[llvm] [AMDGPU] Add pattern to select scalar ops for fshr with uniform operands (PR #165295)

[Jay Foad via llvm-commits]

================
@@ -207,6 +207,17 @@ class DivergentBinFrag<SDPatternOperator Op> : PatFrag <
   let GISelPredicateCode = [{return true;}];
 }
 
+class DivergentTernaryFrag<SDPatternOperator Op> : PatFrag <
----------------
jayfoad wrote:

Can you check whether this is reqlly required? I.e. do any tests fail if you remove this from the V_ALIGNBIT patterns?

https://github.com/llvm/llvm-project/pull/165295