Quoting Oleg Nesterov ([EMAIL PROTECTED]):
> 1. rename reparent_to_init() to reparent_kthread() and export it
> 
> 2. use init_pid_ns.child_reaper instead of child_reaper(current)

Each of these patches looks good to me, but this part in particular
is a must-have bugfix.

Just started some tests, if any failures come back I'll report them
tonight.

thanks Oleg,
-serge

> 3. set ->exit_signal = -1, so init can't see us and we don't use
>    it to reap the task.
> 
> 4. add reparent_kthread() to kthread() and stopmachine()
> 
> See also
> 
>       http://marc.info/?t=117580282200003&r=1
>       http://marc.info/?t=95299284800003&r=1
> 
> Signed-off-by: Oleg Nesterov <[EMAIL PROTECTED]>
> 
>  include/linux/sched.h |    1 +
>  kernel/exit.c         |   16 ++++++++--------
>  kernel/kthread.c      |    1 +
>  kernel/stop_machine.c |    1 +
>  4 files changed, 11 insertions(+), 8 deletions(-)
> 
> --- 2.6.21-rc5/include/linux/sched.h~2_DETACH 2007-04-05 12:18:28.000000000 
> +0400
> +++ 2.6.21-rc5/include/linux/sched.h  2007-04-10 21:52:27.000000000 +0400
> @@ -1401,6 +1401,7 @@ extern void exit_itimers(struct signal_s
>  
>  extern NORET_TYPE void do_group_exit(int);
>  
> +extern void reparent_kthread(void);
>  extern void daemonize(const char *, ...);
>  extern int allow_signal(int);
>  extern int disallow_signal(int);
> --- 2.6.21-rc5/kernel/exit.c~2_DETACH 2007-04-10 21:32:44.000000000 +0400
> +++ 2.6.21-rc5/kernel/exit.c  2007-04-10 21:59:41.000000000 +0400
> @@ -255,7 +255,7 @@ static int has_stopped_jobs(struct pid *
>  }
>  
>  /**
> - * reparent_to_init - Reparent the calling kernel thread to the init task of 
> the pid space that the thread belongs to.
> + * reparent_kthread - Reparent the calling kernel thread to the init task of 
> the pid space that the thread belongs to.
>   *
>   * If a kernel thread is launched as a result of a system call, or if
>   * it ever exits, it should generally reparent itself to init so that
> @@ -264,20 +264,20 @@ static int has_stopped_jobs(struct pid *
>   * The various task state such as scheduling policy and priority may have
>   * been inherited from a user process, so we reset them to sane values here.
>   *
> - * NOTE that reparent_to_init() gives the caller full capabilities.
> + * NOTE that reparent_kthread() gives the caller full capabilities.
>   */
> -static void reparent_to_init(void)
> +void reparent_kthread(void)
>  {
>       write_lock_irq(&tasklist_lock);
>  
>       ptrace_unlink(current);
>       remove_parent(current);
> -     current->parent = child_reaper(current);
> -     current->real_parent = child_reaper(current);
> +     current->parent = init_pid_ns.child_reaper;
> +     current->real_parent = current->parent;
>       add_parent(current);
>  
> -     /* Set the exit signal to SIGCHLD so we signal init on exit */
> -     current->exit_signal = SIGCHLD;
> +     /* make the task auto-reap */
> +     current->exit_signal = -1;
>  
>       security_task_reparent_to_init(current);
>       write_unlock_irq(&tasklist_lock);
> @@ -391,7 +391,7 @@ void daemonize(const char *name, ...)
>       current->files = init_task.files;
>       atomic_inc(&current->files->count);
>  
> -     reparent_to_init();
> +     reparent_kthread();
>  
>       if (!has_rt_policy(current) && (task_nice(current) < 0))
>               set_user_nice(current, 0);
> --- 2.6.21-rc5/kernel/kthread.c~2_DETACH      2007-04-05 12:18:28.000000000 
> +0400
> +++ 2.6.21-rc5/kernel/kthread.c       2007-04-10 22:02:31.000000000 +0400
> @@ -82,6 +82,7 @@ static int kthread(void *_create)
>       sigset_t blocked;
>       int ret = -EINTR;
>  
> +     reparent_kthread();
>       kthread_exit_files();
>  
>       /* Copy data: it's on keventd's stack */
> --- 2.6.21-rc5/kernel/stop_machine.c~2_DETACH 2006-10-22 18:24:03.000000000 
> +0400
> +++ 2.6.21-rc5/kernel/stop_machine.c  2007-04-10 22:04:23.000000000 +0400
> @@ -33,6 +33,7 @@ static int stopmachine(void *cpu)
>       int irqs_disabled = 0;
>       int prepared = 0;
>  
> +     reparent_kthread();
>       set_cpus_allowed(current, cpumask_of_cpu((int)(long)cpu));
>  
>       /* Ack: we are alive */
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Reply via email to