[Devel] Re: [RFC][PATCH] Do not set /proc inode->pid for non-pid-related inodes

[Serge E. Hallyn]

Quoting Eric W. Biederman (ebiederm at xmission.com):
> Dave Hansen <hansendc at us.ibm.com> writes:
> 
> > So, doesn't that problem go away (or at least move to be umount's duty)
> > if we completely disconnect those inodes' lifetime from that of any
> > process or pid namespace?
> 
> If the last process has exited the pid namespace I would like the
> code to continue to behave as it currently does.
> 
> I would like readdir on /proc/ to not even try to show any pids when
> there are no pids or pid related files in the pid namespace.

In (at least one version of) Dave's patches, the /proc your pidns is
automatically used when you use /proc.  In that case a /proc should
just go away when the last task goes away, since noone else can use
that /proc.

I like that behavior, because otherwise (a) we require every new
pid_namespace to start by remounting /proc ere they get undefined
behavior, and (b) to gain anything from it, we would need a way
to refer to another pidspace for the sake of mounting it's proc,
i.e.

	mount -t proc -o init_pid=7501 proc_vserver1 /vserver1/proc

-serge

> I would like /proc/self to completely disappear when the are not
> any pids in the pid namespace.
> 
> I misspoke in when I said that /proc/<pid> was affected.  The function
> is proc_pid_readdir and it is a subset of /proc/ so it gets a little 
> confusing.
_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers