Our local Flang build on PowerPC was broken as
```
llvm/flang/../mlir/include/mlir/IR/ValueRange.h:401:20: error: 'ArrayRef' is deprecated: Use {} or ArrayRef<T>() instead [-Werror,-Wdeprecated-declarations]
401 | : ValueRange(ArrayRef<Value>(std::forward<Arg>(arg))) {}
| ^
llvm/flang/lib/Optimizer/CodeGen/CodeGen.cpp:2243:53: note: in instantiation of function template specialization 'mlir::ValueRange::ValueRange<const std::nullopt_t &, void>' requested here
2243 | /*cstInteriorIndices=*/std::nullopt, fieldIndices,
| ^
llvm/include/llvm/ADT/ArrayRef.h:70:18: note: 'ArrayRef' has been explicitly marked deprecated here
70 | /*implicit*/ LLVM_DEPRECATED("Use {} or ArrayRef<T>() instead", "{}")
| ^
llvm/include/llvm/Support/Compiler.h:244:50: note: expanded from macro 'LLVM_DEPRECATED'
244 | #define LLVM_DEPRECATED(MSG, FIX) __attribute__((deprecated(MSG, FIX)))
| ^
1 error generated.
```
This patch is to fix it.
I'm surprised at how bad the test coverage is here. There is some
overlap with existing tests, but they aren't comprehensive and do
not cover all the ABIs, or all the different types.
Fixes#147935
This patch makes it so that generate_test_report_github.py generates a
test report even when we don't get any test results. This otherwise
created a pretty confusing user experience on the Github side if the
build failed before any tests ran or in cases like running check-libc
where none of the tests are run through lit.
Reviewers: lnihlen, cmtice
Pull Request: https://github.com/llvm/llvm-project/pull/147871
This is another Buildkite relic that we can get rid of now to simplify
things a bit.
Reviewers: cmtice, DavidSpickett, lnihlen
Reviewed By: DavidSpickett
Pull Request: https://github.com/llvm/llvm-project/pull/147791
This patch removes the buildkite_info object from
generate_test_report_lib as we no longer support buildkite so it is dead
code now. Eventually we should set up equivalent functionality on the
Github side for downloading the logs from a convenient link, but for now
clean up the code.
Reviewers: cmtice, DavidSpickett, lnihlen
Reviewed By: DavidSpickett
Pull Request: https://github.com/llvm/llvm-project/pull/147784
This patch switches over uses of use_clang to clang_setup to fix a
potential race condition that has been impacting CI.
This is split from the refactoring to ensure I'm not missing anything
major here on the clang-tools-extra side.
This should fix#145703.
Reviewers: AaronBallman, HighCommander4, HerrCai0907, petrhosek, Keenuts
Reviewed By: petrhosek
Pull Request: https://github.com/llvm/llvm-project/pull/147437
The file is listing build artifacts to ignore, but LLVM has long had the
policy that in-tree builds are not supported, so the ignore rules
shouldn't serve their original purpose anymore.
The rules however are annoying because although they probably intended
only to ignore top-level build artifacts, they lack the leading `/` so
they match any file with the ignored name anywhere under `libclc/`.
When implementing #93430, I failed to consider some cases involving
function templates.
```
struct A {
template <typename T>
void a(this T self);
};
(&A::a<A>)(A{});
```
This fixes that
Previously we only carried the type arguments which caused value-based
costs to be inadvertantly changed into type-based costs.
I'm just using vp.is.fpclass as an example intrinsic for now since the
type based cost seems to differ from the value based cost, and most
normal intrinsics e.g. min/max have the same value + type based cost.
We still need to handle the cost properly for is.fpclass in a second
patch.
This is needed for an upcoming patch to handle the cost of
llvm.experimental.vp.reverse which suffers from the same problem.
After the recent move to work queues, in certain cases when linking in
the fortran runtime built for offload on AMDGPU as required in certain
cases, we'll get missing symbols when linking. This PR tries to address
this issue by encompassing more of the library in
RT_OFFLOAD_API_GROUP_BEGIN, which has the affect of compiling these
functions for AMDGPU, resolving the missing symbols.
This PR should address the following issue:
https://github.com/llvm/llvm-project/issues/145888
The `GlobalTy` helper has been extended to make both the Size and Ptr be
optional. Now `getGlobalMetadataFromDevice`/`Image` is able to write the
size of the global to the struct, instead of just verifying it.
Whilst at first glance there appears to be no native bfloat instructions
to modify the sign bit, this is only the case when FEAT_AFP is
implemented. Without this feature vector FABS/FNEG does not care about
the floating point format beyond needing to know the position of the
sign bit. From what I can see LLVM has no support for FEAT_AFP in terms
of feature detection or ACLE builtins and so I believe the compiler can
work under the assumption the feature is not enabled. In fact, if
FEAT_AFP is enabled then I believe the current isel is likely broken for
half, float and double anyway.
Add the supporting OpenMP Dialect operations, types, and interfaces for
modelling
MLIR Operations:
* omp.newcli
* omp.canonical_loop
MLIR Types:
* !omp.cli
MLIR Interfaces:
* LoopTransformationInterface
As a first loop transformations to be able to use these new operation in
follow-up PRs (#144785)
* omp.unroll_heuristic
* Add spec generation to offload-tblgen tool
* This patch adds generation of Sphinx compatible reStructuedText
utilizing the C domain to document the Offload API directly from the
spec definition `.td` files.
* Add Sphinx HTML documentation target
* Introduces the `docs-offload-html` target when CMake is configured
with `LLVM_ENABLE_SPHINX=ON` and `SPHINX_OUTPUT_HTML=ON`. Utilized
`offload-tblgen -gen-spen` to generate Offload API specification docs.
This PR exposes the backend pass config to plugins via a callback.
Plugin authors can register a callback that is being triggered before
the target backend adds their passes to the pipeline. In the callback
they then get access to the `TargetMachine`, the `PassManager`, and the
`TargetPassConfig`. This allows plugins to call
`TargetPassConfig::insertPass`, which is honored in the subsequent
`addPass` of the main backend. We implemented this using the legacy pass
manager since backends still use it as the default.
We can use NBSL/BSL2N to implement the following operations via the
corresponding identities:
* EON(a, b) = BSL2N(a, a, b) = BSL2N(b, b, a)
* NAND(a, b) = NBSL(a, b, b) = NBSL(b, a, a)
* NOR(a, b) = NBSL(a, b, a) = NBSL(b, a, b)
* ORN(a, b) = BSL2N(a, b, a)
Most of these operations are currently lowered into at least two
instructions because we don't have dedicated Neon/SVE instructions
for them. With the appropriate pattern of NBSL/BSL2N we can lower
them in a single instruction.
We could also use NBSL to implement an unpredicated NOT(a) =
NBSL(a, a, a), but because of the tied register constraint, this
may not always be profitable.
Lower it just like the vector [l]lrint, using vfcvt, with the right
rounding mode. Updating costs to account for this custom-lowering is
left to a companion patch.
Avoid using the underlying type, and be more consistent -- MacroDefStart stores
the result of SourceManager::getNextLocalOffset() which returns the
SourceLocation::UIntTy.
Cascaded GEP (i.e. GEP of GEP) are not handled when determining if it is
ok to runtime unroll loops.
This change simply uses `getUnderlyingObjects` to look through cascaded
GEPs.
This patch updates the Qualification Working Group documentation with
improvements based on our first sync-up meeting in July 2025:
- Added July 2025 meeting slides to `qual-wg/slides/`
- Updated Participation section to include clickable links to Discourse
and Discord
- Clarified contributor recognition to include async contributors
- Added new sections for Meeting Minutes and Presentation Slides
- Linked to the initial RFC and Discord channel in the Contact section
- Added Code of Conduct section
- Minor formatting consistency fixes across the document
This change also updates `GettingInvolved.rst` to add the Qualification
WG with calendar links and Discourse minutes.
These updates aim to improve clarity, traceability, and contributor
onboarding.
---------
Co-authored-by: Wendi Urribarri (Woven by Toyota <wendi.urribarri@woven-planet.global>
According to OpenMP 5.2 (Section 7.8.2), the directive name `declare
target` may be used as a synonym for `begin declare target` only when no
clauses are specified. This clause-less delimited form is now deprecated
and should emit a deprecation warning.
```
// Deprecated usage (should trigger warning):
#pragma omp declare target // deprecated in OpenMP 5.2
void foo1() {
}
#pragma omp end declare target
// Valid usage with clause (should not trigger warning):
#pragma omp declare target enter(foo2)
void foo2() {
}
```
```
// Recommended replacement for deprecated delimited form:
#pragma omp begin declare target
void foo() {
}
#pragma omp end declare target
```
---------
Co-authored-by: urvi-rav <urvi.rav@hpe.com>
Since the function template isn't instantiated before constraint
checking, we'll not be able to find the outer template arguments through
function specialization when evaluating the inner constraint that is
nested within a larger constraint expression.
The only practical solution is to get them back through the code
synthesis context, which also allows us to eliminate an overload of
getTemplateInstantiationArgs.
No release note because it's a regression on trunk.
Fixes https://github.com/llvm/llvm-project/issues/147772
Currently we have slightly different costing for the vp and non-vp
version of the rounding intrinsics.
We can delete this code and use the generic BasicTTIImpl code for the vp
intrinsics which falls back to the non-vp versions.
I'm not sure if the zvfh costing is correct, this should probably be
fixed in a follow up patch. At the moment the non-vp cost is more
important since it is what the loop vectorizer will use.
Make the Visited set a local variable, otherwise we can reject a PHI
(those that do not have a zeroinitializer constant) but mark it as
visited,
and the rest of the function thinks the PHI is ok when it isn't.
This is a bit crude but it's the only fix that consistently worked in my
testing.
Fixes SWDEV-541767
