[clang] [CIR][AMDGPU] Add support for AMDGCN permlane builtins (PR #197526)

[Rana Pratap Reddy via cfe-commits]

================
@@ -211,12 +211,16 @@ CIRGenFunction::emitAMDGPUBuiltinExpr(unsigned builtinId,
     return mlir::Value{};
   }
   case AMDGPU::BI__builtin_amdgcn_permlane16:
-  case AMDGPU::BI__builtin_amdgcn_permlanex16:
+  case AMDGPU::BI__builtin_amdgcn_permlanex16: {
+    llvm::StringRef intrinsicName =
+        builtinId == AMDGPU::BI__builtin_amdgcn_permlane16
+            ? "amdgcn.permlane16"
+            : "amdgcn.permlanex16";
+    return emitBuiltinWithOneOverloadedType<6>(expr, intrinsicName).getValue();
+  }
   case AMDGPU::BI__builtin_amdgcn_permlane64: {
----------------
ranapratap55 wrote:

```suggestion
  case AMDGPU::BI__builtin_amdgcn_permlane64:
      return emitBuiltinWithOneOverloadedType<1>(expr, "amdgcn.permlane64").getValue();
```

https://github.com/llvm/llvm-project/pull/197526