mm/oom_kill: remove the wrong fatal_signal_pending() check in oom_kill_process()
authorOleg Nesterov <oleg@redhat.com>
Fri, 6 Nov 2015 02:48:20 +0000 (18:48 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 6 Nov 2015 03:34:48 +0000 (19:34 -0800)
The fatal_signal_pending() was added to suppress unnecessary "sharing same
memory" message, but it can't 100% help anyway because it can be
false-negative; SIGKILL can be already dequeued.

And worse, it can be false-positive due to exec or coredump.  exec is
mostly fine, but coredump is not.  It is possible that the group leader
has the pending SIGKILL because its sub-thread originated the coredump, in
this case we must not skip this process.

We could probably add the additional ->group_exit_task check but this
patch just removes the wrong check along with pr_info().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Kyle Walker <kwalker@redhat.com>
Cc: Stanislav Kozina <skozina@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/oom_kill.c

index 58f3d2748ceda9f5610b9e0986a36b09839e58cd..c837d06dce5a447787b9965f5bf2103aaab98e64 100644 (file)
@@ -579,11 +579,7 @@ void oom_kill_process(struct oom_control *oc, struct task_struct *p,
                    !(p->flags & PF_KTHREAD)) {
                        if (p->signal->oom_score_adj == OOM_SCORE_ADJ_MIN)
                                continue;
-                       if (fatal_signal_pending(p))
-                               continue;
 
-                       pr_info("Kill process %d (%s) sharing same memory\n",
-                               task_pid_nr(p), p->comm);
                        do_send_sig_info(SIGKILL, SEND_SIG_FORCED, p, true);
                }
        rcu_read_unlock();