- Notifications
You must be signed in to change notification settings - Fork 15.3k
[Flang][OpenMP] Move char box bounds generation for Maps to DirectiveCommons.h #165918
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Flang][OpenMP] Move char box bounds generation for Maps to DirectiveCommons.h #165918
Conversation
| @llvm/pr-subscribers-flang-openmp @llvm/pr-subscribers-flang-fir-hlfir Author: None (agozillon) ChangesCurrently we generate these bounds in the MapInfoFinalization.cpp pass as it seems there's a missing case for character strings/arrays (length parameter) in the DirectiveCommons bounds generation functionality OpenMP uses for it's map operations. This PR tries to add this case to the DirectiveCommons function and remove the need for the bounds generation in the MapInfoFinalization pass, so that we are generating the bounds in the same place most other bounds are generated. Full diff: https://github.com/llvm/llvm-project/pull/165918.diff 2 Files Affected:
diff --git a/flang/include/flang/Lower/DirectivesCommon.h b/flang/include/flang/Lower/DirectivesCommon.h index 2d6906738773a..b564ee1f64423 100644 --- a/flang/include/flang/Lower/DirectivesCommon.h +++ b/flang/include/flang/Lower/DirectivesCommon.h @@ -512,11 +512,19 @@ fir::factory::AddrAndBoundsInfo gatherDataOperandAddrAndBounds( } bool dataExvIsAssumedSize = Fortran::semantics::IsAssumedSizeArray(symRef->get().GetUltimate()); - if (genDefaultBounds && - mlir::isa<fir::SequenceType>(fir::unwrapRefType(info.addr.getType()))) + if (genDefaultBounds && mlir::isa<fir::SequenceType>( + fir::unwrapRefType(info.addr.getType()))) { bounds = fir::factory::genBaseBoundsOps<BoundsOp, BoundsType>( builder, operandLocation, dataExv, dataExvIsAssumedSize, strideIncludeLowerExtent); + } + if (genDefaultBounds && fir::characterWithDynamicLen( + fir::unwrapRefType(info.addr.getType())) || + mlir::isa<fir::BoxCharType>( + fir::unwrapRefType(info.addr.getType()))) { + bounds = {fir::factory::genBoundsOpFromBoxChar<BoundsOp, BoundsType>( + builder, operandLocation, dataExv, info)}; + } asFortran << symRef->get().name().ToString(); } else { // Unsupported llvm::report_fatal_error("Unsupported type of OpenACC operand"); diff --git a/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp b/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp index bd07d7fe01b85..32faada7a3b49 100644 --- a/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp +++ b/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp @@ -987,38 +987,6 @@ class MapInfoFinalizationPass localBoxAllocas.clear(); deferrableDesc.clear(); - // First, walk `omp.map.info` ops to see if any of them have varPtrs - // with an underlying type of fir.char<k, ?>, i.e a character - // with dynamic length. If so, check if they need bounds added. - func->walk([&](mlir::omp::MapInfoOp op) { - if (!op.getBounds().empty()) - return; - - mlir::Value varPtr = op.getVarPtr(); - mlir::Type underlyingVarType = fir::unwrapRefType(varPtr.getType()); - - if (!fir::characterWithDynamicLen(underlyingVarType)) - return; - - fir::factory::AddrAndBoundsInfo info = - fir::factory::getDataOperandBaseAddr( - builder, varPtr, /*isOptional=*/false, varPtr.getLoc()); - - fir::ExtendedValue extendedValue = - hlfir::translateToExtendedValue(varPtr.getLoc(), builder, - hlfir::Entity{info.addr}, - /*continguousHint=*/true) - .first; - builder.setInsertionPoint(op); - llvm::SmallVector<mlir::Value> boundsOps = - fir::factory::genImplicitBoundsOps<mlir::omp::MapBoundsOp, - mlir::omp::MapBoundsType>( - builder, info, extendedValue, - /*dataExvIsAssumedSize=*/false, varPtr.getLoc()); - - op.getBoundsMutable().append(boundsOps); - }); - // Next, walk `omp.map.info` ops to see if any record members should be // implicitly mapped. func->walk([&](mlir::omp::MapInfoOp op) { |
bhandarkar-pranav left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thank you @agozillon
| Thank you @bhandarkar-pranav for the quick review! I'll give it until the middle of next week before landing the PR incase @vzakhari or @razvanlupusoru have any input as we share the same infrastructure for this with OpenACC! |
| Thank you for this fix! Is there an associated test case that you could include? |
razvanlupusoru left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Marking as approved as I will be PTO shortly and won't get another chance to look at this. Please add a test case before final submission. Thank you!
| Thank you very much for the quick review @razvanlupusoru I'll add a test next week before I land the PR! I hope you have a good time on your PTO! |
…Commons.h Currently we generate these bounds in the MapInfoFinalization.cpp pass as it seems there's a missing case for character strings/arrays (length parameter) in the DirectiveCommons bounds generation functionality OpenMP uses for it's map operations. This PR tries to add this case to the DirectiveCommons function and remove the need for the bounds generation in the MapInfoFinalization pass, so that we are generating the bounds in the same place most other bounds are generated.
6c87cd2 to ebc7adb Compare * main: (63 commits) [libc++] Inline vector::__append into resize (llvm#162086) [Flang][OpenMP] Move char box bounds generation for Maps to DirectiveCommons.h (llvm#165918) RuntimeLibcalls: Add entries for vector sincospi functions (llvm#166981) [X86] _mm_addsub_pd is not valid for constexpr (llvm#167363) [CIR] Re-land: Recognize constant aggregate initialization of auto vars (llvm#167033) [gn build] Port d2521f1 [gn build] Port caed089 [gn build] Port 315d705 [gn build] Port 2345b7d [PowerPC] convert memmove to milicode call .___memmove64[PR] in 64-bit mode (llvm#167334) [HLSL] Add internal linkage attribute to resources (llvm#166844) AMDGPU: Update test after e95f6fa [bazel] Port llvm#166980: TLI/VectorLibrary refactor (llvm#167354) [libc++] Split macros related to hardening into their own header (llvm#167069) [libc++][NFC] Remove unused imports from generate_feature_test_macro_components.py (llvm#159591) [CIR][NFC] Add test for Complex imag with GUN extension (llvm#167215) [BOLT][AArch64] Add more heuristics on epilogue determination (llvm#167077) RegisterCoalescer: Enable terminal rule by default for AMDGPU (llvm#161621) Revert "[clang] Refactor option-related code from clangDriver into new clangOptions library" (llvm#167348) [libc++][docs] Update to refer to P3355R2 (llvm#167267) ...
Currently we generate these bounds in the MapInfoFinalization.cpp pass as it seems there's a missing case for character strings/arrays (length parameter) in the DirectiveCommons bounds generation functionality OpenMP uses for it's map operations.
This PR tries to add this case to the DirectiveCommons function and remove the need for the bounds generation in the MapInfoFinalization pass, so that we are generating the bounds in the same place most other bounds are generated.