OpenVZ Forum


Home » Mailing lists » Devel » Re: [PATCH 3/7] Containers (V8): Add generic multi-subsystem API to containers
Re: [PATCH 3/7] Containers (V8): Add generic multi-subsystem API to containers [message #11892] Wed, 11 April 2007 05:01 Go to next message
Srivatsa Vaddagiri is currently offline  Srivatsa Vaddagiri
Messages: 241
Registered: August 2006
Senior Member
[ Sorry abt piece meal reviews, I am sending comments as and when I spot
something ]

On Fri, Apr 06, 2007 at 04:32:24PM -0700, menage@google.com wrote:
> -void container_exit(struct task_struct *tsk)
> +void container_exit(struct task_struct *tsk, int run_callbacks)
> {

[snip]

> + /* Reassign the task to the init_container_group. */
> task_lock(tsk);

[snip]

> + if (tsk->containers != &init_container_group) {

Is this check needed? If we have the check, then:

> + task_unlock(tsk);
> + if (cg)
> + put_container_group(cg);

init_container_group refcount is leaky (fork increments it, but exit
doesnt). Not a big prob perhaps, but ..would be nice to avoid?

--
Regards,
vatsa
Re: [PATCH 3/7] Containers (V8): Add generic multi-subsystem API to containers [message #11901 is a reply to message #11892] Wed, 11 April 2007 08:42 Go to previous message
Paul Menage is currently offline  Paul Menage
Messages: 642
Registered: September 2006
Senior Member
On 4/10/07, Srivatsa Vaddagiri <vatsa@in.ibm.com> wrote:
> [ Sorry abt piece meal reviews, I am sending comments as and when I spot
> something ]

That's no problem.

>
> On Fri, Apr 06, 2007 at 04:32:24PM -0700, menage@google.com wrote:
> > -void container_exit(struct task_struct *tsk)
> > +void container_exit(struct task_struct *tsk, int run_callbacks)
> > {
>
> [snip]
>
> > + /* Reassign the task to the init_container_group. */
> > task_lock(tsk);
>
> [snip]
>
> > + if (tsk->containers != &init_container_group) {
>
> Is this check needed? If we have the check, then:
>
> > + task_unlock(tsk);
> > + if (cg)
> > + put_container_group(cg);
>
> init_container_group refcount is leaky (fork increments it, but exit
> doesnt). Not a big prob perhaps, but ..would be nice to avoid?
>

Yes, I think you're right - over a long uptime that would just cause
problems ...

Paul
Previous Topic: Re: Re: [RFC][PATCH] rename 'struct pid'
Next Topic: Re: [patch 0/8] unprivileged mount syscall
Goto Forum:
  


Current Time: Thu Dec 12 20:19:10 GMT 2024

Total time taken to generate the page: 0.02768 seconds