[CIR] Upstream support ComplexType as return type #164072
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
AmrDeveloper
requested review from
andykaylor,
bcardosolopes,
lanza and
xlauko
as code owners
llvmbot
added
clang
|
@llvm/pr-subscribers-clangir Author: Amr Hesham (AmrDeveloper) ChangesUpstream support ComplexType as a function return type Issue #141365 Full diff: https://github.com/llvm/llvm-project/pull/164072.diff 2 Files Affected:
diff --git a/clang/lib/CIR/CodeGen/CIRGenStmt.cpp b/clang/lib/CIR/CodeGen/CIRGenStmt.cpp
index 5ba64ddb85272..ad8c4d004654f 100644
--- a/clang/lib/CIR/CodeGen/CIRGenStmt.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenStmt.cpp
@@ -475,8 +475,8 @@ mlir::LogicalResult CIRGenFunction::emitReturnStmt(const ReturnStmt &s) {
}
break;
case cir::TEK_Complex:
- getCIRGenModule().errorNYI(s.getSourceRange(),
- "complex function return type");
+ emitComplexExprIntoLValue(rv, makeAddrLValue(returnValue, rv->getType()),
+ /*isInit=*/true);
break;
case cir::TEK_Aggregate:
assert(!cir::MissingFeatures::aggValueSlotGC());
diff --git a/clang/test/CIR/CodeGen/complex.cpp b/clang/test/CIR/CodeGen/complex.cpp
index 083d4383e1f7d..931dc175d59ad 100644
--- a/clang/test/CIR/CodeGen/complex.cpp
+++ b/clang/test/CIR/CodeGen/complex.cpp
@@ -1360,6 +1360,30 @@ void complex_type_argument() {
// OGCG: %[[TMP_ARG:.*]] = load <2 x float>, ptr %[[ARG_ADDR]], align 4
// OGCG: call void @_Z22complex_type_parameterCf(<2 x float> noundef %[[TMP_ARG]])
+float _Complex complex_type_return_type() {
+ return { 1.0f, 2.0f };
+}
+
+// CIR: %[[RET_ADDR:.*]] = cir.alloca !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>, ["__retval"]
+// CIR: %[[RET_VAL:.*]] = cir.const #cir.const_complex<#cir.fp<1.000000e+00> : !cir.float, #cir.fp<2.000000e+00> : !cir.float> : !cir.complex<!cir.float>
+// CIR: cir.store{{.*}} %[[RET_VAL]], %[[RET_ADDR]] : !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>
+// CIR: %[[TMP_RET:.*]] = cir.load %[[RET_ADDR]] : !cir.ptr<!cir.complex<!cir.float>>, !cir.complex<!cir.float>
+// CIR: cir.return %[[TMP_RET]] : !cir.complex<!cir.float>
+
+// TODO(CIR): the difference between the CIR LLVM and OGCG is because the lack of calling convention lowering,
+// LLVM: %[[RET_ADDR:.*]] = alloca { float, float }, i64 1, align 4
+// LLVM: store { float, float } { float 1.000000e+00, float 2.000000e+00 }, ptr %[[RET_ADDR]], align 4
+// LLVM: %[[TMP_RET:.*]] = load { float, float }, ptr %[[RET_ADDR]], align 4
+// LLVM: ret { float, float } %[[TMP_RET]]
+
+// OGCG: %[[RET_ADDR:.*]] = alloca { float, float }, align 4
+// OGCG: %[[RET_VAL_REAL:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[RET_ADDR]], i32 0, i32 0
+// OGCG: %[[RET_VAL_IMAG:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[RET_ADDR]], i32 0, i32 1
+// OGCG: store float 1.000000e+00, ptr %[[RET_VAL_REAL]], align 4
+// OGCG: store float 2.000000e+00, ptr %[[RET_VAL_IMAG]], align 4
+// OGCG: %[[TMP_RET:.*]] = load <2 x float>, ptr %[[RET_ADDR]], align 4
+// OGCG: ret <2 x float> %[[TMP_RET]]
+
void real_on_scalar_bool() {
bool a;
bool b = __real__ a;
|
|
@llvm/pr-subscribers-clang Author: Amr Hesham (AmrDeveloper) ChangesUpstream support ComplexType as a function return type Issue #141365 Full diff: https://github.com/llvm/llvm-project/pull/164072.diff 2 Files Affected:
diff --git a/clang/lib/CIR/CodeGen/CIRGenStmt.cpp b/clang/lib/CIR/CodeGen/CIRGenStmt.cpp
index 5ba64ddb85272..ad8c4d004654f 100644
--- a/clang/lib/CIR/CodeGen/CIRGenStmt.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenStmt.cpp
@@ -475,8 +475,8 @@ mlir::LogicalResult CIRGenFunction::emitReturnStmt(const ReturnStmt &s) {
}
break;
case cir::TEK_Complex:
- getCIRGenModule().errorNYI(s.getSourceRange(),
- "complex function return type");
+ emitComplexExprIntoLValue(rv, makeAddrLValue(returnValue, rv->getType()),
+ /*isInit=*/true);
break;
case cir::TEK_Aggregate:
assert(!cir::MissingFeatures::aggValueSlotGC());
diff --git a/clang/test/CIR/CodeGen/complex.cpp b/clang/test/CIR/CodeGen/complex.cpp
index 083d4383e1f7d..931dc175d59ad 100644
--- a/clang/test/CIR/CodeGen/complex.cpp
+++ b/clang/test/CIR/CodeGen/complex.cpp
@@ -1360,6 +1360,30 @@ void complex_type_argument() {
// OGCG: %[[TMP_ARG:.*]] = load <2 x float>, ptr %[[ARG_ADDR]], align 4
// OGCG: call void @_Z22complex_type_parameterCf(<2 x float> noundef %[[TMP_ARG]])
+float _Complex complex_type_return_type() {
+ return { 1.0f, 2.0f };
+}
+
+// CIR: %[[RET_ADDR:.*]] = cir.alloca !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>, ["__retval"]
+// CIR: %[[RET_VAL:.*]] = cir.const #cir.const_complex<#cir.fp<1.000000e+00> : !cir.float, #cir.fp<2.000000e+00> : !cir.float> : !cir.complex<!cir.float>
+// CIR: cir.store{{.*}} %[[RET_VAL]], %[[RET_ADDR]] : !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>>
+// CIR: %[[TMP_RET:.*]] = cir.load %[[RET_ADDR]] : !cir.ptr<!cir.complex<!cir.float>>, !cir.complex<!cir.float>
+// CIR: cir.return %[[TMP_RET]] : !cir.complex<!cir.float>
+
+// TODO(CIR): the difference between the CIR LLVM and OGCG is because the lack of calling convention lowering,
+// LLVM: %[[RET_ADDR:.*]] = alloca { float, float }, i64 1, align 4
+// LLVM: store { float, float } { float 1.000000e+00, float 2.000000e+00 }, ptr %[[RET_ADDR]], align 4
+// LLVM: %[[TMP_RET:.*]] = load { float, float }, ptr %[[RET_ADDR]], align 4
+// LLVM: ret { float, float } %[[TMP_RET]]
+
+// OGCG: %[[RET_ADDR:.*]] = alloca { float, float }, align 4
+// OGCG: %[[RET_VAL_REAL:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[RET_ADDR]], i32 0, i32 0
+// OGCG: %[[RET_VAL_IMAG:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[RET_ADDR]], i32 0, i32 1
+// OGCG: store float 1.000000e+00, ptr %[[RET_VAL_REAL]], align 4
+// OGCG: store float 2.000000e+00, ptr %[[RET_VAL_IMAG]], align 4
+// OGCG: %[[TMP_RET:.*]] = load <2 x float>, ptr %[[RET_ADDR]], align 4
+// OGCG: ret <2 x float> %[[TMP_RET]]
+
void real_on_scalar_bool() {
bool a;
bool b = __real__ a;
|
Lukacma
pushed a commit
to Lukacma/llvm-project
that referenced
this pull request
Oct 29, 2025
Upstream support ComplexType as a function return type Issue llvm#141365
aokblast
pushed a commit
to aokblast/llvm-project
that referenced
this pull request
Oct 30, 2025
Upstream support ComplexType as a function return type Issue llvm#141365
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Upstream support ComplexType as a function return type
Issue #141365