[llvm] WIP:[LV] Support strided memory accesses with a stride of -1 (PR #128718)

[Mel Chen via llvm-commits]

================
@@ -1308,6 +1309,18 @@ class LoopVectorizationCostModel {
            (SI && TTI.isLegalMaskedScatter(Ty, Align));
   }
 
+  /// Returns true if the target machine can represent \p V as a strided load
+  /// or store operation.
+  bool isLegalStridedLoadStore(Value *V, ElementCount VF) {
----------------
Mel-Chen wrote:

The prototype of isLegalGatherOrScatter is:
```
bool isLegalGatherOrScatter(Value *V, ElementCount VF)
```
And the prototype of isLegalMaskedLoad is:
```
bool isLegalMaskedLoad(Type *DataType, Value *Ptr, Align Alignment)
```
isLegalStridedLoadStore is the same as isLegalGatherOrScatter now.
Do you prefer the prototype of isLegalMaskedLoad?

https://github.com/llvm/llvm-project/pull/128718