[llvm] [X86] Make ISD::ROTL/ROTR vector rotates legal on XOP+AVX512 targets (PR #184587)

[Simon Pilgrim via llvm-commits]

================
@@ -2337,6 +2337,12 @@ bool TargetLowering::SimplifyDemandedBits(
       if (SimplifyDemandedBits(Op1, DemandedAmtBits, DemandedElts, Known2, TLO,
                                Depth + 1))
         return true;
+
+      // rot*(x, 0) --> x
----------------
RKSimon wrote:

I'll take another look - but I was hitting a regression because we were folding the node to VROTLI before the second iteration, and I felt we should be trying to do this generically.

https://github.com/llvm/llvm-project/pull/184587