[CIR] Implement lowering of int-to-bool casts #132996
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
Lowering of int-to-bool casts had been left as NYI because the incubator implemented it by lowering to cir.cmp, which hasn't been upstreamed yet, but there is no reason this cast can't be lowered directly to LLVM's compare operation when we're lowering directly to the LLVM dialect. This change lowers the cast directly to an LLVM compare to zero.
andykaylor
requested review from
xlauko,
bcardosolopes,
erichkeane,
lanza,
AmrDeveloper and
dkolsen-pgi
llvmbot
added
clang
|
@llvm/pr-subscribers-clangir Author: Andy Kaylor (andykaylor) ChangesLowering of int-to-bool casts had been left as NYI because the incubator implemented it by lowering to cir.cmp, which hasn't been upstreamed yet, but there is no reason this cast can't be lowered directly to LLVM's compare operation when we're lowering directly to the LLVM dialect. This change lowers the cast directly to an LLVM compare to zero. Full diff: https://github.com/llvm/llvm-project/pull/132996.diff 3 Files Affected:
diff --git a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
index 30cbee48b4bdc..913c70716c365 100644
--- a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
+++ b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
@@ -385,13 +385,13 @@ mlir::LogicalResult CIRToLLVMCastOpLowering::matchAndRewrite(
break;
}
case cir::CastKind::int_to_bool: {
- assert(!cir::MissingFeatures::opCmp());
- mlir::Type dstType = castOp.getResult().getType();
- mlir::Type llvmDstType = getTypeConverter()->convertType(dstType);
- auto zeroBool = rewriter.create<mlir::LLVM::ConstantOp>(
- castOp.getLoc(), llvmDstType, mlir::BoolAttr::get(getContext(), false));
- rewriter.replaceOp(castOp, zeroBool);
- return castOp.emitError() << "NYI int_to_bool cast";
+ mlir::Value llvmSrcVal = adaptor.getOperands().front();
+ mlir::Value zeroInt = rewriter.create<mlir::LLVM::ConstantOp>(
+ castOp.getLoc(), llvmSrcVal.getType(),
+ mlir::IntegerAttr::get(llvmSrcVal.getType(), 0));
+ rewriter.replaceOpWithNewOp<mlir::LLVM::ICmpOp>(
+ castOp, mlir::LLVM::ICmpPredicate::ne, llvmSrcVal, zeroInt);
+ break;
}
case cir::CastKind::integral: {
mlir::Type srcType = castOp.getSrc().getType();
diff --git a/clang/test/CIR/CodeGen/int-to-bool.cpp b/clang/test/CIR/CodeGen/int-to-bool.cpp
new file mode 100644
index 0000000000000..1b915b643506f
--- /dev/null
+++ b/clang/test/CIR/CodeGen/int-to-bool.cpp
@@ -0,0 +1,74 @@
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -fclangir -emit-cir %s -o %t.cir
+// RUN: FileCheck --input-file=%t.cir %s -check-prefix=CIR
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -fclangir -emit-llvm %s -o %t-cir.ll
+// RUN: FileCheck --input-file=%t-cir.ll %s -check-prefix=LLVM
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -emit-llvm %s -o %t.ll
+// RUN: FileCheck --input-file=%t.ll %s -check-prefix=OGCG
+
+bool f1(unsigned char c) {
+ return c;
+}
+
+// CIR: cir.func @f1
+// CIR: cir.cast(int_to_bool, %{{.*}} : !u8i), !cir.bool
+
+// Note: The full zext/store/load/trunc sequence is checked here to show what
+// CIR is being lowered to. There's no need to check it for every function since
+// the lowering is the same for all of them.
+
+// LLVM: define i1 @f1
+// LLVM: %[[CMP:.*]] = icmp ne i8 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+// LLVM: store i8 %[[ZEXT]], ptr %{{.*}}
+// LLVM: %[[LOAD:.*]] = load i8, ptr %{{.*}}
+// LLVM: %[[TRUNC:.*]] = trunc i8 %[[LOAD]] to i1
+// LLVM: ret i1 %[[TRUNC]]
+
+// OGCG: define{{.*}} i1 @_Z2f1h
+// OGCG: %[[CMP:.*]] = icmp ne i8 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
+
+bool f2(short s) {
+ return s;
+}
+
+// CIR: cir.func @f2
+// CIR: cir.cast(int_to_bool, %{{.*}} : !s16i), !cir.bool
+
+// LLVM: define i1 @f2
+// LLVM: %[[CMP:.*]] = icmp ne i16 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+
+// OGCG: define{{.*}} i1 @_Z2f2s
+// OGCG: %[[CMP:.*]] = icmp ne i16 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
+
+bool f3(unsigned u) {
+ return u;
+}
+
+// CIR: cir.func @f3
+// CIR: cir.cast(int_to_bool, %{{.*}} : !u32i), !cir.bool
+
+// LLVM: define i1 @f3
+// LLVM: %[[CMP:.*]] = icmp ne i32 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+
+// OGCG: define{{.*}} i1 @_Z2f3j
+// OGCG: %[[CMP:.*]] = icmp ne i32 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
+
+bool f4(long l) {
+ return l;
+}
+
+// CIR: cir.func @f4
+// CIR: cir.cast(int_to_bool, %{{.*}} : !s64i), !cir.bool
+
+// LLVM: define i1 @f4
+// LLVM: %[[CMP:.*]] = icmp ne i64 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+
+// OGCG: define{{.*}} i1 @_Z2f4l
+// OGCG: %[[CMP:.*]] = icmp ne i64 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
diff --git a/clang/test/CIR/Lowering/cast.cir b/clang/test/CIR/Lowering/cast.cir
index b2d0a6d42eeb1..6842905dae6a4 100644
--- a/clang/test/CIR/Lowering/cast.cir
+++ b/clang/test/CIR/Lowering/cast.cir
@@ -47,6 +47,9 @@ module {
%21 = cir.load %20 : !cir.ptr<!s16i>, !s16i
%22 = cir.cast(integral, %21 : !s16i), !u64i
// CHECK: %[[TMP:[0-9]+]] = llvm.sext %{{[0-9]+}} : i16 to i64
+ %33 = cir.cast(int_to_bool, %arg1 : !s32i), !cir.bool
+ // CHECK: %[[#ZERO:]] = llvm.mlir.constant(0 : i32) : i32
+ // CHECK: %[[#CMP:]] = llvm.icmp "ne" %arg1, %[[#ZERO]] : i32
// Pointer casts.
cir.store %16, %6 : !s64i, !cir.ptr<!s64i>
|
|
@llvm/pr-subscribers-clang Author: Andy Kaylor (andykaylor) ChangesLowering of int-to-bool casts had been left as NYI because the incubator implemented it by lowering to cir.cmp, which hasn't been upstreamed yet, but there is no reason this cast can't be lowered directly to LLVM's compare operation when we're lowering directly to the LLVM dialect. This change lowers the cast directly to an LLVM compare to zero. Full diff: https://github.com/llvm/llvm-project/pull/132996.diff 3 Files Affected:
diff --git a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
index 30cbee48b4bdc..913c70716c365 100644
--- a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
+++ b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
@@ -385,13 +385,13 @@ mlir::LogicalResult CIRToLLVMCastOpLowering::matchAndRewrite(
break;
}
case cir::CastKind::int_to_bool: {
- assert(!cir::MissingFeatures::opCmp());
- mlir::Type dstType = castOp.getResult().getType();
- mlir::Type llvmDstType = getTypeConverter()->convertType(dstType);
- auto zeroBool = rewriter.create<mlir::LLVM::ConstantOp>(
- castOp.getLoc(), llvmDstType, mlir::BoolAttr::get(getContext(), false));
- rewriter.replaceOp(castOp, zeroBool);
- return castOp.emitError() << "NYI int_to_bool cast";
+ mlir::Value llvmSrcVal = adaptor.getOperands().front();
+ mlir::Value zeroInt = rewriter.create<mlir::LLVM::ConstantOp>(
+ castOp.getLoc(), llvmSrcVal.getType(),
+ mlir::IntegerAttr::get(llvmSrcVal.getType(), 0));
+ rewriter.replaceOpWithNewOp<mlir::LLVM::ICmpOp>(
+ castOp, mlir::LLVM::ICmpPredicate::ne, llvmSrcVal, zeroInt);
+ break;
}
case cir::CastKind::integral: {
mlir::Type srcType = castOp.getSrc().getType();
diff --git a/clang/test/CIR/CodeGen/int-to-bool.cpp b/clang/test/CIR/CodeGen/int-to-bool.cpp
new file mode 100644
index 0000000000000..1b915b643506f
--- /dev/null
+++ b/clang/test/CIR/CodeGen/int-to-bool.cpp
@@ -0,0 +1,74 @@
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -fclangir -emit-cir %s -o %t.cir
+// RUN: FileCheck --input-file=%t.cir %s -check-prefix=CIR
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -fclangir -emit-llvm %s -o %t-cir.ll
+// RUN: FileCheck --input-file=%t-cir.ll %s -check-prefix=LLVM
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -emit-llvm %s -o %t.ll
+// RUN: FileCheck --input-file=%t.ll %s -check-prefix=OGCG
+
+bool f1(unsigned char c) {
+ return c;
+}
+
+// CIR: cir.func @f1
+// CIR: cir.cast(int_to_bool, %{{.*}} : !u8i), !cir.bool
+
+// Note: The full zext/store/load/trunc sequence is checked here to show what
+// CIR is being lowered to. There's no need to check it for every function since
+// the lowering is the same for all of them.
+
+// LLVM: define i1 @f1
+// LLVM: %[[CMP:.*]] = icmp ne i8 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+// LLVM: store i8 %[[ZEXT]], ptr %{{.*}}
+// LLVM: %[[LOAD:.*]] = load i8, ptr %{{.*}}
+// LLVM: %[[TRUNC:.*]] = trunc i8 %[[LOAD]] to i1
+// LLVM: ret i1 %[[TRUNC]]
+
+// OGCG: define{{.*}} i1 @_Z2f1h
+// OGCG: %[[CMP:.*]] = icmp ne i8 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
+
+bool f2(short s) {
+ return s;
+}
+
+// CIR: cir.func @f2
+// CIR: cir.cast(int_to_bool, %{{.*}} : !s16i), !cir.bool
+
+// LLVM: define i1 @f2
+// LLVM: %[[CMP:.*]] = icmp ne i16 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+
+// OGCG: define{{.*}} i1 @_Z2f2s
+// OGCG: %[[CMP:.*]] = icmp ne i16 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
+
+bool f3(unsigned u) {
+ return u;
+}
+
+// CIR: cir.func @f3
+// CIR: cir.cast(int_to_bool, %{{.*}} : !u32i), !cir.bool
+
+// LLVM: define i1 @f3
+// LLVM: %[[CMP:.*]] = icmp ne i32 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+
+// OGCG: define{{.*}} i1 @_Z2f3j
+// OGCG: %[[CMP:.*]] = icmp ne i32 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
+
+bool f4(long l) {
+ return l;
+}
+
+// CIR: cir.func @f4
+// CIR: cir.cast(int_to_bool, %{{.*}} : !s64i), !cir.bool
+
+// LLVM: define i1 @f4
+// LLVM: %[[CMP:.*]] = icmp ne i64 %4, 0
+// LLVM: %[[ZEXT:.*]] = zext i1 %[[CMP]] to i8
+
+// OGCG: define{{.*}} i1 @_Z2f4l
+// OGCG: %[[CMP:.*]] = icmp ne i64 %{{.*}}, 0
+// OGCG: ret i1 %[[CMP]]
diff --git a/clang/test/CIR/Lowering/cast.cir b/clang/test/CIR/Lowering/cast.cir
index b2d0a6d42eeb1..6842905dae6a4 100644
--- a/clang/test/CIR/Lowering/cast.cir
+++ b/clang/test/CIR/Lowering/cast.cir
@@ -47,6 +47,9 @@ module {
%21 = cir.load %20 : !cir.ptr<!s16i>, !s16i
%22 = cir.cast(integral, %21 : !s16i), !u64i
// CHECK: %[[TMP:[0-9]+]] = llvm.sext %{{[0-9]+}} : i16 to i64
+ %33 = cir.cast(int_to_bool, %arg1 : !s32i), !cir.bool
+ // CHECK: %[[#ZERO:]] = llvm.mlir.constant(0 : i32) : i32
+ // CHECK: %[[#CMP:]] = llvm.icmp "ne" %arg1, %[[#ZERO]] : i32
// Pointer casts.
cir.store %16, %6 : !s64i, !cir.ptr<!s64i>
|
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/53/builds/14526 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/172/builds/11341 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/50/builds/12042 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/29/builds/12294 Here is the relevant piece of the build log for the reference |
erichkeane
reviewed
Apr 1, 2025
| @@ -385,13 +385,13 @@ mlir::LogicalResult CIRToLLVMCastOpLowering::matchAndRewrite( | |||
| break; | |||
| } | |||
| case cir::CastKind::int_to_bool: { | |||
| assert(!cir::MissingFeatures::opCmp()); | |||
There was a problem hiding this comment.
Does the corresponding MissingFeatures need to be removed? Or does that have 'more' uses?
There was a problem hiding this comment.
I left it because the feature wasn't implemented, but it looks like it didn't have any other uses. It's being removed in #133159
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.
Lowering of int-to-bool casts had been left as NYI because the incubator implemented it by lowering to cir.cmp, which hasn't been upstreamed yet, but there is no reason this cast can't be lowered directly to LLVM's compare operation when we're lowering directly to the LLVM dialect.
This change lowers the cast directly to an LLVM compare to zero.