[mlir][LLVMIR] Allow !llvm.ptr<ptr> operands in atomicrmw xchg op.
Previously, llvm.atomicrmw only allowed operands that are pointers to
LLVM floats or integers. However, according to the LLVM IR Language
Reference, that op allows pointer to pointer operands in its `xchg`
mode. This patch allows those operands also in MLIR's LLVM dialect and
adapts the tests accordingly.
Reviewed By: gysit
Differential Revision: https://reviews.llvm.org/D153747
diff --git a/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td b/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
index a106592..8536cbf 100644
--- a/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
+++ b/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
@@ -1832,7 +1832,7 @@
// Atomic operations.
//
-def LLVM_AtomicRMWType : AnyTypeOf<[LLVM_AnyFloat, AnyInteger]>;
+def LLVM_AtomicRMWType : AnyTypeOf<[LLVM_AnyFloat, LLVM_AnyPointer, AnyInteger]>;
def LLVM_AtomicRMWOp : LLVM_MemAccessOpBase<"atomicrmw", [
TypesMatchWith<"result #0 and operand #1 have the same type",
diff --git a/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp b/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
index 7bbdb67..b24e2ca 100644
--- a/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
+++ b/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
@@ -2450,7 +2450,7 @@
if (!mlir::LLVM::isCompatibleFloatingPointType(valType))
return emitOpError("expected LLVM IR floating point type");
} else if (getBinOp() == AtomicBinOp::xchg) {
- if (!isTypeCompatibleWithAtomicOp(valType, /*isPointerTypeAllowed=*/false))
+ if (!isTypeCompatibleWithAtomicOp(valType, /*isPointerTypeAllowed=*/true))
return emitOpError("unexpected LLVM IR type for 'xchg' bin_op");
} else {
auto intType = llvm::dyn_cast<IntegerType>(valType);
diff --git a/mlir/test/Dialect/LLVMIR/invalid.mlir b/mlir/test/Dialect/LLVMIR/invalid.mlir
index 787889f..c88ed03 100644
--- a/mlir/test/Dialect/LLVMIR/invalid.mlir
+++ b/mlir/test/Dialect/LLVMIR/invalid.mlir
@@ -633,7 +633,7 @@
// -----
func.func @atomicrmw_expected_ptr(%f32 : f32) {
- // expected-error@+1 {{operand #0 must be LLVM pointer to floating point LLVM type or integer}}
+ // expected-error@+1 {{operand #0 must be LLVM pointer to floating point LLVM type or LLVM pointer type or integer}}
%0 = "llvm.atomicrmw"(%f32, %f32) {bin_op=11, ordering=1} : (f32, f32) -> f32
llvm.return
}
