For example, the proof for this pattern
(('bcsel', ('flt', 'a@32', 0), 'b@32', 'c@32'), ('fcsel_ge', a, c, b)),
would be
bcsel(a < 0, b, c)
bcsel(!(a < 0), c, b)
bcsel(a >= 0, c, b)
fcsel_ge(a, c, b)
However, !(a < 0) => (a >= 0) is well known to produce different
results if `a` is NaN.
Instead of that replacement, use this replacement:
bcsel(a < 0, b, c)
bcsel(-0 < -a, b, c)
bcsel(0 < -a, b, c)
fcsel_gt(-a, b, c)
This is NaN-safe and exact.
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Fixes: 0f5b3c37c5 ("nir: Add opcodes for fused comp + csel and optimizations")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17048>
(cherry picked from commit a2a2fbc510)
This was missing, and the added validation caught it.
Fixes: 708c47e663 ("nir: Validate nir_tex_instr::dest_type bitsize")
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17172>
(cherry picked from commit 4309e09d6f)
point size min/max values are provided through the state vars, so ensure
these are always applied in order to respect ARB_point_parameters
cc: mesa-stable
Acked-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17145>
(cherry picked from commit 4830cc77cb)
As it turns out, MOVs weren't the only instructions that blocked precise
flags propagation in the transition to nir-to-tgsi.
This commit fixes some rendering regressions caused by a4a34cd3.
Fixes: a4a34cd3
Signed-off-by: Italo Nicola <italonicola@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collanora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17144>
(cherry picked from commit 42a1264951)
wezterm in fullscreen 4k was exceeding the xcb max request size
on the put image with llvmpipe. This fixes it to send sub-images,
the Xlib put image used in glx does this internally, but not
the xcb one, so just do it in sections here.
Cc: mesa-stable
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17155>
(cherry picked from commit e6082ac62e)
This was missing the message, breaking UBO-to-push and who-knows-what-else, when
enabling fp16 const buffers.
Fixes: 3dc2095b07 ("pan/bi: Model LD_BUFFER instructions")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17101>
(cherry picked from commit 5ee1179c94)
If you set a name of on a swapchain object, because the base object
struct has not been initialized with a VkDevice,
vk_object_base_finish() will segfault when trying to free the object
name.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: cb1e0db23e ("vulkan/wsi: Make wsi_swapchain inherit from vk_object_base")
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17165>
(cherry picked from commit efc398c722)
Fix the transform to make sure it doesn't disturb the depth range
of the blitted image. Set the Z coordinates of the vertices
by hand instead of relying on the transform to do it.
This is a pre-requisite to Zink always enabling depth clamping.
Fixes: 26c6640835
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16929>
(cherry picked from commit 810135fb42)
Conflicts:
src/gallium/drivers/lima/ci/lima-fails.txt
It seems recent versions of Sphinx has started issuing a warning when
language isn't set to a recognized value. Since we're building with
warnings as errors on CI, this causes issues for people.
We're writing English docs, so we should set this to English. I probably
confused this setting with the highlight_language setting or something
when I set this up initially.
Fixes: 00cd1346bf ("docs: use sphinx")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6711
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17159>
(cherry picked from commit ee9c3d2625)
r300 and r400 have strict rules with swizzles, so we
will need to convert swizzle back.
Operating on 0, 1, H in this case unnecessarily makes
rest of r300 overly complicated.
(also it's not currently able to handle this)
helps with:
deqp-gles2@functional@shaders@random@exponential@fragment@24
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17117>
(cherry picked from commit 6cbb19110b)
Whether a sync object is used cannot depend on where the batch is
submitted from, remove the in_sync and out_sync fields from
panfrost_batch_submit.
Always use an output syncobj, this is required for glFinish to work
correctly. This could be skipped for batches which another batch
depends on, but because of the existence of empty batches which emit
no job, doing so is not trivial.
Never use an input syncobj. There appears to be no point to this, the
kernel driver does implicit sync anyway.
Fixes "seconds per frame" rendering with Neverball; previously, every
batch was submitted with out_sync=0, so DRI's frame throttling could
do nothing. New jobs would keep getting submitted until more than a
thousand were queued in the kernel, which increased rendering latency
for the compositor far beyond acceptable levels.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16966>
(cherry picked from commit d8803c724b)
v2: Do the synchronization in the correct place. Noticed by Curro.
Fixes: b5fa43952a ("intel/fs: Better handle constant sources of FS_OPCODE_PACK_HALF_2x16_SPLIT")
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Tested-by: Felix DeGrood <felix.j.degrood@intel.com> [v1]
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17037>
(cherry picked from commit 676acfe956)
usually inlining is optimal for cpu drivers since the majority of
time is spent in the shaders, and any amount of reduction to shader code
will be optimal
if, however, the shaders are still really big after inlining, this improvement
will be negated by the insane amount of time spent doing stupid llvm optimizer
passes, so check post-inline size to see whether it exceeds a size threshold
lavapipe release build - 1700% improvement
* spec@arb_tessellation_shader@execution@variable-indexing@tcs-output-array-vec4-index-rd-after-barrier
before: 142.15s user 0.42s system 99% cpu 2:23.14 total
after: 8.60s user 0.07s system 99% cpu 8.677 total
fixes#6647
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16977>
it's important to get the full mask in order to accurately provide
dependency info
Acked-by: Dave Airlie <airlied@redhat.com>
Reviewed-By: Tatsuyuki Ishi <ishitatsuyuki@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17111>
by passing through the draw buffers, more accurate barriers can be generated
to ensure synchronization for both the draw buffer scopes and
descriptor binding scopes
ref #6597
Acked-by: Dave Airlie <airlied@redhat.com>
Reviewed-By: Tatsuyuki Ishi <ishitatsuyuki@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17111>
this clamps the layer count to the smallest number of layers rather
than the largest
cc: mesa-stable
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17061>
(cherry picked from commit f74df6205d)
Conflicts:
src/gallium/drivers/zink/zink_framebuffer.h
having this so far down in the BO allocation path meant that slabs were
getting demoted to device-local and then stored as BAR in the pb cache,
which broke the cache pretty badly
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17093>
(cherry picked from commit 5750050432)
the flags in vk_domain_from_heap() are the MINIMUM required flags for
the heap, but there may be other flags present, so ensure those are picked
up as expected
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17093>
(cherry picked from commit d63e04e583)
flush_region is relative to the map, so passing in the offsets again
breaks the flush (if the flush hook is implemented correctly)
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17095>
(cherry picked from commit 28d23c9e26)
If multisample is enabled and alpha testing happens, the
branch can jump out of the fragment shader before the other
samples are generated. Just don't take the branch optimisation
post alpha test if multisample is enabled.
This should fix some rendering bugs in kicad with multisample
enabled.
Cc: mesa-stable
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17049>
(cherry picked from commit 6e5d126a65)
The prior behavior can ignore certain failure result, and might also
clean up queues that are never initialized.
Fixes: ddd7533055 ("venus: initial support for queue/fence/semaphore")
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16731>
(cherry picked from commit cb8dfa4966)
The failure path was never hit though, and will not either.
Fixes: 65abd1d4ae ("venus: implement vn_buffer_cache_entries_create")
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16731>
(cherry picked from commit 01a0bfc3f9)
This was an optimization done a while ago that doesn't seem to be having
much of an impact anymore, and on the other hand, causes all sorts of
breakage with queries, as many of our HW counters don't get incremented
when rasterization is disabled.
This fixes a bunch of issues Zink has with ANV, but more importantly, it
fixes upcoming CTS tests:
dEQP-VK.transform_feedback.primitives_generated_query.*.empty_frag.*
dEQP-VK.transform_feedback.primitives_generated_query.*.no_attachment.*
dEQP-VK.transform_feedback.primitives_generated_query.*.color_write_disable_*
Cc: mesa-stable
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17038>
(cherry picked from commit 4666ef720e)
Conflicts:
src/gallium/drivers/zink/ci/zink-anv-tgl-fails.txt
CI file was deleted as it doesn't exist on 22.1
These are normally only set once because it's constant across the entire
renderpass, but they're trashed by the 3d store path because it needs to
store to CCU instead of GMEM. Therefore we need to save/restore them. Do
it in a way compatible with #5181.
Fixes: b157a5d ("tu: Implement non-aligned multisample GMEM STORE_OP_STORE")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17058>
(cherry picked from commit cba6da2b21)
