On Thu, Feb 12, 2026 at 8:59 AM Maxime Ripard <[email protected]> wrote: > > And the "oh but driver is using the API" is kind of ironic in the > context of the Rust bindings which have globally been in that situation > for years. You can't argue it both ways.
I don't think this is true, if I understand you correctly. To give some context, when Rust was close to land in mainline, we were explicitly asked to cut down what we had out-of-tree to a minimum of APIs, which was fair. So I did exactly that, to a point where you could essentially only declare a module for x86_64 and use `pr_info!` in it (pretty much). I even considered dropping the `pr_info!` too... :) Then we started upstreaming bits piece by piece, getting agreement and input from the relevant maintainers for each bit, and so on, which took years for several reasons, one of which was precisely waiting for actual users for APIs. [ Others were getting maintainers engaged and involved for some of the core APIs/subsystems, getting authors of the original code to submit it upstream via patches was yet another, reworking and cleaning the code with the new, better approaches we found meanwhile (like pin-init), and so on and so forth. ] Cheers, Miguel
