Skip to content
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

Jump to bottom

[CIR][FlattenCFG] Let results of CIR ScopeOp forwarded during FlattenCFG #1147

Merged
merged 1 commit into from
Nov 23, 2024
+36 −6
Merged

[CIR][FlattenCFG] Let results of CIR ScopeOp forwarded during FlattenCFG #1147

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
9 changes: 3 additions & 6 deletions clang/lib/CIR/Dialect/Transforms/FlattenCFG.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -135,13 +135,10 @@ class CIRScopeOpFlattening : public mlir::OpRewritePattern<cir::ScopeOp> {
// Split the current block before the ScopeOp to create the inlining
// point.
auto *currentBlock = rewriter.getInsertionBlock();
auto *remainingOpsBlock =
mlir::Block *continueBlock =
rewriter.splitBlock(currentBlock, rewriter.getInsertionPoint());
mlir::Block *continueBlock;
if (scopeOp.getNumResults() == 0)
continueBlock = remainingOpsBlock;
else
llvm_unreachable("NYI");
if (scopeOp.getNumResults() > 0)
continueBlock->addArguments(scopeOp.getResultTypes(), loc);
bcardosolopes marked this conversation as resolved.
Show resolved Hide resolved

// Inline body region.
auto *beforeBody = &scopeOp.getRegion().front();
Expand Down
33 changes: 33 additions & 0 deletions clang/test/CIR/CodeGen/fullexpr.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,10 @@
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir %s -o %t.cir
// RUN: FileCheck --input-file=%t.cir %s
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir-flat %s -o %t.cir.flat
// RUN: FileCheck --check-prefix=FLAT --input-file=%t.cir.flat %s
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-linux-gnu -fclangir -emit-llvm -o - %s \
bcardosolopes marked this conversation as resolved.
Show resolved Hide resolved
// RUN: | opt -S -passes=instcombine,mem2reg,simplifycfg -o %t.ll
// RUN: FileCheck --check-prefix=LLVM --input-file=%t.ll %s

int go(int const& val);

Expand All @@ -18,3 +23,31 @@ int go1() {
// CHECK-NEXT: cir.yield %[[#RValTmp]] : !s32i
// CHECK-NEXT: }
// CHECK-NEXT: cir.store %[[#RVal]], %[[#XAddr]] : !s32i, !cir.ptr<!s32i>

// FLAT: cir.func @_Z3go1v() -> !s32i
// FLAT: %[[#TmpAddr:]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["ref.tmp0", init] {alignment = 4 : i64}
// FLAT: %[[#XAddr:]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["x", init] {alignment = 4 : i64}
// FLAT: cir.br ^[[before_body:.*]]{{ loc.*}}
// FLAT-NEXT: ^[[before_body]]: // pred: ^bb0
// FLAT-NEXT: %[[#One:]] = cir.const #cir.int<1> : !s32i
// FLAT-NEXT: cir.store %[[#One]], %[[#TmpAddr]] : !s32i, !cir.ptr<!s32i>
// FLAT-NEXT: %[[#RValTmp:]] = cir.call @_Z2goRKi(%[[#TmpAddr]]) : (!cir.ptr<!s32i>) -> !s32i
// FLAT-NEXT: cir.br ^[[continue_block:.*]](%[[#RValTmp]] : !s32i) {{loc.*}}
// FLAT-NEXT: ^[[continue_block]](%[[#BlkArgRval:]]: !s32i {{loc.*}}): // pred: ^[[before_body]]
// FLAT-NEXT: cir.store %[[#BlkArgRval]], %[[#XAddr]] : !s32i, !cir.ptr<!s32i>

// LLVM-LABEL: @_Z3go1v()
// LLVM-NEXT: %[[#TmpAddr:]] = alloca i32, i64 1, align 4
// LLVM: br label %[[before_body:[0-9]+]]
// LLVM: [[before_body]]:
// LLVM-NEXT: store i32 1, ptr %[[#TmpAddr]], align 4
// LLVM-NEXT: %[[#RValTmp:]] = call i32 @_Z2goRKi(ptr %[[#TmpAddr]])
// LLVM-NEXT: br label %[[continue_block:[0-9]+]]

// LLVM: [[continue_block]]:
// LLVM-NEXT: [[PHI:%.*]] = phi i32 [ %[[#RValTmp]], %[[before_body]] ]
// LLVM: store i32 [[PHI]], ptr [[TMP0:%.*]], align 4
// LLVM: [[TMP1:%.*]] = load i32, ptr [[TMP0]], align 4
// LLVM: store i32 [[TMP1]], ptr [[TMP2:%.*]], align 4
// LLVM: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
// LLVM: ret i32 [[TMP3]]