Lionel Landwerlin [Thu, 22 Sep 2022 21:45:55 +0000 (00:45 +0300)]
ci/debian: don't use libclc from the system
Debian (unlike Ubuntu) has a broken libclc package missing files we
would very much like to have in our image, so that intel_clc doesn't
fail. Namely :
/usr/lib/clc/spirv-mesa3d-.spv
/usr/lib/clc/spirv64-mesa3d-.spv
Dropping libclc from the distribution and build int the build & base
test image.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18788>
Lionel Landwerlin [Thu, 22 Sep 2022 15:56:45 +0000 (18:56 +0300)]
ci: build our own version of the LLVM SPIRV translator
Debian stable and Fedora do not package the required version for
intel-clc.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18788>
Lionel Landwerlin [Thu, 22 Sep 2022 10:13:42 +0000 (13:13 +0300)]
ci: disable intel-clc on fedora
Would require Fedora 35.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18788>
Lionel Landwerlin [Thu, 22 Sep 2022 09:26:05 +0000 (12:26 +0300)]
ci: add python3-ply to debian/fedora images
This is needed by Anv to parse GRL (meta opencl kernels).
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18788>
Dave Airlie [Tue, 27 Sep 2022 05:26:42 +0000 (15:26 +1000)]
gallivm/nir: bitcast when non-float ptr type.
This matters more when opaque pointers are used.
Fixes:
203920d4c693 ("gallivm: add atomic 32-bit float support")
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18835>
Dave Airlie [Mon, 26 Sep 2022 19:17:39 +0000 (05:17 +1000)]
gallivm/nir: fix fmin/fmax translation
Fixes:
203920d4c693 ("gallivm: add atomic 32-bit float support")
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18835>
Dave Airlie [Mon, 26 Sep 2022 19:15:26 +0000 (05:15 +1000)]
lavapipe: add fmin/fmax to image lowering.
Fixes:
31695f81c925 ("lavapipe: export VK_KHR_shader_atomic_float")
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18835>
Mike Blumenkrantz [Wed, 21 Sep 2022 16:21:28 +0000 (12:21 -0400)]
zink: also lower 64bit function temps
now that these aren't always lowered from indirects,
64bit function temp arrays may exist in shaders, and they need to use
all the same rewrite machinery
fixes #7310
SoroushIMG <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18728>
Mike Blumenkrantz [Wed, 21 Sep 2022 16:20:56 +0000 (12:20 -0400)]
zink: split up lower_64bit_vars pass
the component functions will be reused; no functional changes
SoroushIMG <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18728>
Mike Blumenkrantz [Wed, 21 Sep 2022 16:19:24 +0000 (12:19 -0400)]
zink: always run optimize_nir after lower_64bit_vars
it's otherwise possible (and likely) that optimizations won't
happen since there's no shader key data active
Fixes:
5b2f850425e ("zink: rewrite 64bit shader i/o as 32bit")
SoroushIMG <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18728>
Mike Blumenkrantz [Wed, 21 Sep 2022 16:18:26 +0000 (12:18 -0400)]
zink: don't flatten 64bit arrays during rewrite
dunno what I was thinking here
Fixes:
5b2f850425e ("zink: rewrite 64bit shader i/o as 32bit")
SoroushIMG <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18728>
Mike Blumenkrantz [Wed, 21 Sep 2022 13:49:11 +0000 (09:49 -0400)]
nir/opt_undef: add a pass to clean up 64bit undefs
somehow 64bit lowering creates patterns like
vec1 64 ssa_1 = undefined
ssa_2 = unpack_64_2x32_split_x ssa_1
and then the 64bit value is never otherwise used. for this case, rewriting
the unpack to just be a 32bit undef allows the 64bit undef to be optimized
out, avoiding spec violations
fixes #6945
SoroushIMG <soroush.kashani@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18728>
Karol Herbst [Sun, 25 Sep 2022 11:43:13 +0000 (13:43 +0200)]
rusticl/device: fix custom device detection
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18809>
Mike Blumenkrantz [Thu, 22 Sep 2022 17:27:14 +0000 (13:27 -0400)]
zink: always set VK_PIPELINE_CREATE_COLOR_ATTACHMENT_FEEDBACK_LOOP_BIT_EXT sometimes
some drivers (mostly desktop) are known to ignore this pipeline flag,
thus it can be set unconditionally in pipeline state to deduplicate
pipeline variants without affecting performance
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Thu, 22 Sep 2022 17:02:17 +0000 (13:02 -0400)]
zink: use feedback loop layout to correctly handle implicit feedback loops
an implicit feedback loop occurs when an app happens to bind the same image
as both a framebuffer attachment and a sampler for the same draw
an explicit feedback loop occurs when an app uses fbfetch to read data back
from the framebuffer using input attachments
fbfetch is already handled, but implicit feedback loops require more work:
* detecting them happens on-the-fly
* pipeline variants are required
this handles implicit feedback loops by detecting them at draw time during
barrier updates and then flagging pipeline state change to trigger variant creation.
the bits are then unset when the framebuffer/sampler binds are removed
fixes #7309
fixes (tu):
KHR-GL46.texture_barrier.disjoint-texels
KHR-GL46.texture_barrier.overlapping-texels
KHR-GL46.texture_barrier.same-texel-rw-multipass
KHR-GL46.texture_barrier_ARB.disjoint-texels
KHR-GL46.texture_barrier_ARB.overlapping-texels
KHR-GL46.texture_barrier_ARB.same-texel-rw-multipass
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Thu, 22 Sep 2022 15:33:20 +0000 (11:33 -0400)]
zink: add a mask of fb attachment idx for resources
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Thu, 22 Sep 2022 15:31:09 +0000 (11:31 -0400)]
zink: reorder zink_resource a little
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Thu, 22 Sep 2022 15:28:51 +0000 (11:28 -0400)]
zink: rename fb_binds -> fb_bind_count
this is more accurate
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Thu, 22 Sep 2022 15:16:14 +0000 (11:16 -0400)]
zink: remove zink_context::new_swapchain
this was my (feeble) attempt at tracking validity for swapchain images,
but it doesn't actually work right, and zink_resource::valid is better
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Wed, 21 Sep 2022 23:14:02 +0000 (19:14 -0400)]
zink: enable VK_ATTACHMENT_LOAD_OP_DONT_CARE for zs in renderpasses
dynamic render path already handled this
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18787>
Mike Blumenkrantz [Tue, 27 Sep 2022 15:58:52 +0000 (11:58 -0400)]
mesa: set normalized_coords for bindless texture buffer samplers
this isn't "used", but it's the default
cc: mesa-stable
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18845>
Adam Jackson [Thu, 22 Sep 2022 19:15:14 +0000 (15:15 -0400)]
egl: Remove a bogus restriction from eglMakeCurrent
The comment here is from the EGL_MESA_configless_context era, and maybe
that was true, but EGL_KHR_no_config_context has no such restriction.
The only restriction is that both surfaces be compatible with the
context, but a no-config context is defined to be compatible with any
surface.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18824>
Adam Jackson [Tue, 20 Sep 2022 21:50:42 +0000 (17:50 -0400)]
egl/dri2: Fix a weird conditional in dri2_make_current
It's not valid to get here with no context but with real surfaces.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18824>
Adam Jackson [Tue, 20 Sep 2022 18:55:37 +0000 (14:55 -0400)]
egl: Factor some common terminate cleanup up to common code
It's a little difficult to see from the diff, but this is effectively
the same calling sequence as before, and more importantly it means the
backend only cleans up backend state rather than needing to call up to
the core.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18824>
Adam Jackson [Tue, 20 Sep 2022 21:51:42 +0000 (17:51 -0400)]
egl: Formatting fix
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18824>
Adam Jackson [Thu, 22 Sep 2022 17:20:18 +0000 (13:20 -0400)]
mesa/fbo: Don't force both read/draw bindings to zero
... when we're only changing one of the two to zero. I could maybe
understand the old behavior here if we really did need both read/draw to
either be both from winsys or both user, but we don't. You can still
get to that half-and-half state even with the old code, you just need to
set whichever of read/draw to 0 that you want first, and then set the
other to the user fbo.
The first code to introduce a read fb to this path was
dbfb375805d94,
which was trying to make sure we restored both values passed to
glXMakeCurrentRead when going back to fb 0. The commit message suggests
that "the API" doesn't allow split read/draw, but that seems to have
ignored the #ifdef EXT_framebuffer_object code immediately nearby which
did exactly that.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18824>
Ruijing Dong [Mon, 26 Sep 2022 15:37:32 +0000 (11:37 -0400)]
radeonsi/vce: using fixed value for vce 52
VCE 52 needs some hard-coded values.
Fixed: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7290
Reviewed-by: Boyuan Zhang <Boyuan.Zhang@amd.com>
Signed-off-by: Ruijing Dong <ruijing.dong@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18823>
LingMan [Sun, 18 Sep 2022 12:59:25 +0000 (14:59 +0200)]
rusticl: Fix compilation if stdout and/or stderr aren't symbols
ISO C only guarantees that `stdout` and `stderr` are expressions of type `FILE*`. It does not
guarantee that they are symbols which can be bound to. In particular, they are macros on the BSDs,
which bindgen cannot deal with (yet?).
Add our own symbol-defining wrapper.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7269
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18810>
David Heidelberg [Mon, 19 Sep 2022 20:53:08 +0000 (22:53 +0200)]
ci/amd: move radeonsi in amd directory
Deduplicate jobs, make easier to use GL+VK scenarios.
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
David Heidelberg [Tue, 2 Aug 2022 22:22:47 +0000 (00:22 +0200)]
ci/radeonsi: add traces for radeonsi Zork
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
David Heidelberg [Mon, 8 Aug 2022 12:00:07 +0000 (14:00 +0200)]
ci/traces: prevent showing wine instalation dialog
Causing piglit timeouts because of an open window instead of running apitrace.
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
David Heidelberg [Tue, 13 Sep 2022 15:18:39 +0000 (17:18 +0200)]
ci: also handle default wine rootfs (for testing)
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
David Heidelberg [Wed, 14 Sep 2022 19:36:51 +0000 (21:36 +0200)]
ci/dxvk: fix DXVK 64-bit only wine setup
Reported: https://github.com/doitsujin/dxvk/issues/2921
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
David Heidelberg [Thu, 4 Aug 2022 14:02:35 +0000 (16:02 +0200)]
ci/lava: prepare wineprefix for GL and DXVK
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
David Heidelberg [Sat, 16 Jul 2022 12:11:16 +0000 (14:11 +0200)]
ci/lava: add wine and apitrace into amd64 rootfs container
Wine is from WineHQ, with stripped i386 part, which we don't use.
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17938>
Samuel Pitoiset [Tue, 20 Sep 2022 10:53:11 +0000 (12:53 +0200)]
radv: fix handling multiview with GPL
When multiview is used, the last pre-rasterization stage should export
the layer to the fragment shader. With GPL, the next stage is
implicitly FS.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18699>
Samuel Pitoiset [Tue, 20 Sep 2022 09:19:13 +0000 (11:19 +0200)]
radv: fix handling primitive ID and clip/cull distances with GPL
When compiling the pre-rasterization stages with GPL, we can't know if
the fragment shader reads the primitive ID or the clip/cull distances
as inputs and we have to export them unconditionally.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18699>
Samuel Pitoiset [Tue, 20 Sep 2022 09:40:35 +0000 (11:40 +0200)]
radv: fix handling ViewportIndex with GPL
When the fragment shader reads the viewport index as input and the last
pre-rasterization stage doesn't export it, it should be implicitly
zero (ie. first viewport). When the next stage is known, it's already
lowered in NIR, but with GPL we can't because we don't know if the FS
reads it. To fix this use AC_EXP_PARAM_DEFAULT_VAL_000.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18699>
Samuel Pitoiset [Tue, 20 Sep 2022 10:46:15 +0000 (12:46 +0200)]
radv: rework multiview lowering in NIR slightly
This will be used to export multiview unconditionally with GPL.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18699>
Iago Toral Quiroga [Fri, 23 Sep 2022 07:10:36 +0000 (09:10 +0200)]
v3dv: expose VK_EXT_image_robustness
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18820>
Iago Toral Quiroga [Mon, 26 Sep 2022 08:47:19 +0000 (10:47 +0200)]
broadcom/compiler: add a lowering for robust image access
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18820>
Iago Toral Quiroga [Fri, 23 Sep 2022 09:19:34 +0000 (11:19 +0200)]
broadcom/compiler: rename static helpers involved with robust buffer access
To make it explicit that they involve buffers, since we will be adding
robust image access shortly.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18820>
Iago Toral Quiroga [Fri, 23 Sep 2022 09:01:45 +0000 (11:01 +0200)]
broadcom/compiler: rename v3d_nir_lower_robust_buffer_access.c
We are going to add code to handle image robustness shortly, so
better rename this to v3d_nir_lower_robust_access.c
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18820>
Erik Faye-Lund [Sat, 17 Sep 2022 11:09:20 +0000 (13:09 +0200)]
docs: fixup broken rst syntax
Seems I messed this up when converting things, and nobody noticed until
now!
Fixes:
d4397c00485 ("docs: use envvar role for envvars")
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18641>
Yonggang Luo [Tue, 27 Sep 2022 02:23:42 +0000 (10:23 +0800)]
ci/docs: Stick to specific version of python packages
Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Acked-by: Eric Engestrom <eric@igalia.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18833>
Erik Faye-Lund [Thu, 8 Sep 2022 09:48:49 +0000 (11:48 +0200)]
mesa/main: don't copy signed rgtc via float
There's no good reason to copy via float instead of snorm, so let's take
that shortcut.
Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18565>
Erik Faye-Lund [Thu, 8 Sep 2022 09:32:17 +0000 (11:32 +0200)]
mesa/main: fix broken indent
Acked-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18565>
Timothy Arceri [Wed, 14 Sep 2022 01:21:00 +0000 (11:21 +1000)]
nir/loop_analyze: remove cost of redundant selects
If we know that a select will be eliminated once the loop is
unrolled than we don't need to count the instruction towards the
cost of the loop.
This change helps 2 loops unroll in an xcom enemy unknown shader
that is loaded full of these redundant selects.
Acked-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18587>
Timothy Arceri [Wed, 14 Sep 2022 00:18:05 +0000 (10:18 +1000)]
nir/loop_analyze: delay instruction cost calculation
Here we move the calculation of the instruction cost of the loop
after we have processed other information such as finding the
induction variables. This is useful because we can use this further
information to find instructions that will be eliminated if the
loop was to unroll and therefore give them a cost of 0.
Acked-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18587>
Karol Herbst [Sun, 25 Sep 2022 12:20:59 +0000 (14:20 +0200)]
rusticl/mem: return errors for OOB accesses
There is a little bit more to do, but this at least resolves crashes due
to OOB accesses and/or assertions.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7346
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18806>
Marek Olšák [Tue, 23 Aug 2022 04:13:37 +0000 (00:13 -0400)]
mesa: make glPopMatrix a no-op if the matrix hasn't changed
This happens a lot with viewperf because glMultMatrixf contains identity
matrices.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Tue, 23 Aug 2022 04:12:59 +0000 (00:12 -0400)]
mesa: use memcmp instead of floating-point comparisons in glMultMatrixf
This is faster.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Tue, 23 Aug 2022 04:12:09 +0000 (00:12 -0400)]
glthread: skip glMultMatrixf if it's identity
This is common with viewperf and it helps.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Tue, 23 Aug 2022 04:09:23 +0000 (00:09 -0400)]
glthread,hud: draw per-frame values and expose the number of batches per frame
This is better because we see the statistics for 1 frame instead of 1 second.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 21 Aug 2022 21:30:18 +0000 (17:30 -0400)]
glthread: use 8-bit GLenum for prim modes where it results in better packing
This is OK because all modes > 14 are invalid.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 21 Aug 2022 19:49:53 +0000 (15:49 -0400)]
glthread: use GLenum16 for enums, but clamp it to 0xffff to get correct errors
0xffff is an invalid enum.
This was initially implemented without clamping and then reverted because
of the concern that we would randomly get correct enums and not report
errors if we just dropped the high bits.
This packs calls better in glthread_batch.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Thu, 11 Aug 2022 08:34:00 +0000 (04:34 -0400)]
mesa: rename GL dispatch initialization functions to make the intent clear
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Fri, 12 Aug 2022 09:20:31 +0000 (05:20 -0400)]
mesa: remove deprecated comments
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Thu, 11 Aug 2022 15:04:13 +0000 (11:04 -0400)]
api_hw_select_init_h.py: simplify the code
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Fri, 12 Aug 2022 08:58:14 +0000 (04:58 -0400)]
glapi: rename 'vtxfmt' to 'beginend' to make it clear
These are really just the functions that execute between Begin/End.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Thu, 11 Aug 2022 05:26:06 +0000 (01:26 -0400)]
mesa: initialize OutsideBeginEnd directly instead of through Exec
Exec is just a mutable pointer to one of the dispatch tables.
OutsideBeginEnd is what we are actually initializing here. This makes
the original intention clear.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Wed, 10 Aug 2022 05:06:40 +0000 (01:06 -0400)]
mesa: remove unused function _mesa_get_dispatch
just use CurrentClientDispatch directly
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Wed, 10 Aug 2022 04:56:26 +0000 (00:56 -0400)]
mesa: don't restore the dispatch at the end of glCallList
No places use the glapi dispatch. All places use ctx directly.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Wed, 10 Aug 2022 04:08:02 +0000 (00:08 -0400)]
vbo: don't call _glapi_set_dispatch for GL_COMPILE_AND_EXECUTE loopback
It uses Exec directly, so this didn't do anything.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Mon, 22 Aug 2022 04:43:36 +0000 (00:43 -0400)]
mesa: reorder and pack gl_buffer_object
The frequently-accessed fields are at the beginning, hoping
we only need to load 1 cache line to get them all.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Mon, 22 Aug 2022 04:30:00 +0000 (00:30 -0400)]
mesa: remove unused gl_buffer_object::Data and Written
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Mon, 22 Aug 2022 05:40:35 +0000 (01:40 -0400)]
mesa: restructure bind_buffer_object for better performance
The CPU time spent in _mesa_BindBuffer_no_error decresed from 11.4% to 9.2%.
I explain in the code what is happening here because it's not obvious.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Mon, 22 Aug 2022 05:37:36 +0000 (01:37 -0400)]
mesa: inline _mesa_handle_bind_buffer_gen and correct things
Changes:
- the function is inlined, which will help the next commit to improve perf
- use no_error inside the function
- if the function returns true, the buffer is non-NULL, so remove NULL
checking in callers
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Mon, 22 Aug 2022 05:31:41 +0000 (01:31 -0400)]
mesa: inline _mesa_reference_buffer_object
The last parameter is always a constant expression, so this simplifies it.
This is also required by later commits.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 21 Aug 2022 17:59:41 +0000 (13:59 -0400)]
mesa: trivial simplifications in _mesa_reference_buffer_object_
The else statement conditions were tautological.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 14 Aug 2022 04:57:11 +0000 (00:57 -0400)]
mesa: clean up st_indirect_draw_vbo interface and callers
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 14 Aug 2022 04:35:41 +0000 (00:35 -0400)]
mesa: remove unused code using _mesa_prim
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 14 Aug 2022 04:21:30 +0000 (00:21 -0400)]
mesa: remove _mesa_draw_gallium_fallback by adapting st_feedback_draw_vbo
This makes st_feedback_draw_vbo implement the DrawGallium.
That's the last users of the fallback functions.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Marek Olšák [Sun, 14 Aug 2022 03:52:46 +0000 (23:52 -0400)]
mesa: replace ALLOC_PRIMS with persistently-allocated context arrays
This is cleaner.
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18828>
Friedrich Vock [Sat, 24 Sep 2022 20:10:02 +0000 (22:10 +0200)]
radv: Build acceleration structures using BVH IR
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Friedrich Vock [Fri, 23 Sep 2022 14:02:37 +0000 (16:02 +0200)]
radv: Create pipelines for conversion meta shaders
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Friedrich Vock [Wed, 21 Sep 2022 13:09:55 +0000 (15:09 +0200)]
radv: Add conversion shader for leaf nodes
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Friedrich Vock [Mon, 19 Sep 2022 21:16:10 +0000 (23:16 +0200)]
radv: Add conversion shader for internal nodes
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Friedrich Vock [Thu, 22 Sep 2022 20:47:11 +0000 (22:47 +0200)]
radv: Add BVH IR types
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Friedrich Vock [Wed, 21 Sep 2022 14:02:36 +0000 (16:02 +0200)]
radv: Split CmdBuildAccelerationStructuresKHR into different functions
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Friedrich Vock [Mon, 19 Sep 2022 21:15:31 +0000 (23:15 +0200)]
radv: Rename internal node shader to lbvh_internal
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18769>
Mike Blumenkrantz [Wed, 14 Sep 2022 18:17:22 +0000 (14:17 -0400)]
zink: only emit runtime array for ssbos if present
it's impossible to accurately determine whether an op will use the base,
sized array or the unsized array due to dynamic indexing, and thus it's
impossible to not violate spec by emitting both arrays at offset=0 so they
can overlap
so instead, just emit the runtime array and blast all bounds checking in
vk drivers into the void
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Wed, 14 Sep 2022 12:18:30 +0000 (08:18 -0400)]
zink: remove color_write_missing driver workaround
this is no longer used
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Fri, 9 Sep 2022 14:42:49 +0000 (10:42 -0400)]
zink: allow reordered clear_buffer calls
rare to hit, but no point in not promoting them when possible
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Wed, 7 Sep 2022 17:21:28 +0000 (13:21 -0400)]
zink: avoid overflow when clamping bufferviews
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Thu, 15 Sep 2022 13:58:08 +0000 (09:58 -0400)]
zink: name bo variables using bitsize
debugging++
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Mon, 12 Sep 2022 15:08:20 +0000 (11:08 -0400)]
zink: fix handling for ssbos that are just runtime arrays
this used to correctly calculate runtime array size as zero, but since the
switch to glsl_get_explicit_size(), zero can no longer be returned
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Mon, 19 Sep 2022 14:14:26 +0000 (10:14 -0400)]
zink: fix sparse queue creation
to use an explicit sparse queue, it has to actually be created along with
the device
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Sun, 25 Sep 2022 19:11:04 +0000 (15:11 -0400)]
zink: always unset var->data.explicit_xfb_buffer
ensure that this isn't accidentally triggering for non-xfb shaders
Fixes:
6d40db84c9d ("zink: handle direct xfb output from output variables")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Fri, 23 Sep 2022 13:46:43 +0000 (09:46 -0400)]
zink: free vertex_state_cache on shutdown
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Fri, 23 Sep 2022 13:46:10 +0000 (09:46 -0400)]
zink: fix some batch tracking leaks
not leaking over time, just leaking on shutdown
Fixes:
83c76b8efbf ("zink: rework batch tracking for resources")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Mon, 26 Sep 2022 17:46:20 +0000 (13:46 -0400)]
zink: add asserts to verify optimal key state
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Mike Blumenkrantz [Mon, 26 Sep 2022 17:46:10 +0000 (13:46 -0400)]
zink: copy optimal key always
this is going away in the future, but for now it's annoying
Fixes:
7336580408a ("zink: unspaghettify some program update code")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18784>
Kenneth Graunke [Fri, 23 Sep 2022 05:50:37 +0000 (22:50 -0700)]
iris: Ignore aux for copy_region source if there's no unresolved color
This mirrors the behavior of iris_resource_texture_aux_usage(), which
bypasses the aux metadata when there's no advantage to using it.
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18790>
Mike Blumenkrantz [Mon, 26 Sep 2022 19:22:35 +0000 (15:22 -0400)]
lavapipe: propagate shader access info across libraries
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18827>
Mike Blumenkrantz [Mon, 26 Sep 2022 18:41:31 +0000 (14:41 -0400)]
lavapipe: set writemask for compute shader buffers
I don't know how this wasn't being set?
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18827>
Sil Vilerino [Mon, 26 Sep 2022 11:59:11 +0000 (07:59 -0400)]
d3d12: Replace size_t with uint64_t in d3d12_video_encoder definitions for cross-plat compatibility
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18821>
Mike Blumenkrantz [Tue, 6 Sep 2022 16:35:43 +0000 (12:35 -0400)]
zink: unspaghettify some program update code
the initial implementation of optimal_keys was added inline, but really
it's an entirely different codepath. by separating these out, it makes the
code more readable, and it also allows for slightly better optimization of
of the optimal_keys codepath
~4-5% improvement for drawoverhead -test 7
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18786>
Mike Blumenkrantz [Fri, 2 Sep 2022 15:13:44 +0000 (11:13 -0400)]
zink: be even more granular with optimal_key program updates
since the bits of each key are easily and efficiently comparable,
the draw-time updating here can be made extremely granular to
update exactly the stages that have changed since the last
time the program was used, further reducing overhead instead of
updating every possible key value
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18786>
Mike Blumenkrantz [Fri, 2 Sep 2022 15:08:46 +0000 (11:08 -0400)]
zink: simplify bit tracking in zink_gfx_program_update()
much nicer now that gfx dirty bits are isolated
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18786>