Multi-plane format textures are allocated as a single buffer
for all planes, so when mapping first plane it will map
all planes.
However, when staging texture is used, only the first
plane will be available.
When asked to map first plane of multi-plane texture,
create a staging texture with matching multi-plane format
and copy all planes on map/unmap.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24389>
There's no legacy register support so we ncessarily write a contiguous vector.
In other words, the write_mask is of the form `(1 << x) - 1`. Meanwhile this
code asserts the write mask is of the form (1 << x)`. Putting it together the
write mask is necessarily always 0x1, writing out a single scalar.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24471>
It's unused now, just need the references to it dropped so we'll still build
when it's all removed.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24471>
Mapping multi-plane surfaces won't work correctly without driver
supporting PIPE_VIDEO_CAP_SUPPORTS_CONTIGUOUS_PLANES_MAP.
Remove the check for ffmpeg process and instead check for
multi-plane format and driver support only.
Reviewed-by: Leo Liu <leo.liu@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24381>
this is stupid, but it makes the default color output "write the output"
instead of zero, which is still wrong, but at least maybe it's wrong
in a way that's useful instead of just being confusing
Acked-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24278>
This heuristic broke when zmike lowered fragcolor using NIR,
This fixes a regression in:
dEQP-GLES31.functional.shaders.framebuffer_fetch.basic.last_frag_data
Fixes: db1371cce1 ("llvmpipe: fix handling of unused color attachments")
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24435>
These instructions could be re-ordered, so depending just on the last one
is not sufficient.
Fixes: d21054b4bc
r600/sfn: Add pass to split address and index register loads
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24454>
Guard util to be standalone by move subdir('util') just behind subdir('android_stub')
texcompress_rgtc_tmp.h is not a standard c header, do not treat it as source
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/19522>
Because these files indeed doesn't belongs to src/util, they are tightly coupled with
src/gallium/auxiliary
Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19522>
As u_pack_color.h is used in vulkan drivers, so decouple it from gallium by this move
And dbghelp.h is included in u_debug_symbol.c and that's resident in src/util/
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/19522>
As pipe_color_union is used in vulkan drivers, so decouple it from gallium by this move
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/19522>
PIPE_MASK_* is used by src/util/*, so do the move to decouple src/util/* from gallium
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/19522>
pipe_swizzle is used by src/util/*, so do the move to decouple src/util/* from gallium
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/19522>
base feedback loop ext is more like a hard requirement, but it's tough to
pin down exactly which version requires it, and also it might "just work"
magically, so...yeah here we are
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24403>
sed + ninja clang-format + fix up spacing for common code.
If you are unhappy that I did not manually change the whitespace of your driver,
you need to enable clang-format for it so the formatting would happen
automatically.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24428>
Not used on Midgard, deals with a -Wunused-function warning.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Italo Nicola <italonicola@collabora.com>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24428>
On platforms with an indirect clear color BO, the clear color used when
sampling is in the form of a pixel. The sampler can interpret the pixel
correctly regardless of the format, so allow fast clears in more cases
within iris_resource_prepare_texture.
On gfx12, this completely removes partial resolves in a couple game
traces I tested:
- 72 partial resolves in a 500 frame trace of CS:GO.
- 1207 partial resolves in a 500 frame trace of Dota 2.
The resolves were triggered when ISL_FORMAT_A8_UNORM surfaces were
sampled as ISL_FORMAT_R8_UNORM.
Reviewed-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24323>
Preferring interlaced results in all surfaces being allocated
as interlaced for H264 decode regardless of whether the decoded
pic is interlaced or progressive, which makes VA postproc
apply deinterlacing even for progressive pics.
Deinterlacing interlaced pics still works because VA frontend
will now reallocate surface as interlaced when decoding interlaced
pics.
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24361>
When decoding interlaced pics, the output surface should
also be allocated as interlaced to make the deinterlace
postproc filter work correctly.
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24361>
When decoding interlaced pics, the output surface should
also be allocated as interlaced to make the deinterlace
postproc filter work correctly.
This makes deinterlacing work for drivers that supports, but
not prefers, interlaced surfaces.
Use extra buffer instead of reallocating as non-interlaced
when exporting interlaced surfaces.
v2: Avoid reallocating in vlVaExportSurfaceHandle
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24361>
MOCS = 0 is a invalid MOCS index on MTL, so it is necessary get a
valid value and set to MI_MATH instructions.
So here the mocs index is set with mi_builder_set_mocs(), it can be
always set but it is required when mi_build will emit MI_MATH
instructions.
The mocs index will only be stored and used in gfx12.5+ platforms
so no changes were are required in crocus or hasvk.
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22508>
gp has explicit load and store reg instructions so it can implement
the new nir_intrinsic_{decl,load,store}_reg intrinsics.
The implementation reuses the gpir_node_find and register_node functions
instead of just emitting load and store instructions in order to
preserve the behavior that there will never be a store_reg followed by
a load_reg of the same register in the same block.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24246>
With the upcoming transition away from nir_register and modifiers in
nir, ppir needs to prepare for those features to be removed.
The most straightforward way for ppir without major reworks at this time
is to use the new nir_legacy helpers so that most of the backend can be
kept untouched (particularly due to use of abs/neg/sat modifiers).
Do this for now to unblock the transition with little impact; ppir can
be optimized at a later moment to have its own passes which will allow
it to drop the nir_legacy use and use just the chasing helpers.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24246>
Even if loads are initially duplicated for each user, with new
optimizations such as folding modifiers, it may happen that loads
end up with only users in different blocks.
Handle that and don't delete or pipeline those nodes.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24246>
