[Mlir-commits] [mlir] [MLIR] Add a utility pass to linearize `memref` (PR #136797)

[Krzysztof Drewniak]

krzysz00 wrote:

*However*, a subview isn't the only possible source of strided memrefs. Someone could give us a `memref<4x4xi32, strided<[8, 1]>>` as a function argument, and that *has* to be a memref<32xi32>` when linearized

So I'd fix the size computation and maybe put the subview logic back 

https://github.com/llvm/llvm-project/pull/136797