[Mlir-commits] [mlir] [mlir][EmitC] Expand the MemRefToEmitC pass - Lowering `CopyOp` (PR #151206)

[Paul Kirth]

================
@@ -8,38 +8,24 @@ func.func @copying(%arg0 : memref<9x4x5x7xf32>, %arg1 : memref<9x4x5x7xf32>) {
 
 // NOCPP: module {
 // NOCPP-NEXT:  emitc.include <"string.h">
-// NOCPP-LABEL:  copying
-// NOCPP-SAME: %[[arg0:.*]]: memref<9x4x5x7xf32>, %[[arg1:.*]]: memref<9x4x5x7xf32>
-// NOCPP-NEXT: %0 = builtin.unrealized_conversion_cast %arg1 : memref<9x4x5x7xf32> to !emitc.array<9x4x5x7xf32>
-// NOCPP-NEXT: %1 = builtin.unrealized_conversion_cast %arg0 : memref<9x4x5x7xf32> to !emitc.array<9x4x5x7xf32>
-// NOCPP-NEXT: %2 = "emitc.constant"() <{value = 0 : index}> : () -> index
-// NOCPP-NEXT: %3 = emitc.subscript %1[%2, %2, %2, %2] : (!emitc.array<9x4x5x7xf32>, index, index, index, index) -> !emitc.lvalue<f32>
-// NOCPP-NEXT: %4 = emitc.apply "&"(%3) : (!emitc.lvalue<f32>) -> !emitc.ptr<f32>
-// NOCPP-NEXT: %5 = emitc.subscript %0[%2, %2, %2, %2] : (!emitc.array<9x4x5x7xf32>, index, index, index, index) -> !emitc.lvalue<f32>
-// NOCPP-NEXT: %6 = emitc.apply "&"(%5) : (!emitc.lvalue<f32>) -> !emitc.ptr<f32>
-// NOCPP-NEXT: %7 = emitc.call_opaque "sizeof"() {args = [f32]} : () -> !emitc.size_t
-// NOCPP-NEXT: %8 = "emitc.constant"() <{value = 1260 : index}> : () -> index
-// NOCPP-NEXT: %9 = emitc.mul %7, %8 : (!emitc.size_t, index) -> !emitc.size_t
-// NOCPP-NEXT: emitc.call_opaque "memcpy"(%6, %4, %9) : (!emitc.ptr<f32>, !emitc.ptr<f32>, !emitc.size_t) -> ()
-// NOCPP-NEXT:    return
-// NOCPP-NEXT:  }
-// NOCPP-NEXT:}
+
 
 // CPP: module {
 // CPP-NEXT:  emitc.include <"cstring">
----------------
ilovepi wrote:

```suggestion
// CPP:  emitc.include <"cstring">
```

https://github.com/llvm/llvm-project/pull/151206