intel/compiler: Re-add .vs_inputs_dual_locations = true

Looks like a rebase mistake.

Fixes: 89fe5190a256 ("intel/compiler: Lower flrp32 on Gen11+")

r600/shader: when using images always load thread id gpr at start (v2)

The delayed loading code was fail if we had control flow.

This fixes:
tests/spec/arb_shader_image_load_store/execution/image_checkerboard.shader_test

v2: don't use temp_reg before setting temp_reg up.

Tested-by: Gert Wollny <gw.fossdev@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

r600: fix whitespace in recent 1d texture commit.

trivial fix.

intel/compiler: Add ICL to test_eu_validate.cpp

With the Align16 tests now disabled, we can run the rest of the tests in
ICL mode (and see them pass!)

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler: Disable Align16 tests on Gen11+

Align16 is no more.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler: Add instruction compaction support on Gen11

Gen11 only differs from SKL+ in that it uses a new datatype index table.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler: Mark line, pln, and lrp as removed on Gen11+

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler: Lower flrp32 on Gen11+

The LRP instruction is no more.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Implement ddy without using align16 for Gen11+

Align16 is no more. We previously generated an align16 ADD instruction
to calculate DDY:

   add(16) g25<1>F  -g23<4>.xyxyF   g23<4>.zwzwF   { align16 1H };

Without align16, we now implement it as:

   add(4) g25<1>F   -g23<0,2,1>F    g23.2<0,2,1>F  { align1 1N };
   add(4) g25.4<1>F -g23.4<0,2,1>F  g23.6<0,2,1>F  { align1 1N };
   add(4) g26<1>F   -g24<0,2,1>F    g24.2<0,2,1>F  { align1 1N };
   add(4) g26.4<1>F -g24.4<0,2,1>F  g24.6<0,2,1>F  { align1 1N };

where only the first two instructions are needed in SIMD8 mode.

Note: an earlier version of the patch implemented this in two
instructions in SIMD16:

   add(8) g25<2>F   -g23<4,2,0>F    g23.2<4,2,0>F  { align1 1N };
   add(8) g25.1<2>F -g23.1<4,2,0>F  g23.3<4,2,0>F  { align1 1N };

but I realized that the channel enable bits will not be correct. If we
knew we were under uniform control flow, we could emit only those two
instructions however.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Simplify ddx/ddy code generation

The brw_reg() constructor just obfuscates things here, in my opinion.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Pass fs_inst to generate_ddx/ddy instead of opcode

In a future patch, generate_ddy will want to inspect inst->exec_size.
Change generate_ddx as well for consistency.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Don't generate integer DWord multiply on Gen11

Like CHV et al., Gen11 does not support 32x32 -> 32/64-bit integer
multiplies.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Implement FS_OPCODE_LINTERP with MADs on Gen11+

The PLN instruction is no more. Its functionality is now implemented
using two MAD instructions with the new native-float type. Instead of

   pln(16) r20.0<1>:F r10.4<0;1,0>:F r4.0<8;8,1>:F

we now have

   mad(8) acc0<1>:NF r10.7<0;1,0>:F r4.0<8;8,1>:F r10.4<0;1,0>:F
   mad(8) r20.0<1>:F acc0<8;8,1>:NF r5.0<8;8,1>:F r10.5<0;1,0>:F
   mad(8) acc0<1>:NF r10.7<0;1,0>:F r6.0<8;8,1>:F r10.4<0;1,0>:F
   mad(8) r21.0<1>:F acc0<8;8,1>:NF r7.0<8;8,1>:F r10.5<0;1,0>:F

... and in the case of SIMD8 only the first pair of MAD instructions is
used.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Return multiple_instructions_emitted from generate_linterp

If multiple instructions are emitted, special handling of things like
conditional mod and NoDDClr/NoDDChk need to be performed.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler/fs: Fix application of cmod and saturate to LINE/MAC pair

This isn't technically broken, but the next patch will make this
function report whether it generated multiple instructions, and that
information will be used to disable the application of conditional mod
by the generic code.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler: Add Gen11+ native float type

This new type exposes the additional precision offered by the
accumulator register and will be used in the next patch to implement the
functionality of the PLN instruction using a pair of MAD instructions.

One weird thing to note: align1 ternary instructions may only have an
accumulator in the dst or src1 normally, but when src0's type is :NF
the accumulator is read.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel/compiler: Add Gen11 register types

The hardware register types' encodings have changed on Gen11. Good thing
we have that superfluous looking brw_reg_type abstraction lying around!

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel: Disable 64-bit extensions on platforms without 64-bit types

Gen11 does not support DF, Q, UQ types in hardware. As a result, we have
to disable some GL extensions until they can be reimplemented.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

intel: Add icl pci id for INTEL_DEVID_OVERRIDE

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>

i965: Warn about preliminary support for Gen11

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

intel: Add a preliminary device for Ice Lake

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Anuj Phogat <anuj.phogat@intel.com>

anv: remove anv_gem_set_context_priority helper

anv_gem_set_context_param is to be used directly instead!

Fixes: 6d8ab53303 "anv: implement VK_EXT_global_priority extension"
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

swr/rast: revert clip distance precision

Fixes piglit tests that broke with 8a64593bde

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

swr/rast: Faster frustum prim culling

Fix clipper validMask setting. We don't need to run frustum rejected
primitives through the clipper.  Perform frustum culling with only
frustum clip codes. Guardband clip codes cannot be used because they
overlap frustum codes.

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

swr/rast: Consolidate TRANSLATE_ADDRESS

Translate is now part of an overloaded LOAD call which required a change to
the code gen to skip the load functions in order to handle them manually
to make them virtual.

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

swr/rast: Code generation cleanup

Generate more compact code from gen_llvm.hpp.

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

swr/rast: Remove draw type from event definitions

- Have the draw type sent to DrawInfoEvent in handlers created in
  archrast.cpp.  The draw type no longer needs to be sent during during
  AR_API_EVENT() call in api.cpp.

- Remove draw type from event defintions in events_private.proto, no
  longer needed

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

swr/rast: whitespace change

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

swr/rast: Fix index buffer overfetch issue for non-indexed draws

Populate pLastIndex, even for the non-indexed case.  An zero pLastIndex
can cause the index offsets inside the fetcher to have non-sensical values
that can be either very large positive or very large negative numbers.

Reviewed-By: Bruce Cherniak <bruce.cherniak@intel.com>

softpipe: don't iterate through PIPE_MAX_SHADER_SAMPLER_VIEWS

We were setting view to NULL if the iteration was larger than i.
But in fact if the view is NULL the code did nothing anyway...

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

cso: don't cycle through PIPE_MAX_SHADER_SAMPLER_VIEWS on context destroy

There's no point, we know the highest non-null one.

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

draw: don't needlessly iterate through all sampler view slots

We already stored the highest (potentially) used number.

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

anv: implement VK_EXT_global_priority extension

v2: add ANV_CONTEXT_REALTIME_PRIORITY (Chris)
    use unreachable with unknown priority (Samuel)

v3: add stubs in gem_stubs.c (Emil)
    use priority defines from gen_defines.h

v4: cleanup, add anv_gem_set_context_param (Jason)

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com> (v2)
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> (v2)
Reviewed-by: Emil Velikov <emil.velikov@collabora.com> (v3)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

i965: use context priority definitions from gen_defines.h

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

intel: add new common header gen_defines.h

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

winsys/amdgpu: request high addresses

We now have hopefully fixed all bugs regarding high addresses on Vega10 and
Raven. Start to use the high range to make room for SVM in the low
range.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

ac/shader: move scanning some info about input PS declarations

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

glsl/linker: fix bug when checking precision qualifier

According to GLSL ES 3.2 spec, see table in 9.2.1 "Linked Shaders"
section, the precision qualifier should match for uniform variables.
This also applies to previous GLSL ES 3.x specs.

This 'if' checks the condition for uniform variables, while for UBOs
it is checked in link_interface_blocks.cpp.

Fixes: b50b82b8a553
("glsl/es31: precision qualifier doesn't need to match in shader interface block members")

Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

anv: set maxResourceSize to the respective value for each generation

v2:
- Add the proper values to gen9+ (Jason)

Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

r600: partly revert disabling tiling for 1d texture.

Previously we had a check for 1d of narrow 2D textures, however
narrow 2d textures caused gpu hangs, but it was correct for 1d
textures.

This fixes a bunch of 1D image piglits for me.

Fixes: 7b8e1c089d (r600/texture: drop lowering 1d/2d images to linear.)
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

nir: fix interger divide by zero crash during constant folding

From the GLSL 4.60 spec Section 5.9 (Expressions):

   "Dividing by zero does not cause an exception but does result in
    an unspecified value."

Fixes: 89285e4d47a6 "nir: add new constant folding infrastructure"
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105271

st/mesa: ensure that images don't try to reference non-existent levels

Ideally the st_finalize_texture call would take care of that, but it
doesn't seem to with KHR-GL45.shader_image_size.advanced-nonMS-*. This
assertion makes sure that no such values are passed to the driver.

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

ac/radv: move load base vertex abi setup to vertex shader.

This was segfaulting:
dEQP-VK.memory.pipeline_barrier.host_write_index_buffer.1024

Fixes: 8de6f797070 (ac/radeonsi: add load_base_vertex() to the abi)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>

ac/shader: fix vertex input with components.

This fixes:
dEQP-VK.glsl.440.linkage.varying.component.*

Fixes: 1c57a6da5e3 (ac/shader: scan vertex inputs usage mask)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

radv: remove device pointer from buffer.

This is never used.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

nir: add lower_ldexp to nir compiler options

Reviewed-by: Marek Olšák <marek.olsak@amd.com>

ac: implement nir_op_ldexp

Reviewed-by: Marek Olšák <marek.olsak@amd.com>

ac: fix nir_op_fdd{x,y} handling

radeonsi, i965 and anv all treat fdd{x,y} opcodes the same as
fdd{x,y}_coarse by default. The SPIR-V spec lets the implementation
decide how it should be handled and radv was previously going
for the higher quality option. Here we change the shared amd
code to match how nir_op_fdd{x,y} is expected to be handled
by the other NIR drivers.

Fixes piglit test:
./bin/arb_shader_texture_lod-texgrad -auto

Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

ac/radeonsi: add load_base_vertex() to the abi

Fixes the following piglit tests:

./bin/arb_shader_draw_parameters-basevertex basevertex -auto -fbo
./bin/arb_shader_draw_parameters-basevertex basevertex-baseinstance -auto -fbo

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

radeonsi: create get_base_vertex() helper

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

radeonsi/nir: disable vertex_id_zero_based lowering

The lowering is incompatible with how the radeonsi backend works.

Fixes piglit test:
./bin/arb_shader_draw_parameters-basevertex vertexid-zerobased -auto

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

ac: add support for handling nir_intrinsic_load_vertex_id

This will be used by radeonsi.

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

ac: fix f2b and i2b for doubles

Without this llvm was asserting in debug builds.

V2: use LLVMConstNull()

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>

intel/ir: Fix invalid type aliasing with undefined behavior in test_eu_compact.

test_fuzz_compact_instruction() was attempting to modify the uint64_t
data array of a brw_inst through a pointer to uint32_t, which has
undefined behavior.  This was causing the test_eu_compact unit test to
fail mysteriously for me on GCC 7 with some additional
harmless-looking changes I had applied to my tree, which happened to
affect the order instructions are emitted by GCC causing the bit
twiddling to be done after the clear_pad_bits() call which is supposed
to overwrite the same data through a pointer of different type,
leading to data corruption.  A similar failure has been reported by
Vinson Lee on the master branch built with GCC 8.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105052
Tested-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>

util/bitset: Make C++ wrapper trivially constructible.

In order to fix a build failure on compilers not implementing
unrestricted unions, which is a C++11 feature.

v2: Provide signed integer comparison and assignment operators instead
    of BITSET_WORD ones to avoid spurious ambiguity warnings on
    comparisons with a signed integer literal.

Fixes: ba79a90fb52e1e81fb "glsl: Switch ast_type_qualifier to a 128-bit bitset."
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105238
Tested-by: Roland Scheidegger <sroland@vmware.com>
Tested-By: George Kyriazis <george.kyriazis@intel.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>

intel/tools: Use gen_device_name_to_pci_device_id in aubinator

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>

intel/common: Add gen_device_name_to_pci_device_id

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>

intel/vulkan: Support INTEL_DEVID_OVERRIDE environment variable

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>

i965: Use gen_get_pci_device_id_override

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>

intel/common: Add gen_get_pci_device_id_override

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>

intel/vulkan: Support INTEL_NO_HW environment variable

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>

android: fix source files path for libmesa_anv_gen11

Signed-off-by: Harish Krupo <harish.krupo.kps@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

meson: avoid changing types for the dri3 option

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

meson: simplify the gbm option code, and avoid changing types

v2: drop gallium comment (Dylan)

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

ac/nir: clean up a hack about rounding 2nd coord component

It's basically just the opposite, and it only makes sense to
round the layer for 2D texture arrays.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nvc0: collapse output slots to have adjacent registers

The hardware skips over unallocated slots, so we have to make sure those
registers are packed together.

Fixes KHR-GL45.enhanced_layouts.fragment_data_location_api

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Tested-by: Karol Herbst <kherbst@redhat.com>

radv: expose async compute on SI

It looks like we had all the pieces in place for this,
just never tested it and turned it on.

I don't see any CTS regressions and the computeshader
demo runs.

Acked-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>

radv: merge tess rings into a single bo

Inspired by a passing commit to radeonsi.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

docs: update calendar, add news and link release notes to 17.3.6

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

docs: add sha256 checksums for 17.3.6

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit b00880973eeab5d29413eb8a27707f62904723ea)

docs: add release notes for 17.3.6

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit b3e5a3f35bfe5b69758ceb90f93fffbee01a5682)

meson: fix building without GL

libgl will be undefined _glx, so move that check inside the
`if with_glx != 'disabled'` block.

v2: - Simplify commit message (Eric, Emil)

Fixes: 5c460337fd9c109 ("meson: Fix GL and EGL pkg-config files with glvnd")
Reported-by: Jason Ekstrand <jason.ekstrand@intel.com>
Signed-off-by: Dylan Baker <dylan.c.baker@intel.com>
CC: Daniel Stone <daniels@collabora.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Untested-by: Jason Ekstrand <jason@jlekstrand.net>
Acked-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

intel: aubinator_error_decode: fix segfault on missing register

Some register might be missing in our genxmls. Don't try to decode
them.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

*-symbol-check: use correct `nm` path when cross-compiling

Inspired-by: a similar patch for libdrm by Heiko Becker
Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

nvir/gm107: consider FILE_FLAGS dependencies in SchedDataCalculatorGM107

currently while insterting barriers, writes and reads to FILE_FLAGS aren't
considered. This can lead to WaR hazards in some situations.

With the previous commit fixes shaders with intstructions like this:
  mad u32 $r2 $r4 $r11 $r2
  mad u32 { $r5 $c0 } $r4 $r10 $r6
  mad (SUBOP:1) u32 $r3 $r4 $r10 $r2 $c0

Affects OpenCL CTS tests on Maxwell+:
basic/test_basic intmath_long
basic/test_basic intmath_long2
basic/test_basic intmath_long4

v2: only put barriers on instructions which actually read flags

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>

nvir/gm107: iterate over all defs in SchedDataCalculatorGM107::findFirstUse

In the sched data calculator we have to track first use of defs by iterating
over all defs of an instruction, not just the first one.

v2: fix minGRP and maxGRP values

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Signed-off-by: Karol Herbst <kherbst@redhat.com>

ac/nir: use ordered float comparisons except for not equal

Original patch from Timothy Arceri, I have just fixed the
not equal case locally.

This fixes one important rendering issue in Wolfenstein 2
(the cutscene transition issue).

RadeonSI uses the same ordered comparisons, so I guess that
what we should do as well.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104302
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104905
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>

android: vulkan/util: add dependency on libnativewindow for O and later

Similar to 90dd6e5 ("Android: egl: add dependency on libnativewindow")

Fixes the following building error:

In file included from out/target/product/x86_64/obj_x86/STATIC_LIBRARIES/libmesa_vulkan_util_intermediates/util/vk_enum_to_str.c:26:
external/mesa/include/vulkan/vk_android_native_buffer.h:22:10: fatal error: 'system/window.h' file not found
         ^~~~~~~~~~~~~~~~~
1 error generated.

Cc: "18.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>

android: anv: add dependency on libnativewindow for O and later

Similar to 90dd6e5 ("Android: egl: add dependency on libnativewindow")

Fixes the following building errors:

In file included from external/mesa/src/intel/vulkan/gen7_cmd_buffer.c:30:
In file included from external/mesa/src/intel/vulkan/anv_private.h:72:
external/mesa/include/vulkan/vk_android_native_buffer.h:22:10: fatal
error: 'system/window.h' file not found
         ^~~~~~~~~~~~~~~~~
1 error generated.
...
In file included from external/mesa/src/intel/vulkan/anv_gem.c:32:
In file included from external/mesa/src/intel/vulkan/anv_private.h:72:
external/mesa/include/vulkan/vk_android_native_buffer.h:22:10: fatal
error: 'system/window.h' file not found
         ^~~~~~~~~~~~~~~~~
1 error generated.

Cc: "18.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>

android: anv/extensions: fix generated sources build

Building rules are aligned to automake ones

The correct script to build anv_extensions.{c,h} is anv_extensions_gen.py
Generation rules for anv_extensions.c requires --out-c option
Generation rules for anv_extensions.h were missing
Necessary include paths are added to avoid following build errors:

cp: cannot stat '.../gen/STATIC_LIBRARIES/libmesa_vulkan_common_intermediates/vulkan/anv_extensions.c':
No such file or directory

In file included from external/mesa/src/intel/vulkan/anv_gem.c:32:
external/mesa/src/intel/vulkan/anv_private.h:75:10: fatal error: 'anv_extensions.h' file not found
         ^~~~~~~~~~~~~~~~~~
1 error generated.

In file included from external/mesa/src/intel/vulkan/anv_batch_chain.c:30:
external/mesa/src/intel/vulkan/anv_private.h:75:10: fatal error: 'anv_extensions.h' file not found
         ^~~~~~~~~~~~~~~~~~
1 error generated.

Fixes: dd088d4bec7 ("anv/extensions: Generate a header file with extension tables")
Cc: "18.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

radeonsi: remove 2 unused user SGPRs from merged TES-GS with 32-bit pointers

The effect of the last 13 commits on user SGPR counts:

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

radeonsi: make SI_SGPR_VERTEX_BUFFERS the last user SGPR input

so that it can be removed and replaced with inline VBO descriptors,
and the pointer can be packed in unused bits of VBO descriptors.
This also removes the pointer from merged TES-GS where it's useless.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

radeonsi: set correct num_input_sgprs for VS prolog in merged shaders

We need to take num_input_sgprs from VS, not the second shader.
No apps suffered from this.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

radeonsi: allow fewer input SGPRs in 2nd shader of merged shaders

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

radeonsi: don't use struct si_descriptors for vertex buffer descriptors

VBO descriptor code will change a lot one day.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

build: Move wayland-scanner check into platform

Also only check for wayland-scanner if building for the Wayland
platform.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Fixes: bfa22266cd4d ("vulkan/wsi/wayland: Add support for zwp_dmabuf")
Cc: Emil Velikov <emil.velikov@collabora.co.uk>
Reported-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105211

build: Move wayland-protocols check into platform

In line with wayland-client and wayland-server, move the check for
wayland-protocols into the wayland platform branch.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Fixes: bfa22266cd4d ("vulkan/wsi/wayland: Add support for zwp_dmabuf")
Cc: Emil Velikov <emil.velikov@collabora.co.uk>
Reported-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105211

vulkan/wsi/wayland: Move Wayland protocol from BUILT_SOURCES

autotools wants to have the BUILT_SOURCES ready as soon as it enters the
directory, even if they are not used. This meant the build failed if
wayland-protocols was not available on the system, even if it was not
enabled.

As BUILT_SOURCES cannot be used in a conditional (cf. 166852ee957f), do
the same thing as EGL and manually encode the dependencies in the
Makefile.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Fixes: bfa22266cd4d ("vulkan/wsi/wayland: Add support for zwp_dmabuf")
Cc: Emil Velikov <emil.velikov@collabora.co.uk>
Reported-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105211

r600: fix tgsi clock last setting

On cayman this was hitting an assert later, which probably wasn't
see on non-cayman due to having the t slot.

Fixes: 9041730d1 (r600: add support for ARB_shader_clock.)

r600: add time lo/hi debugging output.

This just adds the these to the debug prints.

radeonsi/nir: enable lowering of fpow

Lowering fpow in NIR rather than LLVM can be beneficial.

Polaris results:

Totals from affected shaders:
SGPRS: 124928 -> 124896 (-0.03 %)
VGPRS: 68616 -> 68332 (-0.41 %)
Spilled SGPRs: 394 -> 413 (4.82 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Private memory VGPRs: 0 -> 0 (0.00 %)
Scratch size: 0 -> 0 (0.00 %) dwords per thread
Code Size: 3668912 -> 3658368 (-0.29 %) bytes
LDS: 0 -> 0 (0.00 %) blocks
Max Waves: 18575 -> 18593 (0.10 %)
Wait states: 0 -> 0 (0.00 %)

Fixes: d6b753920677 "ac/nir: remove emission of nir_op_fpow"
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

ac: make use of ac_get_llvm_num_components() helper

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>

gallium/tgsi: remove is_msaa_sampler array from tgsi_shader_info

Seems to have not been used since 16be87c90429

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

radeonsi/nir: fix loading of doubles for tess varyings

Reviewed-by: Marek Olšák <marek.olsak@amd.com>

radeonsi/nir: fix lds store in tcs outputs handling

We were ignoring the channel offset.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>

r600: Take ALU_EXTENDED into account when evaluating jump offsets

ALU_EXTENDED needs 4 DWORDS instead of the usual 2, hence if the last ALU
clause within a IF-JUMP or ELSE branch is ALU_EXTENDED the target jump
offset needs to be adjusted accordingly.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104654
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Gert Wollny <gw.fossdev@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>

mesa: Expose EXT_shader_framebuffer_fetch(_non_coherent) on desktop and embedded GL.

Reviewed-by: Plamena Manolova <plamena.manolova@intel.com>

glsl: Silence warnings when reading from a framebuffer fetch output.

Framebuffer fetch outputs are implicitly initialized upon entry to the
fragment shader.

Reviewed-by: Plamena Manolova <plamena.manolova@intel.com>

glsl: Specify framebuffer fetch coherency mode in lower_blend_equation_advanced().

This requires passing an extra argument to the lowering pass because
the KHR_blend_equation_advanced specification doesn't seem to define
any mechanism for the implementation to determine at compile-time
whether coherent blending can ever be used (not even an "#extension
KHR_blend_equation_advanced_coherent" directive seems to be required
in the shader source AFAICT).

In the long run we'll probably want to do state-dependent recompiles
based on the value of ctx->Color.BlendCoherent, but right now there
would be no benefit from that because the only driver that supports
coherent framebuffer fetch is i965 on SKL+ hardware, which are unable
to support the non-coherent path for the moment because of texture
layout issues, so framebuffer fetch coherency is always enabled for
them.

Reviewed-by: Plamena Manolova <plamena.manolova@intel.com>

glsl: Add support for the framebuffer fetch layout(noncoherent) qualifier.

This allows the application to request framebuffer fetch coherency
with per-fragment output granularity.  Coherent framebuffer fetch
outputs (which is the default if no qualifier is present for
compatibility with older versions of the EXT_shader_framebuffer_fetch
extension) will have ir_variable_data::memory_coherent set to true.

Reviewed-by: Plamena Manolova <plamena.manolova@intel.com>