Re: [PATCH RESEND] cgroups: fix incorrect using rcu_dereference() in cgroup_subsys_state()

[Paul Menage]

On Tue, Jan 4, 2011 at 12:18 AM, Lai Jiangshan <laijs@xxxxxxxxxxxxxx> wrote:
> From: Li Zefan <lizf@xxxxxxxxxxxxxx>
> Date: Mon, 25 Aug 2008 11:05:28 +0800
> (Original) Subject: [PATCH] cgroup: fix wrong rcu_dereference()
>
> It is tsk->cgroups which is protected by RCU, not ->subsys[subsys_id].
>
> laijs: updated it(the surrounding code have been changed since these two years).
>
> Signed-off-by: Li Zefan <lizf@xxxxxxxxxxxxxx>
> Signed-off-by: Lai Jiangshan <laijs@xxxxxxxxxxxxxx>
> ---
> diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
> index ed4ba11..a798814 100644
> --- a/include/linux/cgroup.h
> +++ b/include/linux/cgroup.h
> @@ -535,10 +535,11 @@ static inline struct cgroup_subsys_state *cgroup_subsys_state(
>  * cgroup_subsys::attach() methods.
>  */
>  #define task_subsys_state_check(task, subsys_id, __c)                  \
> -       rcu_dereference_check(task->cgroups->subsys[subsys_id],         \
> +       rcu_dereference_check(task->cgroups,                            \
>                              rcu_read_lock_held() ||                   \
>                              lockdep_is_held(&task->alloc_lock) ||     \
> -                             cgroup_lock_is_held() || (__c))
> +                             cgroup_lock_is_held() ||                  \
> +                             (__c))->subsys[subsys_id]

Can you reformat this a bit? It took a fair bit of staring to realize
that (__c))->subsys[subsys_id] wasn't trying to dereference __c.

Maybe make it:

#define task_subsys_state_check(task, subsys_id, __c) ({ \
  struct css_set *_css = rcu_dereference_check(...); \
  _css->subsys[subsys_id]; \
})

Paul
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/