| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
| |
This is based on the code from the GLSL IR pass however unlike the GLSL IR
pass it also supports arrays of arrays.
As well as implementing the logic from the GLSL IR pass we add some
additional intrinsic cases to catch more system values.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
| |
I want this function for nir_gather_info(), and realized it's basically
the same as the ones in nir_lower_io().
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is ported from GLSL and converts
if (cond)
discard;
into
discard_if(cond);
This removes a block, but also is needed by radv
to workaround a bug in the LLVM backend.
v2: handle if (a) discard_if(b) (nha)
cleanup and drop pointless loop (Matt)
make sure there are no dependent phis (Eric)
v3: make sure only one instruction in the then block.
v4: remove sneaky tabs, add cursor init (Eric)
Reviewed-by: Eric Anholt <[email protected]>
Cc: "13.0" <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
| |
We were leaving an undefined value since the ralloc zeroing changes.
Fixes nir_validate() failures on vc4.
v2: Fix the color-index case of drawpixels as well.
Reviewed-by: Rob Clark <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Assuming the hardware is set up to use a screen coordinate system
flipped vertically with respect to the GL's window coordinate system,
the SYSTEM_VALUE_SAMPLE_POS vector will also be flipped vertically
with respect to the value expected by the GL, so we need to give it
the same treatment as gl_FragCoord. Fixes the following CTS tests on
i965:
ES31-CTS.functional.shaders.multisample_interpolation.interpolate_at_offset.at_sample_position.default_framebuffer
ES31-CTS.functional.shaders.sample_variables.sample_pos.correctness.default_framebuffer
when run with any multisample configuration, e.g. rgba8888d24s8ms4.
Cc: <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Anuj Phogat <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
These were broken in e1af20f18a8 when the info field in nir_shader was
turned into a pointer.
Clone was copying the pointer rather than the data and nir_sweep was
cleaning up shader_info rather than claiming it.
Reviewed-by: Eric Anholt <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
No change in behavior. ralloc_size is equivalent to rzalloc_size.
That will change though.
Calls not switched to rzalloc_size:
- ralloc_vasprintf
- glsl_type::name allocation (it's filled with snprintf)
- C++ classes where valgrind didn't show uninitialized values
I switched most of non-glsl stuff to rzalloc without checking whether
it's really needed.
Reviewed-by: Edward O'Callaghan <[email protected]>
Tested-by: Edmondo Tommasina <[email protected]>
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Signed-off-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
As of 59864e8e020 we just use the location assigned by the front-end and
no longer need this for i965.
Since there were some issues in the logic with assigning arrays the same
driver location if they didn't start at the same location just remove it
and let other drivers implement a solution if needed when they add
ARB_enhanced_layouts support.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
When restoring something from shader cache we won't have and don't
want to create a nir_shader this change detaches the two.
There are other advantages such as being able to reuse the
shader info populated by GLSL IR.
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
| |
This will allow use to stop copying values between structs and
will also simplify handling handling these values in the shader cache.
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The previous power-of-two rules didn't catch idiv (because i965 doesn't
set lower_idiv) and imod cases. The udiv and umod cases should have
been caught, but I included them for orthogonality.
This fixes silly code observed from compute shaders with local_size_[xy]
= 1.
Signed-off-by: Ian Romanick <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98299
Reviewed-by: Jordan Justen <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Commit 2ed17d46de045404042f13c6591895a1cf31b167 changed
nir_loop_first_cf_node and friends to return a nir_block instead of a
nir_cf_node. This broke one of the NIR control flow tests.
Signed-off-by: Jason Ekstrand <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98128
|
|
|
|
|
|
|
|
| |
Now that the NIR casting functions have type assertions, we have a bunch of
assertions that aren't needed anymore.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
One of NIR's invariants is that control flow lists always start and end
with blocks. There's no good reason why we should return a cf_node from
these functions since we know that it's always a block. Making it a block
lets us remove a bunch of code.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
This makes calling nir_foo_as_bar a bit safer because we're no longer 100%
trusting in the caller to ensure that it's safe. The caller still needs to
do the right thing but this ensures that we catch invalid casts with an
assert rather than by reading garbage data. The one downside is that we do
use the casts a bit in nir_validate and it's not a validate_assert.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
glsl_print_type() prints arrays of arrays incorrectly. For example,
a type with name float[3][7] would be printed as float[7][3]. (This
is an array of length 3 containing arrays of 7 floats.) cdecl says
that the type name is correct.
glsl_print_type() doesn't really do anything above and beyond printing
type->name, and glsl_print_struct() wasn't used at all. So, drop them.
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
|
| |
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
Tested-by: Mark Janes <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
v2: Delete some stray debug code notice by Iago.
v3: Massive rebase on new ir_function_signature::intrinsic_id mechanism.
Signed-off-by: Ian Romanick <[email protected]>
Reviewed-by: Iago Toral Quiroga <[email protected]> [v1]
Acked-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Otherwise grepping for where atomic_counter_inc and friends are defined
is a very frustrating experience.
Signed-off-by: Ian Romanick <[email protected]>
Reviewed-by: Iago Toral Quiroga <[email protected]>
Acked-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
This intrinsic has no destination, no sources, no variables, and can be
eliminated. In other words, it does nothing and will always get deleted by
dead code elimination. However, it does provide a quick-and-easy way to
temporarily tag a particular location in a NIR shader.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
| |
I found this in a shader that was doing an alpha test when alpha is fixed
at 1.0.
v2: Rebase on master (now the const value is "u32" not "u").
Reviewed-by: Jason Ekstrand <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
VC4 was running into a major performance regression from enabling control
flow in the glmark2 conditionals test, because of short if statements
containing an ffract.
This pass seems like it was was trying to ensure that we only flattened
IFs that should be entirely a win by guaranteeing that there would be
fewer bcsels than there were MOVs otherwise. However, if the number of
ALU ops is small, we can avoid the overhead of branching (which itself
costs cycles) and still get a win, even if it means moving real
instructions out of the THEN/ELSE blocks.
For now, just turn on aggressive flattening on vc4. i965 will need some
tuning to avoid regressions. It does looks like this may be useful to
replace freedreno code.
Improves glmark2 -b conditionals:fragment-steps=5:vertex-steps=0 from 47
fps to 95 fps on vc4.
vc4 shader-db:
total instructions in shared programs: 101282 -> 99543 (-1.72%)
instructions in affected programs: 17365 -> 15626 (-10.01%)
total uniforms in shared programs: 31295 -> 31172 (-0.39%)
uniforms in affected programs: 3580 -> 3457 (-3.44%)
total estimated cycles in shared programs: 225182 -> 223746 (-0.64%)
estimated cycles in affected programs: 26085 -> 24649 (-5.51%)
v2: Update shader-db output.
Reviewed-by: Ian Romanick <[email protected]> (v1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
SPIR-V/Vulkan have a special image type for input attachments
called the subpass type. It has different characteristics than
other images types.
The main one being it can only be an input image to fragment
shaders and loads from it are relative to the frag coord.
This adds support for it to the GLSL types. Unfortunately
we've run out of space in the sampler dim in types, so we
need to use another bit.
v2: Fixup subpass input name (Jason)
Reviewed-by: Jason Ekstrand <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
| |
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Anuj Phogat <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Eric Anholt <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Eric Anholt <[email protected]>
|
|
|
|
|
|
|
|
| |
This is mandatory.
Cc: [email protected]
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Eric Anholt <[email protected]>
|
|
|
|
|
|
|
|
| |
Numeric 2 is actually GLSL_SAMPLER_DIM_3D, which I don't think is what
was intended.
Signed-off-by: Rob Clark <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
| |
I want to re-use this in a different pass, so move to nir.h
Signed-off-by: Rob Clark <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
| |
Turns out it already exists.. so don't duplicate it.
Signed-off-by: Rob Clark <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Unlike the current CSE pass, global value numbering is capable of detecting
common values even if one does not dominate the other. For instance, in
you have
if (...) {
ssa_1 = ssa_0 + 7;
/* use ssa_1 */
} else {
ssa_2 = ssa_0 + 7;
/* use ssa_2 */
}
Global value numbering doesn't care about dominance relationships so it
figures out that ssa_1 and ssa_2 are the same and converts this to
if (...) {
ssa_1 = ssa_0 + 7;
/* use ssa_1 */
} else {
/* use ssa_1 */
}
Obviously, we just broke SSA form which is bad. Global code motion,
however, will repair this for us by turning this into
ssa_1 = ssa_0 + 7;
if (...) {
/* use ssa_1 */
} else {
/* use ssa_1 */
}
This intended to eventually mostly replace CSE. However, conventional CSE
may still be useful because it's less of a scorched-earth approach and
doesn't require GCM. This makes it a bit more appropriate for use as a
clean-up in a late optimization run.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
| |
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Eric Anholt <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Found by inspection. Untested beyond compilation. This also matches the
logic used in nir_lower_alu_to_scalar.
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
Cc: [email protected]
|
|
|
|
|
|
|
|
|
|
|
| |
In aad4f1550, we removed the concept of "fake" edges from NIR. Now, if you
have a block at the end of an infinite loop it really has no predecessors.
This updates the unit tests to match.
Signed-off-by: Jason Ekstrand <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97587
Tested-by: Aaron Watry <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
| |
This was let over from aad4f15506c
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
| |
I accidentally added these with 0dc4cab. Oops!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In 144cbf8 ("nir: Make nir_opt_remove_phis see through moves."), Ken
made nir_opt_remove_phis able to coalesce phi nodes whose sources are
all moves with the same swizzle. However, he didn't add the logic
necessary for handling the fact that the phi may now have multiple
different sources, even though the sources point to the same thing. For
example, if we had something like:
if (...)
a1 = b.yx;
else
a2 = b.yx;
a = phi(a1, a2)
... = a
then we would rewrite it to
if (...)
a1 = b.yx;
else
a2 = b.yx;
... = a1
by picking a random phi source, which in this case is invalid because
the source doesn't dominate the phi. Instead, we need to change it to:
if (...)
a1 = b.yx;
else
a2 = b.yx;
a3 = b.yx;
... = a3;
Fixes 12 CTS tests:
ES31-CTS.functional.tessellation.invariance.outer_edge_symmetry.quads*
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
| |
And re-implement nir_after_cf_node_and_phis() using it.
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When NIR was first introduced, Connor added this fake-edge hack to work
around issues related to unreachable blocks. Thanks to GLSL IR's jump
lowering code, the only unreachable code you can have is a block after an
infinite loop. With SPIR-V, we didn't have the jump lowering code so we
could also end up with the "if (...) { break; } else { continue; }" case
which generates an unreachable block after the if. Because of this, most
of NIR had to be fixed up for handling unreachable blocks. The only
remaining case of not handling unreachable blocks was specifically the
block-after-infinite-loop case in dead_cf which was fixed by the previous
commit. We can now delete the fake edge hack.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
| |
vc4 is about to start using the shader info field to set up discard
handling.
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Jason suggested adding an assert(function->impl) here. All callers
of this function actually want ->impl, so I decided just to change
the API.
We also change the nir_lower_io_to_temporaries API here. All but one
caller passed nir_shader_get_entrypoint(), and with the previous commit,
it now uses a nir_function_impl internally. Folding this change in
avoids the need to change it and change it back.
v2: Fix one call I missed in ir3_compiler (caught by Eric).
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This changes the pass internals to work with a nir_function_impl
directly rather than a nir_function. The next patch will change
the API.
v2: Rebase after framebuffer fetch landed.
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
This requires emitting a series of copies at the top of the program
from each output variable to the corresponding temporary. The initial
copy can be skipped for non-framebuffer fetch outputs whose initial
value is undefined, and the final copy needs to be skipped for
read-only outputs (i.e. gl_LastFragData), since it would be illegal to
emit a store output intrinsic for it.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
| |
The NIR representation of framebuffer fetch is the same as the GLSL
IR's until interface variables are lowered away, at which point it
will be translated to load output intrinsics. The GLSL-to-NIR pass
just needs to copy the bits over to the NIR program.
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In some programs, we can have very deep dominance trees and the recursion
can cause us to risk stack overflows. Instead, we replace the recursion
with a pair of loops, one at the start and one at the end. This is
functionally equivalent to what we had before and it's actually a bit
easier to read in the new form without the recursion.
Signed-off-by: Jason Ekstrand <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97225
Reviewed-by: Connor Abbott <[email protected]>
Reviewed-by: Matt Turner <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Prior to this commit rename_variables_block() is recursively called,
performing a depth-first traversal of the control flow graph. The
function uses a non-trivial amount of stack space for local variables,
which puts us in danger of smashing the stack, given a sufficiently deep
dominance tree.
XCOM: Enemy Within contains a shader with such a dominance tree (1574
nir_blocks in total, depth of at least 143).
Jason tells me that he believes that any walk over the nir_blocks that
respects dominance is sufficient (a DFS might have been necessary prior
to the introduction of nir_phi_builder).
In fact, the introduction of nir_phi_builder made the problem worse:
rename_variables_block(), walks to the bottom of the dominance tree
before calling nir_phi_builder_value_get_block_def() which walks back to
the top of the dominance tree...
In any case, this patch ensures we avoid that problem as well.
Cc: [email protected]
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97225
Reviewed-by: Connor Abbott <[email protected]>
|
|
|
|
|
|
|
| |
Without this the following line will segfault and we don't get to
see the results of the validate_assert() above.
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
| |
Generally you'd see the gl_Color reference first and get some cursor set.
However, in piglit draw-pixel-with-texture we're now seeing the TexCoord
dereferenced first.
Reviewed-by: Rob Clark <[email protected]>
|