[CRIU] [RFC PATCH 3/5] cr-restore: restore root sid before mount namespace

[Stanislav Kinsburskiу]

23 нояб. 2015 г. 20:22 пользователь Pavel Emelyanov <xemul at parallels.com> написал:
>
> On 11/23/2015 08:22 PM, Stanislav Kinsburskiy wrote: 
> > This is a precursor patch for AutoFS mount restore. 
> > To restore AutoFS-nested mount points, we have to created target dentries. 
> > To be able to do so, the process have to be process group leader before AutoFS 
> > mount is called. 
>
> Where does this requirement come from? 
>

It's a way, how kernel restricts autofs management. Kernel doesn't mount anything. It passes request via pipe to user space process instead. Only process, which pgrp is set as a mount property, can create dentries on AutoFS mount.

> > Note: there is another way, how to set process as a group leader. We can fork 
> > a child with CLONE_VM, and use it to mount all the mount points. 
> > 
> > Signed-off-by: Stanislav Kinsburskiy <skinsbursky at virtuozzo.com> 
> > --- 
> >  cr-restore.c |    8 +++++++- 
> >  1 file changed, 7 insertions(+), 1 deletion(-) 
> > 
> > diff --git a/cr-restore.c b/cr-restore.c 
> > index 06fceb6..9aec077 100644 
> > --- a/cr-restore.c 
> > +++ b/cr-restore.c 
> > @@ -1441,7 +1441,10 @@ static int create_children_and_session(void) 
> >  return ret; 
> >  } 
> >  
> > - restore_sid(); 
> > + if (getpid() == 1) 
> > + pr_info("Skipping restore_sid() for init\n"); 
> > + else 
> > + restore_sid(); 
> >  
> >  pr_info("Restoring children in our session:\n"); 
> >  list_for_each_entry(child, &current->children, sibling) { 
> > @@ -1524,6 +1527,9 @@ static int restore_task_with_children(void *_arg) 
> >  if (restore_finish_stage(CR_STATE_RESTORE_NS) < 0) 
> >  goto err; 
> >  
> > + pr_info("Calling restore_sid() for init\n"); 
> > + restore_sid(); 
>
> In the patch comment you say that process has to be group leader, but 
> make it session leader instead. Why? 
>

I'll check for the proper explanation tommorow. Setting it as a group leader before setting as a session leader doesn't. While setting as a session leader is enough.

> > + 
> >  /* 
> >  * We need non /proc proc mount for restoring pid and mount 
> >  * namespaces and do not care for the rest of the cases. 
> > 
> > _______________________________________________ 
> > CRIU mailing list 
> > CRIU at openvz.org 
> > https://lists.openvz.org/mailman/listinfo/criu 
> > . 
> > 
>