[llvm] Revert "AMDGPU: Don't avoid clamp of bit shift in BFE pattern (#115372)" (PR #116091)

[Matt Arsenault via llvm-commits]

================
@@ -150,39 +150,22 @@ define i32 @bzhi32_c4_commutative(i32 %val, i32 %numlowbits) nounwind {
 ; ---------------------------------------------------------------------------- ;
 
 define i32 @bzhi32_d0(i32 %val, i32 %numlowbits) nounwind {
-; GCN-LABEL: bzhi32_d0:
----------------
arsenm wrote:

Should keep the test, and also add a variant with canonical IR 

https://github.com/llvm/llvm-project/pull/116091