a34d72cbed
nak/hw_runner: Use byte_offset()
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30448 >
2024-07-31 21:51:52 +00:00
0116430d39
intel/brw: Handle 16-bit sampler return payloads
...
API requires samplers to return 32-bit even though hardware can handle
16-bit floating point, so we detect that case and make more efficient
use of memory BW. This is helping improve performance of encode and
decode tokens during LLM by at least 5% across multiple platforms.
Thank you Kenneth Graunke for suggesting and guiding me throughout
this implementation.
Signed-off-by: Sushma Venkatesh Reddy <sushma.venkatesh.reddy@intel.com >
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30447 >
2024-07-31 21:26:46 +00:00
ddd9e043dc
intel/brw: Move get_nir_def() higher to avoid UNDEF
...
While extending our backend to handle 16-bit sampler return payloads, we
found that in piglit's arb_texture_view-rendering-formats, the SIMD8 FS
was missing the sampling operation altogether. This was because we were
first emitting the texturing instruction, and then calling
get_nir_def(), which adds an UNDEF instruction when the destination is
smaller than the 32-bit. So the texturing was dead code elimated. Fix
this by calling get_nir_def() earlier.
Thank you to Kenneth Graunke for suggesting and guiding me throughout
this implementation.
Signed-off-by: Sushma Venkatesh Reddy <sushma.venkatesh.reddy@intel.com >
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30447 >
2024-07-31 21:26:46 +00:00
52be72e676
intel: Let compiler set indirect_ubos_use_sampler
...
This option is used for Gfx < 12, elk already set it to true,
so set it in brw and change the drivers to not set it anymore.
Because the dual-compiler support in Iris, the helper function
there had to change to consult the right compiler value instead.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30393 >
2024-07-31 19:26:20 +00:00
f1680c262c
kopper: increment lastStamp during invalidate
...
this is functionally equivalent
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
76757c211d
glx: delete DRI_TEX_BUFFER handling
...
this is always available
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
c4f7853131
loader/dri3: delete texBuffer struct member
...
unused
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
321d576f0c
glx: inline DRI_TEX_BUFFER
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
9fdfebce06
glx: delete check for texbuffer extension
...
this is always available
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
032237c7ec
dri: declare DRI_TEX_BUFFER functions public
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
ef6f874590
dri: delete __DRItexBufferExtension::setTexBuffer
...
no longer used
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
a09c8b577f
glx: delete the only users of __DRItexBufferExtension::setTexBuffer
...
this is pointless since there's only one implementation of DRI_TEX_BUFFER
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
b0d0c1971c
glx: delete releaseTexBuffer
...
this is always null, so none of the code is ever executed
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
0341623f39
dri: delete __DRI_COPY_SUB_BUFFER
...
bye-bye
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
945fe0ef4a
glx: inline DRI_COPY_SUB_BUFFER
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
998d0e3993
drisw: declare DRI_COPY_SUB_BUFFER as PUBLIC
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
34912cc8d1
glx: rename driswCopySubBuffer
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
62847d14b5
glx: delete DRI_KOPPER
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
192febfbbf
glx: switch kopper check for setting up swapinterval/bufferage functions
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
850b0694d6
glx: use local var for LIBGL_KOPPER_DISABLE check
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
2157108e11
dri: declare DRI_KOPPER as PUBLIC
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
38529171d6
dri: delete driCreateNewScreen2
...
unused
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
5d72058556
glx: inline DRI_DRI2
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
5019481a16
dri: delete __DRImesaCoreExtensionRec::createNewScreen
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
63937eca78
glx: inline DRI_SWRAST
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
f5dd5e4b44
glx: delete __DRImesaCoreExtension usage
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
5dba565ada
glx: inline createNewScreen3
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
66e0d12627
glx: inline DRI_CORE functions
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
87720ab42b
loader/dri3: inline DRI_CORE
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
029ec1a20b
dri: declare DRI_CORE functions public
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
e95a462d1f
glx: inline IMAGE_DRIVER functions
...
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
49c3814cd1
dri: declare IMAGE_DRIVER symbols PUBLIC
...
this is a temporary hackaround for a linker issue where dri frontend
functions need to be called directly by (glx/egl/gbm) but can't due
to linkage/visibility
eventually all of these frontends will be a single linkage into a single
library, which will avoid the intermediate linking and solve the problem
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
f9f8652445
loader: split out dri3 into subdir
...
this fixes dependency hell
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
36aaaa968a
meson: move glx subdir after gallium build
...
more dependency hell
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
4095fac72b
loader: move some common dri3 functions out of dri3 loader
...
fixing dependency hell
Reviewed-by: Adam Jackson <ajax@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30400 >
2024-07-31 18:50:38 +00:00
eebb6cd236
anv: stop using 3DSTATE_WM::ForceThreadDispatchEnable
...
Documentation says we should leave this field to the default value
(Normal). Instead we set 3DSTATE_PS_EXTRA::PixelShaderHasUAV when we
see that a fragment shader has side effects.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Tapani Pälli <tapani.palli@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30408 >
2024-07-31 18:18:14 +00:00
c818de7360
aco: combine DPP into v_cvt_f16_f32
...
The label_f2f16 was interfering with the label_dpp.
fossil-db (navi31):
Totals from 36 (0.05% of 79395) affected shaders:
Instrs: 70998 -> 70875 (-0.17%); split: -0.18%, +0.00%
CodeSize: 373184 -> 372244 (-0.25%)
Latency: 1266807 -> 1266752 (-0.00%); split: -0.01%, +0.01%
InvThroughput: 131618 -> 131566 (-0.04%); split: -0.04%, +0.00%
SClause: 1947 -> 1944 (-0.15%)
Copies: 5194 -> 5208 (+0.27%)
VALU: 41019 -> 40921 (-0.24%); split: -0.24%, +0.00%
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com >
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30425 >
2024-07-31 17:26:05 +00:00
f7f343796f
meson: centralize checking for new enough meson for rust support
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:43 +00:00
b3a9c91d9b
meson: centralize bindgen version check
...
It's entirely duplicated and every rust user also wants to use bindgen
anyway. This way bindgen won't show up multiple times in the meson log
as well.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:43 +00:00
03ecda08de
meson: centralize rust handling
...
This bumps the req for everything, but I think that's fine, because one
part restricting to a lower rust version doesn't make much sense as we are
also not doing it generally for C or C++.
This also makes it easier for packagers to know what requires rust.
Reviewed-by: Eric Engestrom <eric@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:43 +00:00
7ff2513c28
rusticl/icd: use ptr::addr_eq
...
It got stabilized with 1.76
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:43 +00:00
19c66754df
rusticl: use ptr::from_ref and ptr::from_mut
...
This is considered safer as it prevents changing the type the pointer
points to.
It got stabilized with 1.76
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:43 +00:00
11e4793b41
rusticl: replace some raw pointer casts
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:42 +00:00
66f7e3a4b2
rusticl: use pointer byte offset API
...
It allows us to be more explicit about the intent here and avoids casting
to u8 or c_void pointers.
It got stabilized with 1.75
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:42 +00:00
40072f57b0
rusticl: use next_multiple_of
...
It got stabilized with 1.73
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:42 +00:00
fb82c253da
rusticl: use div_ceil
...
It got stabilized with 1.73
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30414 >
2024-07-31 16:22:42 +00:00
e2de72f71f
tu: Random non-consequential fixes
...
Few whitespace and harmless , when ; was meant fixes
Reported-by: Steev Klimaszewski <steev@kali.org >
Signed-off-by: Rob Clark <robdclark@chromium.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30431 >
2024-07-31 15:11:48 +00:00
87c889cd8a
freedreno/drm/virtio: Fix issues with 16k (or larger) page sizes
...
Signed-off-by: Rob Clark <robdclark@chromium.org >
Fixes: e6b2785811https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30431 >
2024-07-31 15:11:48 +00:00
7fe3529715
tu: Fix issues with 16k (or larger) page sizes
...
The iova allocations need to be CPU page aligned. (The GPU itself
always supports 4k mappings regardless of the smallest CPU page size,
but GEM buffer allocations must be an integer number of CPU pages.)
Signed-off-by: Rob Clark <robdclark@chromium.org >
Fixes: 63904240f2https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30431 >
2024-07-31 15:11:48 +00:00
745ad42bb0
freedreno: Enable the A735
...
Signed-off-by: Karmjit Mahil <karmjit.mahil@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30379 >
2024-07-31 13:38:52 +00:00
Faith Ekstrand