AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
128,876 Commits 1 Branch 0 Tags
656e428ff4e027d134027df73a0fe13e587011a8
Commit Graph

5418 Commits

Author SHA1 Message Date
Connor Abbott 656e428ff4 nir/opt_if: Remove open-coded nir_ssa_def_rewrite_uses() 
So that we don't have to change these two places later.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6866>
 2020-09-30 15:47:51 +00:00
Connor Abbott c6f871b62e nir/lower_returns: Use nir control flow insertion helpers 
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6866>
 2020-09-30 15:47:51 +00:00
Jason Ekstrand 92a594b154 spirv: Delete the legacy offset/index UBO/SSBO lowering 
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5275>
 2020-09-30 07:20:39 +00:00
Jason Ekstrand 657d49a9ba spirv: Use derefs for push constants 
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5275>
 2020-09-30 07:20:39 +00:00
Jason Ekstrand ac7537f155 nir/lower_io: Add support for push constants 
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5275>
 2020-09-30 07:20:38 +00:00
Jason Ekstrand 7a2b4ce22e nir: Allow creating variables with nir_var_mem_push_const. 
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5275>
 2020-09-30 07:20:38 +00:00
Indrajit Kumar Das 40c1f9883e mesa,glsl: add support for GL_NV_shader_atomic_int64 
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6708>
 2020-09-29 14:24:44 +00:00
Connor Abbott 51e2b31039 nir: Handle per-view io in nir_io_add_const_offset_to_base() 
This isn't strictly necessary for freedreno, since we aren't using it
yet, but I wanted to avoid any problems if we do. If we wanted to handle
this "properly", and handle matrix and array per-view variables, we'd
probably want to encode the "view stride" (number of views per user
location) and base view in the intrinsic, but for now we just don't do
any offsetting and assume the indirect offset is the view.

Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6514>
 2020-09-29 10:32:00 +00:00
Connor Abbott bc8a5c0752 nir: Add per_view to IO semantics 
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6514>
 2020-09-29 10:32:00 +00:00
Connor Abbott 5a88db682e nir/lower_io_arrays: Fix xfb_offset bug 
I noticed this once I started gathering xfb_info after
nir_lower_io_arrays_to_elements_no_indirect.

Fixes: b2bbd978d0 ("nir: fix lowering arrays to elements for XFB outputs")
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6514>
 2020-09-29 10:32:00 +00:00
Connor Abbott df955ce6b6 nir: Count i/o slots correctly for per-view variables 
This function wasn't counting driver slots correctly, resulting in
incorrect driver_location's and input_count. It seems intel doesn't use
this yet.

Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6514>
 2020-09-29 10:32:00 +00:00
Connor Abbott ffe946d7e8 nir: Add nir_lower_multiview pass 
Taken mostly directly from the anv pass. A few anv-specific things that
I could leave in anv aren't included. Specifically on turnip we don't
need to set gl_Layer to 0, and we can handle the case where the FS reads
gl_ViewIndex, so that check is moved into anv.

Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6514>
 2020-09-29 10:31:59 +00:00
Samuel Pitoiset a0e35c7562 nir/lower_io: change nir_io_add_const_offset_to_base to use bitfield modes 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6890>
 2020-09-29 09:40:21 +00:00
Eric Anholt e8c5f8b9d3 nir/lower_clip: Add i/o semantics for load/store intrinsics. 
ir3 looks at the .location on its inputs for handling
non-VARYING_SLOT_POS, but our uninitialized semantics matched that and
threw a compiler assertion failure.

Fixes: 502abfce7f ("nir: save IO semantics in lowered IO intrinsics")
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6716>
 2020-09-28 17:35:30 +00:00
Samuel Pitoiset 39098a2053 nir/lower_memory_model: do not break with global atomic operations 
Global atomics don't have an access flag.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6888>
 2020-09-28 14:47:02 +00:00
Samuel Pitoiset de1409089c nir/lower_memory_model: return progress when visiting instructions 
It never returned progress=TRUE.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6888>
 2020-09-28 14:47:02 +00:00
Rhys Perry ee073cb543 nir/load_store_vectorize: improve vectorization with identical operations 
We might have several identical options to vectorize an entry with, but
only one might be vectorizable because of writes interfering.

An example of this is a pattern found in some CTS tests:
a = load(0)
b = load(4)
store(0, a)
store(4, b)
a = load(0)
b = load(4)
store(0, a)
store(4, b)
...

It might have attempted to vectorize the first load(0) with the second
load(4) without attempting the second load(4) when the first fails. This
changes vectorize_entries() to continue even if the first try_vectorize()
failed.

fossil-db (Navi):
Totals from 117 (0.09% of 137413) affected shaders:
SGPRs: 7040 -> 7088 (+0.68%)
CodeSize: 276504 -> 276308 (-0.07%); split: -0.08%, +0.01%
Instrs: 51152 -> 51111 (-0.08%); split: -0.09%, +0.01%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5415>
 2020-09-28 12:06:30 +00:00
Louis-Francis Ratté-Boulianne f2b94ae085 glsl/linker: Add support for XFB varying lowering in geometry shader 
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6723>
 2020-09-28 07:57:28 +00:00
Jason Ekstrand 8323c03bbf spirv: Add support for OpCopyMemorySized 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>
 2020-09-25 23:48:03 +00:00
Jason Ekstrand a3177cca99 nir: Add a lowering pass to lower memcpy 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>
 2020-09-25 23:48:03 +00:00
Jason Ekstrand b2899f7265 nir: Add a new memcpy intrinsic 
This matches SPIR-V's OpCopyMemorySized

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>
 2020-09-25 23:48:03 +00:00
Jason Ekstrand 8cbb6640f0 spirv: Use the new types helpers 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>
 2020-09-25 23:48:03 +00:00
Jason Ekstrand de295b61da compiler/types: Add glsl_baseN_t_type(bit_size) helpers 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6713>
 2020-09-25 23:48:03 +00:00
Jesse Natalie 5f1855d896 vtn/opencl: Switch fma to conditionally use libclc for 32bit floats 
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 13:20:39 -07:00
Jesse Natalie 8fb25c48a4 vtn/opencl: Only use libclc ldexp when lower_ldexp is set 
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 13:14:46 -07:00
Jesse Natalie 09bca4cb95 vtn/opencl: Switch some nir-sequence ops to use libclc 
All of these are pretty well-defined. Rather than implementing them
as a sequence of nir ops, we can just use the libclc implementation.

v2 (idr): Delete functions that are now unused.

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 13:14:45 -07:00
Jesse Natalie 03fd217d21 vtn/opencl: Switch division-related ops to use libclc 
Specifically, fmod only uses libclc if it was going to be lowered.
Also, add missing half_divide and half_recip handling.

Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 2624faf137 vtn/opencl: Switch exp/pow/log to use libclc 
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 4cc9ad4eeb vtn/opencl: Switch non-native trig to use libclc 
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 6aee49578f vtn/opencl: Hook up OpenCL async copy and group wait opcodes via libclc 
Adds an additional hook for spirv_to_nir to handle a core opcode via
the OpenCL libclc infrastructure, and adds handling for SpvOpGroupAsyncCopy and
SpvOpGroupWaitEvents.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 00261d883d vtn/opencl: Rework handle_instr to be able to handle core SPIR-V opcodes via libclc 
The OpenCL async copy/wait opcodes are core SPIR-V, rather than OpenCL extension opcodes.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie b08fd45be0 vtn/opencl: Implement a lot of opcodes via libclc 
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 6436e3ac18 vtn/opencl: Add infrastructure for calling out to libclc 
This patch adds a function remap table with name mangling, which
can convert a SPIR-V OpenCL extension opcode to a call to the external
libclc shader, which will be lowered/inlined after conversion.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 2013903bc2 vtn/opencl: Rework type handling for CL extension opcodes 
- Change dest type information to be vtn_type rather than glsl_type
 - Add source type information, also in vtn_type form

Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 938be4d6fa nir/vtn: Add handling for SPIR-V event variables 
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Dave Airlie c9a6b94f9e vtn: add an option to create a nir library from spirv 
This adds an options to turn a spir-v library into a set of nir
functions. There is no entry point and all function implentations
are emitted.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 2826a5a923 nir: Add glsl_base_type unsigned -> signed version helper 
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jesse Natalie 93db59e066 nir: Add an internal flag to shader_info 
Don't print the shader if it's marked internal, unless NIR_PRINT
has been explicitly set to 2 (or higher).

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6035>
 2020-09-25 20:09:08 +00:00
Jason Ekstrand 0206fb3941 nir/liveness: Consider if uses in nir_ssa_defs_interfere 
Fixes: f86902e75d "nir: Add an SSA-based liveness analysis pass"
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3428
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Yevhenii Kharchenko <yevhenii.kharchenko@globallogic.com>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6824>
 2020-09-25 14:16:15 +00:00
Rhys Perry a18c84ecce nir/instr_set: hash intrinsic sources 
ministat (CSE only):
Difference at 95.0% confidence
	-9.80325 +/- 0.173089
	-41.4434% +/- 0.461972%
	(Student's t, pooled s = 0.0763653)

ministat (entire run):
Difference at 95.0% confidence
	-3.13667 +/- 0.61519
	-5.11107% +/- 0.990737%
	(Student's t, pooled s = 0.271416)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6860>
 2020-09-25 10:18:36 +00:00
Marek Olšák ea77958fea nir: gather information about fbfetch and dual source color 
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6758>
 2020-09-25 02:29:30 -04:00
Marek Olšák a6abf175ef nir: fix input/output info gathering for lowered IO 
Ooops.

Fixes: 17af07024d - nir: gather all IO info from IO intrinsics

Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6758>
 2020-09-25 02:29:30 -04:00
Marek Olšák ef98c175c0 nir: gather fs.uses_sample_qualifier from lowered IO 
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6758>
 2020-09-25 02:29:30 -04:00
Marek Olšák 7b108e6ac4 nir: set system_values_read for all intrinsics 
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6758>
 2020-09-25 02:29:30 -04:00
Marek Olšák abe9588ff0 nir: gather tess.tcs_cross_invocation info from lowered IO intrinsics 
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6758>
 2020-09-25 02:29:30 -04:00
Marek Olšák 10be706778 nir: gather indirect info from lowered IO intrinsics 
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6758>
 2020-09-25 02:29:30 -04:00
Jason Ekstrand 3173367a47 spirv: vtn_fail with a nice message on unsupported rounding modes 
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6845>
 2020-09-25 01:43:28 +00:00
Kenneth Graunke 140f53e646 Revert "nir: replace lower_ffma and fuse_ffma with has_ffma" 
This reverts commit 939ddf3f67.

Intel has a separate pass for fusing FFMAs selectively.  We split
these flags in commit 1b72c31e1f and
the reasoning still stands.  The patch being reverted was just a
cleanup, so there should be no issue with reverting it.

Acked-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6849>
 2020-09-24 13:11:50 -07:00
Marek Olšák 939ddf3f67 nir: replace lower_ffma and fuse_ffma with has_ffma 
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6756>
 2020-09-24 12:29:11 +00:00
Marek Olšák 771aad3027 nir: split lower_ffma into lower_ffma16/32/64 
AMD wants different behavior for each bit size

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6756>
 2020-09-24 12:29:11 +00:00