Re: [Patch 1/3] Miscellaneous container fixes [message #8680] |
Fri, 01 December 2006 17:25 |
Paul Menage
Messages: 642 Registered: September 2006
|
Senior Member |
|
|
On 12/1/06, Srivatsa Vaddagiri <vatsa@in.ibm.com> wrote:
> This patches fixes various bugs I hit in the recently posted container
> patches.
>
> 1. If a subsystem registers with fork/exit hook during bootup (much
> before rcu is initialized), then the resulting synchronize_rcu() in
> container_register_subsys() hangs. Avoid this by not calling
> synchronize_rcu() if we arent fully booted yet.
>
> 2. If cpuset_create fails() for some reason, then the resulting
> call to cpuset_destroy can trip. Avoid this by initializing
> container->...->cpuset pointer to NULL in cpuset_create().
>
> 3. container_rmdir->cpuset_destroy->update_flag can deadlock on
> container_lock(). Avoid this by introducing __update_flag, which
> doesnt take container_lock().
Ah - this may be the lockup that PaulJ hit.
Thanks for these fixes.
Paul
|
|
|