Dylan Baker [Thu, 19 May 2022 18:06:58 +0000 (11:06 -0700)]
docs: add release notes for 22.0.4
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16617>
Jason Ekstrand [Thu, 19 May 2022 17:47:50 +0000 (12:47 -0500)]
lavapipe: Use the correct ICD path on Win32
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16612>
Kenneth Graunke [Tue, 17 May 2022 23:37:35 +0000 (16:37 -0700)]
intel: Drop Wa_1409226450 (stall before instruction cache invalidation)
Production Tigerlake and DG1 hardware shouldn't need this workaround.
It was only needed on the very first steppings which never went public.
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16575>
Emma Anholt [Thu, 12 May 2022 00:33:11 +0000 (17:33 -0700)]
freedreno/ir3: Lower texture instructions used only for f2f16 to 16-bit.
2.5% improvement in gfxbench vk-5-normal. No obvious change on
gl-5-normal.
shader-db on Rob's android shaders:
total instructions in shared programs: 770644 -> 770595 (<.01%)
instructions in affected programs: 14880 -> 14831 (-0.33%)
total nops in shared programs: 167784 -> 167860 (0.05%)
nops in affected programs: 3351 -> 3427 (2.27%)
total non-nops in shared programs: 602860 -> 602735 (-0.02%)
non-nops in affected programs: 10523 -> 10398 (-1.19%)
total mov in shared programs: 19313 -> 19286 (-0.14%)
mov in affected programs: 365 -> 338 (-7.40%)
total cov in shared programs: 18075 -> 17978 (-0.54%)
cov in affected programs: 566 -> 469 (-17.14%)
total dwords in shared programs: 1612848 -> 1612596 (-0.02%)
dwords in affected programs: 13882 -> 13630 (-1.82%)
total last-baryf in shared programs: 56144 -> 55975 (-0.30%)
last-baryf in affected programs: 482 -> 313 (-35.06%)
total full in shared programs: 36094 -> 36092 (<.01%)
full in affected programs: 10 -> 8 (-20.00%)
total sstall in shared programs: 66986 -> 66923 (-0.09%)
sstall in affected programs: 1392 -> 1329 (-4.53%)
total systall in shared programs: 91244 -> 91072 (-0.19%)
systall in affected programs: 1194 -> 1022 (-14.41%)
total (sy) in shared programs: 4316 -> 4321 (0.12%)
(sy) in affected programs: 19 -> 24 (26.32%)
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16465>
Emma Anholt [Thu, 12 May 2022 02:55:45 +0000 (19:55 -0700)]
freedreno/ir3: Add support for 16-bit nir_texop_lod.
Same basic path, just do the rescaling in half float.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16465>
Emma Anholt [Mon, 2 May 2022 23:22:37 +0000 (16:22 -0700)]
turnip: Make RelaxedPrecision-decorated ALU ops 16-bit.
Improves gfxbench vk-5-normal performance 5.5%.
Fixes: #6346
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16465>
Emma Anholt [Tue, 26 Apr 2022 23:29:04 +0000 (16:29 -0700)]
spirv_to_nir: Cast RelaxedPrecision ALU op dests to mediump.
This is controlled by spirv_to_nir_options.relaxed_precision_alu, because
some drivers don't want it.
This gets us mostly 16-bit math on turnip in vk-5-normal.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16465>
Emma Anholt [Mon, 2 May 2022 22:54:12 +0000 (15:54 -0700)]
spirv: Use nir_vec_scalars() to simplify matrix transpose.
This should emit fewer instructions that need to be copy-propagated away.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16465>
Emma Anholt [Wed, 11 May 2022 23:31:33 +0000 (16:31 -0700)]
freedreno/ir3: Fix 16-bit bit_count.
No need to do the 16-bit lowering if it already is.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16465>
Lionel Landwerlin [Thu, 19 May 2022 10:09:25 +0000 (13:09 +0300)]
u_trace/anv/iris: drop cs argument for recording traces
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16605>
Danylo Piliaiev [Wed, 18 May 2022 10:08:22 +0000 (13:08 +0300)]
docs/u_trace: document u_trace usage
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16579>
Charmaine Lee [Wed, 18 May 2022 22:33:05 +0000 (15:33 -0700)]
svga: fix aa point
Use in_prim from current geometry shader to check for point prim type
when determine if aa point is enabled or not.
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16598>
Charmaine Lee [Wed, 18 May 2022 00:20:53 +0000 (17:20 -0700)]
svga: add need_texcoord_semantic to tgsi_add_point_sprite & tgsi_add_aa_point
Since PIPE_CAP_TGSI_TEXCOORD is now set in SVGA vgpu10 driver,
we need to add a new parameter need_texcoord_semantic to
tgsi_add_point_sprite and tgsi_add_aa_point
to allow setting texcoords using tgsi texcoord semantic.
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16598>
Lionel Landwerlin [Thu, 19 May 2022 07:43:23 +0000 (10:43 +0300)]
intel/ds: fix compilation
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6518
Fixes:
efc2782f970c ("intel/perf: store a copy of devinfo")
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16601>
Dylan Baker [Thu, 19 May 2022 16:03:09 +0000 (09:03 -0700)]
docs: Add calendar entries for 22.1 release.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16610>
Dylan Baker [Wed, 18 May 2022 20:31:26 +0000 (13:31 -0700)]
relnotes: Add sha256sum and fix minor formatting issues
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16610>
Dylan Baker [Wed, 18 May 2022 19:30:26 +0000 (12:30 -0700)]
docs: add release notes for 22.1.0
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16610>
Dylan Baker [Thu, 19 May 2022 16:00:04 +0000 (09:00 -0700)]
docs: update calendar and link releases notes for 22.1.0
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16610>
Alyssa Rosenzweig [Mon, 9 May 2022 20:03:27 +0000 (16:03 -0400)]
pan/va: Use ^ instead of ` to indicate last-use
This syncs the ISA syntax with other Valhall ISA users. It's also somewhat
easier to read.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 18 May 2022 15:57:55 +0000 (11:57 -0400)]
pan/va: Remove DISCARD.f32 destination
It doesn't actually write anything. This is a pointless divergence from Bifrost.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 16 May 2022 13:43:34 +0000 (09:43 -0400)]
pan/va: Handle 2-src blend in lower_split_src
Fixes assertion fail in shaders/dolphin/smg.1.shader_test
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Fri, 13 May 2022 00:21:42 +0000 (20:21 -0400)]
pan/bi: Validate vector widths
Now that our IR is much more strongly typed, and RA code quality depends on
correct typing, add a validation pass to make sure we didn't screw it up. This
pass found a massive number of bugs in early versions of this series.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 19:47:39 +0000 (15:47 -0400)]
pan/bi: Validate preload constraints are satisfied
We tightened the rules around preloading substantially and take advantage of the
rules in RA. The safe helpers it introduced should ensure the rules are
followed, but just in case, add a validation pass to check our work. This pass
found (multiple) bugs in early versions of this series.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Fri, 13 May 2022 15:54:19 +0000 (11:54 -0400)]
pan/bi: See through splits for var_tex fusion
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Thu, 12 May 2022 17:10:57 +0000 (13:10 -0400)]
pan/bi: Optimize split of collect
Required to get decent codegen from UBO pushing.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Thu, 12 May 2022 16:37:00 +0000 (12:37 -0400)]
pan/bi: Don't propagate discard
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 21:12:13 +0000 (17:12 -0400)]
pan/bi: Remove liveness metadata tracking
We don't use it for anything, and with no pass infrastructure it's just an
accident waiting to happen.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 19:39:56 +0000 (15:39 -0400)]
pan/bi: Simplify register precolouring in the IR
In the current IR, any register may be preloaded by reading it anywhere, and any
register may be precoloured by writing it anywhere. This is convenient for
instruction selection, but requires the register allocator to do considerable
gymnastics to ensure it doesn't clobber precoloured registers. It also breaks
the purity of our SSA representation, which complicates optimization passes
(e.g. copyprop).
Let's trade some instruction selection complexity for simplifying register
allocation by constraining how register precolouring works. Under the new model:
* Registers may only be preloaded at the start of the program.
* Precoloured destinations are handled explicitly by RA.
Internally, a stronger invariant is placed for preloading: registers may only be
preloaded by MOV.i32 instructions at the beginning of the block, and these moves
must be unique. These invariants ensure RA can trivially coalesce the moves.
A bi_preload helper is added as a safe version of bi_register respecting these
invariants, allowing a smooth transition for instruction selection.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:21:38 +0000 (12:21 -0400)]
pan/bi: Remove bi_word and bi_word_node
They are no longer used, as offsets are no longer used for normal values (only for
FAU). Keep it like that.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:26:42 +0000 (12:26 -0400)]
pan/bi: Scalarize copyprop
Reduces memory footprint.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:24:31 +0000 (12:24 -0400)]
pan/bi: Scalarize modifier propagation
Reduces memory footprint.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:24:22 +0000 (12:24 -0400)]
pan/bi: Scalarize bi_opt_cse
Reduces memory footprint.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:22:10 +0000 (12:22 -0400)]
pan/bi: Scalarize bi_lower_swizzle
Reduces memory footprint.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:18:36 +0000 (12:18 -0400)]
pan/va: Don't use bi_word in FAU unit test
It will be removed shortly, as the FAU construction helper should be used
instead.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 16:15:54 +0000 (12:15 -0400)]
pan/va: Use split for 64-bit lowering
Written in this way, this pass looks pretty silly...
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Tue, 10 May 2022 16:22:48 +0000 (12:22 -0400)]
pan/bi: Emit collect and split
..Rather than using offsets during instruction selection.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Tue, 10 May 2022 16:23:28 +0000 (12:23 -0400)]
pan/bi: Simplfy BLEND emit
We don't need to collect anything, now that Valhall handles this case correctly.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Tue, 10 May 2022 15:54:03 +0000 (11:54 -0400)]
pan/bi: Lift split/collect cache from AGX
Design based on ACO (and fruitful discussions with Daniel).
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Tue, 10 May 2022 14:03:16 +0000 (10:03 -0400)]
pan/bi: Create COLLECT during isel
This transitions us away from the fake SSA we currently use for vectors.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Tue, 10 May 2022 16:21:58 +0000 (12:21 -0400)]
pan/bi: Expand MAX_DESTS to 4
For splits.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 18 May 2022 16:44:38 +0000 (12:44 -0400)]
pan/bi: Fix mov and pack_32_2x16
Move can take in a vector and write a scalar, depending on the swizzle. We need
to handle this case. Split out mov and pack_32_2x16 so we can specify correct
behaviour for both. Also drop unused 1-bit boolean stuff which obscured the fix.
Fixes:
76cea8e27b3 ("panfrost: Fix pack_32_2x16 implementation")
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 16 May 2022 18:17:51 +0000 (14:17 -0400)]
pan/bi: Lower split/collect before RA
For transitioning to the new scalarized IR.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 11 May 2022 19:39:38 +0000 (15:39 -0400)]
pan/bi: Add bi_before_block cursor
Useful for preloading.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 9 May 2022 21:00:11 +0000 (17:00 -0400)]
pan/bi: Add collect and split instructions
These move-like instructions will be generated during instruction selection and
lowered before/after register allocation.
These need special printer support until we get dynamic sources/destinations.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 9 May 2022 21:28:38 +0000 (17:28 -0400)]
pan/bi: Add source/destination counts
In preparation for dynamic allocation, as needed for phi nodes and parallel
copies. For now, it just serves to simplify the semantics of splits and
collects.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 18 May 2022 14:52:29 +0000 (10:52 -0400)]
pan/bi: Use value-based interference with LCRA
"Revisiting Out-of-SSA Translation for Correctness, Code Quality, and
Efficiency" discusses "value-based interference": two variables interfere if and
only if there exists a point in the program where they are both live *with
different values*. In particular, the source and destination of a move do not
interfere a priori, because they have the same value at that point in the
program. (If a later instruction overwrites one, the required interference will
be added there).
We can use this idea to avoid some extra interferences, avoiding a regression in
moves from split/collect.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 18 May 2022 13:42:41 +0000 (09:42 -0400)]
pan/bi: Lower phis to scalar
If we don't lower phis to scalar, when we go out of SSA, we can get vector
nir_registers. In particular, we can get code like:
r0 = vec2 r0.y, r0.x
This code looks like a move, but is in fact a swap. The trivial lowering of vec2
would not work -- the following fails to swap correctly:
r0.x = r0.y
r0.y = r0.x
Currently, we generate temporaries to handle these cases. It's easy to move the
complexity to NIR, though, and we'll want to scalarize phis for SSA-based RA
anyway.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 18 May 2022 15:02:53 +0000 (11:02 -0400)]
pan/bi: +JUMP can't read same-cycle temp
Minor ISA detail missed in the Bifrost scheduler. I hit this in an early version
of this series (where a move feeding into a blend shader return was not
coalesced). Let's get it fixed in the scheduler.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 9 May 2022 19:56:25 +0000 (15:56 -0400)]
pan/va: Use 64-bit lowering for texturing
Texture instructions on Valhall take 64-bit sources. Now that we have
infrastructure to handle this properly, we don't need to use a non-SSA node to
hack around the optimization.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 9 May 2022 14:17:00 +0000 (10:17 -0400)]
pan/va: Lower split 64-bit sources
This ensures Valhall 64-bit constraints are respected in a simple way. It's not
the most efficient, though. Optimization is deferred until full Valhall support
is upstreamed and the RA is overhauled.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Mon, 9 May 2022 14:21:41 +0000 (10:21 -0400)]
pan/va: Mark more source sizes
This source size information will be consumed by the 64-bit lowering pass, so
ensure it's accurate. That means marking 32-bit and 64-bit sources explicitly on
message passing where it wouldn't match up with the type size suffix of the
instruction.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Alyssa Rosenzweig [Wed, 23 Mar 2022 16:05:10 +0000 (12:05 -0400)]
pan/bi: Update bi_count_write_registers for Valhall
We add some new instructions on Valhall with special register requirements
(texturing, atomics). Handle these appropriately so we can do RA on Valhall.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16585>
Jason Ekstrand [Wed, 18 May 2022 15:05:41 +0000 (10:05 -0500)]
vulkan: Fall back to raw data objects when deserializing if ops == NULL
This can happen if an object is serialized whose object type isn't in
the pipeline cache import ops. In this case, we generate a raw data
object and plan to turn it into the right object type later.
Fixes:
d35e78bb8536 ("vulkan/pipeline_cache: Implement deserialize for raw objects")
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16584>
Rhys Perry [Wed, 18 Aug 2021 13:07:42 +0000 (14:07 +0100)]
radv: validate shaders after linking passes
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5244
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12448>
Rhys Perry [Wed, 18 Aug 2021 12:50:49 +0000 (13:50 +0100)]
radv: add missing NIR_PASS() and switch from NIR_PASS_V()
Unlike NIR_PASS_V(), NIR_PASS() can skip printing the shader when
NIR_DEBUG=print.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Gitlab: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5244
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12448>
Rhys Perry [Thu, 14 Apr 2022 15:13:40 +0000 (16:13 +0100)]
radv: call nir_metadata_preserve in various lowering passes
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/12448>
Rhys Perry [Thu, 14 Apr 2022 15:26:45 +0000 (16:26 +0100)]
nir: call nir_metadata_preserve in nir_lower_memory_model
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/12448>
Rhys Perry [Wed, 18 Aug 2021 13:24:50 +0000 (14:24 +0100)]
nir: call nir_metadata_preserve in nir_vectorize_tess_levels
This is necessary to use this pass with the NIR_PASS() macro.
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/12448>
Rhys Perry [Wed, 18 Aug 2021 13:19:32 +0000 (14:19 +0100)]
nir: call nir_metadata_preserve in nir_io_add_const_offset_to_base
This is necessary to use this pass with the NIR_PASS() macro.
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/12448>
Rhys Perry [Wed, 18 Aug 2021 13:18:15 +0000 (14:18 +0100)]
nir: print file when validation fails
This should make it clear whether a validation failure happens in RADV or
zink.
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/12448>
Rhys Perry [Wed, 18 Aug 2021 12:50:58 +0000 (13:50 +0100)]
nir: allow NIR_PASS(_, )
If a user wants to skip printing the shader if no changes were made
without declaring a dummy variable for the progress.
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/12448>
Pierre-Eric Pelloux-Prayer [Mon, 16 May 2022 19:06:36 +0000 (21:06 +0200)]
radeonsi: wait for PS idle in si_set_framebuffer_state
This is needed to avoid write-after-read hazards in
texture -> render transitions.
This fixes fbo-depth tests that were flaky on GPUs (at
least sienna_cichlid and vega20).
Reviewed-by: Mihai Preda <mhpreda@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16561>
Marek Olšák [Thu, 19 May 2022 09:37:09 +0000 (05:37 -0400)]
amd: rename fishes to Navi21, Navi22, Navi23, Navi24, and Rembrandt
Reviewed-by: Mihai Preda <mhpreda@gmail.com>
Acked-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Martin Roukala <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16604>
Indrajit Kumar Das [Fri, 13 May 2022 11:34:15 +0000 (17:04 +0530)]
mesa/st: clear color buffers using color from a constant buffer
v2: fixed GLCTS failures (mareko)
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15318>
Indrajit Kumar Das [Fri, 13 May 2022 11:33:20 +0000 (17:03 +0530)]
mesa/st: add nir shader to clear color buffers using constant value
v2: use load_uniform instead of load_ubo to fix vc4
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15318>
Indrajit Kumar Das [Fri, 13 May 2022 11:32:14 +0000 (17:02 +0530)]
gallium/u_blitter: clear color buffers using color from a constant buffer
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15318>
Indrajit Kumar Das [Fri, 13 May 2022 11:23:51 +0000 (16:53 +0530)]
radeonsi: save the fs constant buffer to the util blitter context
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15318>
Tomeu Vizoso [Thu, 19 May 2022 05:36:33 +0000 (07:36 +0200)]
Revert "ci: Disable jobs to the Collabora lab"
This reverts commit
224544dc33c60b933b379eef067058ad31ef7bae.
Work has finished and the lab is up and running.
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16600>
Danylo Piliaiev [Wed, 11 May 2022 09:29:52 +0000 (12:29 +0300)]
util/u_trace: Add json output
If we want to load the u_trace output somewhere for analysis it's much
easier to deal with json than to parse strings.
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16475>
Samuel Pitoiset [Mon, 16 May 2022 14:53:05 +0000 (16:53 +0200)]
aco: remove unnecessary intrinsics that are lowered at the ABI level
Fixes:
f553076eaf1 ("aco: Remove now-superfluous intrinsics.")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16577>
Filip Gawin [Tue, 17 May 2022 16:10:21 +0000 (18:10 +0200)]
r300: keep negation if w is an inline constant
(in dataflow swizzles pass)
helps with:
dEQP-GLES2.functional.shaders.random.conditionals.combined.73
on r300 and r400
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16559>
Timothy Arceri [Wed, 18 May 2022 02:00:30 +0000 (12:00 +1000)]
nir/i915g/r300/nv30: skip marking varyings as flat in some drivers
Some older drivers don't support GLSL versions with the concept of flat
varyings and also don't support integers. Here we add a new setting to
make sure we don't use the optimisation that sets varyings to flat.
This setting helps us avoid marking varyings as flat and therefore
potentially having them changed to ints via varying packing.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6500
Fixes:
7647023f3bb5 ("glsl: enable the use of the nir based varying linker")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16573>
Jason Ekstrand [Wed, 18 May 2022 17:49:45 +0000 (12:49 -0500)]
v3dv: Loosen an assert in copy_buffer_to_image_shader
In
f99ac7f2de19 ("v3dv: Don't use color aspects for depth/stencil
images"), we stopped using color aspects for depth/stencil images in a
bunch of cases. This causes us to trigger an assert in
copy_buffer_to_image_shader where it assumes 16-bit is always color but
now it can also be D16_UNORM. The assert isn't protecting us from
anything we weren't already doing before so we can just loosen it a bit.
Fixes:
f99ac7f2de19 ("v3dv: Don't use color aspects for depth/stencil images")
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16592>
Emma Anholt [Tue, 10 May 2022 21:48:28 +0000 (14:48 -0700)]
ci/crosvm: Terminate the previous crosvm after a deqp-runner timeout.
When deqp-runner times out, it kills the deqp process, which in our case
is the previous invocation of our shell script, so the crosvm and socat
cleanup never happened. crosvm has a way to cleanly shut down a previous
crosvm invocation, so let's just use that and do our cleanup when we need
to.
Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16485>
Emma Anholt [Wed, 11 May 2022 17:13:23 +0000 (10:13 -0700)]
ci/crosvm: Rename VSOCK_TEMP_DIR -> VM_TEMP_DIR.
It doesn't contain anything about the vsocks, just files shared between
the host and the guest.
Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16485>
Emma Anholt [Tue, 10 May 2022 21:44:07 +0000 (14:44 -0700)]
ci/crosvm: Simplify the CID setup.
Now that deqp-runner tells us which thread we are, we don't need to go
probing for a spare directory to use.
Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16485>
Jesse Natalie [Wed, 18 May 2022 16:49:46 +0000 (09:49 -0700)]
dzn: Add new parameter to vk_image_view_init call
Fixes:
fc8d2543 ("vulkan,v3dv: Add a driver_internal flag to vk_image_view_init/create")
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16589>
Jesse Natalie [Wed, 18 May 2022 16:13:32 +0000 (09:13 -0700)]
d3d12: MSVC warning around operator precedence causing uint32_t==bool
Fixes:
b171a6ba ("d3d12: Add video encode implementation of pipe_video_codec")
Reviewed-by: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16588>
Jesse Natalie [Wed, 18 May 2022 16:13:14 +0000 (09:13 -0700)]
d3d12: Don't use VLAs
Fixes:
739283da ("d3d12: Improve planar resource support to handle video requirements")
Reviewed-by: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16588>
Jesse Natalie [Wed, 18 May 2022 16:12:39 +0000 (09:12 -0700)]
d3d12: Include windows.h on Windows before dxgicommon.h
The dxgicommon.h header uses UINT types without including windows.h itself.
Fixes:
6dbe05ff ("d3d12: Add util video functions to d3d12_format")
Reviewed-by: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16588>
Samuel Pitoiset [Mon, 16 May 2022 17:00:52 +0000 (19:00 +0200)]
radv: init states from pRasterizationState at only one place
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/16552>
Samuel Pitoiset [Mon, 16 May 2022 16:42:53 +0000 (18:42 +0200)]
radv: init states from pViewportState at only one place
Except the viewport/scissor arrays to avoid useless array copies.
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/16552>
Samuel Pitoiset [Mon, 16 May 2022 16:38:06 +0000 (18:38 +0200)]
radv: init states from pTessellationState at only one place
It's part of the pre-rasterization state of graphics pipeline libs.
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/16552>
Connor Abbott [Tue, 17 May 2022 14:28:30 +0000 (16:28 +0200)]
tu: Implement VK_EXT_pipeline_creation_cache_control
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16562>
Connor Abbott [Mon, 16 May 2022 17:11:42 +0000 (19:11 +0200)]
tu: Implement VK_EXT_pipeline_creation_feedback
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16562>
Connor Abbott [Tue, 17 May 2022 14:06:38 +0000 (16:06 +0200)]
tu: Zero-initialize compute driver key
Fixes: 05329d7 ("tu: Implement pipeline caching with shared Vulkan cache")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16562>
Vadym Shovkoplias [Mon, 16 May 2022 12:49:52 +0000 (15:49 +0300)]
drirc: Set limit_trig_input_range option for glmark2
This fixes jellyfish rendering issues in glmark2 on Intel.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6479
Signed-off-by: Vadym Shovkoplias <vadym.shovkoplias@globallogic.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16525>
Danylo Piliaiev [Tue, 12 Apr 2022 18:04:35 +0000 (21:04 +0300)]
freedreno/a6xx: Add EARLYPREAMBLE flag to all a6xx_sp_xs_ctrl_reg0
Each shader stage has its own "early preamble" flag.
Early preamble is likely an optimization to hide some of latency
when loading UBOs into consts in the preamble.
Early preamble has the following limitations:
- Only shared, a1, and consts regs could be used
(accessing other regs would result in GPU fault);
- No cat5/cat6, only stc/ldc variants are working;
- Values writen to shared regs are not accessible by the rest
of the shader;
- Instructions before shps are also considered to be a part of
early preamble.
Note, for all shaders from d3d11 games blob produced preambles
compatible with early preamble mode.
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15901>
Kenneth Graunke [Tue, 17 May 2022 20:51:49 +0000 (13:51 -0700)]
Revert "st/mesa: Transcode ASTC to BC7 (BPTC) where possible"
This reverts commit
68ef895674b176b4faf875a4d7f4d787b330d4d9.
When trying out transcode_astc=true with BPTC on Asphalt 9, we observed
very poor image quality - to the point that basic UI icons were blocky,
and buttons with a black border had smeared pixels on the edges. Using
DXT5 had no such issues.
I originally suspected there was a bug in the BPTC encoder, but I now
believe the issue is deeper than that. The commit that introduced the
encoder,
17cde55c538009764207bd29b78a909d2c5d14b4, says:
"The compressor is written from scratch and takes a very simple
approach. It always uses a single mode of the BPTC format (4 for
unorm and 3 for half-floats) and picks the two endpoints by dividing
the texels into those which have more or less than the average
luminance of the block and then calculating an average color of the
texels within each division.
It's probably not really sensible to try to use BPTC compression at
runtime because for example with the Nvidia offline compression tool
it can take in the order of an hour to compress a full-screen image.
With that in mind I don't think it's worth having a proper compressor
in Mesa and this approach gives reasonable results for a usage that
is basically a corner case."
In other words, the reason our BPTC compressor was so fast is that it
only implements one of the modes and does a low quality approximation.
This honestly should probably be improved somewhat, but the original
use case was for online-compression, the uncommon but mandatory OpenGL
feature where you can supply uncompressed data and trust the driver to
compress it for you (at unknown and uncontrolled quality and speed).
Unfortunately, the compressor as it stands is simply not usable for
transcoding ASTC data where we want to preserve the underlying image
quality as much as possible.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16566>
Pierre-Eric Pelloux-Prayer [Thu, 12 May 2022 18:09:22 +0000 (20:09 +0200)]
radeonsi: fix glTexBuffer max size handling
The spec says the number of texels must be clamped to the value of
GL_MAX_TEXTURE_BUFFER_SIZE.
Reviewed-by: Qiang Yu <yuq825@gmail.com>
Reviewed-by: Mihai Preda <mhpreda@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16480>
Juan A. Suarez Romero [Mon, 16 May 2022 14:18:06 +0000 (16:18 +0200)]
docs: update extensions for V3D and VC4
Mark some extensions that are already supported by V3D and VC4.
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16528>
Samuel Pitoiset [Wed, 13 Apr 2022 08:13:30 +0000 (10:13 +0200)]
radv: add new pipeline helpers for NIR->ASM compilation
It walks backwards to compile, looks cleaner to me.
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/16553>
Samuel Pitoiset [Wed, 13 Apr 2022 08:23:08 +0000 (10:23 +0200)]
radv: do not try to dump the NIR of the trap handler shader
There is no NIR at all.
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/16553>
Samuel Pitoiset [Wed, 13 Apr 2022 07:55:45 +0000 (09:55 +0200)]
radv: rename shader compile functions to spirv_to_nir/nir_to_asm
For better clarity.
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/16553>
Martin Roukala (né Peres) [Tue, 17 May 2022 06:22:09 +0000 (09:22 +0300)]
radv/ci: enable the experimental support for mesh shaders
The experimental support has been merged for months already, but has
been left unexercised in CI even though dEQP has a mesh_shader.nv
category.
This commit enables the experimental support, which should bring
additional testing \o/.
Suggested-by: Timur Kristóf <timur.kristof@gmail.com>
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16550>
Tomeu Vizoso [Thu, 24 Feb 2022 06:34:09 +0000 (07:34 +0100)]
ci: Disable jobs to the Collabora lab
In anticipation of infrastructure work.
This commit is to be reverted later in the day.
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16576>
Jason Ekstrand [Mon, 2 May 2022 16:43:17 +0000 (11:43 -0500)]
vulkan/pipeline_cache: Implement deserialize for raw objects
When caching NIR, it's cached as a raw object because we cache the
serialized NIR. When it's then loaded from the disk cache later, we
fail to deserialize it because raw objects are a special case. There
are two callers of vk_pipeline_cache_object_deserialize(), one of which
has a special case for raw objects and the other is called only when
we've checked that it isn't a raw object. The special cases are
pointless; raw objects should deserialize themselves.
Fixes:
591da9877900 ("vulkan: Add a common VkPipelineCache implementation")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16281>
Mike Blumenkrantz [Wed, 20 Apr 2022 16:19:52 +0000 (12:19 -0400)]
zink: add ZINK_DEBUG=sync
this is great for debugging
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16544>
Mike Blumenkrantz [Mon, 16 May 2022 15:27:58 +0000 (11:27 -0400)]
zink: add c++ guards for zink_screen.h
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16544>
Mike Blumenkrantz [Tue, 26 Apr 2022 13:02:57 +0000 (09:02 -0400)]
zink: fix non-dynamic vertex stride update flagging
without dynamic vertex input, pipeline vertex state must be recalculated
if buffer strides change or the enabled buffer mask changes in order
to accurately handle dynamic state stride VUs
cc: mesa-stable
fixes:
spec@!opengl 1.1@array-stride
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16563>