st/mesa: use new cso_save/restore_state() functions

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

cso: add new cso_save/restore_state() functions

cso_save_state() takes a bitmask of state items to save.  Calling
cso_restore_state() restores those states.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

cso: remove comment

There's a similar comment just a few lines before.

st/mesa: use new cso_set_viewport_dims() helper

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

cso: add new cso_set_viewport_dims() helper

To simplify some viewport setting code in the state tracker.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

st/mesa: use 'cso' local var instead of st->cso_context

Just a little cleaner.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

st/mesa: consolidate quad drawing code

The glClear, glBitmap and glDrawPixels code now use a new st_draw_quad()
helper function.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

st/mesa: overhaul vertex setup for clearing, glDrawPixels, glBitmap

Define a new st_util_vertex structure which is a bit smaller (9 floats
versus the previous 12 floats per vertex).  Clean up the glClear,
glDrawPixels and glBitmap code that sets up the vertex data and does the
drawing so it's all very similar.  This can lead to more consolidation.

v2: add assertion that vertex buffer slot == 0 to catch possible future
change in cso_get_aux_vertex_buffer_slot() behavior.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

st/mesa: include u_draw.h, not u_draw_quad.h in st_draw.c

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

configure: Bail out on llvm-config component error

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

Handle removal of LLVMAddTargetData in SVN revision 260919

LLVM removed LLVMAddTargetData for the 3.9 release in r260919.  For the two
places in mesa where this is called, only enable the lines when compiling
for less then 3.9.

For the radeon driver, I'm not sure how to check if any other LLVM calls need
to be adjusted.  I think since the target data used is extracted from the
LLVMModule, it isn't necessary to pass it back to LLVM again.

The code does compile, and at least for radeonsi does run OpenGL games.

[ Michel Dänzer: Move #if closer to LLVMAddTargetData in lp_bld_init.c,
  and add HAVE_LLVM < 0x0309 guards around now unused occurrences of TD
  and data_layout ]

Signed-off-by: Matthew Dawson <matthew@mjdsystems.ca>
Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>

i965: Expose logic telling if non-msrt mcs is supported

Alos use the opportunity to mark inputs constant. (Context has to be
given as read-write to intel_miptree_supports_non_msrt_fast_clear()
to support debug output).

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

i965/gen9: Refactor msrt mcs initialization

This will be re-used to initialize auxiliary buffers in lossless
compression case.

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

i965: Add a few assertions on lossless compression

v2 (Ben): Use combination of msaa_layout and number of samples
          instead of introducing explicit type for lossless
          compression (intel_miptree_is_lossless_compressed()).

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

i965: Add a flag telling color resolve pass to ignore CCS_E

v2 (Ben): Use combination of msaa_layout and number of samples
          instead of introducing explicit type for lossless
          compression (intel_miptree_is_lossless_compressed()).

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

i965: Add resolve option for lossless compression

v2 (Ben): Use combination of msaa_layout and number of samples
          instead of introducing explicit type for lossless
          compression (intel_miptree_is_lossless_compressed()).

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

i965: Allow fast clear to be used with lossless compression

v2 (Ben): Use combination of msaa_layout and number of samples
          instead of introducing explicit type for lossless
          compression.
v3 (Ben): Squash with "i965: Resolve color buffer also in
          lossless compression case" and clarify simple
          non-compressed fast clear case.

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

i965: Add helper for detecting lossless compression

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Ben Widawsky <benjamin.widawsky@intel.com>

Revert "i965: Restore vbo after color resolve during brw_try_draw_prims()"

This got pushed accidentally in the first place but wasn't reverted
as it didn't regress piglit but instead fixed one newly introduced
test exercising a corner in case in i965 driver. However, saving and
restoring vertex buffer context is complicated and requires more
thought.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94150

Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Tapani Palli <tapani.palli@intel.com>

nvc0: initial support for GM20x GPUs

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: implement support for maxwell texture headers

Adds support for the new TIC layout that's present on Maxwell GPUs,
heavily based on the code for the existing layout.

This code is required for GM20x support.  While GM10x supports the older
layout still, this commit switches it to use the updated version instead.

Piglit testing shows zero regressions on GM107.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: import maxwell texture header definitions from rnndb

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: split tic format specification

We previously stored texture format information as it would appear in
the TIC.

We're about to support the new TIC layout that appeared with Maxwell,
so it makes more sense to store the data in a split-out format.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: remove nv50_texture.xml.h

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: switch nvc0_tex.c to updated g80_texture.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: switch nvc0_surface.c to updated g80_texture.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: switch nv50_tex.c to updated g80_texture.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: switch nv50_surface.c to updated g80_texture.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: switch nv50_state.c to updated g80_texture.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: switch nv50_formats.c to updated g80_texture.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: import updated g80_texture.xml.h from rnndb

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: remove nv50_defs.xml.h

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: switch nv50_formats.c to updated g80_defs.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: improved macros to handle format specification

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50-: separate vertex formats from surface format descriptions

We've previously had identical naming between vertex and texture
formats, so it mostly made sense to define these together.

However, upcoming patches are going to transition the driver over to
using updated texture header definitions using NVIDIA's naming, and this
will no longer be the case.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: remove unnecessary includes

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: switch nvc0_tex.c to updated g80_defs.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nvc0: switch nvc0_surface.c to updated g80_defs.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: remove unnecessary include

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: switch nv50_transfer.c to g80_defs.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: switch nv50_tex.c to updated g80_defs.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: switch nv50_surface.c to updated g80_defs.xml.h

Verified (binary diff) to produce identical code.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50: import updated g80_defs.xml.h from rnndb

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: count shader images in MaxCombinedShaderOutputResources

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: enable GL image extensions when backend supports them

This enables ARB_shader_image_load_store and ARB_shader_image_size when
the backend claims support for these. It will also implicitly enable the
image component of ARB_shader_texture_image_samples.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

st/mesa: convert GLSL image intrinsics into TGSI

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>

st/mesa: allow st_format.h to be included from C++ files

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

st/mesa: set pipe_image_view layers correctly for 3D textures

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: call st_finalize_texture from image atoms

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: add an image atom for shader images

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

tgsi: show textual format representation

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

gallium: add PIPE_SHADER_CAP_MAX_SHADER_IMAGES

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

gallium: make image views non-persistent objects

Make them akin to shader buffers, with no refcounting/etc. Just used to
pass data about the bound image in ->set_shader_images.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

st/mesa: empty buffer binding if the buffer's not really there

This can happen with 0-sized buffers.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

docs: Document VC4_DEBUG envvar

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Signed-off-by: Eric Anholt <eric@anholt.net>

vc4: Add missing braces in initializer

Silences the following GCC warning:

mesa/src/gallium/drivers/vc4/vc4_qir_schedule.c: In function 'qir_schedule_instructions':
mesa/src/gallium/drivers/vc4/vc4_qir_schedule.c:578:16: warning: missing braces around initializer [-Wmissing-braces]
         struct schedule_state state = { 0 };
                ^

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Signed-off-by: Eric Anholt <eric@anholt.net>

vc4: Correct typo setting 'handled_qinst_cond'

Variable was previously always set to true. Accordingly, the later
assert() served no active purpose.

Found with GCC warning and code inspection:

mesa/src/gallium/drivers/vc4/vc4_qpu_emit.c: In function'vc4_generate_code':
mesa/src/gallium/drivers/vc4/vc4_qpu_emit.c:315:22: warning: variable 'handled_qinst_cond' set but not used [-Wunused-but-set-variable]
                 bool handled_qinst_cond = true;
                      ^

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Signed-off-by: Eric Anholt <eric@anholt.net>

vc4: Don't treat conditional MOVs as raw MOV.

The two consumers want to know that the destination will be exactly the
source, which is not true if we might not set the destination.

Signed-off-by: Eric Anholt <eric@anholt.net>

glsl: warn in GL as well as ES when varying not written

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93339

docs: update GLES 3.1 section for recent nvc0 additions

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>

mesa: need to check resource and set length even if bufSize is 0

This fixes a number of dEQP tests, such as:

dEQP-GLES31.functional.program_interface_query.buffer_limited_query.resource_query

It was expecting the length to be set even in the bufSize == 0 case.
Also _mesa_get_program_resourceiv does some error checking on the
resource which should probably happen even in the bufSize == 0 case as
well although there's no dEQP test for that.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

i965/bxt: Production thread counts

v2: Forgot to squash in the comment removal

Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Reviewed-by: Mark Janes <mark.a.janes@intel.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@intel.com>

egl_dri2: NULL check for xcb_dri2_get_buffers_reply()

Without the check, unsuccessful xcb_dri2_get_buffers_reply(...) causes
segmentation fault in dri2_get_buffers.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Cc: "11.0 11.1" <mesa-stable@lists.freedesktop.org

nv50,nvc0: Remove duplicate logic from nvc0_set_framebuffer_state()

We already have this logic in the gallium/util functions so
lets reduce some entropy while here.

V.2:
  Apply change to nv50 also as suggested by Samuel Pitoiset.

Signed-off-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>

nv50: add missing PIPE_SHADER_CAP_SUPPORTED_IRS

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

i965: Fix gl_DrawID in the vec4 backend.

brw_draw_upload.c uploads VertexID/InstanceID first, then DrawID.
So we need to assign the attribute mapping in that order as well.

Fixes the following Pigit tests with the vec4 backend:
- arb_shader_draw_parameters-drawid vertexid
- arb_shader_draw_parameters-drawid-indirect basevertex

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>

mesa: move assertion in _mesa_cube_face_target()

Fixes piglit arb_texture_view-sampling-2d-array-as-2d-layer regression.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94134
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>

clover: fix build failure since bfd695e

Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>

glsl: Fix overflow of ImageAccess[] array.

The ImageAccess array is statically sized to MAX_IMAGE_UNIFORMS:

   GLenum ImageAccess[MAX_IMAGE_UNIFORMS];

There was no bounds checking ensuring we don't overflow.  Passing in a
shader with too many uniforms would cause writes to extend into other
fields, such as sh->NumImages.

Later linker checks already handle reporting an error when there are too
many images, so just avoid corrupting structures here.

This rearranges the logic a bit to look more like the sampler case.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Tested-by: Jordan Justen <jordan.l.justen@intel.com>

mesa: default FixedSampleLocations to true when using a dummy image

GL_ARB_texture_multisample and GLES 3.1 expect the initial value to be
GL_TRUE. This fixes

dEQP-GLES31.functional.state_query.texture_level.texture_2d_multisample_array.fixed_sample_locations_integer

and a few related tests.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>

nir/types: Add more type constructor functions

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

nir/types: Add a few more glsl_type_is_ functions

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

nir/types: Add helpers for working with sampler and image types

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

nir/types: Add helpers for function types

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

glsl/types: Expose glsl_struct_field and glsl_function_param to C

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

glsl/types: Add a helper for getting image types

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

glsl/types: Add support for function types

SPIR-V has a concept of a function type that's used fairly heavily.  We
could special-case function types in SPIR-V -> NIR but it's easier if we
just add support to glsl_types.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

glsl/types: Add a bare "sampler" type

This is to be used by SPIR-V for representing a sampler that isn't attached
to any particular image.  In SPIR-V, all of the interesting bits such as
dimensionality, sampled type, etc. come from the image, the bare "sampler"
type simply uses a sampled type of VOID and 0 values for the rest.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

glsl/types: Rename sampler_type to sampled_type

It's a bit more descriptive since it is the base type that you get when you
sample from it.  Also, the next commit adds a bare "sampler" type and we
need glsl_type::sampler_type available for a public static member.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

llvmpipe: Do not use barriers if not using threads.

Cc: mesa-stable@lists.freedesktop.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94088
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>

i965: Reupload push and pull constants when we get new shader image unit state.

Fixes several of the
"dEQP-GLES31.functional.image_load_store*load_store*single_layer" dEQP
tests that use image formats we implement using untyped surface
messages.

Cc: mesa-stable@lists.freedesktop.org
Tested-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

i965: fix MAX_COMPUTE_SHARED_SIZE constant value

MAX_COMPUTE_SHARED_SIZE should be set to 32768. This fixes a regression
introduced in be27f77 (mesa: do not use a constant for
MAX_COMPUTE_SHARED_SIZE).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94139
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

nv50/ir: add missing SV_TID and SV_CTAID sysvals on GM107

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50/ir: add MEMBAR emission for GM107

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

docs: document MESA_GLES_VERSION_OVERRIDE envvar

v2: Removed reference to FC not being an allowed suffix (Brian Paul)

Reviewed-by: Brian Paul <brianp@vmware.com>

st/mesa: fix pipe_grid_info initializer

Fixes MSVC build error which doesn't allow empty initializers.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>

trace: add all compute related functions

Changes from v3:
 - dump the TGSI compute program

Changes from v2:
 - remove use of MALLOC()

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: implement limits for ARB_compute_shader

According to the spec, this also increases the following minimum values:
 - MAX_COMBINED_TEXTURE_IMAGE_UNITS     96 (6*16), was 80
 - MAX_UNIFORM_BUFFER_BINDINGS          72 (6*12), was 60

ARB_compute_shader is not enabled by default because images support is
still not implemented yet. If you want to use it you need to set
MESA_EXTENSION_OVERRIDE=GL_ARB_compute_shader.

Changes from v2:
 - make use of the new PIPE_CAP_SHADER_SUPPORTED_IRS cap instead of
   enabling the extension when PIPE_CAP_COMPUTE is enabled.
 - query for PIPE_CAP_COMPUTE first
 - s/shader_supported_irs/compute_supported_irs/
 - disable ARB_compute_shader and add a comment which explains why

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: add compute program dispatch callbacks

This state tracker implements DispatchCompute() and DispatchComputeIndirect().

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: add state validation for compute shaders

This binds atomics, constants, samplers, ssbos, textures and ubos.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: add mappings for compute shader sysvals

LOCAL_INVOCATION_ID, WORK_GROUP_ID and NUM_WORK_GROUPS are respectively
mapped to THREAD_ID, BLOCK_ID and GRID_SIZE.

Changes from v2:
 - add assertions in st_translate_program()

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

st/mesa: keep track of shared memory declarations

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

st/mesa: add intrinsics for shared variables

This adds GLSL intrinsics for load/store and atomic operations.

Changes from v2:
 - use PROGRAM_MEMORY instead of PROGRAM_BUFFER

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: add conversion for compute shaders

According to the spec, there are no predefined inputs nor any
fixed-function outputs.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

st/mesa: add compute shader states

Changes from v2:
 - use as much common code as possible (eg. st_basic_variant)

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

st/mesa: add a second pipeline for compute

Compute needs a new and different validation path.

Changes from v2:
 - make use of unreachable() instead of assert() when the pipeline is
   invalid
 - move the st_pipeline enumeration to st_context.h instead of st_api.h

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

tgsi/ureg: add shared variables support for compute shaders

This introduces TGSI_FILE_MEMORY for shared, global and local memory.
Only shared memory is currently supported.

Changes from v2:
 - introduce TGSI_FILE_MEMORY

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

gallium: add PIPE_SHADER_CAP_SUPPORTED_IRS

This cap indicates the supported representations of programs. It should
be a mask of pipe_shader_ir bits. It will allow to enable
ARB_compute_shader if the underlying driver supports TGSI.

Changes from v2:
 - improve description of PIPE_SHADER_CAP_SUPPORTED_IRS

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

gallium: add indirect compute parameters to pipe_grid_info

Like indirect draw, we need to store a resource and an offset that
needs to be 4 byte aligned. When indirect is used, the size of the
grid (in blocks) is stored with three 32-bit integers.

Changes from v2:
 - s/most values/block sizes/

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

gallium: add a new interface for pipe_context::launch_grid()

This introduces pipe_grid_info which contains all information to
describe a launch_grid call. This will be used to implement indirect
compute in the same fashion as indirect draw.

Changes from v2:
 - correctly initialize pipe_grid_info for nv50/nvc0

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>