This reverts commit 11d09f2fe7.
Reason for revert: Failing roll of Tint to Dawn: https://dawn-review.googlesource.com/c/dawn/+/70100
Original change's description:
> HLSL: force FXC to never unroll loops
>
> Emit the "[loop]" attribute on "for" and "while" so that FXC does not
> attempt to unroll them. This is to work around an FXC bug where it fails
> to unroll loops with gradient operations.
>
> FXC ostensibly unrolls such loops because gradient operations require
> uniform control flow, and loops that have varying iterations may
> possibly not be uniform. Tint will eventually validate that control flow
> is indeed uniform, so forcing FXC to avoid unrolling in these cases
> should be fine.
>
> Bug: tint:1112
> Change-Id: I10077f8b62fbbb230a0003f3864c75a8fe0e1d18
> Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/69880
> Kokoro: Kokoro <noreply+kokoro@google.com>
> Reviewed-by: Ben Clayton <bclayton@google.com>
> Commit-Queue: Antonio Maiorano <amaiorano@google.com>
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: tint:1112
Change-Id: I8e8f3c0abfa6e6bc5d0e67af9428a46ef867d5c1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70540
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Emit the "[loop]" attribute on "for" and "while" so that FXC does not
attempt to unroll them. This is to work around an FXC bug where it fails
to unroll loops with gradient operations.
FXC ostensibly unrolls such loops because gradient operations require
uniform control flow, and loops that have varying iterations may
possibly not be uniform. Tint will eventually validate that control flow
is indeed uniform, so forcing FXC to avoid unrolling in these cases
should be fine.
Bug: tint:1112
Change-Id: I10077f8b62fbbb230a0003f3864c75a8fe0e1d18
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/69880
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Looks like a typo in the test cases I wrote.
Change-Id: Ieb4d8ce28827e47ab0baef7b1178395d97f90ace
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/69841
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
This removes a lot of awkward logic from the MSL writer, and means
that we now handle all module-scope variables with the same transform.
Change-Id: I782e36a4b88dafbc3f8364f7caa7f95c6ae3f5f1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/67643
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
These operators are not defined in the metal namespace when the vector
operands are packed.
Fixed: tint:1121
Change-Id: I2e8f4302e08117ca41bac6c05fb24a70d1215740
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/62480
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
MSL vectors with other widths already match WGSL's rules for alignment
and size.
Change-Id: I237052372463ea8323eab47c3b4ca90c6d8afcc3
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/62600
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
The refactored CanonicalizeEntryPointIO transform makes it much easier
to handle SPIR-V style IO as well, and doing this removes a lot of
duplicated code. Remove all of the SPIR-V transform code for shader IO
and vertex point size.
Bug: tint:920
Change-Id: Id1b97517619b4d2fd09b45d5aee848259f3dfa77
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/60840
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
Auto-Submit: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This is a major reworking of this transform. The old transform code
was getting unwieldy, with part of the complication coming from the
handling of multiple return statements. By generating a wrapper
function instead, we can avoid a lot of this complexity.
The original entry point function is stripped of all shader IO
attributes (as well as `stage` and `workgroup_size`), but the body is
left unmodified. A new entry point wrapper function is introduced
which calls the original function, packing/unpacking the shader inputs
as necessary, and propagates the result to the corresponding shader
outputs.
The new code has been refactored to use a state object with the
different parts of the transform split into separate functions, which
makes it much more manageable.
Fixed: tint:1076
Bug: tint:920
Change-Id: I3490a0ea7a3509a4e198ce730e476516649d8d96
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/60521
Auto-Submit: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Arrays can be extremely large, and having the load and store functions unroll the elements can make the complier explode.
Fixed: chromium:1229233
Change-Id: Ieb5654254e16f5ce724a205d21d954ef9a0cd053
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58382
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Reviewed-by: David Neto <dneto@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
CloneContext::Replace(T* what, T* with) is bug-prone, as complex transforms may want to clone `what` multiple times, or not at all. In both cases, this will likely result in an ICE as either the replacement will be reachable multiple times, or not at all.
The CTS test: webgpu:shader,execution,robust_access:linear_memory:storageClass="storage";storageMode="read_write";access="read";atomic=true;baseType="i32"
Was triggering this brokenness with DecomposeMemoryAccess's use of CloneContext::Replace(T*, T*).
Switch the usage of CloneContext::Replace(T*, T*) to the new function form.
As std::function is copyable, it cannot hold a captured std::unique_ptr.
This prevented the Replace() lambdas from capturing the necessary `BufferAccess` data, as this held a `std::unique_ptr<Offset>`.
To fix this, use a `BlockAllocator` for Offsets, and use raw pointers instead.
Because the function passed to Replace() is called just before the node is cloned, insertion of new functions will occur just before the currently evaluated module-scope entity.
This allows us to remove the "insert_after" arguments to LoadFunc(), StoreFunc(), and AtomicFunc().
We can also kill the icky InsertGlobal() and TypeDeclOf() helpers.
Bug: tint:993
Change-Id: I60972bc13a2fa819a163ee2671f61e82d0e68d2a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58222
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Use the new semantic constant value information to significantly reduce the complex indexing logic emitted for UBO accesses.
This will dramatically reduce the number of `for` loops that are decayed to `while` loops.
Change-Id: I1b0adb5edde2b4ed39c6beafc2e28106b86e0edd
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57701
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Antonio Maiorano
Stephen White
Ben Clayton
James Price