Rhys Perry [Tue, 16 Aug 2022 10:41:15 +0000 (11:41 +0100)]
docs/radv: advertise VK_KHR_global_priority
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/18078>
Joshua Ashton [Sat, 6 Aug 2022 02:15:13 +0000 (02:15 +0000)]
radv: Remove workarounds for D3D9 feedback loops
RADV supports VK_EXT_attachment_feedback_loop_layout now,
and DXVK is using since
63d582a6e7ab63e18503d1238101705a818f5ad0 which
will be in DXVK 2.0, so no need for these anymore.
Older versions of DXVK will still be affected by the issue however.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17920>
PixelyIon [Thu, 11 Aug 2022 14:12:27 +0000 (19:42 +0530)]
tu: Make KGSL wait IOCTLs interrupt-safe
`tu_WaitForFences` and `tu_GetFenceStatus` ended up using `ioctl` instead of
`safe_ioctl` and as a result were responsible for handling `EAGAIN`/`EINTR`
but they weren't handled correctly in wait calls which led to dropping the
wait if the thread received a signal without automatic syscall restarting on
release builds and an assertion due to `errorno` being `EINTR` on debug
builds.
This fixes early wait returns which would lead to UB or cause crashes due to
hitting debug assertions in Skyline Emulator.
Signed-off-by: Mark Collins <mark@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18031>
Samuel Pitoiset [Mon, 15 Aug 2022 14:16:18 +0000 (16:16 +0200)]
radv: advertise VK_KHR_global_priority
Similar to VK_KHR_buffer_device_address which has also been promoted
from EXT.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18061>
Samuel Pitoiset [Wed, 10 Aug 2022 13:53:40 +0000 (15:53 +0200)]
radv: remove radv_graphics_pipeline_info completely
RADV now uses vk_graphics_pipeline_state from Vulkan common code which
is more convenient for eg. graphics pipeline libraries.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Wed, 10 Aug 2022 13:48:19 +0000 (15:48 +0200)]
radv: convert radv_vertex_input_info to vk_vertex_input_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Wed, 10 Aug 2022 08:04:46 +0000 (10:04 +0200)]
radv: convert radv_rendering_info to vk_render_pass_state
This also converts VkAttachmentSampleCountInfoAMD which is part of
vk_render_pass_state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Wed, 10 Aug 2022 07:44:34 +0000 (09:44 +0200)]
radv: convert radv_color_blend_info to vk_color_blend_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 15:50:09 +0000 (17:50 +0200)]
radv: convert radv_multisample_info to vk_multisample_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 14:27:19 +0000 (16:27 +0200)]
radv: convert radv_input_assembly_info to vk_input_assembly_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 14:15:34 +0000 (16:15 +0200)]
radv: convert radv_fragment_shading_rate_info to vk_fragment_shading_rate_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 14:08:25 +0000 (16:08 +0200)]
radv: convert radv_depth_stencil_info to vk_depth_stencil_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 13:23:53 +0000 (15:23 +0200)]
radv: convert radv_rasterization_info to vk_rasterization_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 13:06:44 +0000 (15:06 +0200)]
radv: convert radv_discard_rectangle_info to vk_discard_rectangles_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 12:59:19 +0000 (14:59 +0200)]
radv: convert radv_viewport_info to vk_viewpoert_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 12:55:22 +0000 (14:55 +0200)]
radv: convert radv_tessellation_info to vk_tessellation_state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 12:55:28 +0000 (14:55 +0200)]
radv: call vk_graphics_pipeline_state_fill() when initializing gfx pipelines
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 16:35:52 +0000 (18:35 +0200)]
radv: slightly change the color write enable mask
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Wed, 10 Aug 2022 08:41:30 +0000 (10:41 +0200)]
radv: do not translate primitive topology during gfx info initialization
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 16:14:47 +0000 (18:14 +0200)]
radv: do not translate blend op/factor during gfx info initialization
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 16:00:02 +0000 (18:00 +0200)]
radv: do not translate logic op during gfx info initialization
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 13:41:31 +0000 (15:41 +0200)]
radv: do not translate polygon mode during gfx info initialization
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 13:16:04 +0000 (15:16 +0200)]
radv: simplify the depth clip enable/disable logic
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 13:12:47 +0000 (15:12 +0200)]
radv: store the provoking vertex mode in the graphics pipeline info
For switching to the common graphics pipeline state.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 15:05:09 +0000 (17:05 +0200)]
vulkan: add attachment sample count info AMD support
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 14:34:10 +0000 (16:34 +0200)]
vulkan: fix initializing the primitive topology when it's dynamic
The PSO still contain the topology class and some Vulkan drivers like
RADV need to know this.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Samuel Pitoiset [Tue, 9 Aug 2022 13:01:28 +0000 (15:01 +0200)]
vulkan: initialize discardRectangleCount even if the state is dynamic
Only the array of rectangles should be dynamic.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18015>
Kenneth Graunke [Sat, 13 Aug 2022 08:11:58 +0000 (01:11 -0700)]
intel/compiler: Use named NIR intrinsic const index accessors
In the early days of NIR, you had to prod at inst->const_index[]
directly, but a long while back, we added handy accessor functions
that let you use the actual name of the thing you want instead of
memorizing the exact order of parameters.
Also rewrite a comment I had a hard time parsing.
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18067>
Iago Toral Quiroga [Wed, 27 Jul 2022 12:32:24 +0000 (14:32 +0200)]
v3dv: vkCmdClearAttachments no longer generates its own RCL
We improved this a long time ago and now it emits a clear rect inside
the current subpass job instead of emitting its own job with its own
RCL, so we no longer need to handle this as an exception to the rule.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Wed, 27 Jul 2022 09:42:01 +0000 (11:42 +0200)]
v3dv: add a heuristic for double-buffer mode
For this we add a scoring system that evaluates various aspects of
the draw calls in a job.
If the cost of the geometry side of the pipeline is too high, then
we may pay too high a price in double-buffer mode because with smaller
tile size may will probably have more vertex shader invocations in the
render and binning stages.
On the other hand, if rendering cost is not high enough, we may not
have enough rendering work to hide the latency of tile stores in
double-buffer mode.
Also, because we need to make a decision after we know all the draw
calls in a job, but the double-buffer enable bit comes in the
TILE_BINNING_MODE_CFG that needs to be emitted first in the binning
command list before the draw calls are recorded, if we decide to
enable it we need to rewrite that packet and we need to size the
tile state properly to account for the extra tiles. For this
purpose we delay tile state setup for render pass jobs until we are
finishing a job.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Wed, 27 Jul 2022 08:42:38 +0000 (10:42 +0200)]
v3dv: tell job_compute_frame_tiling whether we want to use double-buffer mode
We want to have control over the double-buffer setting here so we can control
explicitly from the driver when we want to enable this mode.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Wed, 27 Jul 2022 08:21:44 +0000 (10:21 +0200)]
v3dv: postpone tile state allocation for render pass jobs
These are jobs for which we may want to enable double-buffering,
which affects tile state allocation. Since the idea is that we
want to decide about double buffering late, we also want to
postpone allocation of the tile state until we are about to
emit the RCL for the job.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Wed, 27 Jul 2022 08:06:50 +0000 (10:06 +0200)]
v3dv: add a v3dv_job_allocate_tile_state helper
If we enable double-buffer we are reducing the tile size, and thus,
we'll need more tiles and a larger tile state allocation, so we'll
need to call to this helper.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Wed, 27 Jul 2022 07:53:45 +0000 (09:53 +0200)]
v3dv: fix tile state allocation
If we had decided that we only need one layer worth of tile state
we should actually only allocate one layer.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Fri, 22 Jul 2022 07:45:10 +0000 (09:45 +0200)]
v3dv: track if a job is compatible with double-buffer mode
MSAA is not compatible with double-buffer mode. Also, jobs that emit
tile loads or that don't have any stores can't take advantage
of double-buffer mode.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Fri, 22 Jul 2022 07:41:47 +0000 (09:41 +0200)]
v3dv: move check_needs_load/store helpers to unversioned code
These helpers don't depend on the V3D version and we are going
to need them outside v3dvx_cmd_buffer.c soon.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Iago Toral Quiroga [Fri, 22 Jul 2022 06:13:28 +0000 (08:13 +0200)]
broadcom/compiler: track number of TMU operations in prog data
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17854>
Eric Engestrom [Thu, 28 Jul 2022 19:56:13 +0000 (20:56 +0100)]
docs/features: drop unsupported VK_KHR_display & VK_KHR_get_display_properties2 for lavapipe
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:41:12 +0000 (19:41 +0100)]
docs/features: start tracking extensions supported by powervr
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:37:38 +0000 (19:37 +0100)]
docs/features: start tracking extensions supported by panvk
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:32:40 +0000 (19:32 +0100)]
docs/features: start tracking extensions supported by dozen
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6330
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:23:06 +0000 (19:23 +0100)]
docs/features: add missing extensions supported by venus
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:17:37 +0000 (19:17 +0100)]
docs/features: add missing extensions supported by turnip
Reviewed-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:10:48 +0000 (19:10 +0100)]
docs/features: add missing extensions supported by radv
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 18:05:00 +0000 (19:05 +0100)]
docs/features: add missing extensions supported by lavapipe
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Eric Engestrom [Thu, 28 Jul 2022 17:04:02 +0000 (18:04 +0100)]
docs/features: add missing extensions supported by anv
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17796>
Chia-I Wu [Fri, 12 Aug 2022 21:22:34 +0000 (14:22 -0700)]
turnip: handle fence wrapping in autotune
fence values can be 0 after wrapping. We should not treat 0 specially.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18040>
Chia-I Wu [Fri, 12 Aug 2022 20:53:37 +0000 (13:53 -0700)]
turnip: remove unused member in tu_submission_data
buffers_count is set but unused.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18040>
Chia-I Wu [Fri, 12 Aug 2022 20:02:30 +0000 (13:02 -0700)]
turnip: fix a use-after-free in autotune
When removing old histories, check against gpu fence. Otherwise,
pending_results could have dangling pointers to the removed histories.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7055
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18040>
Mike Blumenkrantz [Tue, 9 Aug 2022 12:51:18 +0000 (08:51 -0400)]
zink: use a maybe more accurate wild guess for pcp-less gpl
this is only reachable with an env var, so being wrong is still okay,
but maybe be right slightly more often
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17955>
Mike Blumenkrantz [Tue, 9 Aug 2022 12:42:51 +0000 (08:42 -0400)]
zink: (correctly) require extendedDynamicState2PatchControlPoints for GPL
this is otherwise broken
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17955>
Mike Blumenkrantz [Thu, 11 Aug 2022 12:52:00 +0000 (08:52 -0400)]
zink: require EXT_non_seamless_cube_map for GPL support
it's impossible to precompile without this
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17955>
Yonggang Luo [Fri, 12 Aug 2022 16:41:33 +0000 (00:41 +0800)]
c11: Remove mtx_try from mtx enums as it's not a part of c11 standard
And we support for four types of mtx init enum values that define by C11 standard:
0 mtx_plain - a simple, non-recursive mutex is created.
2 mtx_timed - a non-recursive mutex, that supports timeout, is created.
1 mtx_plain | mtx_recursive - a recursive mutex is created.
3 mtx_timed | mtx_recursive - a recursive mutex, that supports timeout, is created.
musl library also use these enum combination
Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18036>
Georg Lehmann [Sun, 14 Aug 2022 17:25:09 +0000 (19:25 +0200)]
aco: Use v_cmpx pre GFX10.
Foz-DB Vega10:
Totals from 29508 (21.85% of 135041) affected shaders:
CodeSize:
184345656 ->
184345820 (+0.00%)
Instrs:
35906154 ->
35906195 (+0.00%)
Latency:
581696114 ->
581530021 (-0.03%); split: -0.03%, +0.00%
InvThroughput:
245625572 ->
245561351 (-0.03%); split: -0.03%, +0.00%
Copies: 3134925 -> 3278672 (+4.59%)
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
Georg Lehmann [Mon, 15 Aug 2022 10:10:38 +0000 (12:10 +0200)]
aco/assembler: Fix v_cmpx pre GFX10.
The second destination is implicit exec.
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
Georg Lehmann [Sun, 14 Aug 2022 11:56:30 +0000 (13:56 +0200)]
aco: Fix optimizing branching sequence with s_and_saveexec.
This optimization was broken for two reasons:
- s_and_saveexec has two operands, the copy value and exec
- s_and_saveexec has an exec read, so exec_write_used will always be true
before we find branch_exec_val_idx
Foz-DB Navi21:
Totals from 31453 (23.31% of 134913) affected shaders:
CodeSize:
204831260 ->
204831156 (-0.00%)
Instrs:
38157117 ->
38157091 (-0.00%)
Latency:
533708882 ->
531211721 (-0.47%); split: -0.47%, +0.00%
InvThroughput:
107088408 ->
106719188 (-0.34%); split: -0.35%, +0.00%
Copies: 2326179 -> 2502490 (+7.58%)
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
Georg Lehmann [Mon, 15 Aug 2022 09:51:55 +0000 (11:51 +0200)]
aco/assembler: Fix v_cmpx with SDWA.
We need to use the implicit destination.
Fixes:
baab6f18c91 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
Georg Lehmann [Mon, 15 Aug 2022 11:26:23 +0000 (13:26 +0200)]
aco: Check that we don't override exec_val operands during branching sequence optimization.
Fixes:
baab6f18c91 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
Timur Kristóf [Mon, 15 Aug 2022 08:03:58 +0000 (10:03 +0200)]
aco: Check for instructions that inhibit the branching sequence optimization.
Fixes:
baab6f18c91 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
Timur Kristóf [Sun, 14 Aug 2022 13:15:18 +0000 (15:15 +0200)]
aco: Fix invalidated reference in branching sequence optimization.
Inserting in the instructions vector may invalidate the exec_val reference,
so do that last.
Fixes:
baab6f18c91 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
David Heidelberg [Fri, 29 Jul 2022 17:19:45 +0000 (19:19 +0200)]
ci: do not install i386 sub-architecture in VK images
Not required by apitrace.
debian/x86_test-vk image size:
previous: 1.8G
after: 1.0G
v2: added back winehq-stable (missing symlinks)
Acked-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17834>
David Heidelberg [Mon, 15 Aug 2022 08:05:02 +0000 (10:05 +0200)]
ci: use apt-get in x86_test-vk instead of apt
apt shouldn't be used in CI, see:
https://unix.stackexchange.com/questions/590699/should-i-use-apt-or-apt-get-in-shell-scripting
Acked-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17834>
Tomeu Vizoso [Fri, 12 Aug 2022 08:54:14 +0000 (10:54 +0200)]
ci: Report failures to GitLab for LAVA jobs
We were uploading the junit.xml to the artifacts, but weren't telling
GitLab about it.
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18028>
Tomeu Vizoso [Fri, 12 Aug 2022 07:08:01 +0000 (09:08 +0200)]
iris/ci: Set FDO_CI_CONCURRENT for all Chromebook jobs
For some reason we were missing these and that was causing some CPUs to
remain idle, and some boards to have too high load and some tests timing
out occasionally.
Set FDO_CI_CONCURRENT for the number of cores, plus one to account for
GPU and I/O waits.
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18028>
Tomeu Vizoso [Fri, 12 Aug 2022 06:25:04 +0000 (08:25 +0200)]
iris/ci: Test on Jasper Lake Chromebooks
8 such boards have been added to Collabora's lab, let's run all tests
from the selected set.
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18028>
Qiang Yu [Mon, 8 Aug 2022 01:42:16 +0000 (09:42 +0800)]
nir/lower_gs_intrinsics: fix primitive count for points
When primitive is points, EndPrimitive can't be used to count
primitive. Need to use vertex count instead. And it's also not
needed to do vertex per primitive count and overwrite incomplete
primitive work for points.
Fixes:
2be99012e95 ("nir: Add ability to count emitted GS primitives.")
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17805>
Mike Blumenkrantz [Fri, 5 Aug 2022 17:54:58 +0000 (13:54 -0400)]
zink: fix program cache comparisons
I missed this when pipe shader types were merged with mesa types
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:34:57 +0000 (12:34 -0400)]
zink: delete zink_shader_stage()
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:31:44 +0000 (12:31 -0400)]
zink: remove all pipe_shader_type usage
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:20:10 +0000 (12:20 -0400)]
zink: PIPE_SHADER_TYPES -> MESA_SHADER_STAGES
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:19:03 +0000 (12:19 -0400)]
zink: explicitly define ZINK_GFX_SHADER_COUNT
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:17:37 +0000 (12:17 -0400)]
zink: ZINK_SHADER_COUNT -> ZINK_GFX_SHADER_COUNT
more descriptive
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:10:11 +0000 (12:10 -0400)]
zink: use statically sized array for descriptor allocation
this will never be more than 100
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 16:02:34 +0000 (12:02 -0400)]
zink: rename _lazy descriptor stuff
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:43:57 +0000 (11:43 -0400)]
zink: reorder some descriptor structs
should yield slightly better packing
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:42:09 +0000 (11:42 -0400)]
zink: merge batch descriptor data onto batch state
also clean up some related descriptor code
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:32:04 +0000 (11:32 -0400)]
zink: merge program descriptor data onto program struct
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:29:59 +0000 (11:29 -0400)]
zink: merge context descriptor data onto context struct
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:25:36 +0000 (11:25 -0400)]
zink: merge batch descriptor data structs
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:23:31 +0000 (11:23 -0400)]
zink: merge context descriptor data structs
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:21:15 +0000 (11:21 -0400)]
zink: remove descriptor dummy_set
while well-intentioned, this didn't actually do anything
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:20:16 +0000 (11:20 -0400)]
zink: delete some unused descriptor struct members
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:15:39 +0000 (11:15 -0400)]
zink: move lazy descriptor data types into zink_types.h
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 15:05:47 +0000 (11:05 -0400)]
zink: move all types/enums into zink_types.h
this significantly untangles all the header includes and fixes an issue
where zink_descriptors.h couldn't be included by certain files due to
type conflicts
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 13:02:48 +0000 (09:02 -0400)]
zink: delete unused struct types
so old
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 12:52:59 +0000 (08:52 -0400)]
zink: move lazy descriptor functions into zink_descriptors.c
the merginating continues
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 12:27:03 +0000 (08:27 -0400)]
zink: delete a lot of descriptor code
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 12:23:10 +0000 (08:23 -0400)]
zink: remove descriptor vtable
no longer used
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 12:16:18 +0000 (08:16 -0400)]
zink: remove checks for lazy descriptor mode + templates ext
these are always true
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Mike Blumenkrantz [Thu, 4 Aug 2022 12:10:02 +0000 (08:10 -0400)]
zink: remove cached descriptor mode
this just removes the enum(s) for it, making it functionally inaccessible
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18051>
Alyssa Rosenzweig [Sun, 7 Aug 2022 23:17:40 +0000 (19:17 -0400)]
asahi: Identify depth clip mode bit
Setting this bit (at the batch level, not the draw level!) switches to
[-1, 1] clipping instead of Metal's preferred [0, 1] clipping. Using
this bit allows us to drop the clip_halfz lowering we had before, saving
2 instructions in every vertex shader.
Fixes dEQP-GLES2.functional.depth_range.*
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17948>
Alyssa Rosenzweig [Sun, 7 Aug 2022 23:27:06 +0000 (19:27 -0400)]
asahi: Don't advertise ARB_clip_control yet
Our depth handling is pretty busted as it is. Adding clip control into
the mix will only make things more complicated. Best not worry about
that yet. Implementing this ext "properly" has some wrinkles: explain
them.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17948>
Timur Kristóf [Sat, 6 Aug 2022 06:47:22 +0000 (08:47 +0200)]
aco: Add faster code path to store_lds for consecutive write mask.
This makes it more likely to hit the fast path for count == 1
in the split_store_data function.
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/17923>
Eric Engestrom [Fri, 12 Aug 2022 16:51:51 +0000 (17:51 +0100)]
vk/device-select-layer: fix .sType of VkPhysicalDeviceGroupProperties
The validation layers complained:
> Validation Error: [ VUID-VkPhysicalDeviceGroupProperties-sType-sType ] Object 0: VK_NULL_HANDLE, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0xc9edee8b | vkEnumeratePhysicalDeviceGroups: parameter pPhysicalDeviceGroupProperties[0].sType must be VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_GROUP_PROPERTIES The Vulkan spec states: sType must be VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_GROUP_PROPERTIES (https://www.khronos.org/registry/vulkan/specs/1.3-extensions/html/vkspec.html#VUID-VkPhysicalDeviceGroupProperties-sType-sType)
Signed-off-by: Eric Engestrom <eric@igalia.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Fixes:
c196ffaca6d ("vk-device-select: add device group support")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18037>
Lucas Stach [Fri, 12 Aug 2022 17:40:20 +0000 (19:40 +0200)]
etnaviv: don't expose array and 3D texture support on pre-halti GPUs
3D and array texture support is a halti0 GPU feature and
gpu_supports_texture_target() already prevents such textures from being
constructed on older GPUs. Set the screen caps correctly to avoid
exposing EXT_texture_array and OES_texture_3D on those GPUs.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17996>
Lucas Stach [Wed, 10 Aug 2022 20:08:34 +0000 (22:08 +0200)]
etnaviv: allow 3D textures with TS in transfer
The resolve operation to the temporary resource properly iterates
over all layers of the 3D texture, so there is no need to fail the
transfer if we encounter a 3D texture that has a TS BO allocated
due to previous rendering.
Fixes piglit spec@arb_framebuffer_object@fbo-generatemipmap-3d
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17996>
Lucas Stach [Wed, 10 Aug 2022 19:24:03 +0000 (21:24 +0200)]
etnaviv: fix TS buffer allocation for 3D textures
For 3D textures the number of layers is the resource depth. As the TS buffer
allocation only looked at the array size it does not allocate enough TS
storage for 3D textures.
Fixes piglit spec@!opengl 1.2@copyteximage 3d
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17996>
Lucas Stach [Wed, 10 Aug 2022 19:27:27 +0000 (21:27 +0200)]
etnaviv: properly check number of layers in surface creation
3D textures store their number of layers in the resource depth. As the
assert only checks the array_size, it fails on perfectly valid surface
creation requests. Use the proper helper to get the number of layers.
Fixes piglit spec@arb_framebuffer_object@fbo-incomplete
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17996>
Pavel Ondračka [Wed, 10 Aug 2022 07:17:56 +0000 (09:17 +0200)]
r300: fix variables detection for paired ALU and TEX instructions in different branches
TEX instrutions can't write xyz and w to separate registers so we
need to create variables from them first, otherwise we can create
two variables from ALU writing the same register xyz and w in other
branch (this usually works when TEX is not present as the xyz and
w can read/write from different registers).
This fixes regalloc because the variables are later used as a
graph nodes.
The variable order should not matter but it slightly does (leading
to approx 0.3% shader-db temps increase as compared to previous
state), so just sort the variables list afterwards to be as close
to the previous behavior as possible and prevent the regression.
CC: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6936
Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17987>
Axel Davy [Thu, 28 Jul 2022 20:55:40 +0000 (22:55 +0200)]
frontend/nine: Fix ff position_t fallback when w = 0
For post-transformed vertices, w = 0 is similar to
w = 1. Replace the value to fix rcp(w).
It is common for apps to pass w = 0 for
position_t.
cc: mesa-stable
Signed-off-by: Axel Davy <davyaxel0@gmail.com>
Acked-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18021>