[Mesa-dev] Mesa 17.2.4 release candidate

Andres Gomez agomez at igalia.com
Fri Oct 27 17:43:57 UTC 2017

Hello list,

The candidate for the Mesa 17.2.4 is now available. Currently we have:
 - 21 queued
 - 10 nominated (outstanding)
 - and 6 rejected patches

In the current queue we have:

In Mesa Core we have included a change to prevent KOTOR from breaking
when in combination with the ATI fragment shader extension. 
Additionally, NIR has also received a correction.

Mesa's state tracker has gotten a patch to avoid leaks in certain
situations like when resizing a window.

Intel drivers have received several fixes.  The compiler has gotten a
couple, while anv also received one.  i965 got a patch to avoid VA-API, 
Beignet and other contexts in the system to break when in combination
with previous versions of Mesa 17.2.x.

AMD's compiler has received a couple of fixes while radv has also
received another couple, including one to avoid a hang due to an
overflow on huge textures.

Broadcom's vc4 has corrected a problem when compiling with Android's

clover has also seen fixed a problem compiling after a specific clang

Vulkan's WSI has gotten plugged a memory leak for X11.

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 17.2.4 next Sunday (29th of October), 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 1eb4cbc934afd72ea45521eeac763e43564d972b
Author: Dave Airlie <airlied at redhat.com>

    radv/image: bump all the offset to uint64_t.
    (cherry picked from commit 35c66f3e40177a97d74e614e2a324a03e2149c73)

commit cbc081b8711a519dd9aa371debbf68aa0add4076
Author: Kenneth Graunke <kenneth at whitecape.org>

    i965: Revert absolute mode for constant buffer pointers.

    (cherry picked from commit 013d33122028f2492da90a03ae4bc1dab84c3ee9)

commit ce725baa7c90420b46e16087a4f201a7e62b23e5
Author: Matthew Nicholls <mnicholls at feralinteractive.com>

    ac/nir: generate correct instruction for atomic min/max on unsigned images

    (cherry picked from commit 27a0b24bf238342031e0709584e4d71ab228f1ec)

commit 74f19032341748ad257a0c2527624b661e2d2e6f
Author: Jason Ekstrand <jason.ekstrand at intel.com>

    anv/pipeline: Call nir_lower_system_valaues after brw_preprocess_nir

    (cherry picked from commit 279f8fb69cf68d05287e14f60cf67fc025643bc4)


Mesa stable queue

Nominated (10)

Jason Ekstrand (4):
      spirv: Claim support for the simple memory model
      i965/blorp: Use blorp_to_isl_format for src_isl_format in blit_miptrees
      i965/blorp: Use more temporary isl_format variables
      i965/miptree: Take an isl_format in render_aux_usage

Kenneth Graunke (1):
      mesa: Accept GL_BACK in get_fb0_attachment with ARB_ES3_1_compatibility.

Leo Liu (1):
      radeon/video: add gfx9 offsets when rejoin the video surface

Marek Olšák (2):
      radeonsi: add a workaround for weird s_buffer_load_dword behavior on SI
      ac/surface/gfx9: don't allow DCC for the smallest mipmap levels

Nicolai Hähnle (1):
      amd/common/gfx9: workaround DCC corruption more conservatively

Tapani Pälli (1):
      i965: unref push_const_bo in intelDestroyContext

Queued (21)

Andres Gomez (7):
      cherry-ignore: configure.ac: rework llvm detection and handling
      cherry-ignore: glsl: fix derived cs variables
      cherry-ignore: added 17.3 nominations.
      cherry-ignore: radv: Don't use vgpr indexing for outputs on GFX9.
      cherry-ignore: radv: Disallow indirect outputs for GS on GFX9 as well.
      cherry-ignore: mesa/bufferobj: don't double negate the range
      cherry-ignore: broadcom/vc5: Propagate vc4 aliasing fix to vc5.

Bas Nieuwenhuizen (1):
      ac/nir: Fix nir_texop_lod on GFX for 1D arrays.

Dave Airlie (1):
      radv/image: bump all the offset to uint64_t.

Henri Verbeet (1):
      vulkan/wsi: Free the event in x11_manage_fifo_queues().

Jan Vesely (1):
      clover: Fix compilation after clang r315871

Jason Ekstrand (4):
      nir/intrinsics: Set the correct num_indices for load_output
      intel/fs: Handle flag read/write aliasing in needs_src_copy
      anv/pipeline: Call nir_lower_system_valaues after brw_preprocess_nir
      intel/eu: Use EXECUTE_1 for JMPI

Kenneth Graunke (1):
      i965: Revert absolute mode for constant buffer pointers.

Marek Olšák (1):
      Revert "mesa: fix texture updates for ATI_fragment_shader"

Matthew Nicholls (1):
      ac/nir: generate correct instruction for atomic min/max on unsigned images

Michel Dänzer (1):
      st/mesa: Initialize textures array in st_framebuffer_validate
Squashed with
      st/osmesa: include u_inlines.h for pipe_resource_reference

Samuel Pitoiset (1):
      radv: add the draw count buffer to the list of buffers

Stefan Schake (1):
      broadcom/vc4: Fix aliasing issue

Rejected (6)

Emil Velikov (5):
      configure.ac: factor out detection for old and buggy llvm
      configure.ac: rework llvm libs handling for 3.9+
      Travis: add binutils 2.26 for a few more LLVM 3.9 builds
      configure.ac: add llvm_add_optional_component helper
      configure.ac: add missing LLVM components for OpenCL

Reason: Commits are too invasive for 17.2.

Ilia Mirkin (1):
      glsl: fix derived cs variables

Reason: Commit is too big for stable at this point.

More information about the mesa-dev mailing list