| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The num_components value passed into get_mul_for_src is used to only
compose the parts of the swizzle that we know will be used so we don't
compose invalid swizzle components. However, we had a bug where we
passed the number of components of the add all the way through. For the
given source, we need the number of components read from that source.
In the case where we have a narrow add, say 2 components, that is
sourced from a chain of wider instructions, we may not compose all the
swizzles. All we really need to do is pass through the right number of
components at each level.
Fixes: 2231cf0ba3a "nir: Fix output swizzle in get_mul_for_src"
Reviewed-by: Ian Romanick <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GL_ARB_gl_spirv does not provide a sampler deref for e.g. texelFetch(),
so we can't assume that both are present and identical. Simply lower
each if it is present.
Fixes regressions in GL_ARB_gl_spirv tests since I switched everyone to
using this pass. Thanks to Alejandro Piñeiro for catching these.
Fixes: f003859f97c nir: Make gl_nir_lower_samplers use gl_nir_lower_samplers_as_deref
Reviewed-by: Jason Ekstrand <[email protected]>
Tested-by: Alejandro Piñeiro <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Eric Engestrom <[email protected]>
Acked-by: Emil Velikov <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Eric Engestrom <[email protected]>
Acked-by: Emil Velikov <[email protected]>
|
|
|
|
|
| |
Fixes: f6aa9f718516 'anv/pipeline_cache: Add support for caching NIR'
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
|
|
|
|
|
| |
This will allow drivers to pin shader buffers if necessary.
i965 and anv do not need to do this today, but iris will.
Reviewed-by: Jordan Justen <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, BLORP expects drivers to provide two functions for dealing
with buffers: blorp_emit_reloc and blorp_surface_reloc. Both record a
relocation and combine the BO address and offset into a full 64-bit
address. Traditionally, blorp_surface_reloc has written that combined
address to an implicitly-known buffer where surface states are stored.
(In contrast, blorp_emit_reloc returns the value.)
The upcoming Iris driver stores surface states in multiple buffers,
which makes it impossible for blorp_surface_reloc to write the combined
address - it only takes an offset, not the actual buffer to write to.
This commit adds a third function, blorp_get_surface_address, which
combines and returns an address, which is then passed to ISL's surface
state fill functions. Softpin-only drivers can return a real address
here and skip writing it in blorp_surface_reloc. Relocation-based
drivers are have options. They can simply return 0 from the new
function, and continue writing the address from blorp_surface_reloc.
Or, they can return a presumed address from blorp_get_surface_address,
and have other relocation processing write the real value later.
For now, i965 and anv simply return 0.
Reviewed-by: Jordan Justen <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Make sure that the next line starts with spaces so that bullets are
maintained throughout, add `` around a few more special tokens, and fix
SAMPLE_COUNT_TEXTURE -> SAMPLE_COUNT.
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In all GLSL ES versions output variables in fragment shader are allowed
to be invariant.
From Section 4.6.1 ("The Invariant Qualifier") GLSL ES 1.00 spec:
"Only the following variables may be declared as invariant:
...
- Built-in special variables output from the fragment shader."
From Section 4.6.1 ("The Invariant Qualifier") GLSL ES 3.00 spec:
"Only variables output from a shader can be candidates for invariance."
Signed-off-by: Danylo Piliaiev <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107842
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a second level of caching for the pre-lowered NIR that's only
based off of the shader module, entrypoint and specialization constants.
This is enough for spirv_to_nir as well as our first round of lowering
and optimization. Caching at this level should allow for faster shader
recompiles due to state changes.
The NIR caching does not get serialized to disk via either the
VkPipelineCache serialization mechanism or the transparent on-disk
cache. We could but it's usually not that expensive to fall back to
SPIR-V for the odd cache miss especially if it only happens once for
several misses and it simplifies the cache.
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
| |
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
|
|
|
| |
The stuff hashed by anv_pipeline_hash_shader is exactly the inputs to
anv_shader_compile_to_nir so it can be used for NIR caching.
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
|
|
|
| |
They have glsl_sampler_dim enum values of 8 and 9 which don't work when
you & them with 0x7. Fortunately, we have plenty of bits.
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
|
|
|
| |
This lets us make anv_pipeline_compile_to_nir take a device instead of a
pipeline.
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
| |
Brown bag fix...
|
|
|
|
|
|
|
|
|
|
|
|
| |
Thanks to the new NIR load_descriptor intrinsic added by the UBO/SSBO
lowering series, we weren't getting UBO pushing because the UBO range
detection pass couldn't see the constants it needed. This fixes that
problem with a quick round of constant folding. Because we're folding
we no longer need to go out of our way to generate constants when we
lower the vulkan_resource_index intrinsic and we can make it a bit
simpler.
Reviewed-by: Lionel Landwerlin <[email protected]>
|
|
|
|
|
|
|
|
| |
The last round of fixing 3d layer+level layout skipped the tiled case,
since tiled texture support was not in place yet. This finishes the
job.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
This is known when the CSO is created, so no need to patch it in later.
Also, it seems like smaller textures where the first level is small
enough to be linear, it seems like we should set linear tile mode.
See: dEQP-GLES3.functional.texture.format.unsized.rgb_unsigned_byte_3d_pot
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously we'd use format/etc from the primary (z32) buffer for the
stencil (s8), due to confusion about rsc vs psurf. Rework this to drop
extra arg and push down handling of separate stencil case (and make sure
we take the fmt from the right place).
This doesn't completely fix separate-stencil, but at least it avoids the
GPU scribbling over random other cmdstream buffers and causing a bunch
of bogus fails in dEQP.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If nothing else, this will make problems with cmdstream getting blit
over with pixels easier to track down (ie. faults when it first happens
rather than strange failures later from corrupted cmdstream when a
stateobj is later reused).
(NOTE this somewhat depends on the kernel supporting the flag, and the
iommu implementation. But the worst case is just that the cmdstream
ends up writeable as before.)
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Guido Günther <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
|
|
|
|
| |
Cc: Bas Nieuwenhuizen <[email protected]>
Signed-off-by: Eric Engestrom <[email protected]>
Reviewed-by: Samuel Pitoiset <[email protected]> (v1)
|
|
|
|
|
|
|
|
| |
Fixes: 393b59e0772e7bf0426bdf61c740752c4e09dde1
('nir: Rework nir_lower_constant_initializers() to handle functions')
Reviewed-by: Tapani Pälli <[email protected]>
Reviewed-by: Erik Faye-Lund <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
The check for location aliasing was always asuming output variables
but this validation is also called for input variables.
Fixes: e2abb75b0e4 ("glsl/linker: validate explicit locations for SSO programs")
Cc: Iago Toral Quiroga <[email protected]>
Signed-off-by: Andres Gomez <[email protected]>
Reviewed-by: Timothy Arceri <[email protected]>
Reviewed-by: Iago Toral Quiroga <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Patch moves intel_tiled_memcpy[_sse41] libraries to isl, renames some
functions and types and makes the required build system changes for
meson, automake and Android. No functional changes are introduced.
v2: code cleanups, move isl_get_memcpy_type to i965 (Jason)
v3: move isl_mem_copy_fn to priv header, cleanups (Jason, Dylan)
Signed-off-by: Tapani Pälli <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
Reviewed-by: Dylan Baker <[email protected]>
Acked-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
| |
Now that we have software implementations of ARB_gpu_shader_int64 and
ARB_gpu_shader_fp64 we can unconditionally enable these extensions.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
| |
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
| |
Shaders containing software implementations of double-precision
operations can be very large such that we cannot stack-allocate
an array of grf_count*16.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
Shaders containing software implementations of double-precision
operations can be very large such that we have more the 2^16 virtual
registers during optimization.
Move the 'nr' field to the union containing the immediate storage and
expand it to 32-bits.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
The next patch replaces an unsigned bitfield with a plain unsigned,
which triggers gcc to begin warning on signed/unsigned comparisons.
Keeping this patch separate from the actual move allows bisectablity and
generates no additional warnings temporarily.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
| |
A follow on commit will move nr to the same union as the immediate
data, so we should assert these invariants before we overwrite the nr
field.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
A follow on patch will move the 'nr' field to the union containing the
immediate field, so prepare by checking that we're only testing these
assertions if the .file is correct.
The assertions with != ARF were kind of silly to begin with because the
<128 check is specifically only for things in the GRF.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
| |
Reviewed-by: Kenneth Graunke <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
NIR metadata validation verifies that the debug bit was unset (by a call
to nir_metadata_preserve) if a NIR optimization pass made progress on
the shader. With the expectation that the NIR shader consists of only a
single main function, it has been safe to call nir_metadata_preserve()
iff progress was made.
However, most optimization passes calculate progress per-function and
then return the union of those calculations. In the case that an
optimization pass makes progress only on a subset of the functions in
the shader metadata validation will detect the debug bit is still set on
any unchanged functions resulting in a failed assertion.
This patch offers a quick solution (short of a larger scale refactoring
which I do not wish to undertake as part of this series) that simply
unsets the debug bit on unchanged functions.
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
| |
We're going to use it more in a future patch, and this avoids a lot of
gross code.
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
| |
Will be used to communicate that a shader uses 64-bit operations to the
concerned lowering passes.
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
[mattst88]: Found in an old branch of Jason's.
Jason implemented: inot, iand, ior, iadd, isub, ineg, iabs, compare,
imin, imax, umin, umax
Matt implemented: ixor, bcsel, b2i, i2b, i2i8, i2i16, i2i32, i2i64,
u2u8, u2u16, u2u32, u2u64, and fixed ilt
Reviewed-by: Elie Tournier <[email protected]>
|
|
|
|
|
|
|
|
| |
We're going to have multiple functions, so nir_shader_get_entrypoint()
needs to do something a little smarter.
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
| |
Previously it assumed that only a single function (the entrypoint)
existed and attempted to lower constant initializers of shader outputs
for each function, for instance.
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Elie Tournier <[email protected]>
Signed-off-by: Sagar Ghuge <[email protected]>
|