Alyssa Rosenzweig [Tue, 29 Nov 2022 02:30:49 +0000 (21:30 -0500)]
asahi: Fix Z32S8 harder
Fixes dEQP-GLES3.functional.texture.format.sized.2d.depth32f_stencil8_pot after
stencil texturing broke it.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Tue, 29 Nov 2022 01:35:38 +0000 (20:35 -0500)]
asahi: Don't shadow idle resources
Pointless allocation+memcpy.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Wed, 23 Nov 2022 19:32:10 +0000 (14:32 -0500)]
asahi: Model alignment of occlusion query indices
8-byte offsets.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Mon, 28 Nov 2022 04:55:55 +0000 (23:55 -0500)]
asahi: Identify some more fields used with layered
These values depend on the framebuffer width/height and maybe other
stuff. Maybe strides.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Wed, 23 Nov 2022 20:24:05 +0000 (15:24 -0500)]
asahi: Identify XML for anisotropic filtering
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Wed, 23 Nov 2022 19:46:22 +0000 (14:46 -0500)]
asahi: Check-box implement rasterizer discard
Passes dEQP-GLES3.functional.rasterizer_discard.*
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Sat, 26 Nov 2022 03:39:48 +0000 (22:39 -0500)]
asahi: Wire in 1D (array) textures
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Thu, 24 Nov 2022 17:22:05 +0000 (12:22 -0500)]
asahi: Dirty track VS/FS key updates
drawoverhead 1 score doubled to 7668.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Thu, 24 Nov 2022 17:05:09 +0000 (12:05 -0500)]
asahi: Separate VS/FS shader keys
First remove agx_shader_key from asahi_shader_key. It's trivial. agx_shader_key
is going to go away soon now that we lower everything in NIR.
Then everything else is mutually exclusive between stages. That means much
less to hash. drawoverhead test 1 from 2331 to 3443.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Thu, 24 Nov 2022 16:15:44 +0000 (11:15 -0500)]
asahi: Implement invalidate_resource
From Panfrost. This lets us avoid storing depth/stencil attachments at the end
of the frame in GLES.
On my 4K monitor, glmark2 -btexture at fullscreen goes from 705fps to 1150fps. I
assume gains on real workloads will be smaller.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Thu, 24 Nov 2022 15:55:52 +0000 (10:55 -0500)]
asahi: Track batch masks on ZS/blend CSO
Adapted from panfrost, with the work happening at CSO create time instead of
draw time allowing us to do more sophisticated analysis. We'll use these for
accurate masks in a moment.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Thu, 24 Nov 2022 15:57:24 +0000 (10:57 -0500)]
asahi: Dirty track blend state
We'll want this to reduce variant lookups eventually.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Thu, 24 Nov 2022 16:29:39 +0000 (11:29 -0500)]
asahi: Enable dirty tracking
Whoops. drawoverhead test 1 score from 496 -> 2377.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Alyssa Rosenzweig [Tue, 29 Nov 2022 00:21:21 +0000 (19:21 -0500)]
ail: Save level_offsets_compressed_B
So we can bind specific mip levels for rendering into compressed Z/S.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20123>
Aleksey Komarov [Fri, 9 Dec 2022 18:17:42 +0000 (21:17 +0300)]
panfrost: implement clear_depth_stencil
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20238>
David Heidelberg [Fri, 9 Dec 2022 11:50:19 +0000 (12:50 +0100)]
nine: enable on panfrost
Also, enable required kmsro dependencies.
Tested-by: Aleksey Komarov <q4arus@ya.ru>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20238>
David Heidelberg [Fri, 9 Dec 2022 11:06:52 +0000 (12:06 +0100)]
panfrost: implement clear_render_target
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Co-authored-by: Aleksey Komarov <q4arus@ya.ru>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Aleksey Komarov <q4arus@ya.ru>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20238>
David Heidelberg [Fri, 9 Dec 2022 11:06:39 +0000 (12:06 +0100)]
panfrost: Handle resources without depth in batch_to_fb_info
Prevent preloading data from resources which doesn't exist.
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Aleksey Komarov <q4arus@ya.ru>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20238>
David Heidelberg [Fri, 9 Dec 2022 11:06:33 +0000 (12:06 +0100)]
panfrost: Implement GL_EXT_clip_control
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Co-authored-by: Aleksey Komarov <q4arus@ya.ru>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Aleksey Komarov <q4arus@ya.ru>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20238>
Paulo Zanoni [Fri, 2 Dec 2022 18:55:48 +0000 (10:55 -0800)]
intel: add devinfo->has_64bit_float_via_math_pipe
Unusual hardware features that require special hanlding usually get a
devinfo field, so do this for MTL's unordered DF types. This will
guarantee that any platform based on MTL (thus inheriting from
MTL_FEATURES) will automatically be handled in these special cases.
v2: s/has_unordered_64bit_float/has_64bit_float_via_math_pipe/ (Curro).
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Paulo Zanoni [Tue, 29 Nov 2022 01:44:38 +0000 (17:44 -0800)]
intel/compiler: fix intel_swsb_decode for newer platforms
In the previous patch we adjusted the scoreboard pass to take into
consideration a new case of unordered operations for TGL. Fix the
decoding as well.
v2: use intel_device_info_is_mtl() (Curro, Jordan)
v3: the part where we export num_sources_from_inst() is now a separate patch
(Curro).
v4: Work around false positive maybe-unitialized warning since Marge
uses -Werror=maybe-uninitialized (Marge).
Reviewed-by: Francisco Jerez <currojerez@riseup.net> (v3)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Paulo Zanoni [Thu, 8 Dec 2022 23:39:01 +0000 (15:39 -0800)]
intel/compiler: export brw_num_sources_from_inst
We want to call this from brw_disasm.c, so move it out to brw_eu.c
since it's about to become more of a shared utility function than
something specific to the EU validator.
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Paulo Zanoni [Sat, 3 Dec 2022 00:36:21 +0000 (16:36 -0800)]
intel/compiler: avoid 64bit SEL_EXEC on MTL
On MTL, instructions with DF type are unordered, executed in the math
pipe. This means that they require different SWSB dependency handling,
and also that in some cases such as MOVs it's generally faster to
simply use 2 smaller ordered moves than a single unordered MOV.
One problem we have with the current code is that generate_code() is
not setting the proper SWSB dependencies for the generated DF MOVs,
causing some tests to fail.
One solution would be to fix generate_code() by making it set the
appropriate dependencies. This was the first patch I wrote. Another
solution to this problem, pointed to us by Curro, is to change
required_exec_type() so we use UD instructions instead of DF, just
like we do with platforms that don't have 64 bit instructions, which
means there won't be anything to fix in generate_code(). The second
solution is what this patch implements.
This fixes at least:
- dEQP-VK.subgroups.arithmetic.framebuffer.subgroupmin_double_vertex
Thanks to Francisco Jerez for all the major help provided with this
problem.
Credits-to: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Paulo Zanoni [Thu, 21 Apr 2022 23:29:56 +0000 (16:29 -0700)]
intel/compiler: avoid (RegDist, SBID) on DF instructions on MTL
When we use this form there's no way to specify which pipe RegDist
refers to, so there are a few rules to figure this out, which is what
inferred_sync_pipe() implements. But for MTL there's no long pipe and
the documentation does not explicitly explain what should be the
inferred type for its long (DF) instructions - which are out-of-order,
by the way. One way to interpret this is that such case should be
avoided. So add the extra check to entirely avoid this case.
Notice that this is not actually fixing any bug, since returning
TGL_PIPE_LONG (what we do today) will actually make these DF
instructions incompatible with every in-order instruction, so we'll
never opt to use the (RegDist, SBID) form anyway. But still, it's
better to have this case explicitly documented instead of having it
covered by a semi coincidence.
v2: use intel_device_info_is_mtl() (Curro, Jordan)
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Paulo Zanoni [Fri, 1 Apr 2022 22:06:36 +0000 (15:06 -0700)]
intel/compiler: on MTL, DF instructions run in the math pipe
Adjust the scoreboard code to take that into account.
Fixes at least:
- dEQP-VK.glsl.builtin.precision_double.refract.compute.vec3
- dEQP-VK.glsl.builtin.precision_double.matrixcompmult.compute.mat4
v2: use intel_device_info_is_mtl() (Curro, Jordan)
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Francisco Jerez [Sat, 19 Feb 2022 05:58:24 +0000 (21:58 -0800)]
intel/fs: Make the result of is_unordered() dependent on devinfo.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20072>
Lionel Landwerlin [Fri, 9 Dec 2022 13:54:57 +0000 (15:54 +0200)]
Revert "anv: compile anv_acceleration_structure.c"
This reverts commit
74d0be27ae9eb666df948874a20a3b4464df7db1.
Also remove anv_acceleration_structure.c, it was meant to be removed
earlier. There was probably a rebase issue somewhere.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20248>
Chia-I Wu [Fri, 9 Dec 2022 19:07:50 +0000 (11:07 -0800)]
freedreno/a6xx: fix blend all_mrt_write_mask
Fix all_mrt_write_mask when independent_blend_enable is false.
Otherwise, lrz write is always diabled with MRT when
independent_blend_enable is false. This fixes a 2% perf regression for
multiple gfxbench benchmarks.
Fixes:
0132c22de75 ("freedreno/a6xx: Don't disable LRZ for invalid channels")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20254>
Kenneth Graunke [Thu, 29 Sep 2022 18:24:14 +0000 (11:24 -0700)]
iris: Improve direct CPU map heuristics
We were promoting reads with a valid primary to direct CPU maps even
if the mmap mode was IRIS_MMAP_WC, which would mean uncached reads from
VRAM. In that case, GPU blits are in fact useful!
We were also only checking for !DISCARD_RANGE rather than MAP_READ,
which isn't a great idea for image maps, given the discussion in the
previous commit about image map semantics.
The original code was also just confusingly structured. Make a helper
function with clearly defined cases where we want to bail on CPU maps.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Kenneth Graunke [Tue, 4 Oct 2022 21:01:27 +0000 (14:01 -0700)]
iris: Only copy existing data into staging images with PIPE_MAP_READ
When performing transfer maps on images that require staging buffers
(say, for presenting a linear view of tiled memory), we were reading
the existing contents of the buffer into the staging resource on map
unless PIPE_MAP_DISCARD_RANGE was set.
The thinking was to support partial writes. If you map a subrectangle
of an image, but then only write selective pixels - should it preserve
the existing contents of the mapped region? I believed that it should,
unless you pass PIPE_MAP_DISCARD_RANGE to explicitly say that that it's
okay to invalidate the destination region.
However, that does not appear to be the interpretation favored by other
Mesa developers (in particular Michel Dänzer and Marek Olšák). The
radeonsi driver does not do this readback from the destination region
to the staging buffer unless you pass PIPE_MAP_READ. If you want to
do a partial write and preserve contents, you need to pass both flags:
(PIPE_MAP_READ | PIPE_MAP_WRITE). Passing READ is expected to come
with an associated cost.
OpenGL defines GL_MAP_INVALIDATE_RANGE_BIT for mapping buffer objects,
which is translated to PIPE_MAP_DISCARD_RANGE. However, unextended
OpenGL doesn't define mapping textures. There are two main sources of
image maps: our internal MapTextureImage() hook, and gbm_bo_map().
I've audited our internal MapTextureImage() calls, and while some do
pass PIPE_MAP_DISCARD_RANGE, almost all of them wholly overwrite the
mapped region, and those that care about combining with existing image
contents all pass PIPE_MAP_READ. So this should work there.
GBM defines three flags: GBM_BO_TRANSFER_READ, WRITE, and READ_WRITE.
There is no defined "invalidate range" bit. In issue #6020, Matthias
Treydte notes that this extra readback can cause performance problems,
and with iris's current interpretation, there's no way to avoid it.
During that discussion, Michel and Matthias both argued that
GBM_BO_TRANSFER_WRITE should invalidate the destination contents and
avoid the readback, while GBM_BO_TRANSFER_READ_WRITE would preserve it.
This patch makes iris follow that model for image mappings, removing
readback on staging maps for both detiling and stall avoidance, unless
PIPE_MAP_READ is passed. I believe we can change this with impunity.
For buffer objects, Ian Romanick and I both agree that partial writes
should be supported, and GL_MAP_INVALIDATE_RANGE_BIT exists precisely
to indicate that you should spend effort preserving existing contents.
So we continue doing readback for buffers unless PIPE_MAP_DISCARD_RANGE
is flagged, for now. While I think this is work, it also seems to be
undertested in the CTS and Piglit.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6020
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Kenneth Graunke [Fri, 30 Sep 2022 00:56:29 +0000 (17:56 -0700)]
iris: Return idle status from iris_invalidate_buffer, skip busy checks
If we successfully replace the backing storage for a buffer, we know
that it's idle, and the transfer map code can mark it unsynchronized
right away, letting us skip redundant resource_is_busy() checks.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Kenneth Graunke [Fri, 30 Sep 2022 01:06:42 +0000 (18:06 -0700)]
iris: Don't replace backing storage for exported buffers.
We already gave out the old BO...or acquired it from somewhere which
may be affecting it. We simply can't replace the backing store.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Kenneth Graunke [Fri, 30 Sep 2022 00:38:53 +0000 (17:38 -0700)]
iris: Promote DISCARD_RANGE to DISCARD_WHOLE_RESOURCE where possible
This allows us to replace the backing storage for a buffer, which
means we'd have an idle buffer and thus could do an unsynchronized
mapping where we otherwise wouldn't.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Kenneth Graunke [Thu, 29 Sep 2022 23:12:07 +0000 (16:12 -0700)]
iris: Use persistent mappings for pinned memory (userptr)
This is a port of Nicolai's
b52721e3b693e113aa537d163c8a855169e7b75d
from radeonsi. Because GL_AMD_pinned_memory guarantees that mappings
will refer to the same underlying page, we need to avoid using staging
maps. Using a persistent map is a reasonable way to accomplish this.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Kenneth Graunke [Thu, 29 Sep 2022 23:36:01 +0000 (16:36 -0700)]
iris: Delete map->dest_had_defined_contents
Dead since commit
6cc09699cd77286f7601a1afab60dc9fa0bafd48.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19209>
Rhys Perry [Mon, 7 Nov 2022 15:02:53 +0000 (15:02 +0000)]
nir/gather_info: use nir_ssa_scalar_resolved
This lets us skip copies.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19597>
Rhys Perry [Wed, 9 Nov 2022 16:22:27 +0000 (16:22 +0000)]
vtn: add mesh output and task_payload to vtn_mode_is_cross_invocation
This fixes a potential race condition, and removes output loads (which
should not exist in the EXT_mesh_shader).
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7391
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19597>
Rhys Perry [Wed, 9 Nov 2022 16:21:32 +0000 (16:21 +0000)]
nir: add task_payload and shader_out to nir_var_vec_indexable_modes
Since these can be cross-invocation, we need this to write individual
components without race conditions or loads.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Gitlab: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7391
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19597>
Rhys Perry [Mon, 7 Nov 2022 14:28:51 +0000 (14:28 +0000)]
radv: fix task payload lowering when shared_memory_explicit_layout=true
If shared_memory_explicit_layout=true, we would have skipped lowering task
payload variables to explicit types.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19597>
Rhys Perry [Mon, 7 Nov 2022 18:01:29 +0000 (18:01 +0000)]
radv: fix mesh shaders with null winsys
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19597>
Rebecca Mckeever [Fri, 9 Dec 2022 20:08:14 +0000 (14:08 -0600)]
panvk: Delete panvk_CmdSetDeviceMask, panvk_GetDeviceGroupPeerMemoryFeatures
Delete panvk_CmdSetDeviceMask and panvk_GetDeviceGroupPeerMemoryFeatures
so that the vk_common_* version will be used instead. This will avoid
repeated code.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Rebecca Mckeever [Fri, 9 Dec 2022 20:07:59 +0000 (14:07 -0600)]
hasvk: Delete VK_KHR_device_group provided entrypoints
Delete anv_CmdDispatch, anv_CmdSetDeviceMask, and
anv_GetDeviceGroupPeerMemoryFeatures so that the vk_common_*
versions will be used instead. This will avoid repeated code.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Rebecca Mckeever [Fri, 9 Dec 2022 20:07:48 +0000 (14:07 -0600)]
anv: Delete VK_KHR_device_group provided entrypoints
Delete anv_CmdDispatch, anv_CmdSetDeviceMask, and
anv_GetDeviceGroupPeerMemoryFeatures so that the vk_common_*
versions will be used instead. This will avoid repeated code.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Rebecca Mckeever [Fri, 9 Dec 2022 20:07:33 +0000 (14:07 -0600)]
tu: Delete VK_KHR_device_group provided entrypoints
Delete tu_CmdDispatch, tu_CmdSetDeviceMask, and
tu_GetDeviceGroupPeerMemoryFeatures so that the vk_common_*
versions will be used instead. This will avoid repeated code.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Rebecca Mckeever [Fri, 9 Dec 2022 20:07:17 +0000 (14:07 -0600)]
v3dv: Delete VK_KHR_device_group provided entrypoints
Delete v3dv_CmdDispatch, v3dv_CmdSetDeviceMask, and
v3dv_GetDeviceGroupPeerMemoryFeatures so that the vk_common_*
versions will be used instead. This will avoid repeated code.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Rebecca Mckeever [Fri, 9 Dec 2022 20:07:00 +0000 (14:07 -0600)]
radv: Delete VK_KHR_device_group provided entrypoints
Delete radv_CmdDispatch, radv_CmdSetDeviceMask, and
radv_GetDeviceGroupPeerMemoryFeatures so that the vk_common_*
versions will be used instead. This will avoid repeated code.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Rebecca Mckeever [Fri, 9 Dec 2022 20:06:14 +0000 (14:06 -0600)]
vulkan/runtime: Add VK_KHR_device_group provided entrypoints
Add entrypoints vk_common_CmdDispatch, vk_common_CmdSetDeviceMask,
and vk_common_GetDeviceGroupPeerMemoryFeatures in Mesa Vulkan runtime so
that they are available to all drivers.
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20218>
Thong Thai [Tue, 6 Dec 2022 21:02:42 +0000 (16:02 -0500)]
gallium: add new variable for video frame statistics
Video encoder previously reuses the associated_data variable to output
encoding statistics, but it ended up breaking when transcoding. This
commit adds a new variable just for statistics.
Fixes:
2d1bd619dfd ("frontends/va: add ability for encoder to output statistics")
Signed-off-by: Thong Thai <thong.thai@amd.com>
Reviewed-by: Boyuan Zhang <Boyuan.Zhang@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20190>
Rhys Perry [Fri, 9 Dec 2022 16:41:33 +0000 (16:41 +0000)]
radv: remove some unnecessary 64-bit IO handling
nir_lower_io() lowers these to 32-bit.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20137>
Rhys Perry [Fri, 2 Dec 2022 18:05:21 +0000 (18:05 +0000)]
radv: fix streamout with different streams in the same varying slot
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20137>
Rhys Perry [Thu, 8 Dec 2022 19:33:58 +0000 (19:33 +0000)]
aco/ra: don't swap create_vector operand with definition blocker for SGPRs
There is no SGPR swap instruction, we always need 3 XORs.
fossil-db (navi21):
Totals from 76 (0.06% of 135636) affected shaders:
Instrs: 58400 -> 58347 (-0.09%); split: -0.10%, +0.01%
CodeSize: 312580 -> 312368 (-0.07%); split: -0.08%, +0.01%
Latency: 843333 -> 843180 (-0.02%); split: -0.02%, +0.00%
InvThroughput: 126431 -> 126412 (-0.02%)
Copies: 4008 -> 3955 (-1.32%); split: -1.47%, +0.15%
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20240>
Rhys Perry [Thu, 8 Dec 2022 19:32:25 +0000 (19:32 +0000)]
aco/ra: don't swap p_create_vector operand with definition blocker for scc
SCC is 1-bit, and we can't copy a 32-bit value into it.
Fixes dEQP-VK.spirv_assembly.type.scalar.i32.iequal_tesse with
ACO_DEBUG=noopt.
No fossil-db changes.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes:
9476986e6f6 ("aco/ra: special-case get_reg_for_create_vector_copy()")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20240>
Georg Lehmann [Thu, 29 Sep 2022 15:36:55 +0000 (17:36 +0200)]
nir/opt_algebraic: Optimize open coded bfm.
Foz-DB Navi21:
Totals from 1553 (1.15% of 134913) affected shaders:
SpillVGPRs: 2246 -> 2223 (-1.02%); split: -1.42%, +0.40%
CodeSize:
10409156 ->
10410720 (+0.02%); split: -0.03%, +0.04%
Instrs: 1899725 -> 1898773 (-0.05%); split: -0.07%, +0.02%
Latency:
71225814 ->
71118314 (-0.15%); split: -0.21%, +0.06%
InvThroughput:
13384926 ->
13330369 (-0.41%); split: -0.47%, +0.06%
VClause: 38309 -> 38284 (-0.07%); split: -0.17%, +0.11%
SClause: 70743 -> 70706 (-0.05%)
Copies: 167296 -> 167230 (-0.04%); split: -0.28%, +0.24%
Branches: 42446 -> 42444 (-0.00%); split: -0.01%, +0.00%
PreVGPRs: 95191 -> 95188 (-0.00%)
Some minor instructions count regressions in parallel-rdp
because v_bfm_b32 can't use SDWA, but overall an improvement.
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Acked-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18887>
Ruijing Dong [Thu, 8 Dec 2022 21:10:14 +0000 (16:10 -0500)]
frontends/va: fix gst videotestsrc h264 enc fail issue.
problem:
when doing "gst-launch-1.0 -v videotestsrc num-buffer=10 !
vaapih264enc ! fakeink"
The command will fail due to gst will fetch the first
available supported format in the list, it becomes P010_LE
due to the commit in
[
0b02db3007]
frontends/va: fixed av1 decoding 10bit ffmpeg output YUV issue
fix:
move the P010_LE code block to the end of the function, the sequence
of the supported formats restored to its original.
cc: mesa-stable
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/20242>
Yonggang Luo [Mon, 5 Dec 2022 14:25:33 +0000 (09:25 -0500)]
frontend/osmesa: inherit pipe_frontend_drawable instead of allocating separately
This is required by st/mesa now.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Yonggang Luo [Mon, 5 Dec 2022 14:25:33 +0000 (09:25 -0500)]
frontend/hgl: inherit pipe_frontend_drawable instead of allocating separately
This is required by st/mesa now.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 21:46:24 +0000 (16:46 -0500)]
gallium: rename st_framebuffer_iface -> pipe_frontend_drawable, etc.
Also rename:
iface -> drawable
stfb -> drawable (where it means dri_drawable and not st_framebuffer)
stfbi -> drawable or pdrawable (if drawable exists)
pipe_frontend_drawable* is really just dri_drawable* for DRI, and WGL/GLX
have their own variants. This makes it easier to understand what kind of
object is being used.
I always wondered what st_framebuffer_iface, iface, stfbi, iface_stamp,
and iface_ID actually mean. Now those terms are gone forever.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 21:28:14 +0000 (16:28 -0500)]
gallium: remove pipe_frontend_screen::destroy callback, call it directly
This is the only one implemented by mesa/state_tracker.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 21:25:28 +0000 (16:25 -0500)]
gallium: clean up comments in api.h, cosmetic changes
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 21:04:52 +0000 (16:04 -0500)]
gallium: remove st_framebuffer_iface::st_manager_private
It's always equal to self.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 21:02:24 +0000 (16:02 -0500)]
frontend/glx: inherit st_framebuffer_iface instead of allocating separately
All other frontends do it like this. This enables the next cleanup.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 20:37:50 +0000 (15:37 -0500)]
gallium: remove unused ST_CONTEXT_ERROR_*
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 20:26:34 +0000 (15:26 -0500)]
gallium: rename st_manager_private -> st_screen
Oh you didn't know that st_context had a screen object too?
Yeah, this code is full of surprises.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 20:19:57 +0000 (15:19 -0500)]
gallium: remove st_profile_type in favor of gl_api, move out st_context_attribs
st_context_attribs is moved into st_context.h.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 20:11:58 +0000 (15:11 -0500)]
gallium: remove unnecessary ST_PROFILE_*_MASK definitions
This didn't do anything.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 20:00:57 +0000 (15:00 -0500)]
gallium: use PIPE_CONTEXT_* flags instead of ST_CONTEXT_FLAG_*
where PIPE_CONTEXT_* flags are available
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 19:40:28 +0000 (14:40 -0500)]
gallium: move st function declarations from api.h into st_context.h
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 19:33:14 +0000 (14:33 -0500)]
gallium: remove enum st_texture_type
just use GLenum
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 18:40:42 +0000 (13:40 -0500)]
st/mesa: remove st_context callbacks for dri/glx/wgl frontends
Use direct calls.
Other changes:
- kopper called thread_finish twice in the same function.
Remove the second call.
- Context creation had no-op initializations of cso_context and pipe.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 17:59:22 +0000 (12:59 -0500)]
gallium: remove st_context_iface, use st_context directly
st_context_iface was the base class that st_context inherited.
Just use st_context.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 18:15:16 +0000 (13:15 -0500)]
gallium/pp: don't use st_context_iface, use an explicit callback
st_context_iface will be removed.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 18:08:53 +0000 (13:08 -0500)]
gallium/hud: don't use st_context_iface, use an explicit callback
st_context_iface will be removed.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 17:21:45 +0000 (12:21 -0500)]
gallium: rename st_context_iface::st_manager_private to frontend_context
It's really dri_context for DRI.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 17:15:41 +0000 (12:15 -0500)]
gallium: remove duplicated st_context_iface::st_context_private
It's equal to fscreen, which is almost next to it.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Marek Olšák [Sun, 27 Nov 2022 17:11:12 +0000 (12:11 -0500)]
gallium: rename st_manager to pipe_frontend_screen (think dri_screen)
st_manager is really the base class of dri_screen, so let's call it
pipe_frontend_screen. That's a much better name than the vague "st_manager".
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Yonggang Luo <luoyonggang@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20027>
Erik Faye-Lund [Thu, 8 Dec 2022 12:13:33 +0000 (13:13 +0100)]
zink: do not leave needless shader temps around
This used to not matter, but since we started emitting shader-temps
properly, this causes issues where we end up with samplers and images as
shader-temps. That causes asserts while emitting them.
So let's remove the unused vars as well.
This fixes a piglit regression that somehow went unnoticed on CI.
Fixes:
85964945e7c ("zink: emit vars with nir_var_shader_temp mode")
Tested-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20227>
Kenneth Graunke [Thu, 8 Dec 2022 06:34:40 +0000 (22:34 -0800)]
intel/compiler: Delete sampler key handling for planar format stuff
i965 used these, but Gallium drivers do this lowering via a separate
nir_lower_tex call from st/mesa. Vulkan drivers don't use these at all.
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20223>
Kenneth Graunke [Thu, 8 Dec 2022 05:18:02 +0000 (21:18 -0800)]
intel/compiler: Delete key->msaa_16
None of the drivers have used this since we dropped i965, and BLORP
no longer uses it as of the previous commit. We can also drop the
former compressed_multisample_tex_mask (now padding) field so that
things remain 64-bit aligned.
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20223>
Kenneth Graunke [Thu, 8 Dec 2022 03:29:43 +0000 (19:29 -0800)]
intel/blorp: Set key->msaa_16 unconditionally on Gfx9+
This will result in us using the TXF_CMS_W message rather than the
TXF_CMS message on Skylake through Tigerlake for 2/4/8x MSAA blits,
which is technically slightly worse. However, it shouldn't be that
much worse: the TXF_CMS message was removed altogether on Alchemist.
iris and anv set key->msaa_16 unconditionally, to avoid paying the
cost of shader recompiles for a miniscule gain. crocus and hasvk
don't need to set it as they don't support 16x MSAA. BLORP already
recompiles based on the sample count, so it could easily keep doing
this for the minor benefit. But avoiding it will let us drop the
entire msaa_16 key field out of the compiler, which is nice.
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20223>
Kenneth Graunke [Thu, 8 Dec 2022 04:20:12 +0000 (20:20 -0800)]
intel: Drop compressed_multisample_layout_mask from the compiler keys
The compiler looks at this key field to determine whether to perform
an MCS fetch for a txf_ms or samples_identical texture message, if a
nir_tex_src_ms_mcs_intel source wasn't provided. If it isn't set,
it instead uses constant 0 (nothing is compressed).
All of the drivers (iris, crocus, anv, hasvk) unconditionally set this
to ~0 because we don't want to pay for costly shader recompiles (which
can cause nasty stuttering). Most textures are compressed anyway, and
the hardware ignores the l2dms MCS parameter if MCS is disabled.
The only user was BLORP, which sets the key field based on whether the
texture's aux usage has MCS. But if it has MCS, it also does the MCS
fetch itself and supplies it directly. Otherwise, it relies on the
compiler to fill in the 0 value. But it could easily just provide the
0 value itself in that case and not rely on the compiler at all.
With that fixed, we can just drop the key fields entirely. We leave
them as padding for now to avoid repacking structures; we won't need
to after the next commits anyway.
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20223>
Jianxun Zhang [Fri, 11 Nov 2022 22:40:40 +0000 (14:40 -0800)]
intel/common: use format struct in aux mapping
Refactor aux mapping with the new format struct and helpers.
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20167>
Jianxun Zhang [Fri, 11 Nov 2022 22:51:04 +0000 (14:51 -0800)]
intel/common: initialize format of aux mapping on GFX12
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20167>
Jianxun Zhang [Thu, 8 Dec 2022 01:55:57 +0000 (17:55 -0800)]
intel/common: fix style of some comments in intel_aux_map.h
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20167>
Jianxun Zhang [Fri, 11 Nov 2022 20:34:46 +0000 (12:34 -0800)]
intel/common: Add a new struct to describe AUX mapping format
The new struct and some helper functions are for further
refactoring.
Reworks:
* Jordan: Refactor code around aux format array
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20167>
Jianxun Zhang [Wed, 9 Nov 2022 22:27:36 +0000 (14:27 -0800)]
intel/common: Add an enum of formats of AUX mapping
The new enum allows us to support multiple formats in the future.
Signed-off-by: Jianxun Zhang <jianxun.zhang@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20167>
Lionel Landwerlin [Thu, 8 Dec 2022 13:08:30 +0000 (15:08 +0200)]
intel: add MTL performance metrics
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20228>
Väinö Mäkelä [Fri, 18 Nov 2022 06:56:46 +0000 (08:56 +0200)]
hasvk: Allow aliasing with modifiers for WSI images
Ignore ALIAS_BIT when format comes from WSI because
we have the ability to bind the MEMORY_BINDING_PRIVATE
from the other WSI image.
This commit is the same as
f350b78b but for hasvk.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19840>
Gert Wollny [Thu, 8 Dec 2022 11:14:29 +0000 (12:14 +0100)]
r600/sfn: implement GDS op channel mask evaluation
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Thu, 8 Dec 2022 10:44:56 +0000 (11:44 +0100)]
r600/sfn: allocate pinned registers as SSA
Don't allocate the helper_invocation register as pinned, because it is
not an SSA value.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Thu, 8 Dec 2022 10:40:23 +0000 (11:40 +0100)]
r600/sfn: nir_op_vec results don't need channel pinning
This will be handled by the op that uses the vector
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Thu, 8 Dec 2022 10:39:01 +0000 (11:39 +0100)]
r600/sfn: implement TF allowed channel mask override
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Thu, 8 Dec 2022 09:13:39 +0000 (10:13 +0100)]
r600/sfn: clean up TF emission in TCS
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Wed, 7 Dec 2022 17:16:57 +0000 (18:16 +0100)]
r600/sfn: Cleanup copy-prop into vec4 source values
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Tue, 6 Dec 2022 20:30:48 +0000 (21:30 +0100)]
r600/sfn: scheduler allow more lookahead on ALU instr
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Tue, 6 Dec 2022 09:30:36 +0000 (10:30 +0100)]
r600/sfn: Make use of variable length DOT
This frees some alu slots for better group scheduling.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Tue, 6 Dec 2022 09:09:19 +0000 (10:09 +0100)]
r600/sfn: Don't require assignemnt in same block to copy-propagate
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Tue, 6 Dec 2022 07:16:54 +0000 (08:16 +0100)]
r600/sfn: No need to pin LDS dests to chan
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Mon, 5 Dec 2022 16:08:31 +0000 (17:08 +0100)]
r600/sfn: allow copy propagation to LDS read dest
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Mon, 5 Dec 2022 16:03:45 +0000 (17:03 +0100)]
r600/sfn: allow more copy-propagate with dest chan fixed.
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>
Gert Wollny [Mon, 5 Dec 2022 15:29:27 +0000 (16:29 +0100)]
r600/sfn: Allow copy prop into GDS sources values
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20205>