[Ksummit-discuss] [MAINTAINERS SUMMIT] Developing across multiple areas of the kernel

[Linus Torvalds]

On Thu, Jun 29, 2017 at 11:20 AM, Luis R. Rodriguez <mcgrof at kernel.org> wrote:
>
> Linus has said before cross-tree collateral evolutions *could* just be sent to
> him as a set of scripts he could run. It might sound easier said than done
> though, but we can improve on the process by practicing it daily.

I do take cross-tree stuff, and in particular don't mind things if
they are obviously scripted cleanups (with the extreme case of that
being to just send me the script - that makes it *really* obvious to
me that it's scripted ;).

That said, I generally end up considering it a last option. And that's
not because I can't do it, but because things like that tend to be
really painful for stable backporting etc. Cross-tree stuff has a
tendency to also be large and invasive in other ways.

If it's *small* and just cross-tree, that's fine, and I take it all
the time, although because of general laziness I certainly prefer when
it goes through somebody else (often the tip tree or Andrew Morton).

In fact, the real problem with cross-tree things is often not that it
affects multiple maintainers, but that it by definition doesn't have a
maintainer AT ALL that tracks it and pushes it. So the "random
occasional cross-tree updates" often end up languishing because nobody
takes ownership of them.

                        Linus