[CRIU] v2: automatic detection of external bind mounts

Tycho Andersen tycho.andersen at canonical.com
Fri Apr 10 07:36:03 PDT 2015


On Fri, Apr 10, 2015 at 12:30:56PM +0300, Pavel Emelyanov wrote:
> On 04/10/2015 12:13 AM, Tycho Andersen wrote:
> > Hi Pavel,
> > 
> > On Thu, Apr 09, 2015 at 02:56:48PM -0600, Tycho Andersen wrote:
> >>
> >> Hmm, actually I found something wrong with this set. Please let me
> >> debug it.
> > 
> > So the problem is the fsroot_mounted() check; I knew it wasn't in there for a
> > reason :). I just tested the patch on the right (systemd) host and found the
> > problem. For e.g. pstore (but also kernel security and a few others), I have a
> > mountinfo entry in the container that looks like this:
> > 
> > 90 83 0:23 / /sys/fs/pstore rw,nosuid,nodev,noexec,relatime master:11 - pstore pstore rw
> > 
> > and on the host:
> > 
> > 29 17 0:23 / /sys/fs/pstore rw,nosuid,nodev,noexec,relatime shared:11 - pstore pstore rw
> > 
> > So, fsroot_mounted() is true, which means this mountpoint is never checked for
> > external sharing as it should be. I'm not sure what the right answer is here;
> > things work as I want them to if I leave the check out, but semantically the
> > check seems to make sense (at least in the case where there is no sharing).
> > Maybe that's the right answer, to only bail on fsroot_mounted() entries if
> > there is no external sharing?
> 
> Ah, so this is fsroot shared mount %) Phew! Well, I think you were right about removing
> the fsroot_mounted() check. In the end fsroot external mount shouldn't differ from
> non-fsroot mount in this sense.

Great, I resent v4 of the patch without the check in it. I think I've
covered all the review comments now, but please let me know if
I've missed something.

Tycho

> -- Pavel
> 


More information about the CRIU mailing list