[clang] [CIR][X86] Add CIR codegen support for fpclass x86 builtins (PR #172813)

[Jianjian Guan via cfe-commits]

================
@@ -26,3 +26,42 @@ __m512bh test_mm512_undefined_pbh(void) {
   // OGCG: ret <32 x bfloat> zeroinitializer
   return _mm512_undefined_pbh();
 }
+
+__mmask32 test_mm512_mask_fpclass_pbh_mask(__mmask32 __U, __m512bh __A) {
+  // CIR-LABEL: _mm512_mask_fpclass_pbh_mask
+  // CIR: %[[A:.*]] = cir.call_llvm_intrinsic "x86.avx10.fpclass.bf16.512"
+  // CIR: %[[B:.*]] = cir.cast bitcast {{.*}} : !u32i -> !cir.vector<32 x !cir.bool>
----------------
jacquesguan wrote:

Addressed.

https://github.com/llvm/llvm-project/pull/172813