proc, oom: do not report alien mms when setting oom_score_adj
authorMichal Hocko <mhocko@suse.com>
Thu, 21 Feb 2019 06:19:42 +0000 (22:19 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 21 Feb 2019 17:01:00 +0000 (09:01 -0800)
Tetsuo has reported that creating a thousands of processes sharing MM
without SIGHAND (aka alien threads) and setting
/proc/<pid>/oom_score_adj will swamp the kernel log and takes ages [1]
to finish.  This is especially worrisome that all that printing is done
under RCU lock and this can potentially trigger RCU stall or softlockup
detector.

The primary reason for the printk was to catch potential users who might
depend on the behavior prior to 44a70adec910 ("mm, oom_adj: make sure
processes sharing mm have same view of oom_score_adj") but after more
than 2 years without a single report I guess it is safe to simply remove
the printk altogether.

The next step should be moving oom_score_adj over to the mm struct and
remove all the tasks crawling as suggested by [2]

[1] http://lkml.kernel.org/r/97fce864-6f75-bca5-14bc-12c9f890e740@i-love.sakura.ne.jp
[2] http://lkml.kernel.org/r/20190117155159.GA4087@dhcp22.suse.cz

Link: http://lkml.kernel.org/r/20190212102129.26288-1-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Yong-Taek Lee <ytk.lee@samsung.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/proc/base.c

index 633a6346257346595a71843092fb79efd71c14f8..f5ed9512d193a2d93df6fc54cd2202ad3367fb41 100644 (file)
@@ -1086,10 +1086,6 @@ static int __set_oom_adj(struct file *file, int oom_adj, bool legacy)
 
                        task_lock(p);
                        if (!p->vfork_done && process_shares_mm(p, mm)) {
-                               pr_info("updating oom_score_adj for %d (%s) from %d to %d because it shares mm with %d (%s). Report if this is unexpected.\n",
-                                               task_pid_nr(p), p->comm,
-                                               p->signal->oom_score_adj, oom_adj,
-                                               task_pid_nr(task), task->comm);
                                p->signal->oom_score_adj = oom_adj;
                                if (!legacy && has_capability_noaudit(current, CAP_SYS_RESOURCE))
                                        p->signal->oom_score_adj_min = (short)oom_adj;