[clang] [mlir] [CIR] Honor the Direct coercion offset in x86_64 callconv lowering (PR #220724)

[Adam Smith via cfe-commits]

================
@@ -87,13 +94,18 @@ struct ArgClassification {
     return kind == other.kind && coercedType == other.coercedType &&
            indirectAlign == other.indirectAlign &&
            signExtend == other.signExtend && canFlatten == other.canFlatten &&
-           byVal == other.byVal;
+           byVal == other.byVal && directOffset == other.directOffset;
   }
 
-  static ArgClassification getDirect(Type coerced = nullptr) {
+  static ArgClassification getDirect(unsigned offset, Type coerced = nullptr) {
----------------
adams381 wrote:

I think that looks pretty clean. I've gone with it.

https://github.com/llvm/llvm-project/pull/220724