… (#189058)

…188917)

…ardOperands API to BranchOpInterface (#187864)

To simplify the output of the reduction-tree pass, this PR introduces
the eraseRedundantBlocksInRegion. For regions containing multiple
execution paths, this functionality selects the shortest 'interesting'
path. Additionally, this PR adds the getSuccessorForwardOperands API to
BranchOpInterface. This allows us to extract the ForwardOperands for a
specific path chosen from multiple alternatives, enabling the creation
of a cf.br operation for the redirected jump.

…tions (#189113)

Fixes llvm/llvm-project#187716.

…ter (#188924)

…ssorForwardOperands API to BranchOpInterface" (#189150)

Reverts llvm/llvm-project#187864, because it is causing same build bot
failures. See https://lab.llvm.org/buildbot/#/builders/138/builds/27662
and
https://lab.llvm.org/buildbot/#/builders/169/builds/21376/steps/11/logs/stdio
for memory leak issues.

…on index (#188508)

When a dynamic index of -1 (the kPoisonIndex sentinel) was folded into
the static position of a vector.insert op,
foldDenseElementsAttrDestInsertOp would proceed to call
calculateInsertPosition, which returned -1. The subsequent iterator
arithmetic (allValues.begin() + (-1)) was undefined behaviour, causing
an assertion in DenseElementsAttr::get.

Fix by bailing out early in foldDenseElementsAttrDestInsertOp when any
static position equals kPoisonIndex, consistent with how
InsertChainFullyInitialized already guards this case.

Fixes #188404

Assisted-by: Claude Code

…nt (#189163)

When invoking `-test-bytecode-roundtrip=test-dialect-version=X.Y` on a
module that contains no test dialect operations, the reader type
callback in `runTest0` called
`reader.getDialectVersion<test::TestDialect>()` and then immediately
asserted that it succeeded. However, if the test dialect was never
referenced in the bytecode (because no test dialect types appear in the
module), the dialect's version information is not stored in the
bytecode, so `getDialectVersion` legitimately returns failure.

When the test dialect version is unavailable in the bytecode being read,
the module contains no test dialect types, so no "funky"-group overrides
are needed and the callback can safely skip by returning `success()`.

A regression test is added with a module that has no test dialect ops,
exercising the `test-dialect-version=2.0` path that previously crashed.

Fixes #128321
Fixes #128325

Assisted-by: Claude Code

… (#188064)

This PR adds two new field specifiers (`operand` and `attribute`) and
extends the existing one (`result`):
- `default_factory` parameter is added for `result` and `attribute` to
specify default value via a lambda/function
- `kw_only` parameter is added for all these three specifiers, to make a
field a keyword-only parameter (without giving a default value).

```python
def result(
    *,
    infer_type: bool = False,
    default_factory: Optional[Callable[[], Any]] = None,
    kw_only: bool = False,
) -> Any: ...


def operand(
    *,
    kw_only: bool = False,
) -> Any: ...


def attribute(
    *,
    default_factory: Optional[Callable[[], Any]] = None,
    kw_only: bool = False,
) -> Any: ...
```

Examples about how to use them:
```python
class OperandSpecifierOp(TestFieldSpecifiers.Operation, name="operand_specifier"):
    a: Operand[IntegerType[32]] = operand()
    b: Optional[Operand[IntegerType[32]]] = None
    c: Operand[IntegerType[32]] = operand(kw_only=True)

class ResultSpecifierOp(TestFieldSpecifiers.Operation, name="result_specifier"):
    a: Result[IntegerType[32]] = result()
    b: Result[IntegerType[16]] = result(infer_type=True)
    c: Result[IntegerType] = result(
        default_factory=lambda: IntegerType.get_signless(8)
    )
    d: Sequence[Result[IntegerType]] = result(default_factory=list)
    e: Result[IntegerType[32]] = result(kw_only=True)

class AttributeSpecifierOp(
    TestFieldSpecifiers.Operation, name="attribute_specifier"
):
    a: IntegerAttr = attribute()
    b: IntegerAttr = attribute(
        default_factory=lambda: IntegerAttr.get(IntegerType.get_signless(32), 42)
    )
    c: StringAttr["a"] | StringAttr["b"] = attribute(
        default_factory=lambda: StringAttr.get("a")
    )
    d: IntegerAttr = attribute(kw_only=True)
```

---------

Co-authored-by: Rolf Morel <rolfmorel@gmail.com>

Summary:
These were renamed and the aliases removed, fix running the tests.

Signed-off-by: Shikhar Soni <shikharish05@gmail.com>

…89128)

This fixes #186684.

Also fix (not) breaking variables declared on the same line as the
closing brace.

And adapt whitesmith to that changes.

…efs (#188860)

Fixes #188695

…ng and tests (#184365)

Closes #181654

… broadcast from sg-to-wi (#185960)

This PR adds distribution patterns for vector.step, vector.shape_cast &
vector.broadcast in the new sg-to-wi pass

…. (#188721)

If a load and a store have different address spaces, we cannot create a
runtime check. Instead, always copy the data to an alloca matching the
store address space.

Fixes llvm/llvm-project#185236.

PR: llvm/llvm-project#188721

Need to check if the potential bitcast/bswap-like construct is a root of
the reduction, otherwise it cannot represent a bitcast/bswap construct.

Fixes #189184

Add VPlan printing test for
 llvm/llvm-project#186252
 llvm/llvm-project#189022

FormatTest.cpp is too huge, extract some tests to mitigate this a bit.

#184545 default-enables the IO sandbox in assert-builds. This causes
Clang using Polly to crash (#188568).

The issue is that `PassBuilder` uses `vfs::getRealFileSystem()` by
default which is considered a IO sandbox violation in the Clang process.
With this PR store the VFS from the `PassBuilder` from the original
`registerPollyPasses` call for creating other `PassBuilder` instances.

This PR also adds infrastructure for running Polly in `clang` (in
addition in `opt`). `opt` does not enable the sandbox such that we need
separate tests using Clang.

Closes: #188568

…ase (#189202)

On musl, rlimit64 is an alias for rlimit rather than a distinct type
provided by glibc. Add a SANITIZER_MUSL elif branch so that
struct_rlimit64_sz is defined for musl-based Linux targets.

…189199)

Use start + (end - start) / 2 instead of (start + end) / 2 to compute
the midpoint address. The original expression overflows when start + end
exceeds UPTR_MAX, which happens on 32-bit targets whose memory layout
includes regions above 0x80000000.

As a new contributor, it helps to correctly see the right maintainer.

…ion (#189145)

Fixes llvm/llvm-project#163913.

…C (#189214)

ICF's InputSection::replace() calls markDead() on folded sections, so
`!sec->isLive()` already filters them.

The code was accidentaly removed in
intel-restricted/applications.compilers.llvm-project@4d19936#diff-0bcd40d8040a4c485a91c581d1f60f04570fa7d3fc168ff4999db520a33e82af
.

Fixes: Clang::Index/index-builtin-opencl.clcpp 
https://jira.devtools.intel.com/browse/CMPLRLLVM-74121

…682bd (#35866)

Fix unit test failure in StencilTest.DescribeImplicitOperator caused by
commit f4682bd which introduced staged lambda initialization.

In SYCL/CUDA/OpenMP builds, getManglingNumber() is called between
setLambdaContextDecl() and setLambdaNumbering(), which can trigger
linkage computation for init-capture variables while the lambda is only
partially initialized. This causes the cached linkage to differ from the
linkage computed after full initialization, triggering the assertion in
DeduceVariableDeclarationType.

Solution: Invalidate cached linkage after deducing the type. This
ensures
linkage is recomputed with the complete type.

Fixes regressions due to:

* 6df388f 2026-03-13 [sycl-web] Reland
8ce2b9c zahira.ammarguellat@intel.com
* f4682bd 2026-03-13
[Clang][ItaniumMangle] Fix recursive mangling for lambda init-captures
(#182667) ototot@google.com

---------

Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>

Fix the sycl-oneapi-gpu-amdgpu.cpp test that broke after commit
2757b58 which merged changes from 'main' to 'sycl-web'.

The merge introduced ToolChain::normalizeOffloadTriple() which
automatically completes incomplete triples (e.g., "amdgcn" becomes
"amdgcn-amd-amdhsa"). However, the SYCL driver should reject
incomplete triples when directly specified by the user.

Updated Driver.cpp:1389 to avoid normalizing user-provided triples
for SYCL.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

Fix SYCLNativeCPUUtils vecz transform to handle the upstream change in
commit 3604119 that disallows
calling getTerminator() on blocks without terminators.

The rewireDivergentLoopExitBlocks function creates new basic blocks
(newDivergentLE) without terminators, but later operations like
DT->recalculate(), isReachable(), and predecessors() traverse the CFG
and call getTerminator(), which now asserts on blocks without terminators.

Applied the same workaround pattern used in commit 873322287a4312a7
(VPO/Paropt fix): temporarily add UnreachableInst terminators to newly
created blocks before CFG traversal operations, then remove and replace
them with proper terminators in computeNewTargets.

Changes:
- Add temporary UnreachableInst to newDivergentLE after creation
- Use getTerminatorOrNull() instead of getTerminator() where blocks
  may temporarily lack terminators
- Skip blocks without terminators in isReachable() CFG traversal
- Handle placeholder UnreachableInst terminators in computeNewTargets

Fixes: SYCL :: check_device_code/native_cpu/vectorization.cpp

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

…lectives

The unsigned variants (uchar, ushort, uint, ulong) of __clc__get_group_scratch
functions were declared in collectives.cl but never defined in
collectives_helpers.cl for both PTX and AMD targets. This caused linker errors
during device code compilation when group broadcast operations used unsigned types.

Add implementations that cast from the corresponding signed type's scratch memory.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

…160956)"

This reverts commit cc38e42.