svn commit: r367038 - head/sbin/mount_nullfs

Oliver Pinter oliver.pntr at gmail.com
Sun Oct 25 21:46:00 UTC 2020


On Sunday, October 25, 2020, Edward Tomasz Napierala <trasz at freebsd.org>
wrote:

> On 1025T0717, Cy Schubert wrote:
> > In message <CAPjTQNEq2ukvkNoF6bsymTai52XV1tbu3TxW4rrN4n5hASMFqQ at mail.
> gmail.c
> > om>
> > , Oliver Pinter writes:
> > > On Sunday, October 25, 2020, Edward Tomasz Napierala <
> trasz at freebsd.org>
> > > wrote:
> > >
> > > > Author: trasz
> > > > Date: Sun Oct 25 14:09:00 2020
> > > > New Revision: 367038
> > > > URL: https://svnweb.freebsd.org/changeset/base/367038
> > > >
> > > > Log:
> > > >   Remove the check that prevents creating "loops" from
> mount_nullfs(8).
> > > >
> > > >
> > > The simple question is: why?
> >
> > I was about to ask that myself.
> >
> > The why is usually more important than the how.
>
> Two reasons.
>
> First, from what I can tell this is an artificial limitation which serves
> no purpose.  I thought it was to prevent some kind of deadlock between
> vnodes, so I asked kib@, and he mentioned removing that very limitation
> from the kernel code; thus, it looks to me like a historical leftover.
>
> Second, I've stumbled upon this when trying to nullfs-mount '/' on
> '/compat/ubuntu/bsd/', which would be useful for autochroot functionality,
> when (or if) it becomes a thing.


Thank you!


More information about the svn-src-head mailing list