[Mesa-dev] Mesa 13.0.6 release candidate

Emil Velikov emil.l.velikov at gmail.com
Wed Mar 15 18:44:20 UTC 2017

Hello list,

The candidate for the Mesa 13.0.6 is now available. Currently we have:
 - 104 queued
 - 1 nominated (outstanding)
 - and 3 rejected patch(es)

Note: due to a rework in the stable nomination scheme we have more patches
went unnoticed previously. Hence we have a larger amount of fixes than usual.

Some of the changes include:

On the GLX/EGL front we have crash fixes in GLX/DRI3, ResetNotificationStrategy
matches the shared context for EGL, and eglQuerySurface returns correct geometry
when running under X11/DRI3.

There's a number of crash fixes affecting all Gallium drivers. An old regression
fix for r300 on BE hardware been fixed. The radeonsi driver has fixes for
Tessellation shaders on Carrizo and Stoney hardware

While on the nouveau side, compute shader have been improved on some
nvc0 devices.

For the Intel drivers (both GL and Vulkan) we have a diverse bunch of patches -
from CTS fixes for Sandy Bridge, to improved swizzle clears and improved
handling of GPUs without (Last Level Cache) LLC.

The RADV Vulkan driver has also seen over a dozen of fixes and Improvements.

Take a look at section "Mesa stable queue" for more information.

Testing reports/general approval

Any testing reports (or general approval of the state of the branch) will be
greatly appreciated.

The plan is to have 13.0.6 this Friday (17th of March), around or shortly
after 18:00 GMT.

If you have any questions or suggestions - be that about the current patch
queue or otherwise, please go ahead.

Trivial merge conflicts

commit ea6ae74cb0bb6c398eb7412bd50ea0db01116ee2
Author: Marek Olšák <marek.olsak at amd.com>

    radeonsi: fix UNSIGNED_BYTE index buffer fallback with non-zero start (v2)

    (cherry picked from commit a264fee6245856340fab9024e1a428626e966335)

commit f9673f8213ad5790a5219c5562f3b828ff793d10
Author: Nicolai Hähnle <nicolai.haehnle at amd.com>

    winsys/amdgpu: reduce max_alloc_size based on GTT limits

    (cherry picked from commit 550125e1e73e2441989da11495057a20dd9dad44)

commit 64537aa6b481bf853f272c442956bbf93bea4ad9
Author: Nicolai Hähnle <nicolai.haehnle at amd.com>

    radeonsi: fix UINT/SINT clamping for 10-bit formats on <= CIK

    (cherry picked from commit 066a117be77fdc2b29c8eafabb4e2c2fa902a18e)

commit cea0fa44ae65ec6b8b724022008355afec2a2a4c
Author: Emil Velikov <emil.velikov at collabora.com>

    i965: move brw_define.h ifndef guard to the top

    (cherry picked from commit 077078ce77e8653725def01ed291eb486989a9ad)

commit 1d143f0018d29c7f3867a8a7f3249aeeb931c1d4
Author: Dave Airlie <airlied at redhat.com>

    radv: setup llvm target data layout

    (cherry picked from commit b8ee70384adc3286d18febba7a92047118cc0f0f)


Mesa stable queue

Nominated (1)

Jason Ekstrand (1):
      60d1aac nir/intrinsics: Make load_barycentric_input take a
2-component coor

Queued (104)

Alex Smith (2):
      radv: Emit pending flushes before executing a secondary command buffer
      radv: Flush before copying with PKT3_WRITE_DATA in CmdUpdateBuffer

Bartosz Tomczyk (1):
      glsl: fix heap-buffer-overflow

Bas Nieuwenhuizen (8):
      radv: Pass CMASK alignment to application.
      radv: Pass DCC alignment to application.
      radv: Never try to create more than max_sets descriptor sets.
      radv: Reset emitted compute pipeline when calling secondary cmd buffer.
      radv: Only use PKT3_OCCLUSION_QUERY when it doesn't hang.
      radv: Use correct size for availability flag.
      radv: Disable HTILE for textures with multiple layers/levels.
      radv: Emit cache flushes before CP DMA.

Ben Crocker (3):
      gallivm: Improve debug output (V2)
      gallivm: Override getHostCPUName() "generic" w/ "pwr8" (v4)
      gallivm: Reenable PPC VSX (v3)

Brendan King (1):
      egl/dri3: implement query surface hook

Bruce Cherniak (1):
      swr: Prune empty nodes in CalculateProcessorTopology.

Connor Abbott (1):
      anv: fix Get*MemoryRequirements for !LLC

Dave Airlie (13):
      radv: program a default point size.
      radv: handle transfer_write as a dst flag.
      radv/ac: handle nir irem opcode.
      radv/ac: implement txs for buffer textures.
      radv/ac: correctly size shared memory usage.
      radv/ac: avoid the fmask path when doing txs.
      radv: pass FMASK alignment to application
      tgsi: fix memory leak in tgsi sanity check
      radv: fix depth format in blit2d.
      radv: fix txs for sampler buffers
      radv: drop Z24 support.
      radv: disable mip point pre clamping.
      radv: setup llvm target data layout

Emil Velikov (4):
      Revert "get-pick-list.sh: Require explicit "13.0" for nominating
stable patches"
      cherry-ignore: don't pick nir_op_pack_double optimisation fix
      i965: move brw_define.h ifndef guard to the top
      cherry-ignore: add ANV fast clears related fixes

Fredrik Höglund (2):
      radv: fix the dynamic buffer index in vkCmdBindDescriptorSets
      radv/ac: fix multiple descriptor sets with dynamic buffers

George Kyriazis (1):
      swr: Align query results allocation

Grazvydas Ignotas (3):
      r300g: only allow byteswapped formats on big endian
      gallium/u_queue: fix a crash with atexit handlers
      gallium/u_queue: set num_threads correctly if not all threads start

Gregory Hainaut (1):
      glapi: fix typo in count_scale

Ian Romanick (1):
      mesa: Don't advertise GL_OES_read_format in core profile

Ilia Mirkin (8):
      nvc0: increase number of ubo binding points
      nvc0/ir: fix robustness guarantees for constbuf loads on kepler+ compute
      nvc0/ir: fix ubo max clamp, reset file index
      gm107/ir: fix address offset bitfield for ATOMS
      nvc0: set the render condition in the compute object
      st/mesa: don't pass compare mode for stencil-sampled textures
      nvc0: take extra pushbuf space into account for pushbuf_space calls
      nvc0: increase alignment to 256 for texture buffers on fermi

Jacob Lifshay (1):
      vulkan/wsi: Improve the DRI3 error message

Jason Ekstrand (11):
      i965: Use a better guardband calculation.
      intel/blorp: Swizzle clear colors on the CPU
      i965/fs: Remove the inline pack_double_2x32 optimization
      anv: Add an invalidate_range helper
      anv/query: clflush the bo map on non-LLC platforms
      genxml: Make MI_STORE_DATA_IMM more consistent
      anv/query: Perform CmdResetQueryPool on the GPU
      blorp/exec: Use uint32_t for copying varying data
      intel/blorp: Explicitly flush all allocated state
      anv: Accurately advertise dynamic descriptor limits
      anv: Properly handle destroying NULL devices and instances

Jonas Pfeil (1):
      ralloc: Make sure ralloc() allocations match malloc()'s alignment.
Squashed with
      ralloc: don't leave out the alignment factor

Jose Maria Casanova Crespo (1):
      glsl: non-last member unsized array on SSBO must fail
compilation on GLSL ES 3.1

Kenneth Graunke (7):
      i965: Fix fast depth clears for surfaces with a dimension of 16384.
      i965: Use a UW source type for CS_OPCODE_CS_TERMINATE.
      i965: Fix check for negative pitch in can_do_fast_copy_blit().
      i965: Support the force_glsl_version driconf option.
      i965: Combine the Gen6 SF and Clip viewport atoms.
      mesa: Do (TCS && !TES) draw time validation in ES as well.
      egl: Ensure ResetNotificationStrategy matches for shared contexts.

Lionel Landwerlin (3):
      spirv: don't assert with location decorations on non i/o variables
      anv: wsi: report presentation error per image request
      i965/fs: fix uninitialized memory access

Marc Di Luzio (1):
      glsl: correct compute shader checks for memoryBarrier functions

Marek Olšák (10):
      st/mesa: destroy pipe_context before destroying st_context (v2)
      radeonsi: don't invoke DCC decompression in update_all_texture_descriptors
      radeonsi: fix UNSIGNED_BYTE index buffer fallback with non-zero start (v2)
      gallium/util: remove unused u_index_modify helpers
      gallium/u_index_modify: don't add PIPE_TRANSFER_UNSYNCHRONIZED
      gallium/u_queue: fix random crashes when the app calls exit()
      st/mesa: reset sample_mask, min_sample, and render_condition for PBO ops
      st/mesa: set blend state for PBO readbacks
      radeonsi: fix broken tessellation on Carrizo and Stoney
      radeonsi: mark all bound shader buffer ranges as initialized

Nicolai Hähnle (12):
      mesa/main: fix meta caller of _mesa_ClampColor
      radeonsi: fix texture gather on stencil textures
      glsl: split DIV_TO_MUL_RCP into single- and double-precision flags
      glx/dri3: handle NULL pointers in loader-to-DRI3 drawable conversion
      glx/dri3: guard in_current_context against a disappeared drawable
      glx: guard swap-interval functions against destroyed drawables
      dri/common: clear the loaderPrivate pointer in driDestroyDrawable
      winsys/amdgpu: reduce max_alloc_size based on GTT limits
      radeonsi: handle MultiDrawIndirect in si_get_draw_start_count
      radeonsi: fix UINT/SINT clamping for 10-bit formats on <= CIK
      st/glsl_to_tgsi: avoid iterating past the head of the instruction list
      st/mesa: inform the driver of framebuffer changes before compute

Samuel Iglesias Gonsálvez (6):
      glsl: fix heap-use-after-free in ast_declarator_list::hir()
      i965/fs: mark last DF uniform array element as 64 bit live one
      i965/fs: detect different bit size accesses to uniforms to push
them in proper locations
      i965/fs: fix indirect load DF uniforms on BSW/BXT
      i965/fs: fix source type when emitting MOV_INDIRECT to read ICP handles
      i965/fs: emit MOV_INDIRECT with the source with the right register type

Samuel Pitoiset (1):
      winsys/amdgpu: avoid potential segfault in amdgpu_bo_map()

Rejected (3)

Jason Ekstrand (3):
      a4393bd i965/fs: Fix the inline nir_op_pack_double optimization

Reason: Causes a regression since the optimization is 'broken' and removed
with a later commit.

      42b10b1 anv/blorp/clear_subpass: Only set surface clear color
for fast clears
      6b644e5 anv: Stall before fast-clear operations

Reason: Patches address ANV fast clear which landed after the branchpoint.

More information about the mesa-dev mailing list