Re: [PATCH] MAINTAINERS: add Josh Law as reviewer for library code

[Lorenzo Stoakes (Oracle)]

+cc David for mention

On Fri, Mar 13, 2026 at 04:33:49PM +0100, Christian Brauner wrote:
> - first patch Feb 28 under pseudonym "techyguyperplexable", switched to "Josh Law" only when Andrew required a real name
>   in https://lore.kernel.org/all/20260228114939.de7d44de38d907a9b6632480@xxxxxxxxxxxxxxxxxxxx
> - 142+ emails in ~2 weeks — 37 patches in a single day (Mar 1)
> - All trivial/cosmetic — SPDX headers, comment grammar, spacing, const qualifiers. Zero bug fixes, zero new functionality
> - Carpet-bombed multiple subsystems — lib/, arm64/, staging/, input/, etc.
> - within 1 week of first-ever patch, submitted MAINTAINERS: add Josh Law as reviewer for library code covering all of lib/ (locking.c, iov_iter.c, rhashtable.c, etc.)
> - Email identity mismatch — From: hlcj1234567@xxxxxxxxx, Signed-off-by: objecting@xxxxxxxxxxxxx
> - Formatting problems — top-posting, line length violations, patches not applying cleanly
>
> So I mean, one week to Reviewer. Even if we're being very generous here,
> we need to do a lot more due diligence going forward. We can't just hand
> out core components like this and risking our reputation and security
> posture.

Thanks! Agree absolutely that we need to be careful about this. Jia Tan should
be instructive.

Josh - there's nothing personal here to be clear, this is a question of
procedure and caution.

More broadly I think we should avoid assigning new people to catch-all
categories anyway unless they are well established enough to be involved with
_everything_ the catch-all covers.

For instance adding people to the mm/* other than perhaps... David ;) would be
crazy.

Also - Andrew - I think for cases where you are the only maintainer but it
impacts others, you should seek acks proportional to the scope the MAINTAINERS
entry spans - in this case that'd be a _lot_ of people - but that only
underlines that we shouldn't be updating such entries anyway.

In fact - can we just do away with catch-all's and just make sure MAINTAINERS
entries are established for everything?

I have ground to stand on for this as I personally did it for mm, although we do
still have a catch-all (not sure if necessary any more?)

In the case of lib/ a quick fix could be to figure out which files are not
covered by other MAINTAINERS entries and adding them all to what is currently
the catch-all?

Cheers, Lorenzo