freedreno: set modifier when exporting buffer

Fixes an assert we start hitting with kms/gbm:

  #0  0x0000007fbf3d6e3c in raise () from /lib64/libc.so.6
  #1  0x0000007fbf3c4a68 in abort () from /lib64/libc.so.6
  #2  0x0000007fbf3d04e8 in __assert_fail_base () from /lib64/libc.so.6
  #3  0x0000007fbf3d0550 in __assert_fail () from /lib64/libc.so.6
  #4  0x0000007fbf5a73c4 in gbm_dri_bo_create (gbm=0x5820f0, width=2160, height=1440, format=875713112, usage=0, modifiers=0x695e00, count=1) at ../src/gbm/backends/dri/gbm_dri.c:1150
  #5  0x0000007fbf5a49c4 in gbm_bo_create_with_modifiers (gbm=0x5820f0, width=2160, height=1440, format=875713112, modifiers=0x695e00, count=1) at ../src/gbm/main/gbm.c:491
  #6  0x0000007fbbac3d64 in get_back_bo (dri2_surf=0x6f4cc0) at ../src/egl/drivers/dri2/platform_drm.c:258
  #7  0x0000007fbbac4318 in dri2_drm_image_get_buffers (driDrawable=0x704490, format=4098, stamp=0x6fc730, loaderPrivate=0x6f4cc0, buffer_mask=1, buffers=0x7fffffe210) at ../src/egl/drivers/dri2/platform_drm.c:409
  #8  0x0000007fbf5a5318 in image_get_buffers (driDrawable=0x704490, format=4098, stamp=0x6fc730, loaderPrivate=0x70e150, buffer_mask=1, buffers=0x7fffffe210) at ../src/gbm/backends/dri/gbm_dri.c:135
  #9  0x0000007fbe4308c4 in dri_image_drawable_get_buffers (drawable=0x6fc730, images=0x7fffffe210, statts=0x6f2660, statts_count=1) at ../src/gallium/state_trackers/dri/dri2.c:339
  #10 0x0000007fbe430c44 in dri2_allocate_textures (ctx=0x614b30, drawable=0x6fc730, statts=0x6f2660, statts_count=1) at ../src/gallium/state_trackers/dri/dri2.c:466
  #11 0x0000007fbe435580 in dri_st_framebuffer_validate (stctx=0x714160, stfbi=0x6fc730, statts=0x6f2660, count=1, out=0x7fffffe3b8) at ../src/gallium/state_trackers/dri/dri_drawable.c:85
  #12 0x0000007fbe7b2c84 in st_framebuffer_validate (stfb=0x6f2190, st=0x714160) at ../src/mesa/state_tracker/st_manager.c:222
  #13 0x0000007fbe7b4884 in st_api_make_current (stapi=0x7fbf0430d8 <st_gl_api>, stctxi=0x714160, stdrawi=0x6fc730, streadi=0x6fc730) at ../src/mesa/state_tracker/st_manager.c:1074
  #14 0x0000007fbe434f44 in dri_make_current (cPriv=0x703c20, driDrawPriv=0x704490, driReadPriv=0x704490) at ../src/gallium/state_trackers/dri/dri_context.c:301
  #15 0x0000007fbe42c910 in driBindContext (pcp=0x703c20, pdp=0x704490, prp=0x704490) at ../src/mesa/drivers/dri/common/dri_util.c:579
  #16 0x0000007fbbabab40 in dri2_make_current (drv=0x69d170, disp=0x69c6e0, dsurf=0x6f4cc0, rsurf=0x6f4cc0, ctx=0x70cb40) at ../src/egl/drivers/dri2/egl_dri2.c:1456
  #17 0x0000007fbbaa8ef4 in eglMakeCurrent (dpy=0x69c6e0, draw=0x6f4cc0, read=0x6f4cc0, ctx=0x70cb40) at ../src/egl/main/eglapi.c:862
  #18 0x0000007fbf5736ac in InternalMakeCurrentVendor (dpy=dpy@entry=0x614fb0, draw=draw@entry=0x6f4cc0, read=read@entry=0x6f4cc0, context=context@entry=0x70cb40, apiState=apiState@entry=0x6fc940, vendor=0x6975f0) at libegl.c:861
  #19 0x0000007fbf573764 in InternalMakeCurrentDispatch (dpy=0x614fb0, draw=0x6f4cc0, read=0x6f4cc0, context=0x70cb40, vendor=0x6975f0) at libegl.c:630
  #20 0x0000000000403640 in init_egl (egl=0x5805a8 <gl>, gbm=0x580528 <gbm>, samples=0) at ../common.c:263
  #21 0x0000000000403c1c in init_cube_smooth (gbm=0x580528 <gbm>, samples=0) at ../cube-smooth.c:225
  #22 0x0000000000408618 in main (argc=1, argv=0x7fffffe8d8) at ../kmscube.c:145

Fixes: 1ce5d757d04 freedreno: core buffer modifier support
Signed-off-by: Rob Clark <robdclark@gmail.com>

radv: always pass the GFX9 fence data to si_cs_emit_cache_flush()

Remove two useless checks.

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

radv: compute the GFX9 fence VA at allocation time

Instead of doing every time we emit cache flushes.

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

radv: only allocate the GFX9 fence and EOP BOs for the gfx queue

It's invalid to emit a ZPASS_DONE event on the compute queue, and
the fence BO is unused on the compute queue (ie. we don't flush
CB or DB caches).
This saves some space in the upload BO.

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

radv: remove old_fence parameter from si_cs_emit_write_event_eop()

This parameter is actually useless as the immediate value
can always be zero.

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

radv: improve gathering of load_push_constants with dynamic bindings

For example, if a pipeline has two stages VS and FS. And if only
the fragment stage needs dynamic bindings, we shouldn't allocate
an extra user SGPR for the vertex stage. Of course, if the vertex
stage loads constants, it needs an user SGPR.

This should reduce the number of SET_SH_REG packets that are emitted.

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

gallium: Add PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS

In the Intel backend, it makes the most sense to treat gl_TessLevelInner
and gl_TessLevelOuter as ordinary shader inputs.  For Radeon, it makes
more sense to treat them as system values which get special handling.

We already have a compiler option for this, but the Iris driver will
need a capability bit so we can set it appropriately.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>

nv50,nvc0: mark textures dirty on fb update

We may have to flush the cache if there are any textures presently bound
that refer to the outgoing framebuffer. This is only checked at
validation time.

Fixes a number of dEQP-GLES3.functional.fbo.color.repeated_clear.sample.*
tests, which would bind a texture, then clear it while the binding was
in effect, and then render to a different texture. This seems legal
under the "no feedback loops" rule.

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

ac/nir_to_llvm: fix interpolateAt* for structs

This fixes the arb_gpu_shader5 interpolateAt* tests that contain
structs.

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

ac/nir_to_llvm: add bindless support for uniform handles

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

radeonsi/nir: add missing piece for bindless image support

This fixes some piglit tests and is was TGSI does.

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

freedreno: core buffer modifier support

Split out of a patch from Fritz Koenig to decouple from a6xx UBWC
enablement, and added fd_resource_create_with_modifiers().

loader: fix the no-modifiers case

Normally modifiers take precendence over use flags, as they are more
explicit.  But if the driver supports modifiers, but the xserver does
not, then we should fallback to the old mechanism of allocating a buffer
using 'use' flags.

Fixes: 069fdd5f9facbd72fb6a289696c7b74e3237e70f
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org>

freedreno: add query for dmabuf modifiers

freedreno: drm_fourcc.h header include

Add Qualcomm modifier for UBWC

svga: add new gallium formats to the format conversion table

Fixes a static assertion which broke the build.

Fixes: 3ee240890 "gallium: add SINT formats to have exact counterparts to SNORM formats"
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Neha Bhende<bhenden@vmware.com>

radeonsi: rename rfence -> sfence

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: rename rbo, rbuffer to buf or buffer

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: rename rsrc -> ssrc, rdst -> sdst

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: rename rquery -> squery

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: rename r600_resource -> si_resource

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

vulkan: make generated enum to strings helpers available from c++

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>

radeonsi: remove r600 from comments

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

winsys/amdgpu: rename rfence, rsrc, rdst -> afence, asrc, adst

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: rename rview -> sview

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: rename rscreen -> sscreen

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: disable render cond & pipeline stats for internal compute dispatches

radeonsi: use compute for resource_copy_region when possible

v2: marek: fix snorm8 blits

Signed-off-by: Sonny Jiang <sonny.jiang@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>

radeonsi: add compute_last_block to configure the partial block fields

gallium/util: add util_format_snorm8_to_sint8 (from radeonsi)

gallium: add SINT formats to have exact counterparts to SNORM formats

for radeonsi

radeonsi: move PKT3_WRITE_DATA generation into a helper function

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: don't use WRITE_DATA.DST_SEL == MEM_GRBM on >= CIK

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: fix the top-of-pipe fence on SI

SI doesn't have MEM.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: correct WRITE_DATA.DST_SEL definitions

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radeonsi: compile clear and copy buffer compute shaders on demand

same as all other shaders

radeonsi: remove redundant call to emit_cache_flush in compute clear/copy

launch_grid calls it.

radeonsi: use buffer_store_format_x & xy

radeonsi: fix rendering to tiny viewports where the viewport center is > 8K

This fixes an assertion failure with GL CTS when cts-runner is used.
(not a specific test)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108877
Cc: 18.3 <mesa-stable@lists.freedesktop.org>

radeonsi: fix a u_blitter crash after a shader with FBFETCH

This fixes an assertion failure with GL CTS when cts-runner is used.
(not a specific test)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108877
Cc: 18.3 <mesa-stable@lists.freedesktop.org>

winsys/amdgpu: use the new BO list API

anv: Implement transform feedback queries

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

genxml: Add SO_PRIM_STORAGE_NEEDED and SO_NUM_PRIMS_WRITTEN

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

anv: Implement CmdBegin/EndQueryIndexed

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

anv: Implement vkCmdDrawIndirectByteCountEXT

Annoyingly, this requires that we implement integer division on the
command streamer.  Fortunately, we're only ever dividing by constants so
we can use the mulh+add+shift trick and it's not as bad as it sounds.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

anv: Implement the basic form of VK_EXT_transform_feedback

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

anv: Add pipeline cache support for xfb_info

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

anv: Add but do not enable VK_EXT_transform_feedback

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

nir/xfb: distinguish array of structs vs array of blocks

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir/xfb: Properly handle arrays of blocks

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>

nir/xfb: don't assert when xfb_buffer/stride is present but not xfb_offset

In order to allow nir_gather_xfb_info to be used on OpenGL,
specifically ARB_gl_spirv.

So, from OpenGL 4.6 spec, section 11.1.2.1, "Output Variables":

    "outputs specifying both an *XfbBuffer* and an *Offset* are
     captured, while outputs not specifying both of these are not
     captured. Values are captured each time the shader writes to such
     a decorated object."

This implies that are captured if both are present, and not if one of
those are lacking. Technically, it doesn't explicitly point that
having just one or the other is a mistake. In some cases, glslang is
adding some extra XfbBuffer without XfbOffset around, and mentioning
that technically that is not a bug (see issue#1526)

And for the case of Vulkan, as the same glslang issue mentions, it is
not clear if that should be a mistake or not. But even if it is a
mistake, it is not really needed to be checked on the driver, and we
can let the validation layers to check that.

v2: simplify explicit_xfb_buffer and explicit_offset checks (Jason).

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

nir/xfb: Fix offset accounting for dvec3/4

Before, we were double-counting the component slots when we had a dvec3
or dvec4.  Instead, just add them in once and manually offset the
recorded output offset.

Fixes: 19064b8c "nir: Add a pass for gathering transform feedback info"
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>

nir: Preserve offsets in lower_io_to_scalar_early

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>

nir: fix lowering arrays to elements for XFB outputs

If we have a transform feedback output like:

float[2] x2_out (VARYING_SLOT_VAR1.x, 0, 0)

which is lowered by nir_lower_io_arrays_to_elements to,

float x2_out (VARYING_SLOT_VAR1.x, 0, 0)
float x2_out@5 (VARYING_SLOT_VAR2.x, 0, 0)

We have to update the destination offset to avoid overwriting
the same value.

v2 (Jason Ekstrand):
 - Compute the correct offsets for arrays of vectors and/or doubles

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>

nir: do not remove varyings used for transform feedback

When a xfb buffer is explicitely declared on a varying
variable, we shouldn't remove it at link time.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

spirv: Only set interface_type on blocks

Instead of setting interface_type to whatever the per-vertex type is, we
only set it on blocks.  This allows later passes to tell the difference
between variables that are in blocks and those that aren't.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>

spirv: Only split blocks

Instead of splitting every per-vertex struct, just split the ones that
are actually blocks.  The reason for the split is so that we have
separate variables for separate locations, qualifiers, and builtin
decorations.  The vulkan spec only allows these on members of blocks.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>

spirv: Initialize struct member offsets to -1

This is the "no offset specified" value.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

anv: Always emit at least one vertex element

This seems to make the simulator happier.  The early return wasn't
really protecting anything and the code that follows will happily
initialize the dummy element to STORE_0 and emit it.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

configure: EGL requirements only apply if EGL is built

Issue was hit with this configuration:
  --disable-{egl,gbm} --with-platform=drm

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Fixes: 3208fd2e46b ("configure: move platform handling further up")
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

freedreno: a2xx: add partial lower_scalar pass for ir2

Some instructions can only be scalar on a2xx, lower these only

Signed-off-by: Jonathan Marek <jonathan@marek.ca>

freedreno: a2xx: add ir2 copy propagation

Two cases:
* replacing srcs which refer to MOV instructions
* replacing MOVs used to write to exports

Signed-off-by: Jonathan Marek <jonathan@marek.ca>

freedreno: a2xx: insert scalar MOV to allow 2 source scalar

If we want to use a scalar instruction with two sources, both sources have
to be in the same register. This covers a common case by inserting a scalar
MOV into a previous instruction with only a vector alu instruction.

A better method would be to have the sources end up in the same register in
the first place, but when one source is a constant this is the only way.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>

freedreno: a2xx: NIR backend

This patch replaces the a2xx TGSI compiler with a NIR compiler.

It also adds several new features:
-gl_FrontFacing, gl_FragCoord, gl_PointCoord, gl_PointSize
-control flow (including loops)
-texture related features (LOD/bias, cubemaps)
-filling scalar ALU slot when possible

Signed-off-by: Jonathan Marek <jonathan@marek.ca>

nir: cleanup glsl_get_struct_field_offset, glsl_get_explicit_stride

Take away const qualifier from return type of these functions as
-Wignored-qualifiers points out it is ignored for these cases.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

travis: fix autotools build after --enable-autotools switch addition

Fixes: e68777c87ceed02ab199 "autotools: Deprecate the use of autotools"
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Acked-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

spirv: Update the JSON and headers from Khronos master

This corresponds to commit 79b6681aadcb53c27d1052e on GitHub.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: Mark deref UBO and SSBO access as non-scalar

Fixes: 63b9aa2e2574 "spirv: Add support for using derefs for..."
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir/spirv: handle ContractionOff execution mode

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir/vtn: add caps for some cl related capabilities

vtn supports these, so don't squalk if user is happy with enabling
these.

v2: add new members sorted

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

vtn: handle SpvExecutionModelKernel

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

mesa: add MESA_SHADER_KERNEL

used for CL kernels

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

anv/pipeline: Add a pdevice helper variable

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

relnotes: Add newly added Vulkan extensions

Both the Intel and RADV people have been really bad about adding things
to the release notes.  We should start actually paying attention.

Acked-by: Tapani Pälli <tapani.palli@intel.com>

anv: Only parse pImmutableSamplers if the descriptor has samplers

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>

radv: prevent dirtying of dynamic state when it does not change

DXVK often sets dynamic state without actually changing it.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radv: avoid context rolls when binding graphics pipelines

It's common in some applications to bind a new graphics pipeline without
ending up changing any context registers.

This has a pipline have two command buffers: one for setting context
registers and one for everything else. The context register command buffer
is only emitted if it differs from the previous pipeline's.

v2: ensure late scissor emission is done when radv_emit_rbplus_state() is
    called
v2: make use of cmd_buffer->state.workaround_scissor_bug
v3: rename "workaround_scissor_bug" to
    "context_roll_without_scissor_emitted"

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radv: add missed situations for scissor bug workaround

v2: rename "workaround_scissor_bug" to
    "context_roll_without_scissor_emitted"

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

radv: pass radv_draw_info to radv_emit_draw_registers()

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

freedreno: a2xx: sysmem rendering

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@gmail.com>

freedreno: a2xx: fix non-zero texture base offsets

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@gmail.com>

freedreno: a2xx: fix VERTEX_REUSE/DEALLOC on a20x

On a20x, set VGT_VERTEX_REUSE_BLOCK_CNTL to 2 and don't change it. Small
rearrangement on a220 to reduce the size of draw commands.

Only set DEALLOC_CNTL on a20x because the correct a220 value is not known.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@gmail.com>

freedreno: a2xx: fix gmem2mem viewport

Fixes cases where previous viewport values might case gmem2mem to fail.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@gmail.com>

freedreno: a2xx: cleanup REG_A2XX_PA_CL_VTE_CNTL

Doesn't change much, but reduces the size of fd2_emit_state

gmem2mem does not need to change the value: no Z clipping on resolve
mem2gmem now needs to restore the common value after rendering

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@gmail.com>

freedreno: a2xx: cleanup init_shader_const

Only 3 vertices are used so we can drop the data for vertex 4

It doesn't make sense to have 1.1 for some coordinates, use 1.0 instead

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@gmail.com>

nir: add bit_size parameter to system values with multiple allowed bit sizes

v2: add assert to verify we have at least one valid bit_size
v3: fix use of load_front_face in nir_lower_two_sided_color and tgsi_to_nir

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: add legal bit_sizes to intrinsics

With OpenCL some system values match the address bits, but in GLSL we also
have some system values being 64 bit like subgroup masks.

With this it is possible to adjust the builder functions so that depending
on the bit_sizes the correct bit_size is used or an additional argument is
added in case of multiple possible values.

v2: validate dest bit_size
v3: generate hex values in python code
    remove useless imports
    rename and move bit_sizes
v4: add 1 to legal bit_sizes for front_face

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir/validate: allow to check against a bitmask of bit_sizes

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: replace more nir_load_system_value calls with builder functions

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

glsl/lower_output_reads: set invariant and precise flags on temporaries

fixes a couple of deqp tests (on nvc0 and potential other drivers):
dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_1
dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_2
dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_3
dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_1
dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_2
dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_3
dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_1
dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_2
dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_3

CC: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

nv50,nvc0: add missing CAPs for unsupported features

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

nir/spirv: handle SpvStorageClassCrossWorkgroup

v2: rename nir_var_global to nir_var_mem_global

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: rename nir_var_shared to nir_var_mem_shared

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: rename nir_var_ssbo to nir_var_mem_ssbo

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: rename nir_var_ubo to nir_var_mem_ubo

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: rename nir_var_function to nir_var_function_temp

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

nir: rename nir_var_private to nir_var_shader_temp

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

intel/genxml: add missing MI_PREDICATE compare operations

Doesn't save us a great deal of lines but at least they get decoded in
aubinators.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Rafael Antognolli <rafael.antognolli@intel.com>

anv: document cache flushes & invalidations

A little bit of explanation regarding how vkCmdPipelineBarrier()
works.

v2: Avoid referring to data port cache when it's actually sampler
    caches (Jason)
    Complete explanation for indirect draws (Jason)

v3: s/samplers/sampler/ (Jason)
    s/UBOs/data port/
    Add documentation for VK_ACCESS_CONDITIONAL_RENDERING_READ_BIT_EXT (Lionel)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Eric Engestrom <eric.engestrom@intel.com> (v1)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> (v2)

anv: narrow flushing of the render target to buffer writes

In commit 9a7b3199037ac4 ("anv/query: flush render target before
copying results") we tracked all the render target writes to apply a
flushes in the vkCopyQueryResults(). But we can narrow this down to
only when we write a buffer (which is the only input of
vkCopyQueryResults).

v2: Drop newer render target write flags introduce by 1952fd8d2ce905
    ("anv: Implement VK_EXT_conditional_rendering for gen 7.5+")

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> (v1)