[clang] [CIR][AArch64] Lower NEON vrsra_n intrinsics (PR #191129)

[Andrzej Warzyński via cfe-commits]

================
@@ -2275,12 +2296,29 @@ CIRGenFunction::emitAArch64BuiltinExpr(unsigned builtinID, const CallExpr *expr,
   case NEON::BI__builtin_neon_vqshld_n_s64:
   case NEON::BI__builtin_neon_vrshrd_n_u64:
   case NEON::BI__builtin_neon_vrshrd_n_s64:
-  case NEON::BI__builtin_neon_vrsrad_n_u64:
-  case NEON::BI__builtin_neon_vrsrad_n_s64:
     cgm.errorNYI(expr->getSourceRange(),
                  std::string("unimplemented AArch64 builtin call: ") +
                      getContext().BuiltinInfo.getName(builtinID));
     return mlir::Value{};
+  case NEON::BI__builtin_neon_vrsrad_n_u64:
+  case NEON::BI__builtin_neon_vrsrad_n_s64: {
+    cir::IntType int64Type = builtinID == NEON::BI__builtin_neon_vrsrad_n_u64
+                                 ? builder.getUInt64Ty()
+                                 : builder.getSInt64Ty();
+    ops[1] = builder.createBitcast(ops[1], int64Type);
+    ops[2] = builder.createNeg(ops[2]);
+    const StringRef intrName = builtinID == NEON::BI__builtin_neon_vrsrad_n_u64
+                                   ? "aarch64.neon.urshl"
+                                   : "aarch64.neon.srshl";
+
+    llvm::SmallVector<mlir::Value, 2> args = {
+        ops[1], builder.createIntCast(ops[2], builder.getSInt64Ty())};
+    ops[1] =
+        emitNeonCall(cgm, builder, {int64Type, builder.getSInt64Ty()}, args,
+                     intrName, int64Type, getLoc(expr->getExprLoc()));
----------------
banach-space wrote:

This can be simplified as (no extra logic from `emitNeonCall` is required here):
```suggestion
    ops[1] = builder.emitIntrinsicCallOp(loc, intrName, int64Type, args);
```

https://github.com/llvm/llvm-project/pull/191129