[llvm] [AMDGPU] Add v2i32 V_BFI combines for andn2 and orn2 (PR #196325)

[Harrison Hao via llvm-commits]

================
@@ -2735,6 +2735,18 @@ def : AMDGPUPatIgnoreCopies <
               (i32 (EXTRACT_SUBREG VReg_64:$z, sub1))), sub1)
 >;
 
+// v2i32 version
+def : AMDGPUPatIgnoreCopies <
----------------
harrisonGPU wrote:

Thanks, I have updated, I used multiclass to implement both v2i32 and i64.

https://github.com/llvm/llvm-project/pull/196325