The only change since the previous kernel is the updated dtb for
dragonboard-820c, which allows the GPU firmware to be found.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37478>
Instead of hardcoding the A/B slot selection, allow an optional
environment variable (LAVA_FASTBOOT_CMD) for devices that need it.
This is used for sm8350-hdk to select slot A, but can be omitted
for devices that don’t support A/B slots.
Update the lava-trigger container to pull in
1b521e7b38.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37478>
Non-PCI devices are responsible for checking device matches
themselves by overriding can_present_on_device on wsi_device
to take the non-blitting WSI path.
We can allow on-device presentation for both the display
controller, and any panthor node. The simplest way of
achieving that is allowing this for all devices with bus
type platform, as there's nothing else that has this
property.
Reviewed-by: Christoph Pillmayer <christoph.pillmayer@arm.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37543>
This is more in line with similar opcodes like umul_32x16.
Also change its const expr: the masking based on bit size was
unnecessary as it is only defined for 32 bits. Use simple casts instead.
Signed-off-by: Job Noorman <jnoorman@igalia.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37863>
Specify an amd_ip_type when initializing radv_cmd_stream
so that we know which packets are valid for the HW IP type
of a given command stream.
This field should be used instead of radv_cmd_buffer::qf when
emitting packets to a command stream.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37775>
Refactor the functions to make it clear what they do:
copy_memory_to_image -> gfx_or_compute_copy_memory_to_image
copy_image_to_memory -> compute_copy_image_to_memory
copy_image -> gfx_or_compute_copy_image
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37775>
There's no reason to compare the pointer before writing it here any
more. All that does, is make us trigger an unnessecary Valgrind error
with a branch based on an undefined value if the passed pointer wasn't
properly initialized here.
Fixes: b3133e250e ("gallium: add pipe_context::resource_release to eliminate buffer refcounting")
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37842>
arr[0] and arr[1] aren't really the two allocated sampler views any more
at this point. To make it possible to reason about these pointers,
assign the created objects to speparate pointers, so we can free them
from those instead of having to track which array indices they're stored
in.
Fixes: 73da0dcddc ("gallium: eliminate frontend refcounting from samplerviews")
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37839>
Instead of crashing with segfault within upload_shader_desc_info, allow
the caller to handle the alloc failure (they already do).
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37861>
Better not cap at slab_size, otherwise the bo cache is purely wasting
memory as it could only be used for up to slab_size upon re-use. This
change unlocks the pool alloc limit to the actual bo size.
Dropped the redundant pair of parentheses to stay in one line.
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37861>
Video encode feedback queries have a configurable set of feedback bits,
specified in VkQueryPoolVideoEncodeFeedbackCreateInfoKHR::encodeFeedbackFlags.
Only the bits specified should be output when retrieving results.
Fixes: 1d74661dfd ("radv: add encoder queue support pieces and encoder queries.")
Reviewed-by: David Rosca <david.rosca@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37774>
Drivers need to track which feedback bits were set at creation time to
output the correct things in vkGetQueryPoolResults().
Cc: mesa-stable
Reviewed-by: David Rosca <david.rosca@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37774>
vkd3d-proton (DX12) and Zink (GL) have always been enabling features
that require the global BO list to be enabled.
Since DXVK 2.7+ (August 2025), it's also always enabled by default for
DX9-11 games (because it requires BDA now).
The global BO list used to decrease performance in the past mostly
because of bad memory management in AMDGPU, but it seems the situation
slightly improved since. Though, there might still some workloads that
hit the issue, but I think it should be mostly good overall.
This introduces RADV_DEBUG=nobolist to disable the global BO list
when no features require it.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6957
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/2331
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35075>
The implementation must use the data in VkPipelineBinaryInfoKHR when
provided instead of importing binaries from libraries.
This fixes a memleak with shaders found with ASAN.
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37844>
