On Thu, Aug 3, 2017, at 08:00, Ian D. Leroux wrote:
> On Thu, 3 Aug 2017 10:54:30 +0000 (UTC) chris...@astron.com (Christos
> Zoulas) wrote:
> > In article <20170802215811.02ff2faba38001ebe4f53...@fastmail.fm>,
> > Ian D. Leroux <idler...@fastmail.fm> wrote:
> > >The patches stop swap1_stop from blindly unmounting a tmpfs-mounted
> > >/dev/while the system is still running multi-user.
> > [...]
> > Why not just skip /dev?
> [ Ian grumbles that this all feels unclean and worries about device
> nodes on other filesystems ]
Then again, perhaps I'm overthinking this. swap1_stop runs after most
of userland has gone down and only a fraction of a second before
userland disappears altogether. If I read the dependencies correctly,
/usr and /var may already be gone by then, and anything not unmounted as
part of swap1_stop will get the axe a split-second later.
Is there any reasonable scenario in which a device node outside /dev
might still be accessible or needed after swap1_stop runs? If not,
then my objections are mostly theoretical and I'd be happy to see a
pragmatic fix committed. I'll prepare a patch against -current,