platform/upstream/mesa.git
4 years agopan/bi: Add IR iteration macros
Alyssa Rosenzweig [Tue, 3 Mar 2020 19:32:28 +0000 (14:32 -0500)]
pan/bi: Add IR iteration macros

Copypaste from Midgard, for the most part.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add quirks system
Alyssa Rosenzweig [Tue, 3 Mar 2020 19:27:05 +0000 (14:27 -0500)]
pan/bi: Add quirks system

Modeled after the Midgard system. Already we know of two
compiler-visible differences between G52 and G71, so let's keep track so
we can eventually port the compiler to other Bifrost systems.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add high-latency property for classes
Alyssa Rosenzweig [Tue, 3 Mar 2020 18:55:33 +0000 (13:55 -0500)]
pan/bi: Add high-latency property for classes

This is required to know how to schedule legally, and also influences
some issues relating to RA.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add CSEL condition
Alyssa Rosenzweig [Thu, 5 Mar 2020 12:46:00 +0000 (07:46 -0500)]
pan/bi: Add CSEL condition

Along with src_types, this is enough to represent CSEL.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add bi_branch data
Alyssa Rosenzweig [Tue, 3 Mar 2020 18:48:13 +0000 (13:48 -0500)]
pan/bi: Add bi_branch data

For BI_BRANCH, of course. Meshes well with the cfg.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Extract bifrost_branch structure
Alyssa Rosenzweig [Tue, 3 Mar 2020 18:47:49 +0000 (13:47 -0500)]
pan/bi: Extract bifrost_branch structure

It's in the disassembler as bitfields, let's extract to a proper
structure so we can see what's there.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add pred/successors to build CFG
Alyssa Rosenzweig [Tue, 3 Mar 2020 18:47:13 +0000 (13:47 -0500)]
pan/bi: Add pred/successors to build CFG

We'll want this for analysis passes or something, probably.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add constants to bi_clause
Alyssa Rosenzweig [Tue, 3 Mar 2020 18:01:41 +0000 (13:01 -0500)]
pan/bi: Add constants to bi_clause

Scheduling will have to pay attention to this.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add EXTRACT, MAKE_VEC synthetic ops
Alyssa Rosenzweig [Tue, 3 Mar 2020 13:57:03 +0000 (08:57 -0500)]
pan/bi: Add EXTRACT, MAKE_VEC synthetic ops

These allow translating between the vector I/O and scalar ALUs,
facilitated by an RA dance to ensured contiguous registers are used.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add source type for conversions
Alyssa Rosenzweig [Tue, 3 Mar 2020 13:37:15 +0000 (08:37 -0500)]
pan/bi: Add source type for conversions

We should now be able to unambiguously represent conversions.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add swizzles
Alyssa Rosenzweig [Tue, 3 Mar 2020 13:35:51 +0000 (08:35 -0500)]
pan/bi: Add swizzles

Requires a new field on bifrost_instruction, as well as a new class
property and a new class for the dedicated swizzle ops.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Clarify special op scheduling
Alyssa Rosenzweig [Tue, 3 Mar 2020 13:16:50 +0000 (08:16 -0500)]
pan/bi: Clarify special op scheduling

They're encoded on ADD but eat the full cycle.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add clause header fields to bi_clause
Alyssa Rosenzweig [Tue, 3 Mar 2020 13:09:18 +0000 (08:09 -0500)]
pan/bi: Add clause header fields to bi_clause

These will be filled out during scheduling (and possibly RA), to be used
when emitting code.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add class-specific ops
Alyssa Rosenzweig [Tue, 3 Mar 2020 12:58:05 +0000 (07:58 -0500)]
pan/bi: Add class-specific ops

For disambiguating things like min and max within the MINMAX class.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add constant field to bi_instruction
Alyssa Rosenzweig [Tue, 3 Mar 2020 12:47:29 +0000 (07:47 -0500)]
pan/bi: Add constant field to bi_instruction

Now that we can index it.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add special indices
Alyssa Rosenzweig [Tue, 3 Mar 2020 12:45:33 +0000 (07:45 -0500)]
pan/bi: Add special indices

For fixed registers, uniforms, and constants, which bypass the usual SSA
mechanism to map well to the ISA.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add dest_type field to bifrost_instruction
Alyssa Rosenzweig [Tue, 3 Mar 2020 03:03:05 +0000 (22:03 -0500)]
pan/bi: Add dest_type field to bifrost_instruction

A number of opcodes within a class are disambiguated by type/size, and
whether modifiers make sense or not depends on whether the instruction
acts like a float.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add bi_clause, bi_bundle abstractions
Alyssa Rosenzweig [Tue, 3 Mar 2020 03:00:07 +0000 (22:00 -0500)]
pan/bi: Add bi_clause, bi_bundle abstractions

These will be used during and after scheduling.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add PAN_SCHED_* flags
Alyssa Rosenzweig [Tue, 3 Mar 2020 02:53:13 +0000 (21:53 -0500)]
pan/bi: Add PAN_SCHED_* flags

Class (mostly) determines scheduling options.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add bi_load_vary structure
Alyssa Rosenzweig [Tue, 3 Mar 2020 02:48:51 +0000 (21:48 -0500)]
pan/bi: Add bi_load_vary structure

For ld_vary in the IR.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Pull out bifrost_load_var
Alyssa Rosenzweig [Tue, 3 Mar 2020 02:45:47 +0000 (21:45 -0500)]
pan/bi: Pull out bifrost_load_var

We're not using this structure yet but we want everything in the ISA
ready for us.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add bi_load structure
Alyssa Rosenzweig [Tue, 3 Mar 2020 02:19:16 +0000 (21:19 -0500)]
pan/bi: Add bi_load structure

Fills out the class for LD_ATTR, LD_VAR_ADDR

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add bifrost_minmax_mode field
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:53:47 +0000 (20:53 -0500)]
pan/bi: Add bifrost_minmax_mode field

We'll open up a union for class specific data, since this is interesting
only to BI_MINMAX. (And even then...)

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add a bifrost_roundmode field
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:52:36 +0000 (20:52 -0500)]
pan/bi: Add a bifrost_roundmode field

And a class property signaling it's okay to use.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Factor out enum bifrost_minmax_mode
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:51:03 +0000 (20:51 -0500)]
pan/bi: Factor out enum bifrost_minmax_mode

We'll want it from the compiler-side.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add BI_GENERIC property
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:46:37 +0000 (20:46 -0500)]
pan/bi: Add BI_GENERIC property

I don't want to have 20 class-specific structures floating around. So
let's derive them all from a common generic ALU type.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add modifiers to bi_instruction
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:40:52 +0000 (20:40 -0500)]
pan/bi: Add modifiers to bi_instruction

Now that we can check if we support them via the class.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add class properties
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:38:26 +0000 (20:38 -0500)]
pan/bi: Add class properties

We need to keep track of what specific classes support. For now just
track floating point modifiers.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add src/dest fields to bifrost_instruction
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:24:03 +0000 (20:24 -0500)]
pan/bi: Add src/dest fields to bifrost_instruction

...along with some helpers to generate indices. The indexing scheme is
mostly a copypaste from Midgard, except we specifically reserve 0 as the
sentinel (midgard uses ~0 for this which has always been a pain point).

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Add the control flow graph
Alyssa Rosenzweig [Tue, 3 Mar 2020 01:06:34 +0000 (20:06 -0500)]
pan/bi: Add the control flow graph

We're starting to build up the IR data structures in preparation to get
everything piped through.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Stub out new compiler
Alyssa Rosenzweig [Tue, 3 Mar 2020 00:47:11 +0000 (19:47 -0500)]
pan/bi: Stub out new compiler

Just enough to pipe in the NIR shader.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopan/bi: Gut old compiler
Alyssa Rosenzweig [Tue, 3 Mar 2020 00:30:11 +0000 (19:30 -0500)]
pan/bi: Gut old compiler

We're making some pretty dramatic design pivots so this early on it'll
be easier to start from scratch, I think.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agopanfrost: Add note about preloaded varyings
Alyssa Rosenzweig [Tue, 3 Mar 2020 02:32:31 +0000 (21:32 -0500)]
panfrost: Add note about preloaded varyings

There's a magic bit in preload_regs which controls this. It doesn't
appear to be supported on G71 but it is on G52. I'd guess G72 supports
it too but I don't have a way to check this.

Needless to say, we'll need a quirks database for this.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4061>

4 years agoaco: fix image load/store with lod and 1D images
Samuel Pitoiset [Thu, 5 Mar 2020 10:32:06 +0000 (11:32 +0100)]
aco: fix image load/store with lod and 1D images

Make sure to add the lod value if non-null as the 2nd operand.

Fixes dEQP-VK.image.load_store_lod.with_format.1d.* on all gens
except GFX9.

Fixes: 4d49a7ac737 ("aco: handle nir_intrinsic_image_deref_{load,store} with lod")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4060>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4060>

4 years agogitlab-ci: Distribute jobs across more stages
Michel Dänzer [Thu, 27 Feb 2020 17:27:56 +0000 (18:27 +0100)]
gitlab-ci: Distribute jobs across more stages

The stages and mapping of jobs to them are somewhat arbitrary; the goal
is to avoid having to scroll through large numbers of jobs.

v2: (Pierre-Eric Pelloux-Prayer)
* Use even more stages for test jobs
* Give somewhat meaningful names to stages

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3995>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3995>

4 years agogitlab-ci: Drop "test-" prefix from llvmpipe/softpipe job names
Michel Dänzer [Tue, 3 Mar 2020 17:17:12 +0000 (18:17 +0100)]
gitlab-ci: Drop "test-" prefix from llvmpipe/softpipe job names

Redundant.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3995>

4 years agovbo: merge draws even when begin==0 or end==0
Marek Olšák [Fri, 14 Feb 2020 19:18:45 +0000 (14:18 -0500)]
vbo: merge draws even when begin==0 or end==0

Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: merge more primitive types for glBegin/End (v2)
Marek Olšák [Fri, 14 Feb 2020 04:22:44 +0000 (23:22 -0500)]
vbo: merge more primitive types for glBegin/End (v2)

v2: clean it up more

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agomesa: deduplicate draw indirect functions
Marek Olšák [Fri, 14 Feb 2020 19:29:32 +0000 (14:29 -0500)]
mesa: deduplicate draw indirect functions

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agomesa: optimize get_index_size
Marek Olšák [Tue, 3 Mar 2020 20:08:09 +0000 (15:08 -0500)]
mesa: optimize get_index_size

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agomesa: remove _mesa_index_buffer::index_size in favor of index_size_shift
Marek Olšák [Tue, 3 Mar 2020 20:03:28 +0000 (15:03 -0500)]
mesa: remove _mesa_index_buffer::index_size in favor of index_size_shift

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Suggested-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agoRevert "mesa: check for z=0 in _mesa_Vertex3dv()"
Marek Olšák [Thu, 27 Feb 2020 21:06:47 +0000 (16:06 -0500)]
Revert "mesa: check for z=0 in _mesa_Vertex3dv()"

This reverts commit f04d7439a0ad6e13ff2912ff824553b6bcf511a4.

It no longer helps performance and the current vbo implementation is
faster anyway.

The app that hit this was a CAD program called Spazio3D. It made pretty
terrible use of the OpenGL API and we sent them some tips for improvements.
I'm assuming they've fixed this by now.

Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: fold code from vbo_exec_fixup_vertex to vbo_exec_wrap_upgrade_vertex
Marek Olšák [Thu, 27 Feb 2020 20:43:55 +0000 (15:43 -0500)]
vbo: fold code from vbo_exec_fixup_vertex to vbo_exec_wrap_upgrade_vertex

Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: clean up conditional blocks in ATTR_UNION
Marek Olšák [Thu, 27 Feb 2020 20:30:08 +0000 (15:30 -0500)]
vbo: clean up conditional blocks in ATTR_UNION

Move the A != 0 code to the first block.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: handle GS and tess primitive types when splitting Begin/End
Marek Olšák [Wed, 19 Feb 2020 00:38:33 +0000 (19:38 -0500)]
vbo: handle GS and tess primitive types when splitting Begin/End

Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: clean up vbo_copy_vertices
Marek Olšák [Wed, 19 Feb 2020 00:05:17 +0000 (19:05 -0500)]
vbo: clean up vbo_copy_vertices

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: deduplicate copy_vertices functions
Marek Olšák [Tue, 18 Feb 2020 23:37:36 +0000 (18:37 -0500)]
vbo: deduplicate copy_vertices functions

There are some differences in exec, but those look like bug fixes not ported
to vbo_save.

Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agovbo: don't look at the second draw's count when merging 2 glBegin/End draws
Marek Olšák [Fri, 14 Feb 2020 04:15:47 +0000 (23:15 -0500)]
vbo: don't look at the second draw's count when merging 2 glBegin/End draws

Only the first count needs to be aligned.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agomesa: replace some index_size multiplications and divisions with shifts
Marek Olšák [Thu, 13 Feb 2020 20:29:22 +0000 (15:29 -0500)]
mesa: replace some index_size multiplications and divisions with shifts

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agomesa: add index_size_shift = log2(index_size) into _mesa_index_buffer
Marek Olšák [Wed, 12 Feb 2020 23:02:24 +0000 (18:02 -0500)]
mesa: add index_size_shift = log2(index_size) into _mesa_index_buffer

for faster division

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanic@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4052>

4 years agoandroid: r600/sfn: Add GDS instructions
Mauro Rossi [Sat, 22 Feb 2020 15:20:10 +0000 (16:20 +0100)]
android: r600/sfn: Add GDS instructions

Fixes the following building errors:

external/mesa/src/gallium/drivers/r600/sfn/sfn_emitssboinstruction.cpp:59: error: undefined reference to 'r600::GDSInstr::GDSInstr(r600::ESDOp, r600::GPRVector const&, std::__1::shared_ptr<r600::Value> const&, std::__1::shared_ptr<r600::Value> const&, std::__1::shared_ptr<r600::Value> const&, int)'
...
external/mesa/src/gallium/drivers/r600/sfn/sfn_emitssboinstruction.cpp:256: error: undefined reference to 'r600::RatInstruction::RatInstruction(r600::ECFOpCode, r600::RatInstruction::ERatOp, r600::GPRVector const&, r600::GPRVector const&, int, std::__1::shared_ptr<r600::Value> const&, int, int, int, bool)'

Fixes: 32d3435a ("r600/sfn: Add GDS instructions")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
4 years agoandroid: r600/sfn: fix includes and libmesa_nir dependency
Mauro Rossi [Sat, 22 Feb 2020 15:20:10 +0000 (16:20 +0100)]
android: r600/sfn: fix includes and libmesa_nir dependency

Fixes the following building errors:

In file included from external/mesa/src/gallium/drivers/r600/sfn/sfn_debug.cpp:28:
In file included from external/mesa/src/gallium/drivers/r600/sfn/sfn_debug.h:34:
In file included from external/mesa/src/compiler/nir/nir.h:41:
In file included from external/mesa/src/compiler/nir_types.h:36:
external/mesa/src/compiler/glsl_types.h:38:10: fatal error: 'main/config.h' file not found
#include "main/config.h"
         ^~~~~~~~~~~~~~~
1 error generated.

In file included from external/mesa/src/gallium/drivers/r600/sfn/sfn_debug.cpp:28:
In file included from external/mesa/src/gallium/drivers/r600/sfn/sfn_debug.h:34:
external/mesa/src/compiler/nir/nir.h:50:10: fatal error: 'nir_opcodes.h' file not found
#include "nir_opcodes.h"
         ^~~~~~~~~~~~~~~
1 error generated.

Fixes: f718ac62 ("r600/sfn: Add a basic nir shader backend")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
4 years agoandroid: aco: fix PIPE_FORMAT related building errors
Mauro Rossi [Sat, 22 Feb 2020 15:20:10 +0000 (16:20 +0100)]
android: aco: fix PIPE_FORMAT related building errors

Fixes the following building errors:

In file included from external/mesa/src/amd/compiler/aco_dead_code_analysis.cpp:25:
In file included from external/mesa/src/amd/compiler/aco_ir.h:33:
In file included from external/mesa/src/compiler/nir/nir.h:40:
external/mesa/src/util/format/u_format.h:33:10: fatal error: 'pipe/p_format.h' file not found
#include "pipe/p_format.h"
         ^~~~~~~~~~~~~~~~~
...
In file included from external/mesa/src/amd/compiler/aco_dominance.cpp:31:
In file included from external/mesa/src/amd/compiler/aco_ir.h:33:
In file included from external/mesa/src/compiler/nir/nir.h:40:
external/mesa/src/util/format/u_format.h:33:10: fatal error: 'pipe/p_format.h' file not found
#include "pipe/p_format.h"
         ^~~~~~~~~~~~~~~~~
...
In file included from external/mesa/src/amd/compiler/aco_instruction_selection.cpp:31:
In file included from external/mesa/src/amd/common/ac_shader_util.h:32:
In file included from external/mesa/src/compiler/nir/nir.h:40:
external/mesa/src/util/format/u_format.h:33:10: fatal error: 'pipe/p_format.h' file not found
#include "pipe/p_format.h"
         ^~~~~~~~~~~~~~~~~
3 errors generated.

Fixes: 8d07d661 ("glsl,nir: Switch the enum representing shader image formats to PIPE_FORMAT.")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
4 years agonir: Flush to zero with OOB low exponents in ldexp
Jason Ekstrand [Thu, 27 Feb 2020 19:58:34 +0000 (13:58 -0600)]
nir: Flush to zero with OOB low exponents in ldexp

Reviewed-by: Arcady Goldmints-Orlov <agoldmints@igalia.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
4 years agozink. Added storage CISto descriptor pool.
Duncan Hopkins [Wed, 4 Mar 2020 11:24:24 +0000 (11:24 +0000)]
zink. Added storage CISto descriptor pool.
Added storage in descriptor pool for combined image samplers as well as uniform buffers.
Stops some shaders from running through a pools storage faster than zinks internal tracking.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4045>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4045>

4 years agogitlab-ci: Add jobs to be able to test Vulkan
Andres Gomez [Thu, 20 Feb 2020 16:26:30 +0000 (18:26 +0200)]
gitlab-ci: Add jobs to be able to test Vulkan

Also, adds an example job for radv.

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agogitlab-ci: Add gfxreconstruct traces support
Andres Gomez [Wed, 12 Feb 2020 21:21:07 +0000 (23:21 +0200)]
gitlab-ci: Add gfxreconstruct traces support

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agogitlab-ci: Change devices format to <api-vendor-deviceId>
Andres Gomez [Thu, 20 Feb 2020 15:32:08 +0000 (17:32 +0200)]
gitlab-ci: Change devices format to <api-vendor-deviceId>

In preparation to having "vk" (Vulkan) along "gl" (OpenGL/ES).

This is so it is clearer which traces belong to which API and also for
the build jobs.

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agogitlab-ci: build VulkanTools into the Vulkan testing container
Andres Gomez [Thu, 20 Feb 2020 15:09:03 +0000 (17:09 +0200)]
gitlab-ci: build VulkanTools into the Vulkan testing container

In preparation for having automated testing with Vulkan traces.

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agogitlab-ci: build gfxreconstruct into the Vulkan testing container
Andres Gomez [Thu, 20 Feb 2020 12:18:54 +0000 (14:18 +0200)]
gitlab-ci: build gfxreconstruct into the Vulkan testing container

In preparation for having automated testing with Vulkan traces.

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agogitlab-ci: add missing popd to the build-deqp-vk.sh script
Andres Gomez [Thu, 20 Feb 2020 13:30:36 +0000 (15:30 +0200)]
gitlab-ci: add missing popd to the build-deqp-vk.sh script

Since we are at it, replace "cd" with pushd / popd and homogenize how
VK-GL-CTS is built in comparison with other build scripts.

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agotracie: correct typo
Andres Gomez [Thu, 20 Feb 2020 11:45:21 +0000 (13:45 +0200)]
tracie: correct typo

Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
4 years agoetnaviv: fix alpha test on GC3000
Christian Gmeiner [Fri, 17 Jan 2020 09:32:33 +0000 (10:32 +0100)]
etnaviv: fix alpha test on GC3000

Store ref_value in PE_STENCIL_CONFIG_EXT as done by blob.

Fixes following piglits:
 spec@ext_framebuffer_object@fbo-alphatest-formats
 spec@ext_packed_float@fbo-alphatest-formats
 spec@ext_texture_srgb@fbo-alphatest-formats

Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4028>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4028>

4 years agoetnaviv: update headers from rnndb
Christian Gmeiner [Tue, 28 Jan 2020 04:43:06 +0000 (05:43 +0100)]
etnaviv: update headers from rnndb

Update to etna_viv commit 3bc187a.

Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4028>

4 years agoegl/wayland: Don't invalidate buffers on no-op resize
Daniel Stone [Tue, 3 Mar 2020 10:52:32 +0000 (10:52 +0000)]
egl/wayland: Don't invalidate buffers on no-op resize

The Wayland platform's resize_callback is invoked from libwayland-egl
when wl_egl_window_resize() is called. The resize call is the only place
for the application to insert dx/dy arguments to wl_surface_attach().

When modifying the cursor hotspot (as in wayland/wayland#148), we want
to set dx/dy, but leave the surface size the same. If we get
wl_egl_window_resize() with the same width and height argument as we
already have, we do not need to invalidate our existing drawable.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4030>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4030>

4 years agoRevert "glx: convert glx_config_create_list to one big calloc"
Andrii Simiklit [Wed, 15 Jan 2020 10:34:38 +0000 (12:34 +0200)]
Revert "glx: convert glx_config_create_list to one big calloc"

This reverts commit 35fc7bdf0e6ad6547e39099e7060a3d89539b56d.

Unfortunately mentioned commit introduced a memory leak because
`driwindowsMapConfigs` and `createDriMode` functions allocate
small memory portions for each element:
 21,576 (232 direct, 21,344 indirect) bytes in 1 blocks are definitely lost in loss record 1,411 of 1,414
    at 0x483A7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
    by 0x5D4AA09: createDriMode (dri_common.c:291)
    by 0x5D4ABF5: driConvertConfigs (dri_common.c:310)
    by 0x5D58414: dri3_create_screen (dri3_glx.c:945)
    by 0x5D39829: AllocAndFetchScreenConfigs (glxext.c:815)
    by 0x5D39C57: __glXInitialize (glxext.c:941)
    by 0x5D3290A: GetGLXPrivScreenConfig (glxcmds.c:174)
    by 0x5D34F38: glXQueryExtensionsString (glxcmds.c:1307)
    by 0x4F83038: glXQueryExtensionsString (in /usr/local/lib/libGL.so.1.7.0)
    by 0x4F2EA6B: ??? (in /usr/lib/x86_64-linux-gnu/libwaffle-1.so.0.6.0)
    by 0x4F2A0D7: waffle_display_connect (in /usr/lib/x86_64-linux-gnu/libwaffle-1.so.0.6.0)
    by 0x498F42A: wfl_checked_display_connect (piglit-util-waffle.h:74)

There is one more thing which disallow us to easily fix it are different element sizes
for instance: `glx_config_create_list` allocates memory just for `glx_config`,
`driwindowsMapConfigs` for `driwindows_config` and
`createDriMode` for `__GLXDRIconfigPrivate`.
Yes it is possible but size of such fix
will be more big and complex than original one.
So it make sense only if the malloc overhead
really is a big problem there.

Acked-by: Eric Engestrom <eric@engestrom.ch>
Signed-off-by: Andrii Simiklit <andrii.simiklit@globallogic.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3406>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3406>

4 years agozink. Don't set incorrect sType in VkImportMemoryFdInfoKHR struct
Vilya Harvey [Tue, 3 Mar 2020 16:19:29 +0000 (16:19 +0000)]
zink. Don't set incorrect sType in VkImportMemoryFdInfoKHR struct

imfi.sType was being set to an invalid value, triggering a warning in Clang. The only valid value for imfi.sType is VK_STRUCTURE_TYPE_IMPORT_MEMORY_FD_INFO_KHR which is the value it is being given at initialisation time, a few lines earlier. The incorrect value, VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT, is supposed to be used in imfi.handleType instead - and indeed, handleType *is* being set to this value a few lines later.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4034>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4034>

4 years agoturnip: support indirect draw
Hyunjun Ko [Tue, 25 Feb 2020 06:35:01 +0000 (06:35 +0000)]
turnip: support indirect draw

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3976>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3976>

4 years agoandroid: gallium/auxiliary: fix "Unused source files" in tesselator
Mauro Rossi [Sat, 29 Feb 2020 10:33:49 +0000 (11:33 +0100)]
android: gallium/auxiliary: fix "Unused source files" in tesselator

Avoids the following Android Build System error:

FAILED:
build/make/core/binary.mk:1245: error: external/mesa/src/gallium/auxiliary/Android.mk: libmesa_gallium: Unused source files: tessellator/tessellator.hpp
10:24:30 ckati failed with: exit status 1

Fixes: bd0188f ("gallium/auxiliary: add the microsoft tessellator and a pipe wrapper.")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Dave Airlie <airlied@redhat.com>
4 years agoci: Flip db410c back to docker mode.
Eric Anholt [Tue, 3 Mar 2020 19:44:45 +0000 (11:44 -0800)]
ci: Flip db410c back to docker mode.

Turns out there's corporate policy to not deploy AGPL software, so I
have to take down the LAVA lab until we sort out how to do it without
a local server.

Reviewed-by: Daniel Stone <daniels@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4038>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4038>

4 years agointel/gen12+: Disable mid thread preemption.
Rafael Antognolli [Tue, 3 Mar 2020 16:07:32 +0000 (08:07 -0800)]
intel/gen12+: Disable mid thread preemption.

Fixes a GPU hang in Car Chase.

Cc: mesa-stable@lists.freedesktop.org
v2: Add comment explaining why (Jason).

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4035>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4035>

4 years agoRevert "gallium/swr: Fix min/max range index draw"
Krzysztof Raszkowski [Tue, 3 Mar 2020 15:02:20 +0000 (15:02 +0000)]
Revert "gallium/swr: Fix min/max range index draw"

This reverts commit 5e9a2c603f38308f7027d6a5e4575e5fc24c1bd5

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4032>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4032>

4 years agovc4: fix vc4_yuv_blit overwriting fragment constant buffer slot 0
Chris Lord [Fri, 28 Feb 2020 15:18:14 +0000 (15:18 +0000)]
vc4: fix vc4_yuv_blit overwriting fragment constant buffer slot 0

vc4_yuv_blit calls util_blitter_restore_constant_buffer_state without
first calling util_blitter_save_fragment_constant_buffer_slot. This causes
subsequent crashes in vc4_write_uniforms when using fragment shaders that
reference YUV textures.

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2581
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3997>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3997>

4 years agoaco: handle v_add_co_u32_e64 in parse_base_offset()
Rhys Perry [Fri, 21 Feb 2020 12:02:06 +0000 (12:02 +0000)]
aco: handle v_add_co_u32_e64 in parse_base_offset()

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3902>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3902>

4 years agoaco: fix carry-out size for wave32 v_add_co_u32_e64
Rhys Perry [Fri, 21 Feb 2020 12:00:38 +0000 (12:00 +0000)]
aco: fix carry-out size for wave32 v_add_co_u32_e64

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Fixes: e0bcefc3a0a ('aco/wave32: Use lane mask regclass for exec/vcc.')
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3902>

4 years agogallium/swr: fix corruptions in Unigine Heaven
Jan Zielinski [Thu, 27 Feb 2020 16:22:05 +0000 (17:22 +0100)]
gallium/swr: fix corruptions in Unigine Heaven

Few changes to fix the last corruptions in Heaven:
- fix indirect TCS input when vertex/attribute
  index is not the same for each patch
- use the correct functions to build loops in shader
- fix using vmask for writting TCS output

Reviewed-by: Krzysztof Raszkowski <krzysztof.raszkowski@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3980>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3980>

4 years agost/va: GetConfigAttributes: check profile and entrypoint combination
Satyajit Sahu [Thu, 20 Feb 2020 08:35:32 +0000 (14:05 +0530)]
st/va: GetConfigAttributes: check profile and entrypoint combination

Added check if profile is supported or not for the entrypoint in
GetConfigAttributes.

Signed-off-by: Satyajit Sahu <satyajit.sahu@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3889>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3889>

4 years agointel/isl: Implement D16_UNORM workarounds.
Rafael Antognolli [Fri, 21 Feb 2020 20:03:05 +0000 (12:03 -0800)]
intel/isl: Implement D16_UNORM workarounds.

GEN:BUG:14010455700 (lineage 1808121037):

   "To avoid sporadic corruptions “Set 0x7010[9] when Depth Buffer
   Surface Format is D16_UNORM , surface type is not NULL & 1X_MSAA"

Required for fixing ttps://gitlab.freedesktop.org/mesa/mesa/issues/2501.

GEN:BUG:1806527549:

   "Set HIZ_CHICKEN (7018h) bit 13 = 1 when depth buffer is D16_UNORM."

This one could fix a GPU hang in some workloads.

v2: Implement WA in isl and add another similar WA (Jason).
v3: Add flushes before changing chicken registers (Jason)
v4: Depth flush and stall + end of pipe sync when changing registers
(Jason).

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3801>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3801>

4 years agoaco: keep track of which events are used in a barrier
Rhys Perry [Tue, 11 Feb 2020 16:52:20 +0000 (16:52 +0000)]
aco: keep track of which events are used in a barrier

And properly handle unordered events so that they always wait for 0.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: 93c8ebfa780 ('aco: Initial commit of independent AMD compiler')
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3774>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3774>

4 years agost/va/postproc: reallocate interlaced destination buffer
Thong Thai [Fri, 28 Feb 2020 15:56:49 +0000 (10:56 -0500)]
st/va/postproc: reallocate interlaced destination buffer

When the source buffer is progressive source, re-allocate the
destination buffer as progressive if it isn't already - otherwise
transcoding will fail.

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/1418
Signed-off-by: Thong Thai <thong.thai@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4001>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4001>

4 years agopanfrost: fix transform feedback
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 07:04:22 +0000 (03:04 -0400)]
panfrost: fix transform feedback

Fix different use cases for transform feedback by setting:

 - PIPE_CAP_PACKED_STREAM_OUTPUT=0
 - PIPE_CAP_VIEWPORT_TRANSFORM_LOWERED=1
 - PIPE_CAP_PSIZ_CLAMPED=1

This is enough for all dEQP xfb-related test cases to run
successfully.

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com> (Update dEQP expectations)
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agogallium: add PIPE_CAP_PSIZ_CLAMPED
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 07:01:59 +0000 (03:01 -0400)]
gallium: add PIPE_CAP_PSIZ_CLAMPED

This new capability indicates that the point size has been clamped.
This also means that the gl_PointSize has been modified and that
its value should be lowered for transform feedback, if needed.

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agogallium: add PIPE_CAP_VIEWPORT_TRANSFORM_LOWERED
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 06:59:53 +0000 (02:59 -0400)]
gallium: add PIPE_CAP_VIEWPORT_TRANSFORM_LOWERED

This new capability indicates that the nir_lower_viewport_transform
pass is enabled. This also means that the gl_Position value is
modified and should be lowered for transform feedback, if needed.

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agogallium: add PIPE_CAP_PACKED_STREAM_OUTPUT
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 06:46:48 +0000 (02:46 -0400)]
gallium: add PIPE_CAP_PACKED_STREAM_OUTPUT

Setting this cap to 0 (default is 1) should disable packing
optimization for stream output (e.g. GL transform feedback captured
variables).

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agoglsl/linker: add xfb workaround for modified built-in variables
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 05:13:54 +0000 (01:13 -0400)]
glsl/linker: add xfb workaround for modified built-in variables

Some lowering passes modify the value of built-in variables in
order for drivers to work properly. However, modifying such values
will also break transform feedback as the captured value won't
match what's expected.

For example, on some hardware, the vertex shaders are expected to
output gl_Position in screen space. However, the transform
feedback captured value is still supposed to be the world-space
coordinates (see nir_lower_viewport_transform).

To fix that, we create a new variable that contains the
pre-transformation value and use it for transform feedback instead
of the built-in one.

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agoglsl/linker: handle array/struct members for DisableXfbPacking
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 04:19:50 +0000 (00:19 -0400)]
glsl/linker: handle array/struct members for DisableXfbPacking

When varying packing is disabled for transform feedback and a xfb
declaration points to an array element or structure member, the
element/member should be aligned to the start of a slot as well.
If that's not the case, a new varying is created and the
element/member value is copied.

There might a way to further optimize the number of slots allocated
or the number of copies necessary if the performance cost is
problematic. For example, in cases where simply padding the top
level variable might correctly align all the captured values.

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agoglsl/linker: add DisableTransformFeedbackPacking workaround
Louis-Francis Ratté-Boulianne [Sat, 12 Oct 2019 04:05:03 +0000 (00:05 -0400)]
glsl/linker: add DisableTransformFeedbackPacking workaround

Some drivers (e.g. Panfrost) don't support packing of varyings when
used for transform feedback. This new constant ensures that any
varying used for xfb is aligned at the start of a slot and won't be
packed with other varyings.

Scenarios where transform feedback declarations are related to an
array element or a struct member will be handled in a subsequent
patch.

Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com> (Fix order of arguments to varying_matches())
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/2433>

4 years agospirv: fix memory_barrier_tcs_patch emission
Rhys Perry [Fri, 28 Feb 2020 19:31:22 +0000 (19:31 +0000)]
spirv: fix memory_barrier_tcs_patch emission

Shouldn't affect any driver, since all currently implement
memory_barrier_tcs_patch as a no-op. It also looks like optimizations are
fine

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4003>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4003>

4 years agospirv: improve creation of memory_barrier
Rhys Perry [Mon, 2 Mar 2020 13:56:50 +0000 (13:56 +0000)]
spirv: improve creation of memory_barrier

It shouldn't check for atomic counters or return in case we also need to
create a TCS output barrier.

Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4003>

4 years agolima: don't disable tiling if there's linear modifier in list
Vasily Khoruzhick [Tue, 3 Mar 2020 05:42:48 +0000 (21:42 -0800)]
lima: don't disable tiling if there's linear modifier in list

Instead we should disable it if tiling modifier is not here and we
already do that.

Reviewed-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4029>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4029>

4 years agoac: rename min_vgpr_alloc to min_wave64_vgpr_alloc
Samuel Pitoiset [Mon, 2 Mar 2020 08:07:52 +0000 (09:07 +0100)]
ac: rename min_vgpr_alloc to min_wave64_vgpr_alloc

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3975>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3975>

4 years agoac: rename vgpr_alloc_granularity to wave64_vgpr_alloc_granularity
Samuel Pitoiset [Thu, 27 Feb 2020 07:47:54 +0000 (08:47 +0100)]
ac: rename vgpr_alloc_granularity to wave64_vgpr_alloc_granularity

And update the value.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3975>

4 years agoac: rename lds_size_per_cu to lds_size_per_workgroup
Samuel Pitoiset [Thu, 27 Feb 2020 07:45:46 +0000 (08:45 +0100)]
ac: rename lds_size_per_cu to lds_size_per_workgroup

It's more accurate.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3975>

4 years agoturnip: Execute main cs for secondary command buffers
Brian Ho [Thu, 27 Feb 2020 23:48:43 +0000 (18:48 -0500)]
turnip: Execute main cs for secondary command buffers

Previously, we only added the secondary command buffer's draw and
draw epilogue command streams to the primary command buffer on
vkCmdExecuteCommands. However, we also need to merge the primary cs
for non-draw operations like vkCmdCopyBuffer and vkCmdBeginQuery.

Fixes dEQP-VK.memory.pipeline_barrier.host_write_transfer_src.*
and various other tests in dEQP-VK.api.command_buffers.*.

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3988>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3988>

4 years agoturnip: Promote tu_cs_get_size/is_empty to header
Brian Ho [Fri, 28 Feb 2020 15:33:34 +0000 (10:33 -0500)]
turnip: Promote tu_cs_get_size/is_empty to header

These will be used in tu_cmd_buffer.c.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3988>

4 years agonvc0: enable EXT_texture_shadow_lod
Ilia Mirkin [Mon, 2 Mar 2020 03:08:20 +0000 (22:08 -0500)]
nvc0: enable EXT_texture_shadow_lod

This passes all the CTS tests for this extension.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4014>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4014>

4 years agost/mesa: allow TXB2/TXL2 to work with cube array shadow textures
Ilia Mirkin [Mon, 2 Mar 2020 02:55:59 +0000 (21:55 -0500)]
st/mesa: allow TXB2/TXL2 to work with cube array shadow textures

It's a bit asymmetric, but it's such a contrived use-case, and not a lot
of drivers will support it.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4014>

4 years agonv50,nvc0: add newly added PIPE_CAP's to list
Ilia Mirkin [Mon, 2 Mar 2020 00:14:02 +0000 (19:14 -0500)]
nv50,nvc0: add newly added PIPE_CAP's to list

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4014>

4 years agoanv: multiply the scratch space by 4 on gen9-10 like iris and i965
Paulo Zanoni [Sat, 29 Feb 2020 00:03:02 +0000 (16:03 -0800)]
anv: multiply the scratch space by 4 on gen9-10 like iris and i965

My understanding is that there's no reason for the scratch space
allocation to be different between iris, i965 and anv. Let's make all
the functions behave the same.

I don't know if this fixes any specific gen9 bugs, it it might since
it increases the scratch space.

v2: Rebase.
v3: Rebase.
v4: Remove redundant gen 11 check (Jason).

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4006>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4006>

4 years agointel/device: bdw_gt1 actually has 6 eus per subslice
Paulo Zanoni [Fri, 24 Jan 2020 22:41:25 +0000 (14:41 -0800)]
intel/device: bdw_gt1 actually has 6 eus per subslice

Found by inspection, I'm not aware of any bugs caused by this typo.

According to Lionel, it seems we only use this to generate masks
of available EUs for perfromance queries, and it's only used when we
can't query the fused parts of the GPU through DRM_IOCTL_I915_QUERY.
So this patch should help for the corner case where the Kernel is too
old to support the query ioctl.

v2: improve commit message, cc stable (Lionel).

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4006>