[Lxc-users] lxc-attach & setns patch for kernel 2.6.34

[Daniel Lezcano]

On 07/28/2010 10:41 AM, Sebastien Pahl wrote:
> Hi,
>
> I'm trying to adapt the patch for lxc-attach to 2.6.34.
> A few hunks fail but they are trivial to fix.
> There is only one where I'm not sure:
>
> <------------------------------>
>
> Index: 2.6.33/kernel/cgroup.c
> ===================================================================
> --- 2.6.33.orig/kernel/cgroup.c
> +++ 2.6.33/kernel/cgroup.c
> @@ -2453,7 +2453,7 @@ static struct cgroup_pidlist *cgroup_pid
> {
>    struct cgroup_pidlist *l;
>    /* don't need task_nsproxy() if we're looking at ourself */
> -   struct pid_namespace *ns = get_pid_ns(current->nsproxy->pid_ns);
> +   struct pid_namespace *ns =
> get_pid_ns(task_active_pid_ns(current));
>    /*
>     * We can't drop the pidlist_mutex before taking the l->mutex in
> case
>     * the last ref-holder is trying to remove l from the list at the
> same
>
> <------------------------------>
>
> The kernel/cgroup.c file has changed:
>
> Instead of:
>
> struct pid_namespace *ns = get_pid_ns(current->nsproxy->pid_ns);
>
> now there is:
>
> struct pid_namespace *ns = current->nsproxy->pid_ns;
>
> > From what I have seen in the rest of the code I assume I can just
> replace it by this line:
>
> struct pid_namespace *ns = task_active_pid_ns(current);
>
> Am I correct in my assumptions?
>    

Yes, I think so. The get_pid_ns is done later in the code to reduce the 
rollback code (put_pid_ns) in case of error.
As you have the right pid namespace in the 'ns' variable, your 
modification is correct I think.

Thanks
   -- Daniel