1 Mesa 20.0.5 Release Notes / 2020-04-22
2 ======================================
4 Mesa 20.0.5 is a bug fix release which fixes bugs found since the 20.0.4
7 Mesa 20.0.5 implements the OpenGL 4.6 API, but the version reported by
8 glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
9 glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being
10 used. Some drivers don't support all the features required in OpenGL
11 4.6. OpenGL 4.6 is **only** available if requested at context creation.
12 Compatibility contexts may report a lower version depending on each
15 Mesa 20.0.5 implements the Vulkan 1.2 API, but the version reported by
16 the apiVersion property of the VkPhysicalDeviceProperties struct depends
17 on the particular driver being used.
24 2c56a82a28cc924e40ea49752abdf1d701c9952481f53cbc7a080271597f572e mesa-20.0.5.tar.xz
32 - nir: nir_lower_returns can't handle nested loops
33 - Graphic artifacts with Mesa 20.0.4 on intel HD 510 GPU
34 - Mesa 20 regression makes Lightsprint demos crash
35 - Build Fails with Clang Shared Library
36 - dri_common.h:58:8: error: unknown type name '__GLXDRIdrawable'
37 - Graphical glitches on Intel Graphics when Xorg started on Iris driver
38 - SIGSEGV src/compiler/glsl/ast_function.cpp:53
39 - manywin aborts with "i965: Failed to submit batchbuffer: Invalid
41 - manywin aborts with "i965: Failed to submit batchbuffer: Invalid
43 - manywin aborts with "i965: Failed to submit batchbuffer: Invalid
45 - manywin aborts with "i965: Failed to submit batchbuffer: Invalid
47 - v3d: transform feedback issue
49 dEQP-VK.binding_model.descriptorset_random.sets4.noarray.ubolimitlow.sbolimitlow.imglimitlow.noiub.comp.noia.0
52 dEQP-VK.pipeline.timestamp.misc_tests.reset_query_before_copy
53 - https://gitlab.freedesktop.org/mesa/mesa/-/issues/2727
54 - enable storageBuffer16BitAccess feature in radv for SI and CIK
55 - Weston drm-backend.so seems to fail with Mesa master and
56 LIBGL_ALWAYS_SOFTWARE=1
57 - vaapi bob deinterlacer produces wrong output height on AMD
62 Arcady Goldmints-Orlov (1):
64 - nir: Lower returns correctly inside nested loops
66 Bas Nieuwenhuizen (3):
68 - radv: Store 64-bit availability bools if requested.
69 - radv: Consider maximum sample distances for entire grid.
70 - radv: Use correct buffer count with variable descriptor set sizes.
74 - util/sparse_array: don't stomp head's counter on pop operations
78 - EGL: Add eglSetDamageRegionKHR to GLVND dispatch list
82 - st/mesa: Update shader info of ffvp/ARB_vp after translation to NIR
86 - draw: free the NIR IR.
87 - llvmpipe/nir: free the nir shader
91 - .pick_status.json: Update to 089e1fb287eb9b70c191091128ed5ba7edd2960a
92 - .pick_status.json: Update to 65e2eaa4d3a7095ac438fafb09d1e36a4210966e
93 - .pick_status.json: Update to 28d36d26c2212276e1238fad8f0b12caab97fee8
94 - .pick_status.json: Update to acf7e73be54c7f1cc52fcc9be38a9df26849200e
95 - .pick_status.json: Update to 13ce637f1b28381e72470763ff5e39dd3c562476
96 - .pick_status.json: Update to c3c1f4d6bcc210408f8b180727d269838b38193b
100 - glx: set the loader_logger early and for everyone
101 - egl/drm: reinstate (kms\_)swrast support
102 - Revert "egl/dri2: Don't dlclose() the driver on
103 dri2_load_driver_common failure"
104 - glx: omit loader_loader() for macOS
108 - ci: Remove LLVM from ARM test drivers.
112 - docs/relnotes: add sha256sum for 20.0.4
116 - nir: fix wrong assignment to buffer in xfb_varyings_info
120 - nv50: don't try to upload MSAA settings for BUFFER textures
124 - anv/image: Use align_u64 for image offsets
125 - nir/load_store_vectorize: Fix shared atomic info
126 - spirv: Handle OOB vector extract operations
127 - intel: Add \_const versions of prog_data cast helpers
128 - anv: Report correct SLM size
130 Jose Maria Casanova Crespo (1):
132 - v3d: Primitive Counts Feedback needs an extra 32-bit padding.
134 Juan A. Suarez Romero (2):
136 - intel/compiler: store the FS inputs in WM prog data
137 - anv/pipeline: allow more than 16 FS inputs
141 - clover: fix build with single library clang build
142 - Revert "nvc0: fix line width on GM20x+"
144 Lionel Landwerlin (7):
146 - iris: properly free resources on BO allocation failure
147 - iris: share buffer managers accross screens
148 - iris: make resources take a ref on the screen object
149 - i965: store DRM fd on intel_screen
150 - i965: share buffer managers across screens
151 - iris: drop cache coherent cpu mapping for external BO
152 - util/sparse_free_list: manipulate node pointers using atomic
157 - st/mesa: fix a crash due to passing a draw vertex shader into the
160 Mathias Fröhlich (1):
162 - i965: Move down genX_upload_sbe in profiles.
166 - meson: Specify the maximum required libdrm in dri.pc
170 - gitlab-ci/lava: fix handling of lava tags
171 - gitlab-ci: add FILES_HOST_URL and move FILES_HOST_NAME into jobs
172 - gitlab-ci: re-enable mali400/450 and t820 jobs
176 - aco: fix 1D textureGrad() on GFX9
180 - nir: fix definition of imadsh_mix16 for vectors
184 - ci: Split out radv build-testing on arm64
188 - ac/nir: split 8-bit load/store to global memory on GFX6
189 - ac/nir: split 8-bit SSBO stores on GFX6
190 - radv/llvm: enable 8-bit storage features on GFX6-GFX7
191 - ac/nir: split 16-bit load/store to global memory on GFX6
192 - ac/nir: split 16-bit SSBO stores on GFX6
193 - radv/llvm: enable 16-bit storage features on GFX6-GFX7
194 - radv: do not abort with unknown/unimplemented descriptor types
195 - radv/llvm: fix exporting the viewport index if the fragment shader
197 - aco: fix exporting the viewport index if the fragment shader needs it
201 - mesa/st: unbind shader state before deleting it
202 - mesa/st: release variants for active programs before unref
203 - glsl: stop processing function parameters if error happened
204 - mesa/st: initialize all winsys_handle fields for memory objects
208 - gallium/auxiliary/vl: fix bob compute shaders for deint yuv
212 - radeonsi: don't lower constant arrays to uniforms in GLSL IR
216 - meson: Link Gallium Nine with ld_args_build_id
220 - gitlab-ci: Place files from the Mesa repo into the build tarball
221 - gitlab-ci: Serve files for LAVA via separate service
225 - swr/rasterizer: Use private functions for min/max to avoid namespace
227 - swr: Remove Byte Order Mark.
231 - scons/windows: Support build with LLVM 10.