[Mlir-commits] [mlir] [mlir][vector] Verify non-unit strides on `masked/expand/compress` ops and fix `SparseVectorization` miscompilation (PR #210952)

[Andrzej Warzyński]

================
@@ -6190,6 +6190,32 @@ static LogicalResult verifyLoadStoreMemRefLayout(Operation *op,
   return success();
 }
 
+/// Unlike vector.load/store, ops such as vector.maskedload/maskedstore and
+/// vector.expandload/compressstore are also used on memrefs whose most minor
+/// dimension has a dynamic stride that is contiguous at runtime but not
+/// provable as such at the type level (e.g., buffers produced by the
+/// sparsifier).
----------------
banach-space wrote:

This is the first paragraph documenting this method, yet it doesn't say what the method does. Instead, it discusses the context in which it might be used. Please update.

https://github.com/llvm/llvm-project/pull/210952