aco: fix underestimated pressure in spiller when a phi has a killed def

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5805>

panfrost: Use Midgard-specific reloads

v2: Be more explicit about sampler types. Prefer the term "load" to
"resolve" to match VK convention. Generate shaders for MRT 8x.  Blit
shader generation adds about 6ms to startup cost. We could cache thes.
shaders to disk if we needed to (or indeed, ship binaries).

v3: Fallback on u_blitter on Bifrost so Bifrost continues to work.
KHR_partial_update support is mostly no-oped on Bifrost now, but that's
okay for now - compositors are still functional.

v4: Specialize on multisample state as well to enable reloads of MSAA
textures. This requires 2x the shader variants, so I assume we're up to
12ms startup cost for generation. Annoying. Also fix interactions with
depth- or stencil-only clears of combined depth-stencil surfaces.

v5: Cache to the device (screen) instead of the context, reducing
duplicated work in apps that create many contexts (e.g. Chromium)

v6: Squash in KHR_partial_update cleanup to fix intermediate
regressions on a few tests.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5824>

st/mesa: use fragment shader to copy stencil buffer

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5836>

mesa/main: use p_atomic_inc_return instead of locking

There's no good reason for using a mutex here, as we have a simpler
primitive; atomic integers. So let's use that instead, to simplify
things a bit.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5901>

st/mesa: fix corrupted texture levels, when adding more levels than expected

Some of existing texture levels can be corruted,
after calling 'glTexImage' with param 'level' higher than
max expected value 'floor(log2(max(width, height, depth)))'.

To fix we prevent overwriting image buffer pointer
in 'st_texture_object', if it was already allocated
for multiple mip-levels storage.

Fixes piglit test: 'arb_copy_image add-illegal-levels'

Signed-off-by: Yevhenii Kharchenko <yevhenii.kharchenko@globallogic.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5785>

gallium/util: do not use _MTX_INITIALIZER_NP on Windows

We already have another way of initializing these, so it's just a matter
of avoiding _MTX_INITIALIZER_NP here.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5902>

radv/winsys: pass the buffer list via the CS ioctl for less CPU overhead

The legacy path requires one more ioctl to create the buffer
list and this is more costly for the CPU.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5899>

radv/winsys: replace alloca() by malloc() everywhere

To remove the mix of alloca() and malloc().

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5899>

v3d/compiler: Lower geometry output store base into offset src

When generating the VPM write instruction for geometry shader outputs,
emit_store_output_gs ends up adding the base and offset arguments
together with an ADD instruction. The addition was done at the VIR level
after scheduling so it always ends up right next to the corresponding
stvpm instruction. Most of the time the offset is constant but nothing
does any constant folding at the VIR level.

This patch makes it instead fold the addition into the offset at the NIR
level in v3d_nir_lower_io so that the NIR-level constant folding can get
rid of the addition most of the time.

v2: Use nir_iadd_imm to simplify the code. (Eric Anholt)

Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5825>

radeonsi: prevent a gfx10_ngg_calculate_subgroup_info failure for TES+NGG GS

arb_tessellation_shader-tes-gs-max-output -small -scan 1 50 -auto -fbo
doesn't pass, but at least all shaders are compiled successfully.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5700>

ci: bump piglit checkout for dsa tests

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5896>

mesa: change dsa texture error codes for GL 4.6

GL 4.6 changed error code for when the effective target of the
texture is illegal. Since it's not an illegal enum they modified
it to be an illegal operation.

However the CTS test for this is missing support for two cases,
I'm chasing that up, but I expect this will cause a CTS regression
for anyone who runs this test. I'm leaning on the side of being
compliant rather than passing the test until the test is fixed.

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5896>

panfrost: Extract panfrost_batch_reserve_framebuffer

We need to trigger it explicitly for reloads without draws (for Z^S
reload which is an edge case).

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Track surfaces drawn per-batch

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Set zs_samples as necessary

Fixes MSAA Z/S.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Handle per-sample shading

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Add rectangle subtraction algorithm

For better supporting KHR_partial_update.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Identify zs_samples field

For MSAA depth/stencil.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Include sample count in payload estimates

Otherwise we might not reserve enough space.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Add MALI_PER_SAMPLE bit

For gl_SampleID reads.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Expose panfrost_get_blend_shader

It is needed to produce a blend shader for blits.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

panfrost: Force Z/S writeback

This is unfortunately necessary for conformance at this stage.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

pan/mdg: Implement gl_SampleID

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

pan/mdg: Identify per-sample interpolation mode

So this is what .interp0 was this whole time.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

pan/mdg: Bump compiler RT maximum

We don't actually support MRT 8x yet but we would like to soon, so bump
it in the compiler.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5929>

Android: Fixes for Q and R

Fix Android-Q build:
- Use AOSP prebuilt bison by specifying $(BISON) variable
- Use AOSP prebuilt flex by specifying $(LEX) variable

Fix Android-R build:
- Add M4 environmet variable for Android R and higher (See [1])

[1] - https://cs.android.com/android/_/android/platform/build/+/2bfffb9f48a78de12faf5da77424c0cecb70d6eb:Changes.md;dlc=997661002af1282d938e88c3c723037e42e5d283

Signed-off-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Tested-by: Mauro Rossi <issor.oruam@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5894>

llvmpipe/format: fix snorm conversion

This fixes:
GTF-GL45.gtf33.GL3Tests.vertex_type_2_10_10_10_rev.vertex_type_2_10_10_10_rev_conversion
and
piglit
spec/arb_texture_view/rendering-formats/clear gl_rgba8_snorm as gl_r32f

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5820>

gallivm/sample: always square rho before fast log2

The fast log2 works better if rho is squared, i.e.

fast_log2(sqrt(2)) == 0.4
0.5 * fast_log2(2) == 0.5

so just square rho, and always divide by 2 afterwards.

Fixes:
GTF-GL45.gtf30.GL3Tests.sgis_texture_lod.sgis_texture_lod_basic_lod_selection

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5820>

nv50_2d: regenerate envytools-based rnndb headers

The headers hadn't been regenerated from envytools in a long time,
and there were a few minor divergences.

Based on envytools commit c20929ed0f3be18b8419f7332ee22d905feb6589

Among other things, rnndb has changed naming to G80/etc, for now
I've not tackled switching that over and replaced the nvidia
codenames back to the chip ids that mesa uses with the following:

  $ sed -i 's/G80_2D/NV50_2D/g' rnndb/graph/g80_2d.xml.h
  $ sed -i 's/GF100_2D/NVC0_2D/g' rnndb/graph/g80_2d.xml.h

No other modifications of the headergen'd headers was done, which
was helped by the differing #define's being unutilised presently.

Signed-off-by: Rhys Kidd <rhyskidd@gmail.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5920>

radv: destroy the base object if VkCreateInstance() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkAllocateCommandBuffers() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateFence() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateSemaphore() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateEvent() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateBuffer() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateImage() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateRenderPass*() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

radv: destroy the base object if VkCreateQueryPool() failed

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5868>

mesa/st: release debug_output after destroying the context

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3230
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/2218
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5789>

mesa: add bool param to _mesa_free_context_data

The param controls whether _mesa_destroy_debug_output should be called or not.
No functional changes; this will be used by the next commit.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5789>

mesa: rename _mesa_free_errors_data

Use the _mesa_init_XXX / _mesa_destroy_XXX pattern to clearly associate
the 2 functions.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5789>

gitlab-ci: Fix "triggered by Marge for a merge request" rule

The commit below changed the rule such that it accidentally also applied
to the non-MR pipelines created by Marge, resulting in Marge triggering
twice as many jobs as necessary.

Fixes: 549b4a3dd4ab "gitlab-ci: Automatically run pipelines for Marge
                     Bot pre-merge only"

Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5898>

intel/ehl: Add new PCI-IDs

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

intel/ehl: Rename gen_device_info struct

Renaming makes it easier to relate a pciid with device configuration.

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

intel/ehl: Use macro GEN11_LP_FEATURES in device info

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

intel/ehl: Use GEN11_URB_MIN_MAX_ENTRIES in device info

Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

panfrost: Dual source blend support

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5620>

pan/mdg: Skip z/s combining for dual-source writes

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5620>

pan/mdg: Dual source blend input/writeout support

We write to r2, which is preseved through to the blend shader, from
where it is read. We won't worry about MRT to keep things simple.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5620>

pan/mdg: Add a nir pass to reorder store_output intrinsics

Real writeout stores, which break execution, need to be moved to after
dual-source stores, which are just standard register writes.

v2: Don't move stores forward, to avoid moving them to before where
    their source is written.

v3: Only reorder past dual-source stores.

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5620>

gallium: Dual source support in blend_factor_to_shader

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5620>

compiler: Add dual-source factors to blend_factor

Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5620>

freedreno/ir3: DCE unused arrays

Letting unused arrays stick around confuses RA, which assigns vreg names
to the unused arrays, but then does not precolor them (because they are
unused).  This leads to an assert in ra_select_reg_merged():

  skqp: ../src/freedreno/ir3/ir3_ra.c:589: name_to_instr: Assertion '!name_is_array(ctx, name)' failed.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3262
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5907>

freedreno/a6xx: don't enable early-z/lrz if no z-test

But if shader explicitly asks for early-fragment-tests, obey it's
wishes.

Fixes a handful of skia (skqp) CTS fails (9.0_r12)

* gles_bug593049
* gles_circular_arcs_fill
* gles_circular_arcs_stroke_and_fill_square
* gles_clippedcubic2
* gles_complexclip2_path_bw
* gles_complexclip2_rrect_bw
* gles_complexclip3_complex
* gles_complexclip3_simple
* gles_crbug_691386
* gles_cubicclosepath
* gles_cubicpath
* gles_degeneratesegments
* gles_filltypespersp
* gles_innershapes_bw
* gles_inverse_paths
* gles_mixedtextblobs
* gles_onebadarc
* gles_quadclosepath
* gles_quadpath
* gles_rrect_clip_bw
* gles_scale-pixels
* gles_scaledstrokes
* gles_squarehair
* gles_strokes_zoomed
* gles_windowrectangles

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5907>

freedreno/ir3/ra: be better at failing

It doesn't happen much.  But it's annoying when we hit an impossible
condition deep in RA 90% thru a long test run.  Add some ra_assert()/
ra_unreachable() helper macros so we can bail cleanly and fail RA.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5907>

freedreno/a6xx: bail instead of crash for compile fails

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5907>

freedreno/ir3: make compile fails more visible

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5907>

freedreno/ir3: add missing VS driver params

Some of these only used by turnip so far, this is just for clarity.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5907>

zink: enable tgsi texcoord pipe cap

this requires some modifications to the ntv slot remapping, as we definitely
don't want to reserve another 25% of the available slots for the (deprecated)
texcoord varyings

now we remap VARYING_SLOT_TEX[n] to the last available slots, which lets us avoid
needing to do permanent reservation, and we check to make sure that we haven't
seen the corresponding texcoord varying any time we emit a non-texcoord varying in
that slot

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5551>

nouveau: expose HMM

v2: moved caps

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Pierre Moreau <dev@pmoreau.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5906>

nvc0: support PIPE_CAP_RESOURCE_FROM_USER_MEMORY_COMPUTE_ONLY

v2: rework by adding a new buffer status

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5906>

gallium: add PIPE_CAP_RESOURCE_FROM_USER_MEMORY_COMPUTE_ONLY

With the current UAPI we only support user pointers from the compute
engines, so we need a way to express that in gallium.

v2: fix typos
v3: add allows_user_pointers helper

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Pierre Moreau <dev@pmoreau.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5906>

nouveau: enable HMM

v2: move declarations into libdrm
v3: fix typos
    rework handling of how much memory to reserve
v4: remove unused parameter
    unmap cutout on error and when the screen is destroyed
v5: move into screen_create
    enable HMM only if CL gets enabled

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5906>

ci: bump libdrm to 2.4.102

Since version 2.4.101 there are only xz archives hence the bz2 to xz
change.

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5906>

ci: need to install wget in order to download libdrm

Fixes: dcd171f5e9b ("gitlab-ci: More stable URL for kernel and ramdisk artifacts, for LAVA")
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5906>

nir: Support load/store of temps as scratch in nir_lower_explicit_io

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5889>

nir: When nir_lower_vars_to_explicit_types is run on temps, update scratch_size

To allow interop with other scratch ops, append any remaining temp vars
to the end of any already-allocated scratch space.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5889>

nir_lower_io: Add addr_format_is_offset helper

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5889>

freedreno/regs: update a6xx PC regs

Update some registers in the 0x9800-0xa000 range.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5870>

freedreno/regs: update a6xx VPC regs

Update some registers in the 0x9000-0x95ff range.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5870>

freedreno/regs: update a6xx RB regs

Update some registers in the 0x8c00-0x8dff range.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5870>

freedreno/regs: update a6xx GRAS registers

Update some registers in the 0x8000-0x87ff range.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5870>

gitlab-ci: re-enable arm64_a630_vk

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5904>

turnip: remove use of tu_cs_entry for draw states

The tu_cs_entry struct doesn't match well what we want for SET_DRAW_STATE
and CP_INDIRECT_BUFFER (requires extra steps to get iova and size), so
start phasing it out.

Additionally, use newly added tu_cs_draw_state where it doesn't require any
effort (it requires a fixed size, but gets rid of the extra end_sub_stream)

Note this also changes the behavior of CmdBindDescriptorSets for compute to
emit directly in cmd->cs instead of doing through a CP_INDIRECT.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5558>

turnip: fix inconsistencies with tu6_load_state_size

The next patch assumes the correct size is returned in tu6_emit_load_state.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5558>

turnip: emit compute pipeline directly in CmdBindPipeline

There's no need to defer it, and can get rid DIRTY_COMPUTE_PIPELINE.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5558>

turnip: use DIRTY SDS bit to avoid making copies of pipeline load state ib

Some testing showed that the DIRTY bit has the desired behavior, so use it
to make things a bit simpler.

Note in CmdBindPipeline, having the TU_CMD_DIRTY_DESCRIPTOR_SETS behind a
if(pipeline->layout->dynamic_offset_count) was wrong.

Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5558>

zink: try to handle multisampled null buffers

I don't have any tests for this that I've run into yet, so this is mostly
just guessing

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5686>

zink: handle empty attachments

create an empty buffer and surface to reuse for the fb attachment here

this fixes most of the arb_framebuffer_object tests in piglit

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5686>

pan/mdg: Fix non-debug compiliation

Fixes error when the assert is optimized out:

   ../src/panfrost/midgard/midgard_compile.c: In function ‘output_load_rt_addr’:
   ../src/panfrost/midgard/midgard_compile.c:1644:1: error: control reaches end of non-void function [-Werror=return-type]
    }

Closes #3270

Fixes: 7781d2c2ead ("pan/mdg: Support MRT in output load lowering")
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5895>

Revert https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4580

It broke the CI pipeline on master:

https://gitlab.freedesktop.org/mesa/mesa/-/jobs/3604314
https://gitlab.freedesktop.org/mesa/mesa/-/jobs/3604315

Revert for now, to allow other MRs to be merged.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5905>

nouveau: expose HMM

v2: moved caps

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Pierre Moreau <dev@pmoreau.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4580>

nvc0: support PIPE_CAP_RESOURCE_FROM_USER_MEMORY_COMPUTE_ONLY

v2: rework by adding a new buffer status

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4580>

gallium: add PIPE_CAP_RESOURCE_FROM_USER_MEMORY_COMPUTE_ONLY

With the current UAPI we only support user pointers from the compute
engines, so we need a way to express that in gallium.

v2: fix typos
v3: add allows_user_pointers helper

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Pierre Moreau <dev@pmoreau.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4580>

nouveau: enable HMM

v2: move declarations into libdrm
v3: fix typos
    rework handling of how much memory to reserve
v4: remove unused parameter
    unmap cutout on error and when the screen is destroyed
v5: move into screen_create
    enable HMM only if CL gets enabled

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4580>

ci: bump libdrm to 2.4.102

Since version 2.4.101 there are only xz archives hence the bz2 to xz
change.

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4580>

zink: block resolve blits for depth/stencil buffers

"The format features of dstImage must contain VK_FORMAT_FEATURE_COLOR_ATTACHMENT_BIT"
- vkCmdResolveImage spec

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5888>

zink: block vkCmdBlitImage usage for multi sampled blits

this is prohibited by spec

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5888>

zink: try copy_region hook for blits where we can't do a regular blit or resolve

in cases where the formats match, we can likely just pass this through for now

fixes a ton of spec@!opengl 1.1@depthstencil-default_fb-blit piglit tests

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5888>

mesa/main: use call_once instead of open-coding

We already have a utility for this, so let's use that instead.

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5879>

mesa/main: factor out one-time-init into a helper

This will make the next commit a bit cleaner.

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5879>

gv100/ir: set ftz bit on floating point operations

Fixes Unigine Heavens ambient occlusion

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5881>

meson: fix android vulkan build

Android doesn't have `pthread_cancel()` and is unlikely to ever
implement it [1], but `wsi_common_display.c` needs it (or an
alternative).

Let's just disable the platform on Android (as it used to be
before 448eb19158f483d807ef).

[1] https://android-review.googlesource.com/c/platform/bionic/+/1215779/1/docs/status.md

Fixes: 448eb19158f483d807ef ("vulkan: automatically compile the `display` platform when available")
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Acked-by: Nataraj Deshpande <nataraj.deshpande@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5860>

gitlab-ci: Drop dependencies:

Artifacts from jobs listed in needs: are downloaded by default, so
there's no need to list them in dependencies: as well (in fact,
https://docs.gitlab.com/ce/ci/yaml/#artifact-downloads-with-needs says
using dependencies: together with needs: is invalid, so we might have
been getting lucky...).

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5845>

gitlab-ci: Remove indirect dependencies from needs:

Tomeu discovered that GitLab 12.8 fixed the bug where jobs would
spuriously run even though some of their dependency jobs were skipped.
So we don't need to list indirect dependencies anymore.

Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5845>

tu: Don't invalidate irrelevant state when changing pipeline

At least in the future this could let us avoid re-emitting gfx/cs
constants when the other changes. This also matches what the blob does.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5877>

freedreno/a6xx: Add some documentation for shared consts

I'm not convinced we'll actually want to use this, and there may be
another enable bit in SP_UNKNOWN_AB00, but it's nice to at least write
this down in case we want to try using it in the future.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5877>

freedreno/a6xx: Rename and document HLSQ_UPDATE_CNTL

It turns out that this clears CP_LOAD_STATE6 packets, including
disabling any pending loads for SS6_INDIRECT/SS6_BINDLESS (these loads
don't actually happen until the draw itself, and I'm not sure if they
happen if the state is unused by the shader) and marking constants and
UBO descriptors loaded with SS6_DIRECT as invalid. It's used very
differently from HLSQ_UPDATE_CNTL on a4xx from whence the name came, and
unlike on a4xx it's not readable, so this probably doesn't line up with
HLSQ_UPDATE_CNTL on a4xx.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5877>

clover: implements clEnqueueFillBuffer

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5897>

clover: add more cl_mem_object_type to pipe_texture_target mapping

It avoid unnecessary CL_INVALID_VALUE return from clGetSupportedImageFormats

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5897>