[flang][cuda] Remove set_allocator_idx operation #157747
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
llvmbot
added
flang
|
@llvm/pr-subscribers-flang-fir-hlfir Author: Valentin Clement (バレンタイン クレメン) (clementval) ChangesThe allocator index is set from the component genre #157731 . There is no more need of an operation to set it at a later point. Patch is 27.65 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/157747.diff 14 Files Affected:
diff --git a/flang-rt/lib/cuda/descriptor.cpp b/flang-rt/lib/cuda/descriptor.cpp
index d3cc6c2c99e79..aa75d4eff0511 100644
--- a/flang-rt/lib/cuda/descriptor.cpp
+++ b/flang-rt/lib/cuda/descriptor.cpp
@@ -62,15 +62,6 @@ void RTDEF(CUFDescriptorCheckSection)(
}
}
-void RTDEF(CUFSetAllocatorIndex)(
- Descriptor *desc, int index, const char *sourceFile, int sourceLine) {
- if (!desc) {
- Terminator terminator{sourceFile, sourceLine};
- terminator.Crash("descriptor is null");
- }
- desc->SetAllocIdx(index);
-}
-
RT_EXT_API_GROUP_END
}
} // namespace Fortran::runtime::cuda
diff --git a/flang-rt/unittests/Runtime/CUDA/AllocatorCUF.cpp b/flang-rt/unittests/Runtime/CUDA/AllocatorCUF.cpp
index 4b22e29cc8172..f1f931e87a86e 100644
--- a/flang-rt/unittests/Runtime/CUDA/AllocatorCUF.cpp
+++ b/flang-rt/unittests/Runtime/CUDA/AllocatorCUF.cpp
@@ -72,14 +72,3 @@ TEST(AllocatableCUFTest, DescriptorAllocationTest) {
EXPECT_TRUE(desc != nullptr);
RTNAME(CUFFreeDescriptor)(desc);
}
-
-TEST(AllocatableCUFTest, CUFSetAllocatorIndex) {
- using Fortran::common::TypeCategory;
- RTNAME(CUFRegisterAllocator)();
- // REAL(4), DEVICE, ALLOCATABLE :: a(:)
- auto a{createAllocatable(TypeCategory::Real, 4)};
- EXPECT_EQ((int)kDefaultAllocator, a->GetAllocIdx());
- RTNAME(CUFSetAllocatorIndex)(
- a.get(), kDeviceAllocatorPos, __FILE__, __LINE__);
- EXPECT_EQ((int)kDeviceAllocatorPos, a->GetAllocIdx());
-}
diff --git a/flang/include/flang/Lower/CUDA.h b/flang/include/flang/Lower/CUDA.h
index 4a831fd502af4..ab9dde8ad5198 100644
--- a/flang/include/flang/Lower/CUDA.h
+++ b/flang/include/flang/Lower/CUDA.h
@@ -47,10 +47,6 @@ static inline unsigned getAllocatorIdx(const Fortran::semantics::Symbol &sym) {
return kDefaultAllocator;
}
-void initializeDeviceComponentAllocator(
- Fortran::lower::AbstractConverter &converter,
- const Fortran::semantics::Symbol &sym, const fir::MutableBoxValue &box);
-
mlir::Type gatherDeviceComponentCoordinatesAndType(
fir::FirOpBuilder &builder, mlir::Location loc,
const Fortran::semantics::Symbol &sym, fir::RecordType recTy,
diff --git a/flang/include/flang/Optimizer/Builder/Runtime/CUDA/Descriptor.h b/flang/include/flang/Optimizer/Builder/Runtime/CUDA/Descriptor.h
index 43dca65322a62..bdeb7574012c6 100644
--- a/flang/include/flang/Optimizer/Builder/Runtime/CUDA/Descriptor.h
+++ b/flang/include/flang/Optimizer/Builder/Runtime/CUDA/Descriptor.h
@@ -31,10 +31,6 @@ void genSyncGlobalDescriptor(fir::FirOpBuilder &builder, mlir::Location loc,
void genDescriptorCheckSection(fir::FirOpBuilder &builder, mlir::Location loc,
mlir::Value desc);
-/// Generate runtime call to set the allocator index in the descriptor.
-void genSetAllocatorIndex(fir::FirOpBuilder &builder, mlir::Location loc,
- mlir::Value desc, mlir::Value index);
-
} // namespace fir::runtime::cuda
#endif // FORTRAN_OPTIMIZER_BUILDER_RUNTIME_CUDA_DESCRIPTOR_H_
diff --git a/flang/include/flang/Optimizer/Dialect/CUF/CUFOps.td b/flang/include/flang/Optimizer/Dialect/CUF/CUFOps.td
index 23ab882615209..e38738230ffbc 100644
--- a/flang/include/flang/Optimizer/Dialect/CUF/CUFOps.td
+++ b/flang/include/flang/Optimizer/Dialect/CUF/CUFOps.td
@@ -388,25 +388,4 @@ def cuf_StreamCastOp : cuf_Op<"stream_cast", [NoMemoryEffect]> {
let hasVerifier = 1;
}
-def cuf_SetAllocatorIndexOp : cuf_Op<"set_allocator_idx", []> {
- let summary = "Set the allocator index in a descriptor";
-
- let description = [{
- Allocator index in the Fortran descriptor is used to retrived the correct
- CUDA allocator to allocate the memory on the device.
- In many cases the allocator index is set when the descriptor is created. For
- device components, the descriptor is part of the derived-type itself and
- needs to be set after the derived-type is allocated in managed memory.
- }];
-
- let arguments = (ins Arg<fir_ReferenceType, "", [MemRead, MemWrite]>:$box,
- cuf_DataAttributeAttr:$data_attr);
-
- let assemblyFormat = [{
- $box `:` qualified(type($box)) attr-dict
- }];
-
- let hasVerifier = 1;
-}
-
#endif // FORTRAN_DIALECT_CUF_CUF_OPS
diff --git a/flang/include/flang/Runtime/CUDA/descriptor.h b/flang/include/flang/Runtime/CUDA/descriptor.h
index 7555f276ac1de..06e4a4649db1b 100644
--- a/flang/include/flang/Runtime/CUDA/descriptor.h
+++ b/flang/include/flang/Runtime/CUDA/descriptor.h
@@ -41,10 +41,6 @@ void RTDECL(CUFSyncGlobalDescriptor)(
void RTDECL(CUFDescriptorCheckSection)(
const Descriptor *, const char *sourceFile = nullptr, int sourceLine = 0);
-/// Set the allocator index with the provided value.
-void RTDECL(CUFSetAllocatorIndex)(Descriptor *, int index,
- const char *sourceFile = nullptr, int sourceLine = 0);
-
} // extern "C"
} // namespace Fortran::runtime::cuda
diff --git a/flang/lib/Lower/Allocatable.cpp b/flang/lib/Lower/Allocatable.cpp
index 444b5b6c7c4b1..53239cb83c6c0 100644
--- a/flang/lib/Lower/Allocatable.cpp
+++ b/flang/lib/Lower/Allocatable.cpp
@@ -450,9 +450,6 @@ class AllocateStmtHelper {
if (alloc.getSymbol().test(Fortran::semantics::Symbol::Flag::AccDeclare))
Fortran::lower::attachDeclarePostAllocAction(converter, builder,
alloc.getSymbol());
- if (Fortran::semantics::HasCUDAComponent(alloc.getSymbol()))
- Fortran::lower::initializeDeviceComponentAllocator(
- converter, alloc.getSymbol(), box);
}
void setPinnedToFalse() {
diff --git a/flang/lib/Lower/CUDA.cpp b/flang/lib/Lower/CUDA.cpp
index 1293d2c5bd3ae..bb4bdee78f97d 100644
--- a/flang/lib/Lower/CUDA.cpp
+++ b/flang/lib/Lower/CUDA.cpp
@@ -17,95 +17,6 @@
#define DEBUG_TYPE "flang-lower-cuda"
-void Fortran::lower::initializeDeviceComponentAllocator(
- Fortran::lower::AbstractConverter &converter,
- const Fortran::semantics::Symbol &sym, const fir::MutableBoxValue &box) {
- if (const auto *details{
- sym.GetUltimate()
- .detailsIf<Fortran::semantics::ObjectEntityDetails>()}) {
- const Fortran::semantics::DeclTypeSpec *type{details->type()};
- const Fortran::semantics::DerivedTypeSpec *derived{type ? type->AsDerived()
- : nullptr};
- if (derived) {
- if (!FindCUDADeviceAllocatableUltimateComponent(*derived))
- return; // No device components.
-
- fir::FirOpBuilder &builder = converter.getFirOpBuilder();
- mlir::Location loc = converter.getCurrentLocation();
-
- mlir::Type baseTy = fir::unwrapRefType(box.getAddr().getType());
-
- // Only pointer and allocatable needs post allocation initialization
- // of components descriptors.
- if (!fir::isAllocatableType(baseTy) && !fir::isPointerType(baseTy))
- return;
-
- // Extract the derived type.
- mlir::Type ty = fir::getDerivedType(baseTy);
- auto recTy = mlir::dyn_cast<fir::RecordType>(ty);
- assert(recTy && "expected fir::RecordType");
-
- if (auto boxTy = mlir::dyn_cast<fir::BaseBoxType>(baseTy))
- baseTy = boxTy.getEleTy();
- baseTy = fir::unwrapRefType(baseTy);
-
- Fortran::semantics::UltimateComponentIterator components{*derived};
- mlir::Value loadedBox = fir::LoadOp::create(builder, loc, box.getAddr());
- mlir::Value addr;
- if (auto seqTy = mlir::dyn_cast<fir::SequenceType>(baseTy)) {
- mlir::Type idxTy = builder.getIndexType();
- mlir::Value one = builder.createIntegerConstant(loc, idxTy, 1);
- mlir::Value zero = builder.createIntegerConstant(loc, idxTy, 0);
- llvm::SmallVector<fir::DoLoopOp> loops;
- llvm::SmallVector<mlir::Value> indices;
- llvm::SmallVector<mlir::Value> extents;
- for (unsigned i = 0; i < seqTy.getDimension(); ++i) {
- mlir::Value dim = builder.createIntegerConstant(loc, idxTy, i);
- auto dimInfo = fir::BoxDimsOp::create(builder, loc, idxTy, idxTy,
- idxTy, loadedBox, dim);
- mlir::Value lbub = mlir::arith::AddIOp::create(
- builder, loc, dimInfo.getResult(0), dimInfo.getResult(1));
- mlir::Value ext =
- mlir::arith::SubIOp::create(builder, loc, lbub, one);
- mlir::Value cmp = mlir::arith::CmpIOp::create(
- builder, loc, mlir::arith::CmpIPredicate::sgt, ext, zero);
- ext = mlir::arith::SelectOp::create(builder, loc, cmp, ext, zero);
- extents.push_back(ext);
-
- auto loop = fir::DoLoopOp::create(
- builder, loc, dimInfo.getResult(0), dimInfo.getResult(1),
- dimInfo.getResult(2), /*isUnordered=*/true,
- /*finalCount=*/false, mlir::ValueRange{});
- loops.push_back(loop);
- indices.push_back(loop.getInductionVar());
- builder.setInsertionPointToStart(loop.getBody());
- }
- mlir::Value boxAddr = fir::BoxAddrOp::create(builder, loc, loadedBox);
- auto shape = fir::ShapeOp::create(builder, loc, extents);
- addr = fir::ArrayCoorOp::create(
- builder, loc, fir::ReferenceType::get(recTy), boxAddr, shape,
- /*slice=*/mlir::Value{}, indices, /*typeparms=*/mlir::ValueRange{});
- } else {
- addr = fir::BoxAddrOp::create(builder, loc, loadedBox);
- }
- for (const auto &compSym : components) {
- if (Fortran::semantics::IsDeviceAllocatable(compSym)) {
- llvm::SmallVector<mlir::Value> coord;
- mlir::Type fieldTy = gatherDeviceComponentCoordinatesAndType(
- builder, loc, compSym, recTy, coord);
- assert(coord.size() == 1 && "expect one coordinate");
- mlir::Value comp = fir::CoordinateOp::create(
- builder, loc, builder.getRefType(fieldTy), addr, coord[0]);
- cuf::DataAttributeAttr dataAttr =
- Fortran::lower::translateSymbolCUFDataAttribute(
- builder.getContext(), compSym);
- cuf::SetAllocatorIndexOp::create(builder, loc, comp, dataAttr);
- }
- }
- }
- }
-}
-
mlir::Type Fortran::lower::gatherDeviceComponentCoordinatesAndType(
fir::FirOpBuilder &builder, mlir::Location loc,
const Fortran::semantics::Symbol &sym, fir::RecordType recTy,
diff --git a/flang/lib/Lower/ConvertVariable.cpp b/flang/lib/Lower/ConvertVariable.cpp
index c79c9b1ab0f51..b1cd14c56969a 100644
--- a/flang/lib/Lower/ConvertVariable.cpp
+++ b/flang/lib/Lower/ConvertVariable.cpp
@@ -786,62 +786,6 @@ static mlir::Value createNewLocal(Fortran::lower::AbstractConverter &converter,
return res;
}
-/// Device allocatable components in a derived-type don't have the correct
-/// allocator index in their descriptor when they are created. After
-/// initialization, cuf.set_allocator_idx operations are inserted to set the
-/// correct allocator index for each device component.
-static void
-initializeDeviceComponentAllocator(Fortran::lower::AbstractConverter &converter,
- const Fortran::semantics::Symbol &symbol,
- Fortran::lower::SymMap &symMap) {
- if (const auto *details{
- symbol.GetUltimate()
- .detailsIf<Fortran::semantics::ObjectEntityDetails>()}) {
- const Fortran::semantics::DeclTypeSpec *type{details->type()};
- const Fortran::semantics::DerivedTypeSpec *derived{type ? type->AsDerived()
- : nullptr};
- if (derived) {
- if (!FindCUDADeviceAllocatableUltimateComponent(*derived))
- return; // No device components.
-
- fir::FirOpBuilder &builder = converter.getFirOpBuilder();
- mlir::Location loc = converter.getCurrentLocation();
-
- fir::ExtendedValue exv =
- converter.getSymbolExtendedValue(symbol.GetUltimate(), &symMap);
- mlir::Type baseTy = fir::unwrapRefType(fir::getBase(exv).getType());
- if (auto boxTy = mlir::dyn_cast<fir::BaseBoxType>(baseTy))
- baseTy = boxTy.getEleTy();
- baseTy = fir::unwrapRefType(baseTy);
-
- if (fir::isAllocatableType(fir::getBase(exv).getType()) ||
- fir::isPointerType(fir::getBase(exv).getType()))
- return; // Allocator index need to be set after allocation.
-
- auto recTy =
- mlir::dyn_cast<fir::RecordType>(fir::unwrapSequenceType(baseTy));
- assert(recTy && "expected fir::RecordType");
-
- Fortran::semantics::UltimateComponentIterator components{*derived};
- for (const auto &sym : components) {
- if (Fortran::semantics::IsDeviceAllocatable(sym)) {
- llvm::SmallVector<mlir::Value> coord;
- mlir::Type fieldTy =
- Fortran::lower::gatherDeviceComponentCoordinatesAndType(
- builder, loc, sym, recTy, coord);
- mlir::Value base = fir::getBase(exv);
- mlir::Value comp = fir::CoordinateOp::create(
- builder, loc, builder.getRefType(fieldTy), base, coord);
- cuf::DataAttributeAttr dataAttr =
- Fortran::lower::translateSymbolCUFDataAttribute(
- builder.getContext(), sym);
- cuf::SetAllocatorIndexOp::create(builder, loc, comp, dataAttr);
- }
- }
- }
- }
-}
-
/// Must \p var be default initialized at runtime when entering its scope.
static bool
mustBeDefaultInitializedAtRuntime(const Fortran::lower::pft::Variable &var) {
@@ -1164,9 +1108,6 @@ static void instantiateLocal(Fortran::lower::AbstractConverter &converter,
if (mustBeDefaultInitializedAtRuntime(var))
Fortran::lower::defaultInitializeAtRuntime(converter, var.getSymbol(),
symMap);
- if (converter.getFoldingContext().languageFeatures().IsEnabled(
- Fortran::common::LanguageFeature::CUDA))
- initializeDeviceComponentAllocator(converter, var.getSymbol(), symMap);
auto *builder = &converter.getFirOpBuilder();
if (needCUDAAlloc(var.getSymbol()) &&
!cuf::isCUDADeviceContext(builder->getRegion())) {
@@ -1426,9 +1367,6 @@ static void instantiateAlias(Fortran::lower::AbstractConverter &converter,
if (mustBeDefaultInitializedAtRuntime(var))
Fortran::lower::defaultInitializeAtRuntime(converter, var.getSymbol(),
symMap);
- if (converter.getFoldingContext().languageFeatures().IsEnabled(
- Fortran::common::LanguageFeature::CUDA))
- initializeDeviceComponentAllocator(converter, var.getSymbol(), symMap);
}
//===--------------------------------------------------------------===//
diff --git a/flang/lib/Optimizer/Builder/Runtime/CUDA/Descriptor.cpp b/flang/lib/Optimizer/Builder/Runtime/CUDA/Descriptor.cpp
index a6ee98685f3c9..37e4c5a706df2 100644
--- a/flang/lib/Optimizer/Builder/Runtime/CUDA/Descriptor.cpp
+++ b/flang/lib/Optimizer/Builder/Runtime/CUDA/Descriptor.cpp
@@ -47,18 +47,3 @@ void fir::runtime::cuda::genDescriptorCheckSection(fir::FirOpBuilder &builder,
builder, loc, fTy, desc, sourceFile, sourceLine)};
fir::CallOp::create(builder, loc, func, args);
}
-
-void fir::runtime::cuda::genSetAllocatorIndex(fir::FirOpBuilder &builder,
- mlir::Location loc,
- mlir::Value desc,
- mlir::Value index) {
- mlir::func::FuncOp func =
- fir::runtime::getRuntimeFunc<mkRTKey(CUFSetAllocatorIndex)>(loc, builder);
- auto fTy = func.getFunctionType();
- mlir::Value sourceFile = fir::factory::locationToFilename(builder, loc);
- mlir::Value sourceLine =
- fir::factory::locationToLineNo(builder, loc, fTy.getInput(3));
- llvm::SmallVector<mlir::Value> args{fir::runtime::createArguments(
- builder, loc, fTy, desc, index, sourceFile, sourceLine)};
- fir::CallOp::create(builder, loc, func, args);
-}
diff --git a/flang/lib/Optimizer/Dialect/CUF/CUFOps.cpp b/flang/lib/Optimizer/Dialect/CUF/CUFOps.cpp
index ade80716f2561..687007d957225 100644
--- a/flang/lib/Optimizer/Dialect/CUF/CUFOps.cpp
+++ b/flang/lib/Optimizer/Dialect/CUF/CUFOps.cpp
@@ -345,17 +345,6 @@ llvm::LogicalResult cuf::StreamCastOp::verify() {
return checkStreamType(*this);
}
-//===----------------------------------------------------------------------===//
-// SetAllocatorOp
-//===----------------------------------------------------------------------===//
-
-llvm::LogicalResult cuf::SetAllocatorIndexOp::verify() {
- if (!mlir::isa<fir::BaseBoxType>(fir::unwrapRefType(getBox().getType())))
- return emitOpError(
- "expect box to be a reference to class or box type value");
- return mlir::success();
-}
-
// Tablegen operators
#define GET_OP_CLASSES
diff --git a/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp b/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp
index 9834b0499b930..4fe83d41c3b96 100644
--- a/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp
+++ b/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp
@@ -928,34 +928,6 @@ struct CUFSyncDescriptorOpConversion
}
};
-struct CUFSetAllocatorIndexOpConversion
- : public mlir::OpRewritePattern<cuf::SetAllocatorIndexOp> {
- using OpRewritePattern::OpRewritePattern;
-
- mlir::LogicalResult
- matchAndRewrite(cuf::SetAllocatorIndexOp op,
- mlir::PatternRewriter &rewriter) const override {
- auto mod = op->getParentOfType<mlir::ModuleOp>();
- fir::FirOpBuilder builder(rewriter, mod);
- mlir::Location loc = op.getLoc();
- int idx = kDefaultAllocator;
- if (op.getDataAttr() == cuf::DataAttribute::Device) {
- idx = kDeviceAllocatorPos;
- } else if (op.getDataAttr() == cuf::DataAttribute::Managed) {
- idx = kManagedAllocatorPos;
- } else if (op.getDataAttr() == cuf::DataAttribute::Unified) {
- idx = kUnifiedAllocatorPos;
- } else if (op.getDataAttr() == cuf::DataAttribute::Pinned) {
- idx = kPinnedAllocatorPos;
- }
- mlir::Value index =
- builder.createIntegerConstant(loc, builder.getI32Type(), idx);
- fir::runtime::cuda::genSetAllocatorIndex(builder, loc, op.getBox(), index);
- op.erase();
- return mlir::success();
- }
-};
-
class CUFOpConversion : public fir::impl::CUFOpConversionBase<CUFOpConversion> {
public:
void runOnOperation() override {
@@ -1017,8 +989,8 @@ void cuf::populateCUFToFIRConversionPatterns(
const mlir::SymbolTable &symtab, mlir::RewritePatternSet &patterns) {
patterns.insert<CUFAllocOpConversion>(patterns.getContext(), &dl, &converter);
patterns.insert<CUFAllocateOpConversion, CUFDeallocateOpConversion,
- CUFFreeOpConversion, CUFSyncDescriptorOpConversion,
- CUFSetAllocatorIndexOpConversion>(patterns.getContext());
+ CUFFreeOpConversion, CUFSyncDescriptorOpConversion>(
+ patterns.getContext());
patterns.insert<CUFDataTransferOpConversion>(patterns.getContext(), symtab,
&dl, &converter);
patterns.insert<CUFLaunchOpConversion, CUFDeviceAddressOpConversion>(
diff --git a/flang/test/Fir/CUDA/cuda-alloc-free.fir b/flang/test/Fir/CUDA/cuda-alloc-free.fir
index 8b6e7d67931df..31f2ed022b6c4 100644
--- a/flang/test/Fir/CUDA/cuda-alloc-free.fir
+++ b/flang/test/Fir/CUDA/cuda-alloc-free.fir
@@ -94,19 +94,4 @@ func.func @_QQalloc_char() attributes {fir.bindc_name = "alloc_char"} {
// CHECK: %[[BYTES_CONV:.*]] = fir.convert %[[BYTES]] : (index) -> i64
// CHECK: fir.call @_FortranACUFMemAlloc(%[[BYTES_CONV]], %c0{{.*}}, %{{.*}}, %{{.*}}) {cuf.data_attr = #cuf.cuda<device>} : (i64, i32, !fir.ref<i8>, i32) -> !fir.llvm_ptr<i8>
-
-func.func @_QQsetalloc() {
- %0 = cuf.alloc !fir.type<_QMm1Tdt1{a2:!fir.box<!fir.heap<!fir.array<?xf32>>>}> {bindc_name = "d1", data_attr = #cuf.cuda<managed>, uniq_name = "_QFEd1"} -> !fir.ref<!fir.type<_QMm1Tdt1{a2:!fir.box<!fir.heap<!fir.array<?xf32>>>}>>
- %1 = fir.coordinate_of %0, a2 : (!fir.ref<!fir.type<_QMm1Tdt1{a2:!fir.box<!fir.heap<!fir.array<?xf32>>>}>>) -> !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
- cuf.set_allocator_idx %1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>> {data_attr = #cuf.cuda<device>}
- return
-}
-
-// CHECK-LABEL: func.func @_QQsetalloc() {
-// CHECK: %[[DT:.*]] = fir.call @_FortranACUFMemAlloc
-// CHECK: %[[CONV:.*]] = fir.convert %[[DT]] : (!fir.llvm_ptr<i8>) -> !fir.ref<!fir.type<_QMm1Tdt1{a2:!fir.box<!fir.heap<!fir.array<?xf3...
[truncated]
|
wangzpgi
approved these changes
Sep 9, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The allocator index is set from the component genre #157731 . There is no more need of an operation to set it at a later point.