"Serge E. Hallyn" <serue@us.ibm.com> writes:

> Quoting sukadev@us.ibm.com (sukadev@us.ibm.com):
>> Oleg Nesterov [oleg@tv-sign.ru] wrote:
>> | On 12/05, Eric W. Biederman wrote:
>> | > 
>> | > +pid_t pid_vnr(struct pid *pid)
>> | > +{
>> | > +	return pid_nr_ns(pid, current->nsproxy->pid_ns);
>> | > +}
>> 
>> Hmm. current->nsproxy be NULL during process exit ?
>> So this safe as long as pid_vnr() is not called after a
>> process exits its namespaces. Probably no such callers atm. 
>
> Yes I did a little audit for those this morning bc I frankly didn't
> believe there weren't any.  But I couldn't find any :)

Cool.  The only case pid_vnr would make sense in that context is if
we were talking to user space after we had exited our namespaces.
Which is at least as stretch.

Eric
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers