Skip to content

[CIR] Upstream support for lowering cir.switch to LLVM #140425

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
May 22, 2025
Merged

[CIR] Upstream support for lowering cir.switch to LLVM #140425

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
384ec29
Added lowering logic for switch and appropriate tests
Andres-Salamanca May 18, 2025
602f0ca
Apply review
Andres-Salamanca May 20, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 32 additions & 0 deletions clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1038,6 +1038,37 @@ mlir::LogicalResult CIRToLLVMGlobalOpLowering::matchAndRewrite(
return mlir::success();
}

mlir::LogicalResult CIRToLLVMSwitchFlatOpLowering::matchAndRewrite(
cir::SwitchFlatOp op, OpAdaptor adaptor,
mlir::ConversionPatternRewriter &rewriter) const {

llvm::SmallVector<mlir::APInt, 8> caseValues;
if (op.getCaseValues()) {
Copy link
Contributor

@andykaylor andykaylor May 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this if necessary? I think it there are no case values we just won't enter the for loop below.

for (mlir::Attribute val : op.getCaseValues()) {
auto intAttr = dyn_cast<cir::IntAttr>(val);
Copy link
Contributor

@andykaylor andykaylor May 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
auto intAttr = dyn_cast<cir::IntAttr>(val);
auto intAttr = cast<cir::IntAttr>(val);

Make this assert.

caseValues.push_back(intAttr.getValue());
}
}

llvm::SmallVector<mlir::Block *, 8> caseDestinations;
llvm::SmallVector<mlir::ValueRange, 8> caseOperands;

for (mlir::Block *x : op.getCaseDestinations()) {
Copy link
Contributor

@andykaylor andykaylor May 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Don't need braces for loops with single-line bodies.

caseDestinations.push_back(x);
}

for (mlir::OperandRange x : op.getCaseOperands()) {
Copy link
Contributor

@andykaylor andykaylor May 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No braces

caseOperands.push_back(x);
}

// Set switch op to branch to the newly created blocks.
rewriter.setInsertionPoint(op);
rewriter.replaceOpWithNewOp<mlir::LLVM::SwitchOp>(
op, adaptor.getCondition(), op.getDefaultDestination(),
op.getDefaultOperands(), caseValues, caseDestinations, caseOperands);
return mlir::success();
}

mlir::LogicalResult CIRToLLVMUnaryOpLowering::matchAndRewrite(
cir::UnaryOp op, OpAdaptor adaptor,
mlir::ConversionPatternRewriter &rewriter) const {
Expand Down Expand Up @@ -1641,6 +1672,7 @@ void ConvertCIRToLLVMPass::runOnOperation() {
CIRToLLVMGetGlobalOpLowering,
CIRToLLVMGetMemberOpLowering,
CIRToLLVMSelectOpLowering,
CIRToLLVMSwitchFlatOpLowering,
CIRToLLVMShiftOpLowering,
CIRToLLVMStackSaveOpLowering,
CIRToLLVMStackRestoreOpLowering,
Expand Down
10 changes: 10 additions & 0 deletions clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -149,6 +149,16 @@ class CIRToLLVMFuncOpLowering : public mlir::OpConversionPattern<cir::FuncOp> {
mlir::ConversionPatternRewriter &) const override;
};

class CIRToLLVMSwitchFlatOpLowering
: public mlir::OpConversionPattern<cir::SwitchFlatOp> {
public:
using mlir::OpConversionPattern<cir::SwitchFlatOp>::OpConversionPattern;

mlir::LogicalResult
matchAndRewrite(cir::SwitchFlatOp op, OpAdaptor,
mlir::ConversionPatternRewriter &) const override;
};

class CIRToLLVMGetGlobalOpLowering
: public mlir::OpConversionPattern<cir::GetGlobalOp> {
public:
Expand Down
Loading
Loading