Frederic Weisbecker (fweisbec@xxxxxxxxx) wrote:
> When we fetch the css_set of the tasks on cgroup migration, we don't need
> anymore to synchronize against cgroup_exit() that could swap the old one
> with init_css_set. Now that we are using threadgroup_lock() during
> the migrations, we don't need to worry about it anymore.
>
> Signed-off-by: Frederic Weisbecker <fweisbec@xxxxxxxxx>
> Cc: Tejun Heo <tj@xxxxxxxxxx>
> Cc: Li Zefan <lizf@xxxxxxxxxxxxxx>
> Cc: Containers <containers@xxxxxxxxxxxxxxxxxxxxxxxxxx>
> Cc: Cgroups <cgroups@xxxxxxxxxxxxxxx>
> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> Cc: Oleg Nesterov <oleg@xxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Paul Menage <paul@xxxxxxxxxxxxxx>
> Cc: Mandeep Singh Baines <msb@xxxxxxxxxxxx>
Reviewed-by: Mandeep Singh Baines <msb@xxxxxxxxxxxx>
> ---
> kernel/cgroup.c | 21 +++++++++++----------
> 1 files changed, 11 insertions(+), 10 deletions(-)
>
> diff --git a/kernel/cgroup.c b/kernel/cgroup.c
> index 5fdd98c..6b5f693 100644
> --- a/kernel/cgroup.c
> +++ b/kernel/cgroup.c
> @@ -1850,13 +1850,13 @@ static int cgroup_task_migrate(struct cgroup *cgrp, struct cgroup *oldcgrp,
> struct css_set *newcg;
>
> /*
> - * get old css_set. we need to take task_lock and refcount it, because
> - * an exiting task can change its css_set to init_css_set and drop its
> - * old one without taking cgroup_mutex.
> + * get old css_set. We are synchronized through threadgroup_lock()
> + * against PF_EXITING setting such that we can't race against
> + * cgroup_exit() changing the css_set to init_css_set and dropping the
> + * old one.
> */
> - task_lock(tsk);
> + WARN_ON_ONCE(tsk->flags & PF_EXITING);
> oldcg = tsk->cgroups;
> - task_unlock(tsk);
>
> /* locate or allocate a new css_set for this task. */
> if (guarantee) {
> @@ -1875,9 +1875,7 @@ static int cgroup_task_migrate(struct cgroup *cgrp, struct cgroup *oldcgrp,
> return -ENOMEM;
> }
>
> - /* @tsk can't exit as its threadgroup is locked */
> task_lock(tsk);
> - WARN_ON_ONCE(tsk->flags & PF_EXITING);
> rcu_assign_pointer(tsk->cgroups, newcg);
> task_unlock(tsk);
>
> @@ -2177,10 +2175,13 @@ int cgroup_attach_proc(struct cgroup *cgrp, struct task_struct *leader)
> /* nothing to do if this task is already in the cgroup */
> if (tc->cgrp == cgrp)
> continue;
> - /* get old css_set pointer */
> - task_lock(tc->task);
> + /*
> + * get old css_set pointer. threadgroup is locked so this is
> + * safe against concurrent cgroup_exit() changing this to
> + * init_css_set.
> + */
> oldcg = tc->task->cgroups;
> - task_unlock(tc->task);
> +
> /* if we don't already have it in the list get a new one */
> if (!css_set_fetched(cgrp, tc->task, oldcg, &newcg_list))
> if (css_set_prefetch(cgrp, oldcg, &newcg_list))
> --
> 1.7.5.4
>
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linuxfoundation.org/mailman/listinfo/containers
