Re: [GIT PULL] gpio: updates for v5.13

[Al Viro <viro@zeniv.linux.org.uk>]

On Mon, May 03, 2021 at 11:03:57AM -0700, Linus Torvalds wrote:
> Al,
>  would you mind taking a look at this part:
> 
> On Sun, May 2, 2021 at 12:32 PM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
> >
> > You'll notice that we have a bunch of configfs commits in our tree not acked by
> > the configfs maintainers. These commits implement the concept of committable
> > items in configfs - something that was well defined in the documentation for
> > years but has remained unimplemented. Despite the first submission of these
> > patches back in November 2020[1] and repeated pings & resending, configfs
> > maintainers have remained unresponsive. After reviewing these on the GPIO
> > mailing list, we decided to pick them up ourselves and send them your way
> > together with the first user: the new GPIO simulator.
> 
> It doesn't look huge to me, and I don't care all that deeply about
> configfs, and honestly, I'm not seeing huge amounts of actual
> development there, with recent commits all being about cleanup of vfs
> changes (eg things like the new idmapping changes etc).
> 
> That said, I really don't want to pull that with some core sanity checking.
> 
> So Al, do you see anything horrendous in how that configfs thing uses
> a rename to do kind of an "atomic swap" of configfs state?

Give me a few hours; configfs is playing silly buggers with a lot of
structures when creating/tearing down subtrees, and I'd actually
expect more trouble with configfs data structures than with VFS ones.

I'll take a look.