[Mlir-commits] [mlir] [mlir][VectorOps] Extend vector.constant_mask to support 'all true' scalable dims (PR #66638)

[Diego Caballero]

dcaballe wrote:

THanks for the contribution. Really interesting... Is the constant input required to be always lower or equal to the scalable base?

For example, what these two examples would mean:

```
%mask = vector.constant_mask [2, 7] : vector<4x[4]xi1>
%mask = vector.constant_mask [2, 3] : vector<4x[4]xi1>
```

I'm also wondering if it would make sense to distinguish between `[2, 4] : vector<4x[4]xi1>` and `[2, [4]] : vector<4x[4]xi1>` as they may mean completely different things?

https://github.com/llvm/llvm-project/pull/66638