Re: [regression?] escaping commas in overlayfs mount options

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Oct 6, 2023 at 4:03 PM Sebastian Wick <sebastian.wick@xxxxxxxxxx> wrote:
>
> On Fri, Sep 29, 2023 at 07:44:09AM +0300, Amir Goldstein wrote:
> > On Fri, Sep 29, 2023 at 4:08 AM Ryan Hendrickson
> > <ryan.hendrickson@xxxxxxxxxxxx> wrote:
> > >
> > > Up to and including kernel 6.4.15, it was possible to have commas in
> > > the lowerdir/upperdir/workdir paths used by overlayfs, provided they were
> > > escaped with backslashes:
> > >
> > >      mkdir /tmp/test-lower, /tmp/test-upper /tmp/test-work /tmp/test
> > >      mount -t overlay overlay -o 'lowerdir=/tmp/test-lower\,,upperdir=/tmp/test-upper,workdir=/tmp/test-work' /tmp/test
> > >
> > > In 6.5.2 and 6.5.5, this no longer works; dmesg reports that overlayfs
> > > can't resolve '/tmp/test-lower' (without the comma).
> > >
> > > I see that there is a commit between the 6.4 and 6.5 lines titled [ovl:
> > > port to new mount api][1]. I haven't compiled a kernel before and after
> > > this commit to verify, but based on the code it deletes I strongly suspect
> > > that it, or if not then one of the ovl commits committed on the same day,
> > > is responsible for this change.
> > >
> > > [1]: https://github.com/torvalds/linux/commit/1784fbc2ed9c888ea4e895f30a53207ed7ee8208
> > >
> >
> > That's a good guess.
> > It helps to CC the author of the patch in this case ;-)
> >
> > > Does this count as a regression?
> >
> > "used to work, does not work now" is pretty close to a dictionary
> > definition of a regression :)
> >
> > The question is whether we should fix it.
> > The rule of thumb is that if users complain than we need to fix it,
> > but it's a corner case and if the only users that complained are willing
> > to work around the problem (hint hint) then we may not need to fix it.
>
> It would be nice to have this fixed. A more general question: will you
> commit on keeping the escaping stable from now on or do we have to
> expect changes at any point in the future?
>

I prefer that escaping would be handled in userspace, now that the new
mount API allows that, so deferring the question to libmount maintainer.

Thanks,
Amir.

> In that case we would just reject any string contianing characters that
> need escaping.
>
> > > I can't find documentation for this
> > > escaping feature anywhere, even as it pertains to the non-comma characters
> > > '\\' and ':' (which, I've tested, can still be escaped as expected), so
> > > perhaps it was never properly supported? But a search for escaping commas
> > > in overlayfs turns up resources like [this post][2], suggesting that there
> > > are others who figured this out and expect it to work.
> > >
> > > [2]: https://unix.stackexchange.com/a/552640
> > >
> > > Is there a new way to escape commas for overlayfs options?
> > >
> >
> > Deferring the question to Christian.
> >
> > Thanks,
> > Amir.
>




[Index of Archives]     [Linux Filesystems Devel]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux