Tessellation shader which are using indirect
addressing for tesslevels e.g
gl_TessLevelOuter[gl_InvocationID] = tessLevelOuter;
are crashing because gl_TessLevelOuter is now a
compact array variable and nir expects a constant
array index into the compact array variable.
This patch handles such cases.
This fixes MR 21940
Fixes: 84006587d7 ("glsl: Delete the lower_tess_level pass.")
Tested with glretrace
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25773>
(cherry picked from commit 8cfb46e27d)
only certain drivers can successfully run an xserver with implicit modifier
handling, and the rest will have broken rendering
until such time that a vk extension emerges to handle this more widely,
break this interop for drivers where it's already broken
fixes#9819
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25299>
Fix regex pattern in zink_extensions to use proper escaping. While the original
code works, it relies on Python ignoring incorrect syntax and fallbacking to
passing through the escape. Current behaviour results in SyntaxWarning
whenever the code is used.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25881>
Add `PAN_MAX_AFBC_PACKING_RATIO` variable. If the ratio (size of the
packed resource) / (size of the sparse resource) * 100 is above that
value, we don't care about packing it.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
Add a method to dump a specific superblock given its index (header
and few first bytes from the body) and a method to dump a resource
to a ppm file.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
Add `forcepack` flag that will force conversion to AFBC-packed right
after a texture is uploaded when possible. We only pack 2D resources
larger than 32x32 as of now.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
After converting a texture from one modifier to another, there is
no reason to force the modifier to stay constant afterwards. Set
back `modifier_constant` to false because it is changed by
`resource_setup` as it is causing issues when implementing AFBC
packing.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
Make sure we don't convert the texture for nothing by only
legalizing when creating a batch instead of on surface creation.
Also, to avoid recursive blit, we need to legalize the destination
resource before blitting.
Finally, make sure the resource has a sparse memory layout if
AFBC compressed. The GPU doesn't support rendering to a AFBC-packed
texture.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
When the GPU is converting a texture from linear/tiled to compressed
AFBC, it uses a sparse memory layout. That means that the
superblocks are stored starting at intervals equal to the size of an
uncompressed superblock. When memory usage needs to be optimized, it
is possible to pack the resource by trimming each superblock as much
as possible. The GPU will still be able to read from these packed
textures, but won't be able to write directly to them. If the
layout is AFBC-tiled, the packing process will also de-tile as
tiled+packed is not supported by Mali GPUs.
No new modifier flag has been added as the absence of the
`AFBC_FORMAT_MOD_SPARSE` flag means the resource will be packed.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
The compute shader is going through all the AFBC header blocks and
calculating the size of valid data by summing up the subblock
sizes (and taking into account solid color, uncompressed mode and
copy blocks). The result is written into the given buffer (array
of `pan_afbc_block_info`).
The size is rounded up to the alignment parameter directly in the
shader.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
A few compute shaders are needed to support AFBC-packing. Here is
just the boilerplate to create, compile and retrieve the shaders.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
For an upcoming feature (AFBC packing), we're gonna run internal
compute shaders on resources. This commit just split the internal
of `panfrost_launch_grid` so we can run the shader on a given
batch.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25012>
Don't try to handle emulation bytes and only parse first NAL unit
if the packed header has no emulation bytes.
This fixes parsing packed headers from gstreamer.
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25565>
This is used for parsing VA packed headers and those can be without
emulation prevention bytes.
Add emulation_bytes argument to vl_rbsp_init and skip all emulation
bytes handling when set.
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25565>
VCN_2_2_0 should not use tmz context buffer.
Fixes: ffbbf23e
A minor fix for above commit to use the original comparison logic "<"
instead of "<="
Cc: mesa-stable
Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25856>
si_init_cs_preamble_state() function does not exist anymore and
PA_SC_WINDOW_SCISSOR_TL is set to 0,0. Update the comments with
this information.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25148>
We have competent lowering in NIR already available.
Drivers exposing CAP_DOUBLES but not SHADER_CAP_DROUND:
- d3d12 (NIR lowers ~0 if the underlying impl doesn't do floats)
- svga (Now sets the NIR lowering options)
- softpipe (Doesn't do GL4 so you can't use doubles anyway)
- llvmpipe (Lowers dround_even in NIR and passees the rest through
successfully)
- zink (NIR lowers ~0 if the underlying impl doesn't do floats,
otherwise passes things through successfully, except needed
dround_even lowering to avoid lavapipe regression with
native doubles)
- r600 (sets NIR rounding lowering flags, and lowers all fsign)
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25777>
Right now some drivers are doing dsign lowering in GLSL and haven't had to
have a NIR path due to there not being a corresponding vulkan driver. We
want this in NIR now so that we can retire that batch of GLSL lowering
code.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25777>
Avoids a regression when disabling GLSL-level lowering of the dround class
of ops, while passing through doubles to llvm for the other rounding ops.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25777>
Most drivers that can expose GL4 were claiming the cap anyway (llvmpipe,
softpipe, zink, iris, nvc0, radeonsi, r600, freedreno, d3d12), and just
doing lowering in NIR if nessary.
crocus was only claiming the cap for gen8, but the backend compiler
enables NIR lowering regardless.
svga is the only other GL4 driver that didn't set it, and we can just set
the NIR lowering flag.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25777>
Sync xe_drm.h with commit xxxxx ("drm/xe/uapi: Fix naming of XE_QUERY_CONFIG_MAX_EXEC_QUEUE_PRIORITY").
One not so straght forward change is that sync VM binds now don't
require a syncobj anymore, the uAPI will return as soon the VM bind
operations are done.
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25300>
Make intel_aux_map_add_mapping return false if a mapping is attempted
that would conflict with an existing one. If this function doesn't
return false, it will either fail to return or return true.
The Vulkan driver will make use of this feature to opportunistically
enable CCS if a BO's VMA range has not been already mapped.
Reviewed-by: Jianxun Zhang <jianxun.zhang@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25003>
Fixes crashes on at least my system with multiple GPUs.
Fixes: 0c2045553f ("zink: use screen destructor for creation fails")
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25844>
Make use of etna_cmd_stream_mark_end_of_context_init(..) to mark the
end of context init states. If the cmd stream is untouched until
etna_cmd_stream_flush(..) gets called the drm layer will downgrade this
specific cmd stream to a noop.
This safes us from calling into the kernel with an 'empty' cmd stream.
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/25849>
The svga driver contains several buffer managers with gmr* and mob* prefixes,
but they are of similar functionality. The underlying vmwgfx kernel driver
allocates memory of one type - mobs if supported by the virtual hardware,
and gmrs otherwise - and returns an opaque pointer to userspace, so several
managers are reduntant.
This patch reduces the number of buffer managers by unifying similar managers
and performing refactors where necessary.
Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25832>
There is no need to have a separate shareable pool of surfaces, we can make
all surfaces shareable.
This change makes all surfaces shareable when creating them.
Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25832>
The assert in vlVaSetSurfaceContext would very rarely fail because
the mutex was already unlocked when calling this function from
vlVaBeginPicture.
Keep the lock until returning from vlVaBeginPicture, as that's what
other functions are already doing.
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25823>
This was an ill-advised extension. While we advertised SWAP_COPY support,
we might implement it with a back-copy from the front buffer. And we
never advertised EXCHANGE because we couldn't guarantee it. So, if you
actually used this extension to try to reduce app redraws of the back
buffer, you might actually increase the bandwidth you used. Whoops.
Instead, GLX_EXT_buffer_age and the similar EGL extension give you
feedback on what's left in your back buffer, letting you do minimum
redraws.
This reduces our GLX visual+fbconfig count from 1410 to 940 on an llvmpipe
X server. Reducing visual counts will improve test runtime for
visual-iterating tests like piglit's glx-visuals-*.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25650>
