[Mlir-commits] [mlir] [mlir][llvm] Add support for memset.inline (PR #115711)

[Tobias Gysi]

================
@@ -21,6 +21,25 @@ llvm.func @memset() -> i32 {
 
 // -----
 
+// CHECK-LABEL: llvm.func @memset_inline
+llvm.func @memset_inline() -> i32 {
+  // CHECK-DAG: %[[ALLOCA_LEN:.*]] = llvm.mlir.constant(1 : i32) : i32
+  // CHECK-DAG: %[[ALLOCA:.*]] = llvm.alloca %[[ALLOCA_LEN]] x i32
+  // CHECK-DAG: %[[MEMSET_VALUE:.*]] = llvm.mlir.constant(42 : i8) : i8
+  // After SROA, only one i32 will be actually used, so only 4 bytes will be set.
+  %0 = llvm.mlir.constant(1 : i32) : i32
+  %1 = llvm.alloca %0 x !llvm.array<10 x i32> {alignment = 8 : i64} : (i32) -> !llvm.ptr
+  %memset_value = llvm.mlir.constant(42 : i8) : i8
+  // 16 bytes means it will span over the first 4 i32 entries
----------------
gysit wrote:

```suggestion
  // 16 bytes means it will span over the first 4 i32 entries.
```
ultra nit:

https://github.com/llvm/llvm-project/pull/115711