Mesa 20.1.5 Release Notes / 2020-08-05

Mesa 20.1.5 is a bug fix release which fixes bugs found since the 20.1.4 release.

Mesa 20.1.5 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 20.1.5 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

fac1861e6e0bf1aec893f8d86dbfb9d8a0f426ff06b05256df10e3ad7e02c69b  mesa-20.1.5.tar.xz

New features

  • None

Bug fixes

  • panfrost: Register allocation fails for Firefox WebRender shaders
  • VRAM leak with vuilkan external memory + opengl memory objects
  • Possible array out of bounds in brw_vec4_nir.cpp
  • [tgl][bisected][regression][iris] failure on dEQP-EGL.functional.wide_color.pbuffer_8888_colorspace_default
  • Multiply defined symbols compiling with gcc@10.1.0
  • shrinking descriptor pool on intel+vulkan
  • dEQP-VK.renderpass2.dedicated_allocation.attachment.1.12 fails on NAVI14

Changes

Alyssa Rosenzweig (2):

  • pan/mdg: Mask spills from texture write
  • pan/mdg: Test for SSA before chasing addresses

Bas Nieuwenhuizen (3):

  • radv: Fix uninitialized variable in renderpass.
  • radv: Fix host->host signalling with legacy timeline semaphores.
  • mesa/st: Actually free the driver part of memory objects on destruction.

Daniel Schürmann (3):

  • aco: fix scratch loads which cross element_size boundaries
  • aco: don’t split store data if it was already split into more elements
  • aco: prevent infinite recursion in RA for subdword variables

Daryl W. Grunau (1):

  • prevent multiply defined symbols

Eric Engestrom (6):

  • docs/relnotes: add sha256 sums to 20.1.4
  • .pick_status.json: Update to caa98246a0e180a96f3fcdcd3bfcbef0b136bc11
  • .pick_status.json: Mark bd75e9923302a3d389469b7b233968576a46f4de as denominated
  • .pick_status.json: Mark 1b3be07b5faf867f698668080b060a270c5f795e as denominated
  • bin/gen_release_notes: automatically commit release notes
  • docs: add release notes for 20.1.5

Francisco Jerez (1):

  • intel/ir/gen12+: Work around FS performance regressions due to SIMD32 discard divergence.

Frank Binns (1):

  • egl/dri2: only take a dri2_dpy reference when binding a new context/surfaces

Jason Ekstrand (2):

  • intel/eu: Use non-coherent mode (BTI=253) for stateless A64 messages
  • nir/deref: Don’t try to compare derefs containing casts

Kenneth Graunke (1):

  • iris: Delete shader variants when deleting the API-facing shader

Kristian Høgsberg (1):

  • anv: Pass device to setup_gralloc0_usage for error reporting

Lionel Landwerlin (1):

  • anv: fix descriptor set free

Marcin Ślusarz (6):

  • iris: propagate error from gen_perf_begin_query to glBeginPerfQueryINTEL
  • i965: propagate error from gen_perf_begin_query to glBeginPerfQueryINTEL
  • util: fix possible fd leaks in os_socket_listen_abstract
  • util: fix possible buffer overflow in util_get_process_exec_path
  • mesa: fix out of bounds access in glGetFramebufferParameterivEXT
  • intel/vec4: fix out of bounds read

Marek Olšák (4):

  • ac: update register and packet definitions for preemption
  • radeonsi: add missing initialization of registers
  • radeonsi/gfx10: set the correct value for OFFCHIP_BUFFERING
  • radeonsi: disable SDMA on gfx9

Mauro Rossi (1):

  • radv: fix build on Android 7 (v2)

Yevhenii Kolesnikov (1):

  • nine: fix incorrect calculation of layer count for 3D textures