xref: /aosp_15_r20/external/mesa3d/docs/relnotes/24.0.5.rst (revision 6104692788411f58d303aa86923a9ff6ecaded22)
1Mesa 24.0.5 Release Notes / 2024-04-10
2======================================
3
4Mesa 24.0.5 is a bug fix release which fixes bugs found since the 24.0.4 release.
5
6Mesa 24.0.5 implements the OpenGL 4.6 API, but the version reported by
7glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
8glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
9Some drivers don't support all the features required in OpenGL 4.6. OpenGL
104.6 is **only** available if requested at context creation.
11Compatibility contexts may report a lower version depending on each driver.
12
13Mesa 24.0.5 implements the Vulkan 1.3 API, but the version reported by
14the apiVersion property of the VkPhysicalDeviceProperties struct
15depends on the particular driver being used.
16
17SHA256 checksum
18---------------
19
20::
21
22    38cc245ca8faa3c69da6d2687f8906377001f63365348a62cc6f7fafb1e8c018  mesa-24.0.5.tar.xz
23
24
25New features
26------------
27
28- None
29
30
31Bug fixes
32---------
33
34- anv: vkd3d-proton test_stress_suballocation failure
35- d3d12: Zwift renders with bad textures/lighting
36- NVK: Misrendering with Civilization 6
37- radv: RDR2 might need zerovram
38- Issues rendering gtk4 window decorations on v3d on Fedora-40/mesa-24.0
39- clc: Failure when linking with llvm+clang 18.1 (-Dshared-llvm=disabled)
40- LLVM-18 build issue
41
42
43Changes
44-------
45
46Axel Davy (5):
47
48- frontend/nine: Fix ff ps key
49- frontend/nine: Fix programmable vs check
50- frontend/nine: Fix missing light flag check
51- frontend/nine: Fix destruction race
52- frontend/nine: Reset should EndScene
53
54Connor Abbott (2):
55
56- freedreno/a7xx: Add CP_CCHE_INVALIDATE
57- tu: Implement CCHE invalidation
58
59Dave Airlie (1):
60
61- mesa: reorder st context teardown
62
63David Heidelberg (7):
64
65- r300: add missing licence to the r300_public.h
66- r300: add missing copyright header
67- docs: we support EGL 1.5 for a long time
68- ci/amd: drop old PIGLIT_REPLAY_DESCRIPTION_FILE surpassed by PIGLIT_TRACES_FILE
69- r600: add license header to r600_formats.h
70- r600: add license info to the r600_opcodes.h
71- r600: add license information to the sfn_shader_gs.h
72
73David Stern (1):
74
75- vulkan/wsi/x11: Explicitly discard errors from xcb_present_pixmap.
76
77Eric Engestrom (5):
78
79- docs: add sha256sum for 24.0.4
80- .pick_status.json: Update to 3d68dd78d07b30cefe90d76af681075f4ed6b33d
81- .pick_status.json: Update to fcb568a5d5a52db75fa2f6d04579bb404ca7f597
82- .pick_status.json: Update to 078fe5454e97d073feb18bcdcf7ed1874e8b4835
83- .pick_status.json: Update to 2c1cb65949933a05eedb2eacc15cd893ecaef8aa
84
85Eric R. Smith (2):
86
87- panfrost: mark indirect compute buffer as read
88- gallium: handle copy_image of depth textures
89
90Faith Ekstrand (2):
91
92- nvk: Add a _pad field to nvk_cbuf
93- nvk: Add a _pad field to nvk_fs_key
94
95Georg Lehmann (2):
96
97- aco: don't combine mul+add_clamp to mad_clamp
98- aco/ra: use SDWA for 16bit instructions when the second byte is blocked
99
100Iago Toral Quiroga (2):
101
102- v3d: implement fix for GFXH-1602
103- broadcom/compiler: fix workaround for GFXH-1602
104
105Ian Romanick (3):
106
107- intel/brw: Clear write_accumulator flag when changing the destination
108- intel/brw: Use enums for DPAS source regioning
109- nir: intel/brw: Change the order of sources for nir_dpas_intel
110
111Jesse Natalie (1):
112
113- glsl: Use a stable attr sort for VS in / FS out
114
115Jordan Justen (1):
116
117- intel/dev: Add 0x56be and 0x56bf DG2 PCI IDs
118
119José Roberto de Souza (4):
120
121- anv: Fix calculation of syncs required in Xe KMD
122- iris: Wait for drm_xe_exec_queue to be idle before destroying it
123- anv: Create protected engine context when i915 supports vm control
124- intel: Enable Xe KMD support by default
125
126Juston Li (1):
127
128- Revert "zink: store last pipeline directly for zink_gfx_program::last_pipeline"
129
130Karol Herbst (1):
131
132- meson: fix link failure with llvm-18
133
134Kenneth Graunke (2):
135
136- intel/brw: Fix generate_mov_indirect to check has_64bit_int not float
137- intel/brw: Fix lower_regioning for BROADCAST, MOV_INDIRECT on Q types
138
139Konstantin Seurer (1):
140
141- nir/serialize: Encode data for temporaries
142
143Lionel Landwerlin (7):
144
145- anv: fix protected memory allocations
146- anv: disable protected content around surface state copies
147- anv: disable generated draws in protected command buffers
148- anv: update protection fault property
149- anv: add missing data flush out of L3 for transform feedback writes
150- anv: mark descriptors & pipeline dirty after blorp compute
151- isl: set NullPageCoherencyEnable for depth/stencil sparse surfaces
152
153Lucas Stach (2):
154
155- etnaviv: fix depth writes without testing
156- etnaviv: rs: take src dimensions into account when increasing height alignment
157
158Mike Blumenkrantz (12):
159
160- zink: only check that CUBE_COMPATIBLE for images doesn't subtract flags
161- zink: don't use set_foreach_remove with dmabuf_exports
162- zink: make descriptor pool creation more robust
163- zink: fix shaderdb pipeline compile
164- zink: don't clobber indirect array reads with missing components
165- zink: fix add_derefs case for compact arrays
166- llvmpipe: fix DRAW_USE_LLVM=0
167- glsl: handle xfb resources for spirv before running varying opts
168- mesa: clamp binary pointer in ShaderBinary if length==0
169- glsl: set PSIZ bit in outputs_written when injecting a 1.0 psiz write
170- nir/lower_clamp_color_outputs: fix use with lowered io
171- nir/texcoord_replace: fix scalarized io handling
172
173Nikita Popov (1):
174
175- Pass no-verify-fixpoint option to instcombine in LLVM 18
176
177Patrick Lerda (1):
178
179- r300: fix constants_remap_table memory leak related to the dummy shader path
180
181Paul Gofman (3):
182
183- glsl: allow out arrays in #110 with allow_glsl_120_subset_in_110
184- driconf: add a workaround for Joe Danger 2
185- driconf: add a workaround for Joe Danger
186
187Paulo Zanoni (2):
188
189- anv/xe: don't leak xe_syncs during trtt submission
190- anv, iris: add missing CS_STALL bit for GPGPU texture invalidation
191
192Samuel Pitoiset (3):
193
194- radv: fix conditional rendering with mesh+task and multiview (again)
195- radv: enable radv_zero_vram for Red Dead Redemption 2
196- radv: make sure the heap budget is less than or equal to the heap size
197
198Tapani Pälli (1):
199
200- anv: disable fcv optimization on >= gfx125
201
202Yonggang Luo (1):
203
204- util: Fixes futex_wait on win32
205
206Zack Rusin (1):
207
208- svga: Fix instanced draw detection
209
210Zan Dobersek (1):
211
212- tu: fix memory leaks in tu_shader
213