[llvm] [X86] Align f128 and i128 to 16 bytes when passing on x86-32 (PR #138092)

[Phoebe Wang via llvm-commits]

================
@@ -272,7 +272,7 @@ define <64 x i8> @allones_v64i8() nounwind {
 ;
 ; AVX512-LABEL: allones_v64i8:
 ; AVX512:       # %bb.0:
-; AVX512-NEXT:    vpternlogd $255, %zmm0, %zmm0, %zmm0
+; AVX512-NEXT:    vpternlogd {{.*#+}} zmm0 = -1
----------------
phoebewang wrote:

This is due to comment change for vpternlogd. Not related to the change here.

https://github.com/llvm/llvm-project/pull/138092