[llvm] [RISCV] Improve performCONCAT_VECTORCombine stride matching (PR #68726)

[Michael Maitland via llvm-commits]

================
@@ -13821,6 +13822,56 @@ static SDValue performCONCAT_VECTORSCombine(SDNode *N, SelectionDAG &DAG,
     Align = std::min(Align, Ld->getAlign());
   }
 
+  // If the load ptrs can be decomposed into a common (Base + Index) with a
+  // common constant stride, then return the constant stride. This matcher
+  // enables some additional optimization since BaseIndexOffset is capable of
+  // decomposing the load ptrs to (add (add Base, Index), Stride) instead of
+  // (add LastPtr, Stride) or (add NextPtr, Stride) that matchForwardStrided and
+  // matchReverseStrided use respectively.
+  auto matchConstantStride = [&DAG, &N](ArrayRef<SDUse> Loads) {
+    // Initialize match constraints based on the first load. Initialize
+    // ConstStride by taking the difference between the offset of the first two
+    // loads.
+    if (Loads.size() < 2)
+      return SDValue();
+    BaseIndexOffset BaseLdBIO =
----------------
michaelmaitland wrote:

I think I understand, `getPtrDiff` would return the canonicalized representation of the non-constant value. We wouldn't be able to use BaseIndexOffset here. Do you have any ideas on what we could use?

https://github.com/llvm/llvm-project/pull/68726