[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

[Git][xorg-team/lib/mesa][debian-experimental] 197 commits: docs: Add sha256 sum for 22.2.3

Title: GitLab

Timo Aaltonen pushed to branch debian-experimental at X Strike Force / lib / mesa

Commits:

  • ce33e68f
    by Dylan Baker at 2022-11-07T10:16:36-08:00 
    docs: Add sha256 sum for 22.2.3
  • 6bba416e
    by Dylan Baker at 2022-11-07T10:32:30-08:00 
    .pick_status.json: Update to fd8ec189e575d5220d8b4c30647fa6de57928e07
  • 7dee5d7d
    by Alyssa Rosenzweig at 2022-11-07T10:32:47-08:00 
    panfrost: Don't copy resources if replaced

If a synchronized transfer_map is going to overwrite an entire resource,
there's no need to memcpy in the original contents ahead-of-time. This
memcpy is particularly bad for large buffers where it's copying WC->WC,
although that could be mitigated with threaded_context's cpu_storage in
the future if needed.

Prevents a performance regression in glmark2's buffer scenes from the
next patch, hence the Cc.

Cc: mesa-stable
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19361>
(cherry picked from commit 0b26a9f773956fc00a77b0d4a7aafee5795ce935)
  • 9a92cd91
    by Alyssa Rosenzweig at 2022-11-07T10:32:48-08:00 
    panfrost: Replace resource shadowing flush

The entire point of resource shadowing is to avoid unnecessary flushing.
Flushing readers after shadowing is counterproductive. A refresher on
how resource shadowing is supposed to work:

First, we determine if it's beneficial to shadow resources. If so, we
create a new backing buffer object. We flush the current writer of the
resource, if there is one, so the current contents become known to the
CPU. If we are not discarding the original resource, we then copy the
existing contents of the buffer to the new shadow buffer on the CPU.
Finally, we swap the resource's backing buffer for our shadow. Any batch
that reads the resource will continue to read the old copy of the
resource, and any future draw calls will see the new copy with the
change implemented.

Where did we go wrong?

In 988d5aae74d ("panfrost: Flush resources when shadowing"), we started
flushing all readers. We didn't actually need to flush, we just needed
to avoid dangling references on the batches reading the old copy of the
resource. But that's easily enough avoided: just remove the references.
The batches still hold a reference to the underlying BO, which will be
freed at the right time regardless.

Originally motivated by glmark2 -bbuffer:update-method=subdata, which
has some pathological access paterns.

Firefox is a lot faster anecdotally (now scrolling at 60fps in firefox).

But what actually motivated this is an apitrace from Duckstation's GLES
renderer. With this patch, the in-game portion is improved 3fps to 21fps.

Closes: #4028
Fixes: 988d5aae74d ("panfrost: Flush resources when shadowing")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19361>
(cherry picked from commit 2d8f28df731638e1270b0ae273c7bfd2b29b7993)
  • 842759cf
    by Karol Herbst at 2022-11-07T10:32:48-08:00 
    iris: invalidate sysvals if grid dimension changes

Cc: mesa-stable
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18581>
(cherry picked from commit 9ccdd86b9040b265893712f4125d30552e0fe1cd)
  • 7ade4ab5
    by Bas Nieuwenhuizen at 2022-11-07T10:32:49-08:00 
    radv: Fix compute scratch buffer emission.

Copied wrong from radeonsi. The registers following the scratch
buffer address are the shader rsrc1/rsrc2. Not the user SGPR0
containing the ring resource word 1.

Fixes: 278e533ec9c ("radv: update scratch buffer registers on GFX11")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19488>
(cherry picked from commit b8865ad04643cd7e8faa8a863d81fedbc1e29231)
  • 117ac5ca
    by Bas Nieuwenhuizen at 2022-11-07T10:32:50-08:00 
    radv: Make the compute scratch waves per SE as well.

Fixes: 278e533ec9c ("radv: update scratch buffer registers on GFX11")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19488>
(cherry picked from commit d876ddc920f1aaf48219f2060de6cc99c003bc5d)
  • b2e9eaf7
    by Eric Engestrom at 2022-11-07T10:39:29-08:00 
    vk/runtime: drop incorrect UNUSED annotation

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit c6c5949ff70a47c47795fe9161a7514173b5be24)
  • 79c19062
    by Bas Nieuwenhuizen at 2022-11-08T14:59:33+00:00 
    radv: Do not try to run on GFX11 with mesa 22.2.

It isn't going to work.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19577>
  • 6c2e9191
    by Dylan Baker at 2022-11-08T12:58:54-08:00 
    .pick_status.json: Update to def56b531c86f529bc32d1834ccb479457717db7
  • 7f7586f8
    by Iago Toral Quiroga at 2022-11-08T13:00:39-08:00 
    v3dv: do a better job at cleaning up the device on init failure

These leaks on device creation failure have been there before, but
were only exposed as CTS failures after the recent event refactoring.

Partially fixes:
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit b78fd50e90c700466213e4b3b40ce10547b1e14e)
  • 7d5c727f
    by Iago Toral Quiroga at 2022-11-08T13:00:40-08:00 
    v3dv: handle allocation failure during pipeline initialization

Fixes (with disk cache disabled):
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 1f5966397aa208eb574394a4d0a592b66b776b44)
  • 9855c364
    by Iago Toral Quiroga at 2022-11-08T13:00:41-08:00 
    v3dv: do better cleanup on failure during pipeline cache operation

Fixes (with disk cache enabled):
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 5e97150e21a75920c3521717f70f9f92702729b4)
  • 0971b56a
    by Samuel Pitoiset at 2022-11-08T13:02:05-08:00 
    radv: invalidate L2 instead of only writeback L2 when using DCC stores

It seems INV_L2 is the right thing to do, especially for RDNA2 chips
with non-coherent RBs (NAVI22 is one of these). This fixes DCC
corruption.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6476
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7507
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19516>
(cherry picked from commit 06adf6ad117e92a832c5073326bd83b638b3ca37)
  • 90058545
    by Alex Brachet at 2022-11-08T13:02:06-08:00 
    nir: Fix qsort comparator function

`pred` is a pointer, for sufficiently large numbers these
being cast to int were both > 0 regardless of the order
of `data1` and `data2`.

Fixes: 523a28d3fe0d ("nir: add an instruction set API")

Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19539>
(cherry picked from commit c987a727a76eda1feada07e2eca6a5597dbddd93)
  • dde0dab8
    by Marek Olšák at 2022-11-08T15:11:32-08:00 
    radeonsi/gfx11: fix compute scratch buffer - WAVES is always per SE

Fixes: ba02ed91a60 - ac/gfx11: fix the scratch buffer

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19477>
(cherry picked from commit bdfacd0a24e023515fb7b7fae4a279cff0fbac4e)
  • ef4460c2
    by Francisco Jerez at 2022-11-08T15:11:34-08:00 
    nir/lower_int64: Fix float16 to int64 conversions.

Currently float16 to int64 conversions don't work correctly, because
the "div" variable has an infinite value, since 2^32 isn't
representable as a 16-bit float, which causes the result of of rem(x,
div) to be NaN for all inputs, leading to an incorrect result.  Since
no values of magnitude greater than 2^32 are representable as a
float16 we don't actually need to do the fdiv/frem operations, the
conversion is equivalent to f2u32 with the result padded to 64 bits.

Rework:
 * Jordan: Handle f16 in if/else rather than conditional

Fixes: 936c58c8fcc ("nir: Extend nir_lower_int64() to support i2f/f2i lowering")
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19391>
(cherry picked from commit e14f85366ebbc57f45d2561e0d3f0804f8adb549)
  • 74e3535b
    by Marek Vasut at 2022-11-08T15:11:35-08:00 
    etnaviv: Use old set of state registers for PE configuration on GC880

While the GC880 is HALTI0, it still uses the old set of state registers
for PE pipe configuration. This is another specialty of the GC880, readd
the missing handling for this GPU otherwise e.g. Qt5 cube example suffers
from rendering corruption with both eglfs and wayland backends.

Fixes: 7c46a488362 ("etnaviv: use new PE pipe address states on >= HALTI0")
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Marek Vasut <marex@denx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19562>
(cherry picked from commit 20984aab0f6717f5fbf79fc21c2c5f442472b605)
  • 3f1fe4b6
    by Mauro Rossi at 2022-11-08T15:11:36-08:00 
    Android.mk: Fix gnu++14 related build failures

This patch filters-out '-std=gnu++14' from the cflags obtained
from AOSP/KATI dummy target output to avoid the following building errors:

FAILED: src/gallium/drivers/r600/45f68e3@@r600@sta/sfn_sfn_assembler.cpp.o
...
clang++ ... -std=c++17 ... -std=gnu++14
...
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.cpp:27:
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.h:32:
In file included from ../src/gallium/drivers/r600/sfn/sfn_shader.h:31:
../src/gallium/drivers/r600/sfn/sfn_instr.h:369:56: error: no template named 'is_base_of_v' in namespace 'std'; did you mean 'is_base_of'?
template <typename T, typename = std::enable_if_t<std::is_base_of_v<Instr, T>>>
                                                  ~~~~~^~~~~~~~~~~~
                                                       is_base_of
/home/utente/pie-x86_kernel/external/libcxx/include/type_traits:1412:29: note: 'is_base_of' declared here
struct _LIBCPP_TEMPLATE_VIS is_base_of
                            ^
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.cpp:27:
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.h:32:
In file included from ../src/gallium/drivers/r600/sfn/sfn_shader.h:31:
../src/gallium/drivers/r600/sfn/sfn_instr.h:369:51: error: template argument for non-type template parameter must be an _expression_
template <typename T, typename = std::enable_if_t<std::is_base_of_v<Instr, T>>>
                                                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/utente/pie-x86_kernel/external/libcxx/include/type_traits:439:16: note: template parameter is declared here
template <bool _Bp, class _Tp = void> using enable_if_t = typename enable_if<_Bp, _Tp>::type;
               ^
2 errors generated.

Cc: "22.2" "22.3" mesa-stable
Reviewed-by: Roman Stratiienko <r.stratiienko@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19563>
(cherry picked from commit fd8ec189e575d5220d8b4c30647fa6de57928e07)
  • b7a7e78a
    by Ian Romanick at 2022-11-08T15:11:37-08:00 
    intel/fs: Fix constant propagation into 32x16 integer multiplication

Don't copy propagate the constant in situations like

    mov(8)          g8<1>D          0x7fffffffD
    mul(8)          g16<1>D         g8<8,8,1>D      g15<16,8,2>W

On platforms that only have a 32x16 multiplier, this will result in
lowering the multiply to

    mul(8)          g15<1>D         g14<8,8,1>D     0xffffUW
    mul(8)          g16<1>D         g14<8,8,1>D     0x7fffUW
    add(8)          g15.1<2>UW      g15.1<16,8,2>UW g16<16,8,2>UW

On Gfx8 and Gfx9, which have the full 32x32 multiplier, it results in

    mul(8)          g16<1>D         g15<16,8,2>W    0x7fffffffD

Volume 2a of the Skylake PRM says:

    When multiplying a DW and any lower precision integer, the
    DW operand must on src0.

See also https://gitlab.freedesktop.org/mesa/crucible/-/merge_requests/104.

Previous to INTEL_shader_integer_functions2 (in Vulkan or OpenGL), I
don't think it would be possible to create a situation where this could
occur.  I discovered this via some optimizations that can determine that
the non-constant source must be able to fit in 16-bits.  The case listed
above came from piglit's "ext_transform_feedback-order arrays points"
with those optimizations in place.

No shader-db or fossil-db changes on any Intel platform.

Fixes: de6c0f84879 ("intel/fs: Implement support for NIR opcodes for INTEL_shader_integer_functions2")
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17718>
(cherry picked from commit db204121686726c74dd0aba2d1c1790d40e7baba)
  • 40bdc68a
    by Dylan Baker at 2022-11-09T09:33:39-08:00 
    .pick_status.json: Update to 4ca61b5420ed26683891664aa35dc5f1a3bad063
  • 3801b943
    by Caio Oliveira at 2022-11-09T09:33:45-08:00 
    nir: Don't reorder volatile intrinsics

Fixes issue with "is helper invocation" that in recent SPIR-V is mapped to
a volatile Load.  The CSE was catching the loads before they were transformed
in the new is_helper_invocation intrinsic (that is not reorderable).

Fixes: 729df14e452 ("nir: Handle volatile semantics for loading HelperInvocation builtin")
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: M Henning <drawoc@darkrefraction.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19432>
(cherry picked from commit 8ab628ab2e4d4f460e2eabdb11876997c0ab13bc)
  • 2a0c174d
    by Alyssa Rosenzweig at 2022-11-09T09:33:49-08:00 
    panfrost: Copy resources when necessary

If the map doesn't set MAP_DISCARD_RANGE, we do have to copy the existing
contents over. MAP_WRITE on its only gives permission to replace the contents,
unfortunately it does not require that the application actually do so.

Closes: #7640
Fixes: 0b26a9f7739 ("panfrost: Don't copy resources if replaced")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reported-by: Roman Elshin
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19576>
(cherry picked from commit cf7a3906b0629b2b6aadd3b18a39eae06b10fd6a)
  • b24cac23
    by Timo Aaltonen at 2022-11-09T19:56:00+02:00 
    rules: Enable building lima drivers on riscv64. (LP: #1995137)
  • 41de25b2
    by Eric Engestrom at 2022-11-09T21:21:58+00:00 
    .pick_status.json: Update to 1174f376096ed6ceebb0fb2810456f1501a68df7
  • 76906af3
    by Eric Engestrom at 2022-11-09T21:22:05+00:00 
    zink: add missing build_id linker args

Fixes: 4e14da056d623ababf7c ("zink: Enable mesa/st frontend shader caching.")
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19485>
(cherry picked from commit dfa8600a2d39aabc88e5a5b87ab86b2b0ea30191)
  • ec5bc021
    by Alyssa Rosenzweig at 2022-11-09T21:22:05+00:00 
    panfrost: Don't copy resources if replaced

If a synchronized transfer_map is going to overwrite an entire resource,
there's no need to memcpy in the original contents ahead-of-time. This
memcpy is particularly bad for large buffers where it's copying WC->WC,
although that could be mitigated with threaded_context's cpu_storage in
the future if needed.

Prevents a performance regression in glmark2's buffer scenes from the
next patch, hence the Cc.

Cc: mesa-stable
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19361>
(cherry picked from commit 0b26a9f773956fc00a77b0d4a7aafee5795ce935)
  • 4be9ceb5
    by Alyssa Rosenzweig at 2022-11-09T21:22:05+00:00 
    panfrost: Replace resource shadowing flush

The entire point of resource shadowing is to avoid unnecessary flushing.
Flushing readers after shadowing is counterproductive. A refresher on
how resource shadowing is supposed to work:

First, we determine if it's beneficial to shadow resources. If so, we
create a new backing buffer object. We flush the current writer of the
resource, if there is one, so the current contents become known to the
CPU. If we are not discarding the original resource, we then copy the
existing contents of the buffer to the new shadow buffer on the CPU.
Finally, we swap the resource's backing buffer for our shadow. Any batch
that reads the resource will continue to read the old copy of the
resource, and any future draw calls will see the new copy with the
change implemented.

Where did we go wrong?

In 988d5aae74d ("panfrost: Flush resources when shadowing"), we started
flushing all readers. We didn't actually need to flush, we just needed
to avoid dangling references on the batches reading the old copy of the
resource. But that's easily enough avoided: just remove the references.
The batches still hold a reference to the underlying BO, which will be
freed at the right time regardless.

Originally motivated by glmark2 -bbuffer:update-method=subdata, which
has some pathological access paterns.

Firefox is a lot faster anecdotally (now scrolling at 60fps in firefox).

But what actually motivated this is an apitrace from Duckstation's GLES
renderer. With this patch, the in-game portion is improved 3fps to 21fps.

Closes: #4028
Fixes: 988d5aae74d ("panfrost: Flush resources when shadowing")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19361>
(cherry picked from commit 2d8f28df731638e1270b0ae273c7bfd2b29b7993)
  • cfa0a223
    by Yonggang Luo at 2022-11-09T21:22:05+00:00 
    glx: Fixes apple/apple_visual.c that include non-exist util/debug.h

Fixes: aa4ac5ff8bb ("utils: Merge util/debug.* into util/u_debug.* and remove util/debug.*")

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Illia Abernikhin <illia.abernikhin@globallogic.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19486>
(cherry picked from commit 287435134c6685a71f5c247778017c8b66a93627)
  • bf46cebc
    by Karol Herbst at 2022-11-09T21:22:05+00:00 
    iris: invalidate sysvals if grid dimension changes

Cc: mesa-stable
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18581>
(cherry picked from commit 9ccdd86b9040b265893712f4125d30552e0fe1cd)
  • a6bf5204
    by Bas Nieuwenhuizen at 2022-11-09T21:22:05+00:00 
    radv: Fix compute scratch buffer emission.

Copied wrong from radeonsi. The registers following the scratch
buffer address are the shader rsrc1/rsrc2. Not the user SGPR0
containing the ring resource word 1.

Fixes: 278e533ec9c ("radv: update scratch buffer registers on GFX11")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19488>
(cherry picked from commit b8865ad04643cd7e8faa8a863d81fedbc1e29231)
  • b4ba437e
    by Bas Nieuwenhuizen at 2022-11-09T21:22:05+00:00 
    radv: Make the compute scratch waves per SE as well.

Fixes: 278e533ec9c ("radv: update scratch buffer registers on GFX11")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19488>
(cherry picked from commit d876ddc920f1aaf48219f2060de6cc99c003bc5d)
  • 32a7d9b8
    by Lionel Landwerlin at 2022-11-09T21:22:05+00:00 
    anv: Reduce RHWO optimization (Wa_1508744258)

Implement Wa_1508744258:

   Disable RHWO by setting 0x7010[14] by default except during resolve
   pass.

Disable the RCC RHWO optimization at all times except when resolving
single sampled color surfaces.

v2: Move stalling to genX(cmd_buffer_apply_pipe_flushes) for clarity (Mark)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Mark Janes <markjanes@swizzler.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19450>
(cherry picked from commit ba0336ab3f69f958e4346c1465c2fe0859d4ac59)
  • d34e913a
    by Samuel Pitoiset at 2022-11-09T21:22:05+00:00 
    radv: fix fallback for extreme geometry with tessellation on GFX11

It would assert anyways. Found by inspection.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19495>
(cherry picked from commit fab87b0f4115bd360ed9334aa7a72518613c2014)
  • 8ebbb2f0
    by Eric Engestrom at 2022-11-09T21:22:05+00:00 
    vk/runtime: drop incorrect UNUSED annotation

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit c6c5949ff70a47c47795fe9161a7514173b5be24)
  • dea1b807
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: remove unnecessary check for NULL

We are initializing the device, so we know this will be NULL.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit c793d384c12ebaa13487aeee35532e52b888c049)
  • 424e3c10
    by Eric Engestrom at 2022-11-09T21:22:05+00:00 
    v3dv: avoid freeing already-freed memory

Fixes: ecb01d53fdb413121ce2 ("v3dv: refactor events")
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit aff368fe83bd4bcb4620a1a19e9b71eb67a05af7)
  • 0f99923b
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: do a better job at cleaning up the device on init failure

These leaks on device creation failure have been there before, but
were only exposed as CTS failures after the recent event refactoring.

Partially fixes:
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit b78fd50e90c700466213e4b3b40ce10547b1e14e)
  • abedacb9
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: fix incorrect return type

Fixes: ecb01d53fdb413121ce2 ("v3dv: refactor events")

Partially fixes:
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 7f905a8117e9dc85088fe23429594537ee240c1e)
  • 78b5a254
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: handle allocation failure during pipeline initialization

Fixes (with disk cache disabled):
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 1f5966397aa208eb574394a4d0a592b66b776b44)
  • ef0210c5
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: do better cleanup on failure during pipeline cache operation

Fixes (with disk cache enabled):
dEQP-VK.api.device_init.create_instance_device_intentional_alloc_fail.basic
dEQP-VK.api.object_management.alloc_callback_fail.device
dEQP-VK.api.object_management.alloc_callback_fail.device_group

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 5e97150e21a75920c3521717f70f9f92702729b4)
  • 7f3657fb
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: increase limit for active event objects

Fixes: ecb01d53fdb413121ce2 ("v3dv: refactor events")
Fixes: dEQP-VK.api.command_buffers.execute_large_primary

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 4c861cf22a63b892cec74bc132313dcd17d77584)
  • 7331880a
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: always check VK_ACCESS_2_MEMORY_READ_BIT for read accesses

Fixes: a981ac053964 ('v3dv: skip binning sync if binning shaders don't access external resources')

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 67e82fd1f2c334da2e888553967ebae04c08146a)
  • 614ae3ba
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: make the helper to emit pipeline barriers public to other files

Fixes: ecb01d53fdb413121ce2 ("v3dv: refactor events")
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit 8113f973b3946e465d508627a4796d94cec1f1d8)
  • ebd65d01
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: fix event synchronization

Since we now implement events in the GPU we need to be more careful
and insert barriers to honor the dependencies provided by the API
as well as ensuring we are synchronizing these with the compute
queue, since that is how we implement GPU event functionality.

Fixes: ecb01d53fdb413121ce2 ("v3dv: refactor events")
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19458>
(cherry picked from commit e6884df088a1e172cd87956fe9afdac63069eb2c)
  • d4a9ef1b
    by Connor Abbott at 2022-11-09T21:22:05+00:00 
    tu: tu: Clear patchControlPoints dirty state with static patchControlPoints

Noticed by inspection after the previous issue.

Fixes: 68f3c38c801 ("tu: Implement extendedDynamicState2PatchControlPoints")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18912>
(cherry picked from commit 4466f9aa02dec99c6879e5e357788843c3b573e5)
  • ab1a7fe3
    by Friedrich Vock at 2022-11-09T21:22:05+00:00 
    radv: Use spirv1.5 instead of vulkan1.2

Ubuntu 20.04 ships a glslangValidator that doesn't know about vulkan1.2 yet.

Fixes: 27186537 ("radv: Add PLOC shader")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19479>
(cherry picked from commit 95ed0330667f726474d24d8c56e478e346622b0b)
  • 15aae04d
    by Mauro Rossi at 2022-11-09T21:22:05+00:00 
    hasvk: fix android build and reported API version

anv_device.c for vulkan.intel_hasvk requires changes to be compiled
and behave correctly for android target

Fixes the following building error:

FAILED: src/intel/vulkan_hasvk/libanv_hasvk_common.a.p/anv_device.c.o
...
../src/intel/vulkan_hasvk/anv_device.c:143:19: error: use of undeclared identifier 'ANV_API_VERSION_1_3'
   *pApiVersion = ANV_API_VERSION_1_3;
                  ^
../src/intel/vulkan_hasvk/anv_device.c:1822:44: error: use of undeclared identifier 'ANV_API_VERSION_1_3'
      .apiVersion = pdevice->use_softpin ? ANV_API_VERSION_1_3 : ANV_API_VERSION_1_2,
                                           ^
../src/intel/vulkan_hasvk/anv_device.c:1822:66: error: use of undeclared identifier 'ANV_API_VERSION_1_2'
      .apiVersion = pdevice->use_softpin ? ANV_API_VERSION_1_3 : ANV_API_VERSION_1_2,
                                                                 ^
3 errors generated.

Cc: "22.3" mesa-stable
Fixes: 00eefdc ("hasvk: stop advertising Vk 1.3 on non-softpin")
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19452>
(cherry picked from commit 814b822fe065ab88409a2833b1c7f331f714b26e)
  • cdecc19a
    by Samuel Pitoiset at 2022-11-09T21:22:05+00:00 
    radv: fix setting MIN_LOD for texture descriptors on GFX11

Found by inspection because the MIN_LOD bits were moved.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19496>
(cherry picked from commit e891e84f4be6271cf34f62768b085ee99b6db86c)
  • 5a5c33ee
    by Martin Roukala (né Peres) at 2022-11-09T21:22:05+00:00 
    zink/ci: mark another test as fixed after the frontend caching series

Just like on Lavapipe, Emma Anholt's !19124 also fixed the
restore-sso-program test on RADV.

Fixes: 4e14da056d62 ("zink: Enable mesa/st frontend shader caching.")
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19468>
(cherry picked from commit dc08875415064033ea1e344a522273bc1255a3b8)
  • 29c6d79a
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: vkCmdWaitEvents2 takes an array of VkDependencyInfo

We have been incorrectly assuming there was just one for all the
events, apparently CTS never uses more than one event.

Fixes: e6884df088 ('v3dv: fix event synchronization')
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19518>
(cherry picked from commit 36ef75b6eb924cf5ebfb09834067d2519d080ca8)
  • 6b4e3cce
    by Iago Toral Quiroga at 2022-11-09T21:22:05+00:00 
    v3dv: use vk_alloc instead of malloc

Fixes: e6884df088 ('v3dv: fix event synchronization')
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19518>
(cherry picked from commit 22789d34be0fb6d81012e2e4d53b738384eb98db)
  • 4acfc6a9
    by Martin Roukala (né Peres) at 2022-11-09T21:22:05+00:00 
    zink/ci: add another subtest to the list of known failures for radv

This test as a whole does not seem to work anywhere, even lavapipe, but
one particular subtest was passing until a recent change
(!19438 - zink: polygon mode fixes?).

After consideration by @kusma, it appears that the subtest was passing
by accident due to zink generating the wrong values. Given that this is
not something that users would ever experience as a regression, we
simply document this new failure along with all the others for this
test.

Fixes: 53721827ea1c ("zink: correct depth-bias enable condition")
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19517>
(cherry picked from commit d7ad9e701424e4b100e11f08f2fa416b228b19d3)
  • 68b469ba
    by Samuel Pitoiset at 2022-11-09T21:22:05+00:00 
    radv: make the GDS/GDS OA buffer objects resident

GDS is used for NGG queries/streamout (GFX10+ only) and the BOs were
only added to the graphics queue because compute doesn't need them.
Though, the kernel emits a GDS switch when a queue submission doesn't
use GDS. That means that submitting jobs on the compute queue without
GDS can reset the state of the graphics queue and lead to GPU hangs.

The only viable solution for now is to make the GDS BOs resident to
avoid resetting the state between queues. This shouldn't introduce
more syncs between queues because GDS BOs are similar for both.

This fixes a GPU hang with Warhammer Chaosbane during loading time and
possibly some spurious random GPU hangs. Note that this GPU hang was
workarounded on the Steam side with RADV_DEBUG=nongg.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19466>
(cherry picked from commit 26c8fedc1bb12fa8f3d6c646308f4b46756d77c7)
  • c0c453e5
    by Samuel Pitoiset at 2022-11-09T21:22:05+00:00 
    radv: invalidate L2 instead of only writeback L2 when using DCC stores

It seems INV_L2 is the right thing to do, especially for RDNA2 chips
with non-coherent RBs (NAVI22 is one of these). This fixes DCC
corruption.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6476
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7507
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19516>
(cherry picked from commit 06adf6ad117e92a832c5073326bd83b638b3ca37)
  • 7ba025d5
    by Alex Brachet at 2022-11-09T21:22:06+00:00 
    nir: Fix qsort comparator function

`pred` is a pointer, for sufficiently large numbers these
being cast to int were both > 0 regardless of the order
of `data1` and `data2`.

Fixes: 523a28d3fe0d ("nir: add an instruction set API")

Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19539>
(cherry picked from commit c987a727a76eda1feada07e2eca6a5597dbddd93)
  • 7110d632
    by Konstantin Seurer at 2022-11-09T21:22:06+00:00 
    radv/rt: Restore prev barycentrics when rejecting hits

Closes: #6348
cc: mesa-stable

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19534>
(cherry picked from commit e5b3efe5825aec5c9163fa0d9b20418d2876fdd2)
  • f4fac39d
    by Marek Olšák at 2022-11-09T21:22:06+00:00 
    radeonsi/gfx11: fix compute scratch buffer - WAVES is always per SE

Fixes: ba02ed91a60 - ac/gfx11: fix the scratch buffer

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19477>
(cherry picked from commit bdfacd0a24e023515fb7b7fae4a279cff0fbac4e)
  • a7482cfa
    by Francisco Jerez at 2022-11-09T21:22:06+00:00 
    nir/lower_int64: Fix float16 to int64 conversions.

Currently float16 to int64 conversions don't work correctly, because
the "div" variable has an infinite value, since 2^32 isn't
representable as a 16-bit float, which causes the result of of rem(x,
div) to be NaN for all inputs, leading to an incorrect result.  Since
no values of magnitude greater than 2^32 are representable as a
float16 we don't actually need to do the fdiv/frem operations, the
conversion is equivalent to f2u32 with the result padded to 64 bits.

Rework:
 * Jordan: Handle f16 in if/else rather than conditional

Fixes: 936c58c8fcc ("nir: Extend nir_lower_int64() to support i2f/f2i lowering")
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19391>
(cherry picked from commit e14f85366ebbc57f45d2561e0d3f0804f8adb549)
  • a23360cd
    by Erik Faye-Lund at 2022-11-09T21:22:06+00:00 
    docs/zink: fix and cleanup rst syntax

This new section didn't use the correct RST syntax, and ended up
with a broken section in the rendered docs.

Fix the syntax, and clean things up a bit to avoid overly long lines.

Fixes: be235edfe2b ("zink: add profile documentation")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19481>
(cherry picked from commit 6b3b6333915e8ca64c531a4decb79c0256e1e4f2)
  • 19711e41
    by Marek Vasut at 2022-11-09T21:22:06+00:00 
    etnaviv: Use old set of state registers for PE configuration on GC880

While the GC880 is HALTI0, it still uses the old set of state registers
for PE pipe configuration. This is another specialty of the GC880, readd
the missing handling for this GPU otherwise e.g. Qt5 cube example suffers
from rendering corruption with both eglfs and wayland backends.

Fixes: 7c46a488362 ("etnaviv: use new PE pipe address states on >= HALTI0")
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Marek Vasut <marex@denx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19562>
(cherry picked from commit 20984aab0f6717f5fbf79fc21c2c5f442472b605)
  • 935aaef3
    by Samuel Pitoiset at 2022-11-09T21:22:06+00:00 
    radv: re-emit the guardband state when restoring meta operations

Meta operations change dynamic states like viewports and previously,
the guardband state was also always re-emitted because it relied on
dynamic viewport/scissor changes.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7577
Fixes: 40d8df72808 ("radv: emit the guardband state separately from the scissor state")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19521>
(cherry picked from commit 33d60bda9d385cca56e3d515472ca75fde53e126)
  • 13915644
    by Samuel Pitoiset at 2022-11-09T21:22:06+00:00 
    ac/nir,radv: rework and fix NGG queries enables for VS/TES

XFB queries need to be enabled with NGG streamout and VS/TES.
Previously, the NGG lowering code relied on has_prim_query for XFB.

This fixes failures with RADV_PERFTEST=ngg_streamout on GFX10.3 with
the vkd3d-proton testsuite. Vulkan CTS is missing TES tests with XFB
queries apparently.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19493>
(cherry picked from commit 505290dc44a5b728dbc6ec9ed038300ab158c323)
  • 4eea1cb8
    by Mauro Rossi at 2022-11-09T21:22:06+00:00 
    Android.mk: Fix gnu++14 related build failures

This patch filters-out '-std=gnu++14' from the cflags obtained
from AOSP/KATI dummy target output to avoid the following building errors:

FAILED: src/gallium/drivers/r600/45f68e3@@r600@sta/sfn_sfn_assembler.cpp.o
...
clang++ ... -std=c++17 ... -std=gnu++14
...
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.cpp:27:
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.h:32:
In file included from ../src/gallium/drivers/r600/sfn/sfn_shader.h:31:
../src/gallium/drivers/r600/sfn/sfn_instr.h:369:56: error: no template named 'is_base_of_v' in namespace 'std'; did you mean 'is_base_of'?
template <typename T, typename = std::enable_if_t<std::is_base_of_v<Instr, T>>>
                                                  ~~~~~^~~~~~~~~~~~
                                                       is_base_of
/home/utente/pie-x86_kernel/external/libcxx/include/type_traits:1412:29: note: 'is_base_of' declared here
struct _LIBCPP_TEMPLATE_VIS is_base_of
                            ^
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.cpp:27:
In file included from ../src/gallium/drivers/r600/sfn/sfn_assembler.h:32:
In file included from ../src/gallium/drivers/r600/sfn/sfn_shader.h:31:
../src/gallium/drivers/r600/sfn/sfn_instr.h:369:51: error: template argument for non-type template parameter must be an _expression_
template <typename T, typename = std::enable_if_t<std::is_base_of_v<Instr, T>>>
                                                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/utente/pie-x86_kernel/external/libcxx/include/type_traits:439:16: note: template parameter is declared here
template <bool _Bp, class _Tp = void> using enable_if_t = typename enable_if<_Bp, _Tp>::type;
               ^
2 errors generated.

Cc: "22.2" "22.3" mesa-stable
Reviewed-by: Roman Stratiienko <r.stratiienko@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19563>
(cherry picked from commit fd8ec189e575d5220d8b4c30647fa6de57928e07)
  • 87e7794d
    by Ian Romanick at 2022-11-09T21:22:06+00:00 
    intel/fs: Fix constant propagation into 32x16 integer multiplication

Don't copy propagate the constant in situations like

    mov(8)          g8<1>D          0x7fffffffD
    mul(8)          g16<1>D         g8<8,8,1>D      g15<16,8,2>W

On platforms that only have a 32x16 multiplier, this will result in
lowering the multiply to

    mul(8)          g15<1>D         g14<8,8,1>D     0xffffUW
    mul(8)          g16<1>D         g14<8,8,1>D     0x7fffUW
    add(8)          g15.1<2>UW      g15.1<16,8,2>UW g16<16,8,2>UW

On Gfx8 and Gfx9, which have the full 32x32 multiplier, it results in

    mul(8)          g16<1>D         g15<16,8,2>W    0x7fffffffD

Volume 2a of the Skylake PRM says:

    When multiplying a DW and any lower precision integer, the
    DW operand must on src0.

See also https://gitlab.freedesktop.org/mesa/crucible/-/merge_requests/104.

Previous to INTEL_shader_integer_functions2 (in Vulkan or OpenGL), I
don't think it would be possible to create a situation where this could
occur.  I discovered this via some optimizations that can determine that
the non-constant source must be able to fit in 16-bits.  The case listed
above came from piglit's "ext_transform_feedback-order arrays points"
with those optimizations in place.

No shader-db or fossil-db changes on any Intel platform.

Fixes: de6c0f84879 ("intel/fs: Implement support for NIR opcodes for INTEL_shader_integer_functions2")
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17718>
(cherry picked from commit db204121686726c74dd0aba2d1c1790d40e7baba)
  • e54150d6
    by Lionel Landwerlin at 2022-11-09T21:22:06+00:00 
    anv: fix missing VkPhysicalDeviceExtendedDynamicState3PropertiesEXT handling

Fixes: 13c422e1b2ed ("anv: toggle on EXT_extended_dynamic_state3")
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19573>
(cherry picked from commit 97b3dd34c1aa11d98df7e26b4aae50be0bd62cf0)
  • f4a7f286
    by Caio Oliveira at 2022-11-09T21:22:06+00:00 
    nir: Don't reorder volatile intrinsics

Fixes issue with "is helper invocation" that in recent SPIR-V is mapped to
a volatile Load.  The CSE was catching the loads before they were transformed
in the new is_helper_invocation intrinsic (that is not reorderable).

Fixes: 729df14e452 ("nir: Handle volatile semantics for loading HelperInvocation builtin")
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: M Henning <drawoc@darkrefraction.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19432>
(cherry picked from commit 8ab628ab2e4d4f460e2eabdb11876997c0ab13bc)
  • e56be98d
    by Yonggang Luo at 2022-11-09T21:22:06+00:00 
    ci: Fixes macos.yml

Stick to macos-11 to prevent accident broken
always install meson with pip to prevent pull new version of python

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19615>
(cherry picked from commit 81b4af28494c065c5646e3f61a72afd829d978c6)
  • d812245d
    by Alyssa Rosenzweig at 2022-11-09T21:22:06+00:00 
    panfrost: Fix build with Perfetto (again)

Sync UAPI for the upstream fix.

Upstream commit: https://cgit.freedesktop.org/drm-misc/commit/?h=drm-misc-fixes&id=c4299907c09a638c0a30f029338d07941c049d73

Closes: #7195
Fixes: 6a4532cbabf ("panfrost: Sync panfrost_drm.h from drm-misc-next")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Tested-by: Chris Healy <healych@amazon.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19581>
(cherry picked from commit d40af879098bea90176619dc11facbd68c555f32)
  • 6c24336e
    by Alyssa Rosenzweig at 2022-11-09T21:22:06+00:00 
    panfrost: Copy resources when necessary

If the map doesn't set MAP_DISCARD_RANGE, we do have to copy the existing
contents over. MAP_WRITE on its only gives permission to replace the contents,
unfortunately it does not require that the application actually do so.

Closes: #7640
Fixes: 0b26a9f7739 ("panfrost: Don't copy resources if replaced")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reported-by: Roman Elshin
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19576>
(cherry picked from commit cf7a3906b0629b2b6aadd3b18a39eae06b10fd6a)
  • dd860e4b
    by Mario Kleiner at 2022-11-09T21:22:06+00:00 
    vulkan/wsi/display: Reset connector state in vkReleaseDisplay().

If an application was transitioning out of fullscreen exclusive
display mode, the wsi_display_connector->active state was not
reset in vkReleaseDisplay() from fullscreen. When the app then
later tried to go to fullscreen display mode again on the same
display output with the same video mode, this caused
_wsi_display_queue_next() to skip a required drmModeSetCrtc()
during the first vkQueuePresent() after entering direct display
mode.

While this often worked by pure luck on a single-display setup,
it goes sideways on a multi-display setup where the viewport
of the associated crtc does not have a (x,y) offset of (0,0).
E.g., XOrg/X11 RandR output leasing of an output whose viewport
starts at x = 1920:

1. X-Server has RandR outputs viewport at x = 1920, in a shared
   framebuffer, shared across all crtc's on a X-Screen.

2. Application leases that output for direct display mode,
   1st vkQueuePresent() triggers drmModeSetCrtc() of output
   to (x,y) = 0,0, as required for Vulkan/wsi/direct framebuffer
   setup.

3. Application does rendering and presenting.

4. Application vkReleaseDisplay() the output, terminates the
   RandR lease. X-Server takes over again.

5. X-Server modesets to reconfigure output back to viewport
   with (x,y) = 1920, 0.

6. Application leases same output again later on, and tries
   vkQueuePresent() again. Because of the bug fixed in this
   commit, the required drmModeSetCrtc() to (x,y) = 0,0 is
   erroneously skipped due to the stale cached connector state.

7. drmModePageflip() fails due to the wrong crtc viewport
   (x,y) = 1920, 0, mismatched for the need of the Vulkan
   framebuffer of (x,y) = 0,0. Kernel returns -ENOSPACE,
   Swapchain goes into permanent VK_ERROR_SURFACE_LOST state.
   Destroying and recreating the swapchain, as recommended
   by the Vulkan spec for error handling won't help. Game over!

Resetting wsi_display_connector->active = false; fixes the
problem of wrong / stale connector state and Vulkan/wsi/display
clients are happy on multi-display setups again, as tested
in various single- and multi-display configurations.

This bug affects all Mesa releases with Vulkan/WSI/Display
support and should therefore be backported.

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Fixes: 352d320a0745 ("vulkan: Add EXT_direct_mode_display [v2]")
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19484>
(cherry picked from commit 24094ee03d625fbcd2d154e8c2dd5434ba88f166)
  • d580ab88
    by Jason Ekstrand at 2022-11-09T21:22:06+00:00 
    intel/lower_mem_access_bit_sizes: Compute alignments automatically

Because dup_mem_intrinsic() retains the SSA offset from the original
intrinsic and only modifies it by adding a constant, we can compute the
alignment based on the original alignment and the constant offset.  This
is both easier and more accurate.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19580>
(cherry picked from commit 85685cf932c25fc53cae5e36b5d75f9d6a79c765)
  • 7701bf12
    by Jason Ekstrand at 2022-11-09T21:22:06+00:00 
    intel: Don't cross DWORD boundaries with byte scratch load/store

The back-end swizzles dwords so that our indirect scratch messages match
the memory layout of spill/fill messages for better cache coherency.
The swizzle happens at a DWORD granularity.  If a read or write crosses
a DWORD boundary, the first bit will get correctly swizzled but whatever
piece lands in the next dword will not because the scatter instructions
assume sequential addresses for all bytes.  For DWORD writes, this is
handled naturally as part of scalarizing.  For smaller writes, we need
to be sure that a single write never escapes a dword.

Fixes: fd04f858b0aa ("intel/nir: Don't try to emit vector load_scratch instructions")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7364
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19580>
(cherry picked from commit 25c180b50974b55e007dbbff18be1d831cd06551)
  • 80b0483e
    by Iago Toral Quiroga at 2022-11-09T21:22:06+00:00 
    broadcom/compiler: avoid using ldvary sequence to hide latency of branching

This can cause us to stomp the contents of r5 before we have a chance to read
it, like this:

0x3d103186bb800000 nop                           ; nop                         ; ldvary.r0
0x3d105686bbf40000 nop                           ; mov rf26, r5                ; ldvary.r1
0x020000ef0000d000 bu.allna  232, r:unif (0x0000001c / 0.000000)
0x3d1096c6bbf40000 nop                           ; mov rf27, r5                ; ldvary.r2

Here, the MOV in the last instruction is supposed to read r5 produced from
ldvary.r0, but because we have inserted the bu instruction in between now
that read happens at the same time that ldvary.r1 updates r5, stomping the
value we were supposed to read.

Fix this by disallowing injection of a branch instruction in between an ldvary
instruction and its write to the r5 register 2 instructions later.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7062
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19616>
(cherry picked from commit 1174f376096ed6ceebb0fb2810456f1501a68df7)
  • 38cb64bf
    by Eric Engestrom at 2022-11-09T21:24:41+00:00 
    VERSION: bump for 22.3.0-rc2
  • caf17bfa
    by Dylan Baker at 2022-11-10T11:41:42-08:00 
    .pick_status.json: Update to ecc2dfc503dcf62573fa539b49caa874f908c1cf
  • b2a3f0d4
    by Iago Toral Quiroga at 2022-11-10T12:02:09-08:00 
    broadcom/compiler: avoid using ldvary sequence to hide latency of branching

This can cause us to stomp the contents of r5 before we have a chance to read
it, like this:

0x3d103186bb800000 nop                           ; nop                         ; ldvary.r0
0x3d105686bbf40000 nop                           ; mov rf26, r5                ; ldvary.r1
0x020000ef0000d000 bu.allna  232, r:unif (0x0000001c / 0.000000)
0x3d1096c6bbf40000 nop                           ; mov rf27, r5                ; ldvary.r2

Here, the MOV in the last instruction is supposed to read r5 produced from
ldvary.r0, but because we have inserted the bu instruction in between now
that read happens at the same time that ldvary.r1 updates r5, stomping the
value we were supposed to read.

Fix this by disallowing injection of a branch instruction in between an ldvary
instruction and its write to the r5 register 2 instructions later.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7062
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19616>
(cherry picked from commit 1174f376096ed6ceebb0fb2810456f1501a68df7)
  • 91c675a0
    by Karol Herbst at 2022-11-10T12:02:10-08:00 
    glsl: fix buffer texture type

Fixes: 3ace6b968b3 ("compiler/types: Add a texture type")
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19381>
(cherry picked from commit 87526f79db68a13ebd448cfd6b1be4b25616c801)
  • e1fe4a64
    by Alyssa Rosenzweig at 2022-11-10T12:02:11-08:00 
    panfrost: Require 64-byte alignment on imports

While Panfrost allocates linear images with strides that are a multiple of 64
bytes, other dma-buf producers on the system may not satisfy this requirement.
However, at least on v7 and newer, any image with a regular format must have a
stride that is a multiple of 64 bytes.

This fixes a real bug in an application that created a linear R8_UNORM image
with stride 480 bytes, imported it as an EGL_image, and then tried to texture
from it with the GPU. Previously, the driver allowed this situation but it
resulted in an imprecise fault from the GPU. This patch corrects the driver to
reject the import as invalid due to the unaligned stride, ensuring we never
attempt to texture from such a resource.

To implement, we add some new layout queries to centralize knowledge about the
stride alignment requirements, and we sprinkle in asserts to show how the
invariant is upheld throughout the lifecycle of image creation to texturing.

Cc: mesa-stable
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19620>
(cherry picked from commit 811f8a19469722bea32f3c539b8cf0939fe3b057)
  • 80570915
    by Yonggang Luo at 2022-11-11T00:17:26+00:00 
    ci: Fixes macos.yml

Stick to macos-11 to prevent accident broken
always install meson with pip to prevent pull new version of python

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Eric Engestrom <eric@igalia.com>
Acked-by: Dylan Baker <dylan.c.baker@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19640>
  • 1867648d
    by Guilherme Gallo at 2022-11-11T02:29:52+00:00 
    ci: Fix kernel+rootfs.* jobs

The rootfs generation is failing due to issues with the deqp and crosvm
build stages.

== crosvm ==

This week, crates.io released the bindgen cargo package at version
0.61.0, but this version could not be installed via `cargo install
bindgen ...`, setting the version to the previous one to avoid breaking
the Mesa rootfs builds.

See also related failed job:
https://gitlab.freedesktop.org/gallo/mesa/-/jobs/30046963

== deqp ==

The deqp build is failing due to the missing archive of an old zlib
release version, which was deleted due to a CVE, see zlib 1.2.13 release
notes.

As the deqp uprev to 1.3.4.0, which contains the fix, was not
straightforward, let's only apply the necessary patch to fix zlib
source code download link and then remove this indirection in an
eventual deqp uprev.

Example of a failed kernel+rootfs build job:
  https://gitlab.freedesktop.org/gallo/mesa/-/jobs/30045324

Solved Conflicts:
 	.gitlab-ci/image-tags.yml

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
(cherry picked from commit db2c14887bb52ec6fd05c5f8b5363f1689b4b45c)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19604>
  • 44807902
    by Guilherme Gallo at 2022-11-11T02:29:52+00:00 
    ci: Update piglit with s3 support

With new S3 support, we can use JWT-only server interaction via the
removal of `role-session` and `minio-host` arguments from PIGLIT_ARGS in
YAML.
This parameter change will come in a later commit.

Solved Conflicts:
	.gitlab-ci/container/build-piglit.sh
	.gitlab-ci/image-tags.yml

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
(cherry picked from commit 70ce1dcacc92a816322082c8695569b6a91a1810)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19604>
  • 2d1b544a
    by Guilherme Gallo at 2022-11-11T02:29:52+00:00 
    ci: Update piglit-traces tests expectations

Found some:
- crashes in zink, softpipe
- fails in a630-restricted
- unexpectedpass in broadcom
    - fixed by https://gitlab.freedesktop.org/mesa/piglit/-/merge_requests/730

More details in the test expectations files comments.

Solved Conflicts:
	src/gallium/drivers/zink/ci/zink-lvp-skips.txt

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
(cherry picked from commit a108e4f70cc61a6b9119cd766d51c9b596e07c7f)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19604>
  • 4827c3c6
    by Guilherme Gallo at 2022-11-11T02:29:52+00:00 
    ci: Update ci-fairy in CI, rootfs and containers

ci-fairy is pulverized in possible different versions at Mesa CI.
This commit updates all of them to the version that migrates minio to
s3.
Also, trigger the build of base and test containers, as both uses
ci-fairy as well

Solved Conflicts:
	.gitlab-ci/image-tags.yml

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
(cherry picked from commit a04ed2f971dccbf8b24f084ac9baaf8b299944a0)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19604>
  • 2ae65f57
    by Benjamin Tissoires at 2022-11-11T02:29:52+00:00 
    CI: convert to use the new S3 server instead of the legacy minio

We don't need to login anymore, but we can't use plain minio commands
now. `ci-fairy` got a helper as `s3cp` to keep an almost identical
API.

Solved Conflicts:
	.gitlab-ci/common/init-stage2.sh
	.gitlab-ci/container/lava_build.sh
	.gitlab-ci/prepare-artifacts.sh
	src/amd/ci/traces-amd.yml
	src/freedreno/ci/traces-freedreno.yml
	src/gallium/frontends/lavapipe/ci/traces-lavapipe.yml

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
(cherry picked from commit 67cee534a88c95a8eb6839f7bcf28a5e6dac8fbf)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19604>
  • 130bfdde
    by Guilherme Gallo at 2022-11-11T02:29:52+00:00 
    docs: Update link releases notes for 22.2.x

Some previous commits changed .gitlab-ci.yml, so the `test-docs-mr` was
triggered. It failed due to the lack of toctree reference in
documentation, even when the relnotes for 22.2.0 to 22.2.3 exist.

This commit fixes that to make the CI pipeline green again.

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19604>
  • ae07c08f
    by Nanley Chery at 2022-11-11T11:57:17+02:00 
    iris: Reduce use of RHWO optimization (Wa_1508744258)

Implement Wa_1508744258:

   Disable RHWO by setting 0x7010[14] by default except during resolve
   pass.

Disable the RCC RHWO optimization at all times except when resolving
single sampled color surfaces. MCS partial resolves are done via
software (i.e., not via a HW bit) and so are not expected to need this
workaround.

Reviewed-by: Mark Janes <markjanes@swizzler.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit 0fa540ef6137aa235f62a0ba60886f64d754566d)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19671>
  • c3884066
    by Lionel Landwerlin at 2022-11-11T11:57:26+02:00 
    anv: Reduce RHWO optimization (Wa_1508744258)

Implement Wa_1508744258:

   Disable RHWO by setting 0x7010[14] by default except during resolve
   pass.

Disable the RCC RHWO optimization at all times except when resolving
single sampled color surfaces.

v2: Move stalling to genX(cmd_buffer_apply_pipe_flushes) for clarity (Mark)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Mark Janes <markjanes@swizzler.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: mesa-stable
(cherry picked from commit ba0336ab3f69f958e4346c1465c2fe0859d4ac59)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19671>
  • 1c54e9a8
    by Dylan Baker at 2022-11-14T12:00:12-08:00 
    .pick_status.json: Update to 9fd1d47aa0f19fac30d29e6ae40ed0383c1930ee
  • cce6fffe
    by Mario Kleiner at 2022-11-14T12:00:20-08:00 
    vulkan/wsi/display: Reset connector state in vkReleaseDisplay().

If an application was transitioning out of fullscreen exclusive
display mode, the wsi_display_connector->active state was not
reset in vkReleaseDisplay() from fullscreen. When the app then
later tried to go to fullscreen display mode again on the same
display output with the same video mode, this caused
_wsi_display_queue_next() to skip a required drmModeSetCrtc()
during the first vkQueuePresent() after entering direct display
mode.

While this often worked by pure luck on a single-display setup,
it goes sideways on a multi-display setup where the viewport
of the associated crtc does not have a (x,y) offset of (0,0).
E.g., XOrg/X11 RandR output leasing of an output whose viewport
starts at x = 1920:

1. X-Server has RandR outputs viewport at x = 1920, in a shared
   framebuffer, shared across all crtc's on a X-Screen.

2. Application leases that output for direct display mode,
   1st vkQueuePresent() triggers drmModeSetCrtc() of output
   to (x,y) = 0,0, as required for Vulkan/wsi/direct framebuffer
   setup.

3. Application does rendering and presenting.

4. Application vkReleaseDisplay() the output, terminates the
   RandR lease. X-Server takes over again.

5. X-Server modesets to reconfigure output back to viewport
   with (x,y) = 1920, 0.

6. Application leases same output again later on, and tries
   vkQueuePresent() again. Because of the bug fixed in this
   commit, the required drmModeSetCrtc() to (x,y) = 0,0 is
   erroneously skipped due to the stale cached connector state.

7. drmModePageflip() fails due to the wrong crtc viewport
   (x,y) = 1920, 0, mismatched for the need of the Vulkan
   framebuffer of (x,y) = 0,0. Kernel returns -ENOSPACE,
   Swapchain goes into permanent VK_ERROR_SURFACE_LOST state.
   Destroying and recreating the swapchain, as recommended
   by the Vulkan spec for error handling won't help. Game over!

Resetting wsi_display_connector->active = false; fixes the
problem of wrong / stale connector state and Vulkan/wsi/display
clients are happy on multi-display setups again, as tested
in various single- and multi-display configurations.

This bug affects all Mesa releases with Vulkan/WSI/Display
support and should therefore be backported.

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Fixes: 352d320a0745 ("vulkan: Add EXT_direct_mode_display [v2]")
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19484>
(cherry picked from commit 24094ee03d625fbcd2d154e8c2dd5434ba88f166)
  • 06c5579f
    by Dylan Baker at 2022-11-14T12:00:29-08:00 
    .pick_status.json: Mark 25c180b50974b55e007dbbff18be1d831cd06551 as backported
  • ef204ae8
    by Emma Anholt at 2022-11-14T12:00:32-08:00 
    ir3/ra: Make sure we don't pick a preferred reg overflowing the file.

If we're in handle_collect()'s dst allocation and are part of a merge set
near the end of the file, our check for reg_elem_size(reg) would let us
use the preferred reg when that would immediately lead to
allocate_dst_fixed() creating an interval extending thruogh reg_size(reg)
that overflows the file.

Avoids a regression on gfxbench5/gl_5_high_off/17.shader_test in the next
commit.  No change on shader-db.

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18946>
(cherry picked from commit a39113b616099a7bcce9e62337731f040c7cad64)
  • fb86c074
    by Tapani Pälli at 2022-11-14T13:37:58-08:00 
    iris: setup stage bitmask for Wa_22011440098

Fixes: 40b66a44998 ("anv, iris: Add Wa_22011440098 for DG2")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19636>
(cherry picked from commit 210d95bdb35c2c172f0a3a9527b181f8988a627f)
  • b8e787da
    by Tapani Pälli at 2022-11-14T13:37:58-08:00 
    anv: setup stage bitmask for Wa_22011440098

Fixes: 40b66a44998 ("anv, iris: Add Wa_22011440098 for DG2")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19636>
(cherry picked from commit ecd4517560f135f64abf6e40acc48807b400ca41)
  • 2406718c
    by Iago Toral Quiroga at 2022-11-14T13:37:58-08:00 
    v3dv: fix debug dump on BO free

We were resetting the BO struct right before dumping its data. Fix
this by moving the reset later.

Fixes: 44fa8304d45 ('v3dv: add a refcount mechanism to BOs')
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19672>
(cherry picked from commit 2c44597181e28b05d4b269ee73d4f3a59bc0a806)
  • 7d0c7c57
    by Timothy Arceri at 2022-11-14T13:37:58-08:00 
    nir: fix typo in lower_double options handling

Seems the intention was to check that both flags were not enabled
instead we were checking that the floor flag was both set and not
set so the result would always be false.

Fixes: 3749a6ecd282 ("nir: honor lower_double options for ffloor and ffract")

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19642>
(cherry picked from commit 34c52d8cb98e74a9debcd2605b377170c025582a)
  • fca0249e
    by Lionel Landwerlin at 2022-11-14T13:37:58-08:00 
    anv: bump pool bucket max allocation size

Age of Empire IV generates a shader of ~2.3Mb on DG2 which is above
the limit we currently have.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19669>
(cherry picked from commit ae76bba34a878dbd299f911bceaa9acdf01eb38a)
  • 71be5b25
    by Jason Ekstrand at 2022-11-14T15:50:40-08:00 
    vulkan: Handle VK_SUBPASS_EXTERNAL at the end of a subpass

Fixes: 1d726940d288 ("vulkan: Add a common CmdBegin/EndRederPass implementation")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19599>
(cherry picked from commit 4ff5051944e7a7a8bf003c331af1a73ac6c7df61)
  • e514b07e
    by Jason Ekstrand at 2022-11-14T15:50:42-08:00 
    vulkan: Unconditionally add barriers for missing external subpass deps

This is a very scorched-earth approach which doesn't take into account
whether or not there are any explicitly provided dependencies.  We could
take a finer-grained approach in theory but it's unlikely to matter in
practice since you usually stall in Begin/EndRenderPass anyway.

Fixes: 1d726940d288 ("vulkan: Add a common CmdBegin/EndRederPass implementation")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6203
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7650
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19599>
(cherry picked from commit 11b2a063bf1f18b3be9542be8c229427a33c92f0)
  • c9e65aca
    by Caio Oliveira at 2022-11-14T15:50:45-08:00 
    intel/compiler: Fix dynarray usage in intel_clc

The code builds up the dynamic array of objects (spirv_objs) and
collect pointers to each of them into another dynamic
array (spirv_ptr_objs).

If the growth of the first array cause a reallocation, it is
possible that the previous pointers end up invalid.

Fixes: 77e929a5273 ("intel/clc: allow multiple CL files to be compiled together")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19730>
(cherry picked from commit 9fd1d47aa0f19fac30d29e6ae40ed0383c1930ee)
  • 6922ca94
    by Dylan Baker at 2022-11-15T15:41:40-08:00 
    .pick_status.json: Update to 51457b19cea9608f6bb2be81430d37da5333d2fe
  • 2828fae0
    by Caio Oliveira at 2022-11-15T15:41:48-08:00 
    intel/compiler: Fix missing tie-breaker in brw_nir_analyze_ubo_ranges() ordering code

Per Ken suggestion, use ascending order for the start offset.

Fixes: 6d28c6e52cf ("i965: Select ranges of UBO data to be uploaded as push constants.")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19731>
(cherry picked from commit 494e2edb90d06f645e09b10d2c3f6b8d274d4dba)
  • c0c5eed5
    by Dave Airlie at 2022-11-15T15:41:50-08:00 
    gallivm: add coro malloc hooks earlier and always.

This fixes GALLIVM_DEBUG=asm for compute shaders, changing
the hooks after dumping causes a segfault because the
memory has already been finalised. Just add the hooks always,
and before dumping anything.

Fixes: f511d2a55337 ("gallivm: rework coroutine malloc/free callouts.")
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19714>
(cherry picked from commit fb7de303ba7cbdb4237d2ae93011bc2845114d35)
  • af4c9789
    by Qiang Yu at 2022-11-15T15:41:51-08:00 
    ac/nir/ngg: fix nogs culling with nuw add

We should not use "nuw" here as negative add positive may wrap
around (negative is 0xffffff??).

This problem can be observed with LLVM15 (I can't see when LLVM14):
  %.neg = mul nsw i32 %31, -4
  %163 = add nuw nsw i32 %.neg, 16
  %164 = lshr i32 257, %.neg
  %165 = lshr i32 %164, %163

LLVM just assume %.neg is possitive, so pre-shift 0x01010101 by 16.
This get wrong value because we can't get back the shifted bits with
a negative shift right.

Fixes: 75dbb404393 ("ac/nir: Remove byte permute from prefix sum of the repack sequence.")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19718>
(cherry picked from commit 982b523769a75c99039deac7f832a1e10260e916)
  • 838c1249
    by Qiang Yu at 2022-11-15T15:41:51-08:00 
    ac/nir/ngg: remove nuw for negative value add

Add negative value is possible to wrap around. I haven't seen this
"nuw" causes any problem yet, but let's remove it for safe.

Fixes: 60ac5dda82e ("ac: Add NIR lowering for NGG GS.")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19718>
(cherry picked from commit c21e184cc5cc818285c5590f36b170362defaf27)
  • 7b1b097d
    by Qiang Yu at 2022-11-15T15:41:52-08:00 
    ac/llvm: fix gfx11 fs input load for 16bit varying

Otherwise we get empty output.

Fixes: b07204d7804 ("radeonsi/gfx11: interp changes for 16bit")
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19637>
(cherry picked from commit 88b1bb326dae2d4cc7e069849f82c08f35c407d5)
  • da12da77
    by Lucas Stach at 2022-11-15T15:41:53-08:00 
    etnaviv: switch to late Z when linear PE is used

In linear PE mode the early and late depth stage do not only disagree
about the cache layout, but they seem to fundamentally disagree about
the buffer layout. When Z was written via the late stage, early tests
always show spurious zfails, even if they are not in the same draw
call. Cache flushing and pipe stalls don't help in that case.

The only option to get reliable Z tests with linear render targets is
to move all Z handling into the PE stage. Even when early Z writes
are possible, we don't know if any other draw to the same surface
needs late Z handling, so we must never use the early stage.

Fixes: 53445284a427 ("etnaviv: add linear PE support")
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19686>
(cherry picked from commit 7fe91c9f660f3b76e2c08c0824d226417231c822)
  • 25c5a09f
    by Erik Faye-Lund at 2022-11-15T15:41:53-08:00 
    docs: do not mention EGL_MESA_drm_display

This extension was removed back in 2016, but it seems we left a mention
of it in the docs.

The entire section with this extension seems kinda pointless now, so
let's drop it entirely.

Fixes: f3e23ead536 ("egl: remove remnants of MESA_drm_display")
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19719>
(cherry picked from commit 9f2c9e4c3fa57deba3da30dfa8e022c0a49c80f0)
  • aea45677
    by Erik Faye-Lund at 2022-11-15T15:41:54-08:00 
    docs: remove stale envvar-reference

This reference was left over when the envvar was removed.

Fixes: 231ccb6100e ("docs: Remove no-longer-accurate text about the xlib driver")

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19593>
(cherry picked from commit 9bfa939a5e6377efe7b610dd9b6846edeb0fdafc)
  • bed2fd36
    by Yonggang Luo at 2022-11-15T15:41:55-08:00 
    clover: empty soversion when on win32

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7675
Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Prodea Alexandru-Liviu <liviuprodea@yahoo.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19653>
(cherry picked from commit 14eef1414725692a22fa24e027e8b1fd618d57ed)
  • c2362cf4
    by Yonggang Luo at 2022-11-15T15:41:56-08:00 
    util: Test __PPC64__ for getting PIPE_ARCH_PPC_64 respond to __PPC64__ take effect

Fixes: e737a99a6fb ("Fix PPC detection on darwin")

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19674>
(cherry picked from commit 7710cc8506be571ce83751c8f7afbe6fac256b36)
  • d6031c68
    by Dylan Baker at 2022-11-16T15:02:11-08:00 
    docs: add release notes for 22.2.4
  • 80df10f9
    by Dylan Baker at 2022-11-16T15:02:11-08:00 
    VERSION: update to 22.2.4
  • dc7b3498
    by Eric Engestrom at 2022-11-17T14:04:58+00:00 
    .pick_status.json: Update to 9bd11f65238ce101bf846f5528f9088630e983f7
  • 674e20d1
    by Timothy Arceri at 2022-11-17T14:05:02+00:00 
    mesa: fix typo from adding glGetObjectLabelEXT

Fixes: 	675bcbb7a1c0 ("mesa: add EXT_debug_label support")

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19607>
(cherry picked from commit e295ee778bd6f136bdca67121f8b0fcb50b3e3d4)
  • 0dad87e4
    by Samuel Pitoiset at 2022-11-17T14:05:02+00:00 
    ac/nir: do not convert GS outputs to the expected variable size on GFX11

Outputs are always considered 32-bits.
Found by inspection.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19612>
(cherry picked from commit d2563e6600fd74dc000fdb031d17d54971ff67cb)
  • 5f582123
    by Karol Herbst at 2022-11-17T14:05:02+00:00 
    glsl: fix buffer texture type

Fixes: 3ace6b968b3 ("compiler/types: Add a texture type")
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19381>
(cherry picked from commit 87526f79db68a13ebd448cfd6b1be4b25616c801)
  • 57e8d21f
    by Alyssa Rosenzweig at 2022-11-17T14:05:02+00:00 
    panfrost: Require 64-byte alignment on imports

While Panfrost allocates linear images with strides that are a multiple of 64
bytes, other dma-buf producers on the system may not satisfy this requirement.
However, at least on v7 and newer, any image with a regular format must have a
stride that is a multiple of 64 bytes.

This fixes a real bug in an application that created a linear R8_UNORM image
with stride 480 bytes, imported it as an EGL_image, and then tried to texture
from it with the GPU. Previously, the driver allowed this situation but it
resulted in an imprecise fault from the GPU. This patch corrects the driver to
reject the import as invalid due to the unaligned stride, ensuring we never
attempt to texture from such a resource.

To implement, we add some new layout queries to centralize knowledge about the
stride alignment requirements, and we sprinkle in asserts to show how the
invariant is upheld throughout the lifecycle of image creation to texturing.

Cc: mesa-stable
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19620>
(cherry picked from commit 811f8a19469722bea32f3c539b8cf0939fe3b057)
  • 91dfc025
    by Lionel Landwerlin at 2022-11-17T14:05:02+00:00 
    anv: fixup invalid enum for nir environment

Also switching away from PIPE_

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 8c4c4c3ee1a2 ("anv: Add softtp64 workaround")
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19638>
(cherry picked from commit 68fd9d28294ee91033690b2288c55480c4125577)
  • fd23e5b2
    by Erik Faye-Lund at 2022-11-17T14:05:02+00:00 
    docs: fixup broken link syntax

Seems I got this slightly wrong when I fixed up the previous syntax
issue. Whoops, let's fix that!

Fixes: 6b3b6333915 ("docs/zink: fix and cleanup rst syntax")
Reviewed-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19574>
(cherry picked from commit ef05d28aa276aebbcf48400a5bef11de20e8b1c9)
  • 077e04bd
    by Emma Anholt at 2022-11-17T14:05:02+00:00 
    ir3/ra: Make sure we don't pick a preferred reg overflowing the file.

If we're in handle_collect()'s dst allocation and are part of a merge set
near the end of the file, our check for reg_elem_size(reg) would let us
use the preferred reg when that would immediately lead to
allocate_dst_fixed() creating an interval extending thruogh reg_size(reg)
that overflows the file.

Avoids a regression on gfxbench5/gl_5_high_off/17.shader_test in the next
commit.  No change on shader-db.

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18946>
(cherry picked from commit a39113b616099a7bcce9e62337731f040c7cad64)
  • aececb9c
    by Dylan Baker at 2022-11-17T14:05:02+00:00 
    util/glsl2spirv: fix type error in argument handling

args.Olib is set to `store_true`, which means it will always be `True`
or `False`, this means that the we always, unconditionally, add
`--keep-uncalled` to the command line.

fixes: 9786d9ef2abb45a4e832cf1347581e3ca3aae9f0

Reviewed-by: Luis Felipe Strano Moraes <luis.strano@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19449>
(cherry picked from commit 76e3b482be125cc0a2edd75bdeb14d4a6e9232fb)
  • 986a55f9
    by Dylan Baker at 2022-11-17T14:05:02+00:00 
    util/glsl2spirv: fix appending extra flags

The variable is called `extra`, but what's written is `extra - flags`,
and `flags` is undefined, so if the variable was ever passed there would
be an uncaught exception.

fixes: 9786d9ef2abb45a4e832cf1347581e3ca3aae9f0

Reviewed-by: Luis Felipe Strano Moraes <luis.strano@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19449>
(cherry picked from commit 4ffa8a9ac010f3db79b1e872f7e664a4470cb981)
  • f68e2531
    by Yonggang Luo at 2022-11-17T14:05:02+00:00 
    radv: Fixes prototypes

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19610>
(cherry picked from commit e5656f07c7a65b1d143684671cd8235af099a84c)
  • 7db1912a
    by Yonggang Luo at 2022-11-17T14:05:02+00:00 
    meson: Refactoring shared gen_vs_module_defs_normal_command out

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19610>
(cherry picked from commit f03421702cc35059e4ead86839906ddfcf05bdc7)
  • cf6a9137
    by Yonggang Luo at 2022-11-17T14:05:02+00:00 
    meson: fixes mingw-clang32 building

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19610>
(cherry picked from commit ab20ec9da76d77ae3942ecf1031366d29a4dff7b)
  • e127adf5
    by Yonggang Luo at 2022-11-17T14:05:02+00:00 
    clover: Rename *OpenCL.def to *OpenCL.def.in

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19610>
(cherry picked from commit a3b26e2758d7fce890717e2679bd0e25ca1330d2)
  • e1415b0d
    by Yonggang Luo at 2022-11-17T14:05:02+00:00 
    clover: Fixes building with mingw-x86

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19610>
(cherry picked from commit f25d483e1bd493fdf6f0f8148259ed085d705617)
  • 782af9b0
    by Yonggang Luo at 2022-11-17T14:05:02+00:00 
    meson: Fixes name_prefix for clover on mingw

Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19610>
(cherry picked from commit 67627c84a71b4c795efa6ddcffa5633ce0239f65)
  • a90e68cb
    by Samuel Pitoiset at 2022-11-17T14:05:02+00:00 
    radv: re-emit NGG culling settings when conservative rast mode is dynamic

Found by inspection.

Fixes: fbed3aed4aa ("radv: add support for dynamic conservative rasterization mode")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19651>
(cherry picked from commit 08b346e81ac262578ed0437464148be6500ea4de)
  • d9d4c97c
    by Luis Felipe Strano Moraes at 2022-11-17T14:05:03+00:00 
    meson: only enable intel-clc for x86_64 builds

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19662>
(cherry picked from commit d983827a62c77713452a9e420fd20ed39c68ef2c)
  • 238c58e7
    by Gert Wollny at 2022-11-17T14:05:03+00:00 
    nir/algeraic_opt: use double options too for lowering ftrunc@64

ftrunc@64 also might need lowering on fp64 only, especially now
that it might be introduced by nir_lower_int64.

Fixes: 29da9856826fa6a4b5117c43c78b4301a49bc6dd
   nir/lower_int64: Enable lowering of 64-bit float to 64-bit integer conversions.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19657>
(cherry picked from commit 917d992b320740d548ee0eb442e927c7ac5184fa)
  • d7ca6cce
    by Lionel Landwerlin at 2022-11-17T14:05:03+00:00 
    anv: split internal surface states from descriptors

On Intel HW we use the same mechanism for internal operations surfaces
as well as application surfaces (VkDescriptor).

This change splits the surface pool in 2, one part dedicated to
internal allocations, the other to application VkDescriptors.

To do so, the STATE_BASE_ADDRESS::SurfaceStateBaseAddress points to a
4Gb area, with the following layout :
   - 1Gb of binding table pool
   - 2Gb of internal surface states
   - 1Gb of bindless surface states

That way any entry from the binding table can refer to both internal &
bindless surface states but none of the driver allocations interfere
with the allocation of the application.

Based off a change from Sviatoslav Peleshko.

v2: Allocate image view null surface state from bindless heap (Sviatoslav)
    Removed debug stuff (Sviatoslav)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7110
Cc: mesa-stable
Tested-by: Sviatoslav Peleshko <sviatoslav.peleshko@globallogic.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19275>
(cherry picked from commit 4ceaed7839afd724b2a2f10f6879f54199c041ad)
  • 61ad5a81
    by Tapani Pälli at 2022-11-17T14:05:03+00:00 
    iris: setup stage bitmask for Wa_22011440098

Fixes: 40b66a44998 ("anv, iris: Add Wa_22011440098 for DG2")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19636>
(cherry picked from commit 210d95bdb35c2c172f0a3a9527b181f8988a627f)
  • fc57b9ac
    by Tapani Pälli at 2022-11-17T14:05:03+00:00 
    anv: setup stage bitmask for Wa_22011440098

Fixes: 40b66a44998 ("anv, iris: Add Wa_22011440098 for DG2")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19636>
(cherry picked from commit ecd4517560f135f64abf6e40acc48807b400ca41)
  • e0867504
    by noasakurajin at 2022-11-17T14:05:03+00:00 
    disable zinks shader cache when the needed functions do not exist

 Fixes: 4e14da056d6 ("zink: Enable mesa/st frontend shader caching.")
 Reviewed-by: Eric Engestrom <eric@igalia.com>

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19592>
(cherry picked from commit 7666988937ee9057b575e61437480e23a254c62a)
  • f7756794
    by Iago Toral Quiroga at 2022-11-17T14:05:03+00:00 
    v3dv: fix debug dump on BO free

We were resetting the BO struct right before dumping its data. Fix
this by moving the reset later.

Fixes: 44fa8304d45 ('v3dv: add a refcount mechanism to BOs')
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19672>
(cherry picked from commit 2c44597181e28b05d4b269ee73d4f3a59bc0a806)
  • 18a8b0a1
    by Timothy Arceri at 2022-11-17T14:05:03+00:00 
    nir: fix typo in lower_double options handling

Seems the intention was to check that both flags were not enabled
instead we were checking that the floor flag was both set and not
set so the result would always be false.

Fixes: 3749a6ecd282 ("nir: honor lower_double options for ffloor and ffract")

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19642>
(cherry picked from commit 34c52d8cb98e74a9debcd2605b377170c025582a)
  • aa63bf18
    by Lucas Stach at 2022-11-17T14:05:03+00:00 
    etnaviv: update headers from rnndb

Update to etna_viv commit 6939cfeba30c.

Cc: 22.3 mesa-stable
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19571>
(cherry picked from commit 804bf2eba2af7f3d6446361127cbc5521ae72b15)
  • 0bcf214d
    by Lucas Stach at 2022-11-17T14:05:03+00:00 
    etnaviv: fix late Z with MSAA active

On RA_WRITE_DEPTH GPUs the RA stage needs to be told that MSAA is active
when the PE Z/S stage is needed. Not sure what it does exactly, but this
fixes broken late Z on those GPUs when performing MSAA rendering.

Cc: 22.3 mesa-stable
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19571>
(cherry picked from commit 2f85d9095e3559597bc1cf0051ab6d7bc18faecd)
  • 58240683
    by Yiwei Zhang at 2022-11-17T14:05:03+00:00 
    venus: handle VK_QUERY_TYPE_PRIMITIVES_GENERATED_EXT

This change adds some docs for the query size, and has been tested with
dEQP-VK.transform_feedback.primitives_generated_query.* on supported
implementations.

Fixes: 8f7b5bf34b4 ("venus: add VK_EXT_primitives_generated_query support")

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Juston Li <justonli@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19684>
(cherry picked from commit f7d7e558c958d4057cf88dfa37f80d150d62d87f)
  • 97a017ed
    by Lionel Landwerlin at 2022-11-17T14:05:03+00:00 
    anv: bump pool bucket max allocation size

Age of Empire IV generates a shader of ~2.3Mb on DG2 which is above
the limit we currently have.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19669>
(cherry picked from commit ae76bba34a878dbd299f911bceaa9acdf01eb38a)
  • fcb0d2fa
    by Gert Wollny at 2022-11-17T14:05:03+00:00 
    r600/sfn: Fix source modifiers for ffract64

Fixes: 79ca456b4837b3bc21cf9ef3c03c505c4b4909f6
    r600/sfn: rewrite NIR backend

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19707>
(cherry picked from commit fd27b18631bff77ce864968238c4e6dbe37606a6)
  • c06dff60
    by Gert Wollny at 2022-11-17T14:05:03+00:00 
    r600/sfn: Fix f2u32 and remove backend lowring of f2u64 and f2i64

The two conversion ops are now handled in nir_lower_int64,
but the fixup for the input to f2u32 has to be handled there
and not in f2u64.

Fixes: 29da9856826fa6a4b5117c43c78b4301a49bc6dd
   nir/lower_int64: Enable lowering of 64-bit float to 64-bit integer conversions.

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19707>
(cherry picked from commit fee004215755ee56a427e2ea96a5ab6dfa1ce823)
  • 8747b272
    by Jason Ekstrand at 2022-11-17T14:05:03+00:00 
    vulkan: Handle VK_SUBPASS_EXTERNAL at the end of a subpass

Fixes: 1d726940d288 ("vulkan: Add a common CmdBegin/EndRederPass implementation")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19599>
(cherry picked from commit 4ff5051944e7a7a8bf003c331af1a73ac6c7df61)
  • 84381034
    by Jason Ekstrand at 2022-11-17T14:05:03+00:00 
    vulkan: Unconditionally add barriers for missing external subpass deps

This is a very scorched-earth approach which doesn't take into account
whether or not there are any explicitly provided dependencies.  We could
take a finer-grained approach in theory but it's unlikely to matter in
practice since you usually stall in Begin/EndRenderPass anyway.

Fixes: 1d726940d288 ("vulkan: Add a common CmdBegin/EndRederPass implementation")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6203
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7650
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19599>
(cherry picked from commit 11b2a063bf1f18b3be9542be8c229427a33c92f0)
  • 5abbb413
    by Samuel Pitoiset at 2022-11-17T14:05:03+00:00 
    radv: enable lowering of subgroup shuffle in NIR on GFX11+

VGPR allocation changed on GFX11 and this might have changed how
shared VGPRs work, so it's probably more secure to lower in NIR.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19679>
(cherry picked from commit d255bc8f05dd5f85f7154b0a3d7aaf381b47c199)
  • 6e4a46e2
    by Caio Oliveira at 2022-11-17T14:05:03+00:00 
    intel/compiler: Fix dynarray usage in intel_clc

The code builds up the dynamic array of objects (spirv_objs) and
collect pointers to each of them into another dynamic
array (spirv_ptr_objs).

If the growth of the first array cause a reallocation, it is
possible that the previous pointers end up invalid.

Fixes: 77e929a5273 ("intel/clc: allow multiple CL files to be compiled together")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19730>
(cherry picked from commit 9fd1d47aa0f19fac30d29e6ae40ed0383c1930ee)
  • be102fed
    by Caio Oliveira at 2022-11-17T14:05:03+00:00 
    intel/compiler: Fix missing tie-breaker in brw_nir_analyze_ubo_ranges() ordering code

Per Ken suggestion, use ascending order for the start offset.

Fixes: 6d28c6e52cf ("i965: Select ranges of UBO data to be uploaded as push constants.")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19731>
(cherry picked from commit 494e2edb90d06f645e09b10d2c3f6b8d274d4dba)
  • ae63ea06
    by Dave Airlie at 2022-11-17T14:05:03+00:00 
    gallivm: add coro malloc hooks earlier and always.

This fixes GALLIVM_DEBUG=asm for compute shaders, changing
the hooks after dumping causes a segfault because the
memory has already been finalised. Just add the hooks always,
and before dumping anything.

Fixes: f511d2a55337 ("gallivm: rework coroutine malloc/free callouts.")
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19714>
(cherry picked from commit fb7de303ba7cbdb4237d2ae93011bc2845114d35)
  • 2a28762d
    by Qiang Yu at 2022-11-17T14:05:04+00:00 
    ac/nir/ngg: fix nogs culling with nuw add

We should not use "nuw" here as negative add positive may wrap
around (negative is 0xffffff??).

This problem can be observed with LLVM15 (I can't see when LLVM14):
  %.neg = mul nsw i32 %31, -4
  %163 = add nuw nsw i32 %.neg, 16
  %164 = lshr i32 257, %.neg
  %165 = lshr i32 %164, %163

LLVM just assume %.neg is possitive, so pre-shift 0x01010101 by 16.
This get wrong value because we can't get back the shifted bits with
a negative shift right.

Fixes: 75dbb404393 ("ac/nir: Remove byte permute from prefix sum of the repack sequence.")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19718>
(cherry picked from commit 982b523769a75c99039deac7f832a1e10260e916)
  • 5421dd52
    by Qiang Yu at 2022-11-17T14:05:04+00:00 
    ac/nir/ngg: remove nuw for negative value add

Add negative value is possible to wrap around. I haven't seen this
"nuw" causes any problem yet, but let's remove it for safe.

Fixes: 60ac5dda82e ("ac: Add NIR lowering for NGG GS.")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19718>
(cherry picked from commit c21e184cc5cc818285c5590f36b170362defaf27)
  • de5ebe43
    by Qiang Yu at 2022-11-17T14:05:04+00:00 
    ac/llvm: fix gfx11 fs input load for 16bit varying

Otherwise we get empty output.

Fixes: b07204d7804 ("radeonsi/gfx11: interp changes for 16bit")
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19637>
(cherry picked from commit 88b1bb326dae2d4cc7e069849f82c08f35c407d5)
  • f100c77d
    by Iago Toral Quiroga at 2022-11-17T14:05:04+00:00 
    v3dv: ignore imported BOs when tracking BO memory usage

Imported BOs are not allocated by the device so we don't
update BO stats when they are imported. Therefore, we should
not be updating them when they are freed either.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19675>
(cherry picked from commit f14e2ca099cbc1c732020b6e1c30aaff4f652d1b)
  • e51b0b10
    by Alejandro Piñeiro at 2022-11-17T14:05:04+00:00 
    v3dv/bo: reset bo and then call gem close

After 'v3dv: fix debug dump on BO free' we changed the order, and this
lead to the following test
dEQP-VK.api.object_management.multithreaded_per_thread_resources.device_memory_small

v2: Expanded comment just before the reset, explaining that we need to
do the reset before we free the BO from the kernel (Iago)

Raising this assertion:
deqp-vk: ../src/broadcom/vulkan/v3dv_bo.c:281: v3dv_bo_alloc: Assertion `bo && bo->handle == 0' failed.

Fixes: 2c44597181e2 ('v3dv: fix debug dump on BO free')

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19693>
(cherry picked from commit ec1cdc13d5cf6026692bf3765be3aeceb511e6d6)
  • f3c55ddd
    by Guilherme Gallo at 2022-11-17T14:05:04+00:00 
    ci: Update piglit with s3 support

With new S3 support, we can use JWT-only server interaction via the
removal of `role-session` and `minio-host` arguments from PIGLIT_ARGS in
YAML.
This parameter change will come in a later commit.

Solved Conflicts:
	.gitlab-ci/container/build-piglit.sh
	.gitlab-ci/image-tags.yml

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
(cherry picked from commit 70ce1dcacc92a816322082c8695569b6a91a1810)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19734>
  • 03d1edf3
    by Guilherme Gallo at 2022-11-17T14:05:04+00:00 
    ci: Update piglit-traces tests expectations

Found some:
- crashes in zink, softpipe
- fails in a630-restricted
- unexpectedpass in broadcom
    - fixed by https://gitlab.freedesktop.org/mesa/piglit/-/merge_requests/730

More details in the test expectations files comments.

Solved Conflicts:
	src/gallium/drivers/zink/ci/zink-lvp-skips.txt

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
(cherry picked from commit a108e4f70cc61a6b9119cd766d51c9b596e07c7f)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19734>
  • 935ba13e
    by Guilherme Gallo at 2022-11-17T14:05:04+00:00 
    ci: Update ci-fairy in CI, rootfs and containers

ci-fairy is pulverized in possible different versions at Mesa CI. This
commit updates all of them to the version that migrates minio to s3.
Also, trigger the build of base and test containers, as both uses
ci-fairy as well.

Solved Conflicts:
	.gitlab-ci/image-tags.yml

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
(cherry picked from commit a04ed2f971dccbf8b24f084ac9baaf8b299944a0)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19734>
  • 76dc28e3
    by Benjamin Tissoires at 2022-11-17T14:05:04+00:00 
    CI: convert to use the new S3 server instead of the legacy minio

We don't need to login anymore, but we can't use plain minio commands
now. `ci-fairy` got a helper as `s3cp` to keep an almost identical
API.

Solved Conflicts:
	.gitlab-ci/common/init-stage2.sh
	.gitlab-ci/container/lava_build.sh
	.gitlab-ci/prepare-artifacts.sh
	src/amd/ci/traces-amd.yml
	src/freedreno/ci/traces-freedreno.yml
	src/gallium/frontends/lavapipe/ci/traces-lavapipe.yml

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
(cherry picked from commit 67cee534a88c95a8eb6839f7bcf28a5e6dac8fbf)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19734>
  • 34319c7d
    by David Heidelberg at 2022-11-17T14:05:04+00:00 
    ci/freedreno: disable antichambers trace

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7668
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19627

Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
(cherry picked from commit f562e37c9325cc107d4f02026946acc14c75b323)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19734>
  • 053f0f59
    by Guilherme Gallo at 2022-11-17T14:05:04+00:00 
    freedreno/ci: Skip civilization-v/CivilizationV-trim trace

It is been flaking, the following jobs are expecting the same checksum,
but produced different ones.

- https://mesa.pages.freedesktop.org/-/mesa/-/jobs/31762457/artifacts/results/summary/results/trace@freedreno-a630@civilization-v@CivilizationV-trim--s705-761-f762-v20201203-v2.trace.html
- https://mesa.pages.freedesktop.org/-/mesa/-/jobs/31763571/artifacts/results/summary/results/trace@freedreno-a630@civilization-v@CivilizationV-trim--s705-761-f762-v20201203-v2.trace.html

Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19734>
  • 4734048d
    by David Heidelberg at 2022-11-17T14:05:04+00:00 
    ci/update_traces_checksum.py: check if checksum is in the array, not it's value

Fixes: 45eda069531a ("ci: introduce update_traces_checksum.py")

Reviewed-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19510>
(cherry picked from commit e1d40d11f5a6083c19582c2e138f453cc2c7d4e4)
  • f55b090c
    by Lucas Stach at 2022-11-17T14:05:04+00:00 
    etnaviv: fix shader register control with MSAA

Apparently MSAA doesn't only add another input, but it also increases
required temporaries by one. Simple programs where the register demand
is given by the number of inputs did work fine, while more complex ones,
where register demand is given by the number of temporaries exhibit
rendering issues without this fix.

Cc: 22.3 mesa-stable
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19582>
(cherry picked from commit e65d266529f1e95252bacd529a830b9d4d50000f)
  • 2aae392b
    by Lucas Stach at 2022-11-17T14:05:04+00:00 
    etnaviv: switch to late Z when linear PE is used

In linear PE mode the early and late depth stage do not only disagree
about the cache layout, but they seem to fundamentally disagree about
the buffer layout. When Z was written via the late stage, early tests
always show spurious zfails, even if they are not in the same draw
call. Cache flushing and pipe stalls don't help in that case.

The only option to get reliable Z tests with linear render targets is
to move all Z handling into the PE stage. Even when early Z writes
are possible, we don't know if any other draw to the same surface
needs late Z handling, so we must never use the early stage.

Fixes: 53445284a427 ("etnaviv: add linear PE support")
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19686>
(cherry picked from commit 7fe91c9f660f3b76e2c08c0824d226417231c822)
  • 86e7eb18
    by Erik Faye-Lund at 2022-11-17T14:05:04+00:00 
    docs: do not mention EGL_MESA_drm_display

This extension was removed back in 2016, but it seems we left a mention
of it in the docs.

The entire section with this extension seems kinda pointless now, so
let's drop it entirely.

Fixes: f3e23ead536 ("egl: remove remnants of MESA_drm_display")
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19719>
(cherry picked from commit 9f2c9e4c3fa57deba3da30dfa8e022c0a49c80f0)
  • 62a6fd39
    by Erik Faye-Lund at 2022-11-17T14:05:04+00:00 
    docs: remove stale envvar-reference

This reference was left over when the envvar was removed.

Fixes: 231ccb6100e ("docs: Remove no-longer-accurate text about the xlib driver")

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19593>
(cherry picked from commit 9bfa939a5e6377efe7b610dd9b6846edeb0fdafc)
  • 559530f1
    by Yonggang Luo at 2022-11-17T14:05:04+00:00 
    clover: empty soversion when on win32

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7675
Cc: mesa-stable

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Prodea Alexandru-Liviu <liviuprodea@yahoo.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19653>
(cherry picked from commit 14eef1414725692a22fa24e027e8b1fd618d57ed)
  • 64b775c9
    by Samuel Pitoiset at 2022-11-17T14:05:04+00:00 
    aco: fix p_interp_gfx11 to not overwrite SCC

s_wqm_b64 clobbers SCC.
Found this while working on dual source blending.

Fixes: 6113ee650a2 ("aco/gfx11: fix FS input loads in quad-divergent control flow")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19747>
(cherry picked from commit 369c9b642524cb543c59d006ad7e3ce089fa6879)
  • 7f9e9239
    by Sil Vilerino at 2022-11-17T14:05:04+00:00 
    frontend/va: Fix WIN32 VA_DRIVER_INIT_FUNC declaration. Remove declspec as it uses .def file

Fixes: b557ceb7 ("frontends/va: Add windows VA frontend support via vl_winsys_win32 and libva-win32")
Closes: #7702

Signed-off-by: Sil Vilerino <sivileri@microsoft.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19754>
(cherry picked from commit d81e40b20c386a79d643fc6c14c3cfb8351daa09)
  • e39b25b7
    by Yonggang Luo at 2022-11-17T14:05:04+00:00 
    util: Test __PPC64__ for getting PIPE_ARCH_PPC_64 respond to __PPC64__ take effect

Fixes: e737a99a6fb ("Fix PPC detection on darwin")

Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19674>
(cherry picked from commit 7710cc8506be571ce83751c8f7afbe6fac256b36)
  • a44264e3
    by Sil Vilerino at 2022-11-17T14:05:04+00:00 
    ci: Add va frontend to windows-build-rules

Fixes: 2d504bc5 ("CI: Add gallium-va and video-codecs in windows-vs2019 and debian-mingw32-x86_64")

Signed-off-by: Sil Vilerino <sivileri@microsoft.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19767>
(cherry picked from commit 912c72eda5a4284bf8ef32f93a708c65083f7722)
  • d09f780b
    by José Roberto de Souza at 2022-11-17T14:05:04+00:00 
    iris: Set priority for replaced engine context

The replace_kernel_ctx() code path was not setting back the context
priority.

Fixes: 5c4c8bdc4c54 ("iris/batch: Add support for engines contexts")
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19687>
(cherry picked from commit cd159c7d6c18d1bbde019fda68cc17922389a767)
  • 449eb079
    by Alyssa Rosenzweig at 2022-11-17T14:05:04+00:00 
    panfrost: Fix reference counting with batch->resources

Refactor accesses to batch->resources to happen through safe helpers
that update the appropriate bookkeeping. This makes it obvious that (in
particular) reference counts are updated when they should be.

The functional change is that we are now correctly unreferencing
resources during shadowing, fixing a leak of shadowed resources.

Closes: #7362
Fixes: 2d8f28df731 ("panfrost: Replace resource shadowing flush")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reported-by: Mastodon, apparently
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19753>
(cherry picked from commit 42212a9bfdab8381beb9206b5d2551344c71d584)
  • 09623eb2
    by Alyssa Rosenzweig at 2022-11-17T14:05:04+00:00 
    panfrost: Use PIPE_CAP_VERTEX_ATTRIB_ELEMENT_ALIGNED_ONLY

..instead of 4BYTE_ALIGNED_ONLY. This is more correct and avoids
needless repacking. Noticed in Firefox, which was hitting the vbuf
translate path.

Fixes: e03622e50fc ("panfrost: Set STRIDE_4BYTE_ALIGNED_ONLY")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19761>
(cherry picked from commit c567e5528fd11db461f076fccd63ec1b01b96b36)
  • a5c8473a
    by Samuel Pitoiset at 2022-11-17T14:05:04+00:00 
    aco: fix missing SCC for p_interp_gfx11 in emit_interp_mov_instr()

Fixes: 369c9b64252 ("aco: fix p_interp_gfx11 to not overwrite SCC")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19777>
(cherry picked from commit 5a3cc2d453149954923abf3d1455e8fe44e5788a)
  • 30ea71b9
    by Gert Wollny at 2022-11-17T14:05:04+00:00 
    r600/sfn: Honor shader key w.r.t. atomic counter layout

Fixes: 79ca456b4837b3bc21cf9ef3c03c505c4b4909f6
        r600/sfn: rewrite NIR backend

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19755>
(cherry picked from commit 0ccf7ed7916bee67c8028bf251ca483b1be07d57)
  • df117562
    by Gert Wollny at 2022-11-17T14:05:04+00:00 
    r600/sfn: Fix location for reading cube array image dimensions

Fixes: 79ca456b4837b3bc21cf9ef3c03c505c4b4909f6
        r600/sfn: rewrite NIR backend

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19755>
(cherry picked from commit 35d6e290e4119e020e8072ce4ad97bc4e9cd269b)
  • 578f8437
    by Samuel Pitoiset at 2022-11-17T14:05:04+00:00 
    aco: add p_dual_src_export_gfx11 for dual source blending on GFX11

Dual source blending must be in strict WQM mode.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19643>
(cherry picked from commit bb90d29660bb44326188809da2deec0675c8264a)
  • 1f23f529
    by Samuel Pitoiset at 2022-11-17T14:05:04+00:00 
    aco: fix dual source blending on GFX11

Assembly looks similar to LLVM.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19643>
(cherry picked from commit fb781bfb0a5a401b5210d613479bbdfb90e94790)
  • 7d19be3e
    by Jonathan Gray at 2022-11-17T14:05:05+00:00 
    util: include sys/time.h for timespec functions

When the futex code moved it removed an include which broke the build
on OpenBSD.

Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Fixes: 095dfc6caa2 ("util: Move the implementation of futex_wake and futex_wait from futex.h to futex.c")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19735>
(cherry picked from commit fe851d7759fca4c61245b142bb90dd561ae818f9)
  • e566cd6a
    by Michel Dänzer at 2022-11-17T14:05:05+00:00 
    Revert "egl/glx: add fallback for zink loading"

This reverts commit 2569215f43f6ce71fb8eb2181b36c6cf976bce2a.

Conflicts:
	src/egl/main/eglapi.c
	src/glx/glxext.c

It broke the fallback to swrast in some cases where zink can't work.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7333
Fixes: 2569215f43f6 ("egl/glx: add fallback for zink loading")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19681>
(cherry picked from commit 71a0a386b5df53431f1fbc9a02a78f7af22611a6)
  • bbcde41c
    by Samuel Pitoiset at 2022-11-17T14:05:05+00:00 
    aco: fix FS inputs loads in WQM with 16-bit

p_wqm needs to use the same size.

Fixes: 16d2c7ad557 ("aco/gfx11: perform FS input loads in WQM")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19788>
(cherry picked from commit 50fe37070fb6b35a1468297f48021d58a9c94463)
  • 7964efaa
    by Samuel Pitoiset at 2022-11-17T14:05:05+00:00 
    radv: suspend/resume XFB queries with NGG for meta operations

XFB queries enable primitives generated queries with NGG and meta
operations shouldn't be counted.

Reproduced on GFX10.3 by forcing NGG streamout.

Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19785>
(cherry picked from commit 62356e71f431684008efc7e471bddb1ada9801a3)
  • 3de72d74
    by Eric Engestrom at 2022-11-17T14:05:05+00:00 
    ci: avoid triggering vc4 & v3d tests on v3dv-only MRs

There are a lot of vulkan-only MRs, so we can save a lot of CI resources
by not running GL tests as well.

Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Reviewed-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19680>
(cherry picked from commit bc286e8586976013936dea71dd66c4e3a3ff4605)
  • 76e3938f
    by Eric Engestrom at 2022-11-17T17:18:07+00:00 
    VERSION: bump for 22.3.0-rc3
  • 6dec3d47
    by Timo Aaltonen at 2022-11-18T10:59:56+02:00 
    rules: Disable building llvmpipe on riscv64. (Closes: #1004039)
  • aa7bf18e
    by Timo Aaltonen at 2022-11-18T11:00:38+02:00 
    Merge branch 'upstream-unstable' into debian-unstable
  • 142c3e1d
    by Timo Aaltonen at 2022-11-18T11:01:08+02:00 
    version bump
  • 915ff4ec
    by Timo Aaltonen at 2022-11-18T11:45:52+02:00 
    patches: Refreshed.
  • 6dd314b0
    by Timo Aaltonen at 2022-11-18T12:03:18+02:00 
    release to sid
  • adbb7d6e
    by Timo Aaltonen at 2022-11-18T12:05:13+02:00 
    Merge branch 'upstream-experimental' into debian-experimental
  • 1115584d
    by Timo Aaltonen at 2022-11-18T12:05:50+02:00 
    version bump
  • b5013db6
    by Timo Aaltonen at 2022-11-18T12:07:22+02:00 
    Merge tag 'mesa-22.2.4' into debian-experimental

mesa-22.2.4
  • 9ead4ee4
    by Timo Aaltonen at 2022-11-18T12:19:03+02:00 
    Merge branch 'debian-unstable' into debian-experimental
  • 449a3ff0
    by Timo Aaltonen at 2022-11-18T12:23:35+02:00 
    fix-symbols.diff: Dropped, upstream.
  • 90b79f80
    by Timo Aaltonen at 2022-11-18T12:56:43+02:00 
    rules: Limit anv ray-tracing support to amd64.
  • 5d7efb07
    by Timo Aaltonen at 2022-11-18T13:00:36+02:00 
    release to experimental

30 changed files:

The diff was not included because it is too large.

—
View it on GitLab.
You're receiving this email because of your account on salsa.debian.org. Manage all notifications · Help

Reply to: