false
 if with_platform_x11
+  dep_xcb = dependency('xcb')
+  dep_xcb_xrandr = dependency('xcb-randr')
   if with_glx == 'xlib'
     dep_x11 = dependency('x11')
     dep_xext = dependency('xext')
-    dep_xcb = dependency('xcb')
-    dep_xcb_xrandr = dependency('xcb-randr')
   elif with_glx == 'dri'
     dep_x11 = dependency('x11')
     dep_xext = dependency('xext')

src/amd/addrlib/src/amdgpu_asic_addr.h

@@ -100,6 +100,8 @@
 #define AMDGPU_NAVI33_RANGE     0x10, 0x20 //# 16 <= x < 32
 #define AMDGPU_GFX1150_RANGE    0x01, 0x40 //# 1 <= x < 64
 #define AMDGPU_GFX1151_RANGE    0xC0, 0xFF //# 192 <= x < 255
+#define AMDGPU_GFX1152_RANGE    0x40, 0x50 //# 64 <= x < 80
+
 #define AMDGPU_GFX1103_R1_RANGE 0x01, 0x80 //# 1 <= x < 128
 #define AMDGPU_GFX1103_R2_RANGE 0x80, 0xC0 //# 128 <= x < 192
 
@@ -172,6 +174,8 @@
 #define ASICREV_IS_NAVI33_P(r)         ASICREV_IS(r, NAVI33)
 #define ASICREV_IS_GFX1150(r)          ASICREV_IS(r, GFX1150)
 #define ASICREV_IS_GFX1151(r)          ASICREV_IS(r, GFX1151)
+#define ASICREV_IS_GFX1152(r)          ASICREV_IS(r, GFX1152)
+
 #define ASICREV_IS_GFX1103_R1(r)       ASICREV_IS(r, GFX1103_R1)
 #define ASICREV_IS_GFX1103_R2(r)       ASICREV_IS(r, GFX1103_R2)
 

src/amd/ci/gitlab-ci-inc.yml

@@ -69,7 +69,7 @@
     - !reference [.vulkan-manual-rules, rules]
     - changes:
         *amd_common_file_list
-      when: on_success
+      when: manual
     - changes:
         *radv_file_list
       when: manual

src/amd/ci/radv-navi21-llvm-fails.txt

@@ -1,8 +1,4 @@
 
-dEQP-VK.spirv_assembly.instruction.compute.float_controls.fp32.input_args.tanh_denorm_flush_to_zero,Fail
-dEQP-VK.spirv_assembly.instruction.graphics.float_controls.fp32.input_args.tanh_denorm_flush_to_zero_frag,Fail
-dEQP-VK.spirv_assembly.instruction.graphics.float_controls.fp32.input_args.tanh_denorm_flush_to_zero_vert,Fail
-
 dEQP-VK.pipeline.monolithic.multisample.storage_image.64x64_1.r32g32b32a32_sfloat.samples_8,Fail
 dEQP-VK.pipeline.monolithic.multisample.storage_image.64x64_1.r8g8b8a8_unorm.samples_8,Fail
 dEQP-VK.pipeline.monolithic.multisample.storage_image.64x64_4.r32g32b32a32_sfloat.samples_8,Fail
@@ -11,18 +7,3 @@ dEQP-VK.pipeline.monolithic.multisample.storage_image.79x31_1.r32g32b32a32_sfloa
 dEQP-VK.pipeline.monolithic.multisample.storage_image.79x31_1.r8g8b8a8_unorm.samples_8,Fail
 dEQP-VK.pipeline.monolithic.multisample.storage_image.79x31_4.r32g32b32a32_sfloat.samples_8,Fail
 dEQP-VK.pipeline.monolithic.multisample.storage_image.79x31_4.r8g8b8a8_unorm.samples_8,Fail
-
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_double_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_f16vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_i16vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_i64vec2_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_i64vec3_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_i64vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_i8vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_int64_t_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_u16vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_u64vec3_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_u64vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_u8vec4_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_uint64_t_fragment,Fail
-dEQP-VK.subgroups.vote.frag_helper.subgroupallequal_vec4_fragment,Fail

src/amd/ci/radv-tahiti-aco-flakes.txt

@@ -1,6 +1 @@
-dEQP-VK.pipeline.fast_linked_library.sampler.*.view_type.*
-dEQP-VK.pipeline.monolithic.sampler.*.view_type.*
-dEQP-VK.pipeline.pipeline_library.sampler.*.view_type.*
-dEQP-VK.pipeline.shader_object_.*.sampler.*.view_type.*
-
 dEQP-VK.pipeline.shader_object_unlinked_binary.interface_matching.decoration_mismatch.out_noperspective_in_none_loose_variable_vert_tesc_tese_geom_out_frag_in

src/amd/common/ac_gpu_info.c

@@ -355,6 +355,14 @@ static intptr_t readlink(const char *path, char *buf, size_t bufsiz)
 
 #define CIK_TILE_MODE_COLOR_2D 14
 
+static bool has_syncobj(int fd)
+{
+   uint64_t value;
+   if (drmGetCap(fd, DRM_CAP_SYNCOBJ, &value))
+      return false;
+   return value ? true : false;
+}
+
 static bool has_timeline_syncobj(int fd)
 {
    uint64_t value;
@@ -899,6 +907,7 @@ bool ac_query_gpu_info(int fd, void *dev_p, struct radeon_info *info,
       case FAMILY_GFX1150:
          identify_chip(GFX1150);
          identify_chip(GFX1151);
+         identify_chip(GFX1152);
          break;
       }
 
@@ -1054,7 +1063,9 @@ bool ac_query_gpu_info(int fd, void *dev_p, struct radeon_info *info,
    info->memory_freq_mhz_effective *= ac_memory_ops_per_clock(info->vram_type);
 
    info->has_userptr = true;
+   info->has_syncobj = has_syncobj(fd);
    info->has_timeline_syncobj = has_timeline_syncobj(fd);
+   info->has_fence_to_handle = info->has_syncobj;
    info->has_local_buffers = true;
    info->has_bo_metadata = true;
    info->has_eqaa_surface_allocator = info->gfx_level < GFX11;
@@ -1320,6 +1331,12 @@ bool ac_query_gpu_info(int fd, void *dev_p, struct radeon_info *info,
 
    info->has_export_conflict_bug = info->gfx_level == GFX11;
 
+   /* When LLVM is fixed to handle multiparts shaders, this value will depend
+    * on the known good versions of LLVM. Until then, enable the equivalent WA
+    * in the nir -> llvm backend.
+    */
+   info->needs_llvm_wait_wa = info->gfx_level == GFX11;
+
    /* Convert the SDMA version in the current GPU to an enum. */
    info->sdma_ip_version =
       (enum sdma_version)SDMA_VERSION_VALUE(info->ip[AMD_IP_SDMA].ver_major,
@@ -1552,6 +1569,8 @@ bool ac_query_gpu_info(int fd, void *dev_p, struct radeon_info *info,
    info->max_scratch_waves = MAX2(32 * info->min_good_cu_per_sa * info->max_sa_per_se * info->num_se,
                                   max_waves_per_tg);
    info->num_rb = util_bitcount64(info->enabled_rb_mask);
+   info->has_scratch_base_registers = info->gfx_level >= GFX11 ||
+                                      (!info->has_graphics && info->family >= CHIP_GFX940);
    info->max_gflops = (info->gfx_level >= GFX11 ? 256 : 128) * info->num_cu * info->max_gpu_freq_mhz / 1000;
    info->memory_bandwidth_gbps = DIV_ROUND_UP(info->memory_freq_mhz_effective * info->memory_bus_width / 8, 1000);
    info->has_pcie_bandwidth_info = info->drm_minor >= 51;
@@ -1902,7 +1921,9 @@ void ac_print_gpu_info(const struct radeon_info *info, FILE *f)
    fprintf(f, "Kernel & winsys capabilities:\n");
    fprintf(f, "    drm = %i.%i.%i\n", info->drm_major, info->drm_minor, info->drm_patchlevel);
    fprintf(f, "    has_userptr = %i\n", info->has_userptr);
+   fprintf(f, "    has_syncobj = %u\n", info->has_syncobj);
    fprintf(f, "    has_timeline_syncobj = %u\n", info->has_timeline_syncobj);
+   fprintf(f, "    has_fence_to_handle = %u\n", info->has_fence_to_handle);
    fprintf(f, "    has_local_buffers = %u\n", info->has_local_buffers);
    fprintf(f, "    has_bo_metadata = %u\n", info->has_bo_metadata);
    fprintf(f, "    has_eqaa_surface_allocator = %u\n", info->has_eqaa_surface_allocator);
@@ -1959,6 +1980,7 @@ void ac_print_gpu_info(const struct radeon_info *info, FILE *f)
    fprintf(f, "    wave64_vgpr_alloc_granularity = %i\n", info->wave64_vgpr_alloc_granularity);
    fprintf(f, "    max_scratch_waves = %i\n", info->max_scratch_waves);
    fprintf(f, "    attribute_ring_size_per_se = %u\n", info->attribute_ring_size_per_se);
+   fprintf(f, "    has_scratch_base_registers = %i\n", info->has_scratch_base_registers);
 
    fprintf(f, "Render backend info:\n");
    fprintf(f, "    pa_sc_tile_steering_override = 0x%x\n", info->pa_sc_tile_steering_override);

src/amd/common/ac_gpu_info.h

@@ -120,6 +120,9 @@ struct radeon_info {
    bool sdma_supports_compression; /* Whether SDMA supports DCC and HTILE. */
    bool has_set_context_pairs_packed;
    bool has_set_sh_pairs_packed;
+   bool needs_llvm_wait_wa; /* True if the chip needs to workarounds based on s_waitcnt_deptr but
+                             * the LLVM version doesn't work with multiparts shaders.
+                             */
 
    /* conformant_trunc_coord is equal to TA_CNTL2.TRUNCATE_COORD_MODE, which exists since gfx11.
     *
@@ -209,7 +212,9 @@ struct radeon_info {
    uint32_t max_submitted_ibs[AMD_NUM_IP_TYPES];
    bool is_amdgpu;
    bool has_userptr;
+   bool has_syncobj;
    bool has_timeline_syncobj;
+   bool has_fence_to_handle;
    bool has_local_buffers;
    bool has_bo_metadata;
    bool has_eqaa_surface_allocator;
@@ -259,6 +264,7 @@ struct radeon_info {
    uint32_t wave64_vgpr_alloc_granularity;
    uint32_t max_scratch_waves;
    uint32_t attribute_ring_size_per_se;
+   bool has_scratch_base_registers;
 
    /* Render backends (color + depth blocks). */
    uint32_t r300_num_gb_pipes;

src/amd/common/ac_nir_lower_tess_io_to_mem.c

@@ -353,14 +353,18 @@ hs_output_lds_map_io_location(nir_shader *shader,
 {
    if (!per_vertex) {
       const uint64_t tf_mask = tcs_lds_tf_out_mask(shader, st);
-      if (BITFIELD64_BIT(loc) & TESS_LVL_MASK)
+      if (loc == VARYING_SLOT_TESS_LEVEL_INNER || loc == VARYING_SLOT_TESS_LEVEL_OUTER) {
+         assert(tf_mask & BITFIELD64_BIT(loc));
          return util_bitcount64(tf_mask & BITFIELD64_MASK(loc));
+      }
 
       const uint32_t patch_out_mask = tcs_lds_per_patch_out_mask(shader);
+      assert(patch_out_mask & BITFIELD_BIT(loc - VARYING_SLOT_PATCH0));
       return util_bitcount64(tf_mask) +
              util_bitcount(patch_out_mask & BITFIELD_MASK(loc - VARYING_SLOT_PATCH0));
    } else {
       const uint64_t per_vertex_mask = tcs_lds_per_vtx_out_mask(shader);
+      assert(per_vertex_mask & BITFIELD64_BIT(loc));
       return util_bitcount64(per_vertex_mask & BITFIELD64_MASK(loc));
    }
 }