Timur Kristóf [Thu, 26 Aug 2021 14:36:28 +0000 (16:36 +0200)]
aco: Emit zero for the derivatives of uniforms.
Observed in a shader from Resident Evil Village.
This also helps prevent emitting invalid IR.
Cc: mesa-stable
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12599>
Mike Blumenkrantz [Fri, 30 Jul 2021 17:14:20 +0000 (13:14 -0400)]
zink: only update inlinable constants when they change
Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12577>
Daniel Schürmann [Thu, 5 Aug 2021 13:52:08 +0000 (15:52 +0200)]
aco/optimizer: combine v_pk_mul_u16 + v_pk_add_u16 -> v_pk_mad_u16
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11678>
Daniel Schürmann [Wed, 30 Jun 2021 17:20:49 +0000 (19:20 +0200)]
aco/optimizer: fuse v_mul_f64 + v_add_f64 -> v_fma_f64
No fossil-db changes.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11678>
Daniel Schürmann [Thu, 1 Jul 2021 16:48:09 +0000 (18:48 +0200)]
aco/optimizer: combine v_mul_lo_u16 + v_add_u16 -> v_mad_u16
Totals from 192 (0.13% of 150170) affected shaders: (GFX10.3)
CodeSize: 1027224 -> 1019872 (-0.72%)
Instrs: 174784 -> 173863 (-0.53%)
Latency: 4235742 -> 4232177 (-0.08%); split: -0.11%, +0.03%
InvThroughput: 1777026 -> 1775945 (-0.06%); split: -0.09%, +0.03%
Copies: 34098 -> 34099 (+0.00%); split: -0.03%, +0.03%
PreVGPRs: 4920 -> 4850 (-1.42%)
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11678>
Daniel Schürmann [Thu, 1 Jul 2021 10:42:08 +0000 (12:42 +0200)]
aco: refactor nir_op_imul selection
Previously, the optimization to use v_mul_lo_u16 for
32bit multiplications was done in instruction_selection.
This was moved to the optimizer to ease some case distinctions.
The mixed results are due to increased use of SDWA.
Totals from 2616 (1.74% of 150170) affected shaders: (GFX10.3)
VGPRs: 143888 -> 143872 (-0.01%); split: -0.02%, +0.01%
CodeSize: 5604032 -> 5604080 (+0.00%); split: -0.01%, +0.01%
Instrs: 1086798 -> 1083915 (-0.27%); split: -0.27%, +0.01%
Latency: 8215793 -> 8213023 (-0.03%); split: -0.10%, +0.07%
InvThroughput:
20765157 ->
20773766 (+0.04%); split: -0.02%, +0.06%
VClause: 35256 -> 35260 (+0.01%); split: -0.02%, +0.03%
SClause: 29021 -> 29024 (+0.01%); split: -0.00%, +0.01%
Copies: 74163 -> 74306 (+0.19%); split: -0.05%, +0.24%
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11678>
Daniel Schürmann [Thu, 1 Jul 2021 13:03:36 +0000 (15:03 +0200)]
aco: only apply extract if not used more than 4 times
Totals from 61 (0.04% of 150170) affected shaders: (GFX10.3)
CodeSize: 1087732 -> 1087380 (-0.03%); split: -0.22%, +0.18%
Instrs: 192343 -> 192205 (-0.07%); split: -0.16%, +0.09%
Latency: 7231670 -> 7148073 (-1.16%); split: -1.19%, +0.04%
InvThroughput: 3436715 -> 3394926 (-1.22%); split: -1.25%, +0.04%
VClause: 4831 -> 4833 (+0.04%)
Copies: 50130 -> 49934 (-0.39%); split: -0.67%, +0.28%
Branches: 5945 -> 5948 (+0.05%)
PreSGPRs: 3486 -> 3472 (-0.40%)
PreVGPRs: 5154 -> 5152 (-0.04%)
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11678>
Mike Blumenkrantz [Fri, 27 Aug 2021 17:23:34 +0000 (13:23 -0400)]
freedreno: ci updates
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12600>
Mike Blumenkrantz [Fri, 27 Aug 2021 14:54:44 +0000 (10:54 -0400)]
freedreno: remove primconvert
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12600>
Mike Blumenkrantz [Fri, 27 Aug 2021 14:54:35 +0000 (10:54 -0400)]
freedreno: export supported primtypes
Acked-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12600>
Jeremy Newton [Wed, 25 Aug 2021 21:03:23 +0000 (17:03 -0400)]
Fix building AMD MM/GL with EL7
This partially reverts commit
4d9acfa5331be6140e84badff7a232dbf07b69b4.
The original patch said:
"Python 3 handles unicode strings by default, so we can drop all that."
But this breaks building on RHEL 7 (or similiar) since python3 support
on is much more limited than newer distros. Backporting all the needed
python 3 libraries to EL7 is a pretty big task, and isn't very easy to
maintain.
For workstation purposes, we need the AMD MM/GL driver building on RHEL
7, so only src/util/driconf_static.py needs to be reverted.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Jeremy Newton <Jeremy.Newton@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12549>
Erik Faye-Lund [Wed, 25 Aug 2021 19:51:44 +0000 (21:51 +0200)]
zink: avoid generating nonsensical code
With this code, we end up generating code such as:
if (!strcmp(extensions[i].extensionName, "VK_KHR_maintenance1")) {
if (VK_MAKE_VERSION(1,2,0) >= screen->vk_version) {
info->have_KHR_maintenance1 = true;
} else {
info->have_KHR_maintenance1 = true;
}
}
That's clearly nonsense, as it does the same thing in the true and false
case. So let's instead try to walk the Vulkan versions up to the one
we're using in a separate pass, and add all extensions that were made core
in that version.
CID: 1473289
Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12593>
Erik Faye-Lund [Wed, 25 Aug 2021 21:46:18 +0000 (23:46 +0200)]
zink: reduce scope of version-struct hack
Without this, we'll end up checking against the wrong version when
enabling core features in the next commit.
Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12593>
Erik Faye-Lund [Thu, 26 Aug 2021 09:12:23 +0000 (11:12 +0200)]
zink: clean up const-value handling for get_ssbo_size
nir_src_as_const_value can return null pointers, and in other places we
at least assert for this. So let's do that here as well, which makes
Coverity a bit less paranoid.
While we're at it, avoid duplicating the nir_src_as_const_value call to
the same source.
CID: 1444291
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12594>
Erik Faye-Lund [Thu, 26 Aug 2021 09:07:38 +0000 (11:07 +0200)]
zink: remove incorrect ASSERTED macro
The documentation for ASSERTED in macros.h says:
> Use ASSERTED to indicate that an identifier is unused outside of an `assert()`,
> so that assert-free builds don't get "unused variable" warnings.
We're using this variable outside of assert, so this shouldn't apply
here.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12594>
Erik Faye-Lund [Thu, 26 Aug 2021 09:03:25 +0000 (11:03 +0200)]
zink: remove needless scope
There's no variables declared in here, so there's no point in having
this slightly awkward scope here.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12594>
Erik Faye-Lund [Thu, 26 Aug 2021 08:24:30 +0000 (10:24 +0200)]
zink: give each major intrinsic it's own emit function
It's so much easier to follow this code if there's not any big blocks of
emitting in the middle of the code that figures out exactly what to do.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12594>
Erik Faye-Lund [Wed, 25 Aug 2021 21:55:46 +0000 (23:55 +0200)]
zink: avoid checking if src is const twice
nir_src_as_const_value also checks if the src is const, and not
checking that it returned null makes Coverity go paranoid thinking we
could dereference a null-pointer. But because it's slightly better to
check once than to check twice, let's rewrite this to avoid the
double-check, making it obvious what's going on here.
CID: 1485624
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12594>
Erik Faye-Lund [Thu, 26 Aug 2021 09:36:21 +0000 (11:36 +0200)]
zink: avoid memcmping null pointers
memcmping NULL pointers is not allowed, and would lead to a crash here.
So let's check that the first isn't NULL; we've already checked that
they're not *both* NULL, so checking one is enough.
CID: 1484801, 1484810
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Thu, 26 Aug 2021 09:27:33 +0000 (11:27 +0200)]
zink: remove needless NULL-check
This NULL-check makes Coverity paranoid because we just dereferenced the
pointer a few lines above.
But we never call this function with a NULL-pointer here, so the NULL
check isn't needed. Let's just remove it, to calm Coverity down a bit.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Thu, 26 Aug 2021 09:24:37 +0000 (11:24 +0200)]
zink: return false on failure
We do this in the other pipe_buffer_map_range failure case, so it makes
sense that we need to do it here as well. If we don't, we'll end up
taking a crash in the check_query_results function, which will
dereference that pointer.
We also need to unmap the buffer if we fail, otherwise we'll leak.
CID: 1475925
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Thu, 26 Aug 2021 09:17:21 +0000 (11:17 +0200)]
zink: pctx can't be null here
We're checking if pctx is null here, but that can't be true. If it
could, then the code that follows would have immediately crashed.
A quick peek at other drivers seems to indicate that this is a safe
assumption.
CID: 1474410, 1474554
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Wed, 25 Aug 2021 22:09:44 +0000 (00:09 +0200)]
zink: do not dereference null-pointer
The "locations" pointer can be null here, and memcpying from a null
pointer is not okay.
CID: 1485978
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Wed, 25 Aug 2021 22:02:11 +0000 (00:02 +0200)]
zink: do not try to dereference null-key
We can't do any of this logic if key is NULL, because that means we'll
dereference memory close to a NULL-pointer.
While we're at it, add some asserts to the zink_fs_key and zink_vs_key
functions who would otherwise be responsible for giving us invalid
non-null pointers out of null-pointers.
CID: 1475973, 1475983
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Wed, 25 Aug 2021 19:44:46 +0000 (21:44 +0200)]
zink: avoid overflow when calculating size
If we multiply before we (implicitly) cast the result to the target
type, we needlessly risk overflowing the result.
CID: 1490790, 1475922
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Wed, 25 Aug 2021 19:38:17 +0000 (21:38 +0200)]
zink: initialize pQueueFamilyIndices
This silences a Coverity warning about an uninitialized variable.
CID: 1490800
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12559>
Erik Faye-Lund [Fri, 27 Aug 2021 08:05:52 +0000 (10:05 +0200)]
zink: do not warn about rare features until used
We currently cause the following scary warning to be printed on
start-up for *every* application regardless if that feature is being
used or not when run on top of ANV:
> WARNING: Some incorrect rendering might occur because the selected
> Vulkan device (Intel(R) UHD Graphics 620 (KBL GT2)) doesn't support
> base Zink requirements: line_rast_feats.stippledRectangularLines
> line_rast_feats.stippledSmoothLines
There's no need to scare the users about this, as most applications
don't care about these combinational features. So let's instead emit a
warning when these features are attempted (but failed) to be used
instead.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12587>
Erik Faye-Lund [Fri, 27 Aug 2021 08:32:41 +0000 (10:32 +0200)]
Revert "zink: always init bordercolor value for sampler"
This reverts commit
336dea90f09c5cefc46de5240da28950fdca0723.
This change was incorrect for two reasons:
1. We already initialize this field on line 334
2. Unconditionally setting this to
VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK breaks rendering with e.g
opaque white borders, because we've already matched those to a
non-custom enum value first.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12591>
Lepton Wu [Sat, 14 Aug 2021 22:52:47 +0000 (15:52 -0700)]
i965: Enable RGBX8888_SRGB format.
This is required by Android. Some Android games like nier reincarnation
show a black screen without this.
Signed-off-by: Lepton Wu <lepton@chromium.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12393>
Timur Kristóf [Tue, 24 Aug 2021 06:57:39 +0000 (08:57 +0200)]
aco: Consider maximum number of workgroups per CU/WGP on Navi.
No Fossil DB changes.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12517>
Timur Kristóf [Tue, 24 Aug 2021 06:44:54 +0000 (08:44 +0200)]
aco: Consider LDS usage by PS inputs in MaxWaves calculation.
Before PS waves are launched, PS inputs are moved from PC to LDS
and the corresponding part of the PC is deallocated.
Each PS input occupies 3 * vec4 (3 * 16 = 48 bytes) of LDS space.
See Figure 10.3 in the GCN3 ISA manual.
These limit occupancy the same way as other stages' LDS usage does.
Note that PS can request additional LDS space via EXTRA_LDS_SIZE,
so that also must be taken into account here.
No Fossil DB changes.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12517>
Mike Blumenkrantz [Fri, 6 Aug 2021 14:54:05 +0000 (10:54 -0400)]
zink: remove extra program ref from cached descriptor updates
this happens in draw/compute now
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12582>
Mike Blumenkrantz [Fri, 6 Aug 2021 14:52:25 +0000 (10:52 -0400)]
zink: avoid pulling in unused push descriptors for cached ubo0
instead of just reading the template bufferinfo data, ensure that only
buffers which will be read are added to the set to avoid stale cache entries
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12582>
Tomeu Vizoso [Fri, 27 Aug 2021 14:55:22 +0000 (16:55 +0200)]
iris/ci: Correctly set freq governors to max
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12598>
Tomeu Vizoso [Fri, 27 Aug 2021 14:55:07 +0000 (16:55 +0200)]
freedreno/ci: Correctly set freq governors to max
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12598>
Boris Brezillon [Fri, 27 Aug 2021 11:00:20 +0000 (13:00 +0200)]
panvk: Fix panvk_copy_fb_desc()
We should not skip the copy when the batch is attached a framebuffer
descriptor, quite the opposite. Let's drop the check instead of reversing
it since we are guaranteed to have an FB attached when
panvk_copy_fb_desc() is called.
Fixes:
792a0ab0b146 ("panvk: Prepare per-gen split")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12590>
Boris Brezillon [Fri, 27 Aug 2021 08:30:38 +0000 (10:30 +0200)]
panvk: Make the per-arch static lib depend on panvk_entrypoints.h
The panvk_entrypoints.h header is included by all panvk_vX_xxx.c
source files, without this dependency the build can fail.
Fixes:
792a0ab0b146 ("panvk: Prepare per-gen split")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Tested-by: Fabio Pedretti <pedretti.fabio@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12590>
Mike Blumenkrantz [Mon, 2 Aug 2021 17:44:30 +0000 (13:44 -0400)]
zink: more effectively utilize batch_usage for query destruction
there's no need to track the number of batches that a query is referenced on,
as all the tracking is already done by the batch_usage mechanism, so this
can be simplified to manage destruction based on whether batch_usage exists
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12578>
Mike Blumenkrantz [Mon, 2 Aug 2021 14:52:20 +0000 (10:52 -0400)]
zink: don't try to sync previous timestamp query qbo values
this makes no sense, so don't explode the qbo by trying
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12578>
Mike Blumenkrantz [Mon, 2 Aug 2021 14:51:15 +0000 (10:51 -0400)]
zink: move time query ending out to zink_end_query
time queries only need to be ended when the api ends them, not per-cmdbuf
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12578>
Mike Blumenkrantz [Mon, 2 Aug 2021 16:01:54 +0000 (12:01 -0400)]
zink: improve threadsafe qbo access
these should be staging resources since they're being read from often,
and this allows dropping the UNSYNCHRONIZED flag from map since it should
be inferred
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12578>
Mike Blumenkrantz [Tue, 15 Jun 2021 15:52:20 +0000 (11:52 -0400)]
zink: make zink_gfx_pipeline_state::vertices_per_patch a bitfield
this is clamped to MAX_PATCH_VERTICES
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Tue, 15 Jun 2021 15:52:08 +0000 (11:52 -0400)]
zink: repack zink_gfx_pipeline_state
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Tue, 15 Jun 2021 15:35:02 +0000 (11:35 -0400)]
zink: convert rasterizer pipeline components to bitfield
this reduces the hashed pipeline key size by 53 bits
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Thu, 15 Jul 2021 13:20:27 +0000 (09:20 -0400)]
zink: steal a bit from rast_samples in pipeline state
zink only handles values up to 64, so this still has an extra bit
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Fri, 18 Jun 2021 16:47:29 +0000 (12:47 -0400)]
zink: add clip_halfz to rasterizer hw state
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Tue, 15 Jun 2021 14:23:38 +0000 (10:23 -0400)]
zink: repack zink_rasterizer_hw_state
this is now 11 bits
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Tue, 15 Jun 2021 13:51:13 +0000 (09:51 -0400)]
zink: zero viewport and scissor count in pipeline with dynamic state1
this is illegal
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Tue, 15 Jun 2021 13:50:26 +0000 (09:50 -0400)]
zink: move viewport count into dynamic state1 part of pipeline hash
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Mike Blumenkrantz [Tue, 15 Jun 2021 13:47:50 +0000 (09:47 -0400)]
zink: move dynamic state1 pipeline members into substruct
this is a bit easier to manage
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12586>
Boris Brezillon [Tue, 18 May 2021 14:03:29 +0000 (16:03 +0200)]
panfrost: v7 does not support RGB32_UNORM textures
Cc: mesa-stable
Fixes:
c6bdd976e611 ("panfrost: Split out v6/v7 format tables")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12588>
Yevhenii Kharchenko [Mon, 22 Feb 2021 22:15:15 +0000 (00:15 +0200)]
iris: fix layer calculation for TEXTURE_3D ReadPixels() on mip-level>0
Fixes assert when ReadPixels() called to read from FBO to
GL_PIXEL_PACK_BUFFER, on mip-level > 0, since num_layers
wasn't properly calculated with mip-level.
v2: patched 'iris_create_sampler_view' function instead of
'resolve_sampler_views'. Just like it was suggested in this
function's comment.
The logic of fix is similar to one in 'update_image_surface' function
of i965 driver, which is introduced in commit
f9fd0cf4790cb2a530e75d1a2206dbb9d8af7cb2.
With a slight change: setting array_len=1, like it was done in
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5808 ,
since minifying depth fails KHR-GLES2.texture_3d.filtering tests.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4145
Fixes:
3c979b0e ('iris: add some draw resolve hooks')
Signed-off-by: Yevhenii Kharchenko <yevhenii.kharchenko@globallogic.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9253>
Samuel Pitoiset [Thu, 26 Aug 2021 09:00:33 +0000 (11:00 +0200)]
radv: remove unecessary radv_finishme() for invalid color formats
Something really bad happen (likely driver bug) if this is triggered.
Replace with some assertions to catch an eventual issue in debug build.
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/12556>
Samuel Pitoiset [Thu, 26 Aug 2021 08:56:04 +0000 (10:56 +0200)]
radv: remove useless check about number of samples in the HW resolve path
Although this can likely hang, this is invalid and should be caught
by the validation layers. There is many ways to hang the GPU with VK,
this check alone is useless.
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/12556>
Samuel Pitoiset [Thu, 26 Aug 2021 08:54:53 +0000 (10:54 +0200)]
radv: remove outdated radv_finishme() in the HW resolve path
Resolving layered MSAA images is actually implemented by the HW
resolve path but never used because the driver uses the compute path.
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/12556>
Mike Blumenkrantz [Fri, 11 Jun 2021 15:20:30 +0000 (11:20 -0400)]
zink: set primitive restart with extended dynamic state2
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Fri, 11 Jun 2021 14:04:24 +0000 (10:04 -0400)]
zink: bump dynamic pipeline state count
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Fri, 11 Jun 2021 14:04:17 +0000 (10:04 -0400)]
zink: template for VK_EXT_extended_dynamic_state2
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Fri, 11 Jun 2021 14:03:59 +0000 (10:03 -0400)]
zink: hook up VK_EXT_extended_dynamic_state2
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Mon, 14 Jun 2021 18:24:32 +0000 (14:24 -0400)]
zink: no-op prim changes for pipeline recalc
this is no longer part of pipeline hash since we're using dynamic state
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Mon, 14 Jun 2021 18:23:05 +0000 (14:23 -0400)]
zink: consolidate pipeline hash tables
with dynamic prim type, pipelines can now be grouped by base prim type
instead of requiring a different pipeline for overall prim type
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Mon, 14 Jun 2021 18:21:39 +0000 (14:21 -0400)]
zink: use dynamic prim type
this is part of dynamic state but wasn't used since it required
actual work to effectively make use of it
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12585>
Mike Blumenkrantz [Fri, 16 Jul 2021 13:44:51 +0000 (09:44 -0400)]
zink: batch mem barrier hooks
memory barriers are redundant, so batch them and apply based on actual
usage to be slightly more efficient
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12576>
Mike Blumenkrantz [Thu, 15 Jul 2021 18:26:04 +0000 (14:26 -0400)]
zink: slim down streamout component of mem barrier hook
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12576>
Mike Blumenkrantz [Thu, 15 Jul 2021 14:42:39 +0000 (10:42 -0400)]
zink: remove query flush from memory barrier hook
qbos are separate buffers with their own barriers, so this just stalls
pointlessly
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12576>
Mike Blumenkrantz [Fri, 11 Jun 2021 13:50:09 +0000 (09:50 -0400)]
zink: use ctx gfx prim mode for draw comparisons
just being consistent
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Fri, 25 Jun 2021 13:32:44 +0000 (09:32 -0400)]
zink: init ctx->gfx_prim_mode to nonzero value to trigger pipeline changes
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Fri, 11 Jun 2021 13:32:35 +0000 (09:32 -0400)]
zink: reorder gfx program/pipeline/descriptor binds if dynamic state is present
this enables deferring the heavy lifting until the rest of the state updates are
done
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Wed, 28 Jul 2021 19:09:11 +0000 (15:09 -0400)]
zink: remove extra unsetting of ctx->vertex_state_changed
this is already managed in zink_get_gfx_pipeline
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Thu, 10 Jun 2021 12:17:33 +0000 (08:17 -0400)]
zink: pass current program's shader array, not ctx array
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Thu, 10 Jun 2021 12:17:09 +0000 (08:17 -0400)]
zink: remove attachment count from pipeline hash
this is redundant
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Thu, 10 Jun 2021 10:43:34 +0000 (06:43 -0400)]
zink: declare ctx var during blend state bind
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12572>
Mike Blumenkrantz [Tue, 10 Aug 2021 14:17:39 +0000 (10:17 -0400)]
zink: assert precise queries are occlusion queries
this should always be the case
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12583>
Mike Blumenkrantz [Tue, 10 Aug 2021 14:16:15 +0000 (10:16 -0400)]
zink: require occlusionQueryPrecise for occlusion queries
ensure this is present to avoid driver explosions
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12583>
Mike Blumenkrantz [Mon, 21 Jun 2021 13:53:35 +0000 (09:53 -0400)]
zink: always init bordercolor value for sampler
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12574>
Mike Blumenkrantz [Thu, 15 Jul 2021 12:33:23 +0000 (08:33 -0400)]
zink: implement PIPE_QUERY_GPU_FINISHED
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12573>
Ella-0 [Mon, 2 Aug 2021 22:13:25 +0000 (22:13 +0000)]
v3dv: Implement VK_EXT_pipeline_creation_feedback
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12371>
Erik Faye-Lund [Wed, 25 Aug 2021 12:16:52 +0000 (14:16 +0200)]
microsoft/compiler: trivial fixes to error-handling
We're really bad at making sure we report errors when we fail to
allocate memory. This makes us a bit better...
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12541>
Erik Faye-Lund [Wed, 25 Aug 2021 12:22:04 +0000 (14:22 +0200)]
microsoft/compiler: return errors from get_n_src
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12541>
Erik Faye-Lund [Wed, 25 Aug 2021 09:45:13 +0000 (11:45 +0200)]
microsoft/compiler: remove needless error-returns
There's no root error-conditions in this code, just code that assumes
they exist and tries to handle them.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12541>
Filip Gawin [Wed, 25 Aug 2021 21:19:36 +0000 (23:19 +0200)]
nir: fix ifind_msb_rev by using appropriate type
As you can see comparion "x < 0" doesn't make
sense if x is unsigned.
Fixes:
a5747f8a ("nir: add opcodes for *find_msb_rev and lowering ")
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12548>
Filip Gawin [Tue, 24 Aug 2021 19:28:19 +0000 (21:28 +0200)]
nir: fix shadowed variable in nir_lower_bit_size.c
Fixes:
6d792989924 ("nir/lower_bit_size: fix lowering of {imul,umul}_high")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12527>
Yiwei Zhang [Sun, 22 Aug 2021 22:26:54 +0000 (22:26 +0000)]
venus: conditionally enable async descriptor set allocation
When VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT is not used to
create the pool, set allocation is guaranteed to not return
VK_ERROR_FRAGMENTED_POOL, and we can safely move set allocation to async
after doing resource tracking in the driver.
Enable after fully tested with assert(false) in the failure case.
Tested with:
- dEQP-VK.api.descriptor*
- dEQP-VK.api.object_management.*
- dEQP-VK.binding_model.descriptor*
- dEQP-VK.descriptor_indexing.*
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Fri, 20 Aug 2021 21:50:14 +0000 (21:50 +0000)]
venus: check descriptor allocations against pool resource
Only kick in when async_set_allocation is enabled.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Wed, 25 Aug 2021 01:17:58 +0000 (01:17 +0000)]
venus: descriptor set to track descriptor count of last binding
Track the descriptor count to be used instead of the variable descriptor
count to avoid duplicate checks in later accounting.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Fri, 20 Aug 2021 18:53:04 +0000 (18:53 +0000)]
venus: descriptor pool to track pool state
It also tracks whether async set allocation is enabled.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Tue, 24 Aug 2021 20:57:23 +0000 (20:57 +0000)]
venus: layout to track variable descriptor count binding info
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Fri, 20 Aug 2021 18:57:43 +0000 (18:57 +0000)]
venus: descriptor layout to track more binding infos
Rename existing max_binding to last_binding to be consistent.
1. layout to track last binding index
2. binding to track descriptor type
3. binding to track descriptor count
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Sat, 21 Aug 2021 00:27:04 +0000 (00:27 +0000)]
venus: add vn_descriptor_set_layout_init
Just a refactoring without functional changes.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Yiwei Zhang [Fri, 20 Aug 2021 21:06:25 +0000 (21:06 +0000)]
venus: refactor failure path for sets allocation
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12501>
Sergii Melikhov [Wed, 25 Aug 2021 12:13:21 +0000 (15:13 +0300)]
dri2: Fix Null pointer dereferences
Fix defect reported by Coverity Scan CID-1490794
Fixes:
0d42033b26 ("glx/dri2: Require the driver to support v4 of __DRI_DRI2")
Signed-off-by: Sergii Melikhov <sergii.v.melikhov@globallogic.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12563>
Christian Gmeiner [Wed, 23 Jun 2021 05:58:01 +0000 (07:58 +0200)]
etnaviv: remove primconvert
This is no longer used.
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12516>
Christian Gmeiner [Wed, 23 Jun 2021 05:56:43 +0000 (07:56 +0200)]
etnaviv: export supported prim types
This is now handled by gallium, so the related codepaths can be dropped.
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Acked-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12516>
Juan A. Suarez Romero [Thu, 26 Aug 2021 08:36:48 +0000 (10:36 +0200)]
broadcom/qpu: remove duplicated opcode variable
`opcode` is wrote twice.
Fixes CID 1490798: Incorrect expression (EVALUATION_ORDER).
Fixes:
8a5f2228dbb ("broadcom/qpu: add new lookup opcode description helper")
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12554>
Timur Kristóf [Wed, 11 Aug 2021 08:09:04 +0000 (10:09 +0200)]
aco: Use workgroup size from input shader info.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>
Timur Kristóf [Wed, 11 Aug 2021 06:53:55 +0000 (08:53 +0200)]
radv: Remove superfluous workgroup size calculations.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>
Timur Kristóf [Wed, 11 Aug 2021 06:54:28 +0000 (08:54 +0200)]
radv: Calculate workgroup sizes in radv_pipeline.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>
Timur Kristóf [Wed, 11 Aug 2021 06:57:04 +0000 (08:57 +0200)]
ac: Calculate workgroup sizes of HW stages that operate in workgroups.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12321>
Samuel Pitoiset [Wed, 25 Aug 2021 12:36:30 +0000 (14:36 +0200)]
ci: update the list of skipped tests for Fiji/RADV
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12553>
Tomeu Vizoso [Mon, 5 Jul 2021 08:00:54 +0000 (10:00 +0200)]
panvk: Initialize timestamp for disk cache
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12552>