Messages in this thread | | | Date | Tue, 26 Aug 2008 16:25:26 -0500 | From | "Serge E. Hallyn" <> | Subject | Re: [PATCH 1/4] pid_ns: zap_pid_ns_processes: fix the ->child_reaper changing |
| |
Quoting Oleg Nesterov (oleg@tv-sign.ru): > zap_pid_ns_processes() sets pid_ns->child_reaper = NULL, this is wrong. > > Yes, we have already killed all tasks in this namespace, and sys_wait4() > doesn't see any child. But this doesn't mean ->children list is empty, > we may have EXIT_DEAD tasks which are not visible to do_wait(). In that > case the subsequent forget_original_parent() will crash the kernel because > it will try to re-parent these tasks to the NULL reaper. > > Even if there are no childs, it is not good that forget_original_parent() > uses reaper == NULL. > > Change the code to set ->child_reaper = init_pid_ns.child_reaper instead. > We could use pid_ns->parent->child_reaper as well, I think this does not > really matter. These EXIT_DEAD tasks are not visible to the new ->parent > after re-parenting, they will silently do release_task() eventually. > > Note that we must change ->child_reaper, otherwise forget_original_parent() > will use reaper == father, and in that case we will hit the (correct) > BUG_ON(!list_empty(&father->children)). > > Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Well while it looked correct to me all along, I couldn't get the testcase to cause an oops. But I see where do_exit() calls zap_pid_ns_processes() (through group_exit()) before forget_original_parent(), which can/should end up dereferencing the null pid_ns, so clearly this is needed and correct.
Acked-by: Serge Hallyn <serue@us.ibm.com>
Thanks, Oleg.
-serge
> --- 2.6.27-rc4/kernel/pid_namespace.c~1_ZAP_DONT_CLEAR_REAPER 2008-07-30 13:12:49.000000000 +0400 > +++ 2.6.27-rc4/kernel/pid_namespace.c 2008-08-24 17:22:59.000000000 +0400 > @@ -179,9 +179,12 @@ void zap_pid_ns_processes(struct pid_nam > rc = sys_wait4(-1, NULL, __WALL, NULL); > } while (rc != -ECHILD); > > - > - /* Child reaper for the pid namespace is going away */ > - pid_ns->child_reaper = NULL; > + /* > + * We can not clear ->child_reaper or leave it alone. > + * There may by stealth EXIT_DEAD tasks on ->children, > + * forget_original_parent() must move them somewhere.
Actually this comment is a little bit misleading - the null deref will happen regardless of whether there are children, right?
> + */ > + pid_ns->child_reaper = init_pid_ns.child_reaper; > acct_exit_ns(pid_ns); > return; > }
| |