xref: /aosp_15_r20/external/mesa3d/docs/relnotes/20.3.3.rst (revision 6104692788411f58d303aa86923a9ff6ecaded22)

Mesa 20.3.3 Release Notes / 2021-01-13
======================================

Mesa 20.3.3 is a bug fix release which fixes bugs found since the 20.3.2 release.

Mesa 20.3.3 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 20.3.3 implements the Vulkan 1.2 API, but the version reported by
the apiVersion property of the VkPhysicalDeviceProperties struct
depends on the particular driver being used.

SHA256 checksum
---------------

::

    f74e212d4838e982a10c203ffa998817d1855c5cf448ae87b58f96edea61d156  mesa-20.3.3.tar.xz


New features
------------

- None


Bug fixes
---------

- \[RADV/ACO/SIENNA_CICHLID\] Into the game Shadow of the Tomb Raider the flickering artifacts are present on brushes.
- DOOM crashes on startup with OpenGL on RX 6800
- Space Engineers rendering regression after 5f79e4e6 which triggers incorrect optimizations from 053be9f0
- star conflict crashes on iris, but loads fine on i965, on HD 5500


Changes
-------

Andrii Simiklit (1):

- st/mesa: don't affect original st_CompressedTexSubImage parameters

Bas Nieuwenhuizen (3):

- radv/winsys: Fix inequality for sparse buffer remapping.
- radv/winsys: Fix offset in range merging.
- radv: Add Android module info to linker script.

Danylo Piliaiev (2):

- nir: account for point-coord origin when lowering it
- nir: fix missing nir_lower_pntc_ytransform.c in the makefile

Dylan Baker (6):

- docs: Add sha256 sum for 20.3.2
- .pick_status.json: Update to 9ef2c44ce682b7981bd7a68c65f338f1e33bb001
- .pick_status.json: Update to b2d000513e4a9b06d3b073913741dae47a00526e
- .pick_status.json: Update to 96ceca33c1dd69a1feed13b0e19bfc38e6f7d979
- .pick_status.json: Update to d9c8422c4168c7433f68b97065283ebf26631d56
- .pick_status.json: Mark 52b6adfbfb51a3b1a54964cbcb2adb0299d1252b as backported

Eric Anholt (4):

- gallium/draw: Fix intermittent failure to bind new geometry shaders.
- gallium/ntt: Fix emitting UBO declarations.
- gallium/ntt: Fix leak of the per-instr liveness information.
- mesa/st: Free the NIR builtins TGSI tokens after passing to the driver.

Erik Faye-Lund (2):

- gallium/util: do not perform n^2 stencil blits
- zink: dot leak dummy_buffer

Georg Lehmann (2):

- vulkan/device-select: fix vkGetInstanceProcAddr self-resolving
- vulkan/overlay: fix vkGetInstanceProcAddr self-resolving

Ian Romanick (3):

- spir-v: Mark floating point comparisons exact
- Revert "nir: Replace an odd comparison involving fmin of -b2f"
- nir/algebraic: Fix broken NaN and -0.0 behavior

Icecream95 (1):

- pan/mdg: Fix promoted uniform moves with 64-bit types

James Park (1):

- util: Disable memstream for Apple builds

Marek Olšák (2):

- st/mesa: don't do glCopyPixels via blit if depth bounds test is enabled
- util: add AMD CPU family enums and enable L3 cache pinning on Zen3

Marek Vasut (1):

- etnaviv: Fix rework ZSA into a derived state

Pavel Asyutchenko (1):

- vulkan/overay: fix violation of VUID-VkDeviceCreateInfo-pNext-00373

Pierre-Eric Pelloux-Prayer (1):

- drirc: radeonsi workaround for CS:GO

Rhys Perry (6):

- aco: fix incorrect address calculation for load_barycentric_at_sample
- ac/nir: use llvm.readcyclecounter for LLVM9+
- nir/load_store_vectorize: don't ignore subgroup memory barriers
- aco: fix unreachable() for uniform 8/16-bit nir_op_mov from VGPR
- radv: add RADV_DEBUG=invariantgeom
- radv: set invariantgeom for Shadow of the Tomb Raider

Samuel Pitoiset (4):

- nir: fix determining if an addition might overflow for phi sources
- radv: disable TC-compat HTILE in GENERAL for Detroit: Become Human
- aco: fix creating the dest vector when 16-bit vertex fetches are splitted
- radv/llvm,aco: always split typed vertex buffer loads on GFX6 and GFX10+

nia (1):

- util: Avoid pthread_setaffinity_np on NetBSD