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

[Federico Bruzzone]

FedericoBruzzone wrote:

**The current state**

| Op | unit-stride check | negative-stride check |
|---|---|---|
| `vector.load` / `vector.store` | ✅ pre-existing (strict) | ✅ (#204611) |
| `vector.maskedload` / `maskedstore` | ✅ **this PR** (lenient: dynamic OK) | ✅ (#205869) |
| `vector.expandload` / `compressstore` | ✅ **this PR** (lenient: dynamic OK) | ✅ **this PR** |
| `vector.gather` / `scatter` | ❌ still permissive (we need to discuss it) | ✅ (#205869) |

Related prior work in this series
- #204309 
- #204611
- #205869

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