[Mlir-commits] [mlir] [Linalg] Update Conv Decomposition patterns to work with generic convolution ops as well (PR #174196)

[Renato Golin]

rengolin wrote:

> To be clear, this is not about generailzation or specialization. This is a PR that extends the existing patterns to handle conv decompisition in generic form. It is a pure improvement for these patterns.

My comment was only related to @banach-space comment, not the PR itself. I agree with your statement above.

> My claim is that upstream patterns provide transformations and downstream users build the passes with the patterns. Does adding a control function to keep the same behavior sound better?

This is what we've been doing and it "works for now". Once we bring this into canonicalization (my comment above), we'll have to change that approach to provide stronger guarantees and forms directly upstream, not just a "bag of tools". But that is orthogonal to this PR.

> t is a little weird to me though, because new users would be confused about why the default behavior is filtering out conv ops in generic op form. If it were me, I'd apply it to all the ops that meet requirement, and have a note in PR description about how to restore to the old behavior. What do you think?

I'd leave this discussion for a future PR.

https://github.com/llvm/llvm-project/pull/174196