[llvm] [InstCombine] Pull shuffles out of binops with splatted ops (PR #137948)

[Philip Reames via llvm-commits]

================
@@ -2271,6 +2271,27 @@ Instruction *InstCombinerImpl::foldVectorBinop(BinaryOperator &Inst) {
     }
   }
 
+  // Similar to the combine above, but handles the case for scalable vectors
+  // where both shuffle(V1, 0) and C are splats.
+  //
+  // Op(shuffle(V1, 0), (splat C)) -> shuffle(Op(V1, (splat C)), 0)
+  if (isa<ScalableVectorType>(Inst.getType()) &&
----------------
preames wrote:

Code structure wise: Remove the scalable vector check, move this *above* the other transform.  That way fixed vector splats go through this too, and we have a much higher chance of finding any bugs.

https://github.com/llvm/llvm-project/pull/137948