Without this application could allocate protected memory buffers
and use it in non-protected queues.
Fixes: c6a91f1695 ("anv: add new heap/pool for descriptor buffers")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28769>
These were added, likely by mistake, in 54050d8844, and never used, but
they are causing problems because they overwrite the proper files from
src/gallium/drivers/llvmpipe/ci/, causing CI failures because the wrong
file is used.
Fixes: 54050d8844 ("svga/ci: land vmware mesa-ci lava farm")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28773>
KHR_video_encode_queue is dependent on KHR_video_queue.
Not exposing this makes encode-only usecases not functional,
eg. SteamVR Link.
Signed-off-by: Joshua Ashton <joshua@froggi.es>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28779>
b-frames can be considered as reference, so the NAL type
should refer to reference type and either RASL or TRAIL
depending on the irap_pic_flag.
Fixes: 72f52329c ("vulkan/video: add a nal_unit lookup for hevc")
Reviewed-by: Hyunjun Ko <zzoon@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28657>
Re-enable the farm after detecting that job failures were caused
by a new CI host to be deployed announcing rpi4 devices from other
host.
Signed-off-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28776>
We needed to halve the number of tests in `zink-tu-a618` due to the
reduced number of a618-limozeen in Collabora's farm, let's create a job
in the nightly pipeline to keep track of zink tests.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28770>
Recently, `sc7180-trogdor-lazor-limozeen-cbg-2` was retired from the
farm, so we need to rebalance the jobs again. The current flow is
causing the job queue to be huge.
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28770>
The ffloor@64 case, which lowers to use ffract, is already ignored if
nir_lower_dfract is set. Do the same thing for ftrunc@64 and ffract@64
and let nir_lower_doubles take care of them directly instead.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28702>
For purposes of address reports, it makes far more sense to report the
actually bound range rather than the full bo_size. RMV code used
effective size, so reproduce that here.
No other code looks at bo_size, so this should be quite safe.
Also fixes a theoretical correctness issue where plane aspect for
DISJOINT image was not passed to GetImageMemoryRequirements2 in internal
code.
Signed-off-by: Hans-Kristian Arntzen <post@arntzen-software.no>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/10996
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28718>
panfrost_initialize_surface is called when a surface is written to,
and marks that surface as valid. If the surface is a depth buffer
with a separate stencil, that separate stencil should also be marked
as valid; otherwise, readpixel will skip reading it (and hence the
stencil data will be read as uninitialized). This only affects
DEPTH32F_STENCIL8 formats.
Cc: mesa-stable
Signed-off-by: Eric R. Smith <eric.smith@collabora.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28738>
This prepares ac_nir_lower_ngg_ms for the possibility of packing
two 16-bit outputs into the same 32-bit output slot, by handling
the high_16bits flag in NIR IO semantics.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28704>
Previously, it would hit an assertion when used with scalarized
IO, because the scalarization will split the primitive indices
store into smaller, per-component stores.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28704>
This was an NV_mesh_shader-only feature and we should have already
removed it. We don't want to carry it forward anymore, because it
would needlessly complicate implementing new features.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28704>
Enable MSAA by default and update CI expectation.
The spec@ext_framebuffer_multisample test is buggy and switches from fail
to skip once we actually expose multisample capabilities, while it should
also have skipped before.
Some of the *copyteximage and spec@!opengl 1.1@depthstencil-default_fb*
tests require multisample resolve blits with parameters that can not be
supported natively with the RS engine and can also not be emulated by
u_blitter without texture multisampling, which is not available in the
GC2000 hardware. Accept those failures on this HW generation.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <cgmeiner@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22439>
These pNext structs are part of VK_KHR_swapchain which is core Vulkan
1.1 but they were missing. Based on ANV, RADV and NVK.
Signed-off-by: Valentine Burley <valentine.burley@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28307>
We must disable features before we fill our internal etna_specs
struct with etna_get_specs(..).
Fixes: d600b45ccc ("etnaviv: Switch to etna_core APIs")
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28757>
We take advantage of the needs_quad_helper_invocation information to
only enable the PER_QUAD TMU access on Fragment Shaders when it is needed.
PER_QUAD access is also disabled on stages different to fragment shader.
Being enabled was causing MMU errors when TMU was doing indexed by vertexid
reads on disabled lanes on vertex stage. This problem was exercised by some
shaders from the GTK new GSK_RENDERER=ngl that were accessing a constant buffer
offset[6], but having PER_QUAD enabled on the TMU access by VertexID was
doing hidden incorrect access to not existing vertex 6 and 7 as TMU was
accessing the full quad.
cc: mesa-stable
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28740>
Checking for the lib flags is unnecessary because in this case
next_stage is NONE and it's already handled correctly by the helper.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28672>
José Roberto de Souza
Mike Blumenkrantz
Connor Abbott
Eric Engestrom
Samuel Pitoiset
Joshua Ashton
Stéphane Cerveau
Jose Maria Casanova Crespo
Constantine Shablia
Guilherme Gallo
Rob Clark
David Heidelberg
Surafel Assefa
Iván Briano
Rhys Perry
Hans-Kristian Arntzen
Eric R. Smith
Timur Kristóf
Juan A. Suarez Romero
Lucas Stach
Valentine Burley
Christian Gmeiner