[clang] [llvm] [CIR][X86] Implement lowering for sqrt builtins (PR #169310)

[Andy Kaylor via llvm-commits]

================
@@ -781,9 +781,21 @@ mlir::Value CIRGenFunction::emitX86BuiltinExpr(unsigned builtinID,
   case X86::BI__builtin_ia32_sqrtsh_round_mask:
   case X86::BI__builtin_ia32_sqrtsd_round_mask:
   case X86::BI__builtin_ia32_sqrtss_round_mask:
+  case X86::BI__builtin_ia32_sqrtpd256:
+  case X86::BI__builtin_ia32_sqrtpd:
+  case X86::BI__builtin_ia32_sqrtps256:
+  case X86::BI__builtin_ia32_sqrtps:
+  case X86::BI__builtin_ia32_sqrtph256:
+  case X86::BI__builtin_ia32_sqrtph:
+    errorNYI("Unimplemented builtin");
+    return {};
   case X86::BI__builtin_ia32_sqrtph512:
   case X86::BI__builtin_ia32_sqrtps512:
-  case X86::BI__builtin_ia32_sqrtpd512:
+  case X86::BI__builtin_ia32_sqrtpd512: {
+    mlir::Location loc = getLoc(expr->getExprLoc());
+    mlir::Value arg = ops[0];
+    return builder.create<cir::SqrtOp>(loc, arg.getType(), arg).getResult();
----------------
andykaylor wrote:

```suggestion
    return cir::SqrtOp::create(builder, loc, arg.getType(), arg).getResult();
```
Co-pilot was wrong about this. Its suggested change is a format this is now deprecated.

https://github.com/llvm/llvm-project/pull/169310