[llvm] [AArch64] Fix heuristics for folding "lsl" into load/store ops. (PR #86894)

[David Green via llvm-commits]

================
@@ -1298,7 +1297,6 @@ def TuneNeoverseV1 : SubtargetFeature<"neoversev1", "ARMProcFamily", "NeoverseV1
                                       "Neoverse V1 ARM processors", [
                                       FeatureFuseAES,
                                       FeatureFuseAdrpAdd,
-                                      FeatureAddrLSLFast,
----------------
davemgreen wrote:

+FeatureAddrLSLSlow14

https://github.com/llvm/llvm-project/pull/86894