On Mon 23-05-16 19:59:30, Tetsuo Handa wrote: > Michal Hocko wrote: > > On Sat 21-05-16 11:01:30, Tetsuo Handa wrote: > > > Currently, oom_scan_process_thread() returns OOM_SCAN_SELECT if > > > oom_task_origin() returned true. But this might cause OOM livelock. > > > > > > If the OOM killer finds a task with oom_task_origin(task) == true, > > > it means that that task is either inside try_to_unuse() from swapoff > > > path or unmerge_and_remove_all_rmap_items() from ksm's run_store path. > > > > > > Let's take a look at try_to_unuse() as an example. Although there is > > > signal_pending() test inside the iteration loop, there are operations > > > (e.g. mmput(), wait_on_page_*()) which might block in unkillable state > > > waiting for other threads which might allocate memory. > > > > > > Therefore, sending SIGKILL to a task with oom_task_origin(task) == true > > > can not guarantee that that task shall not stuck at unkillable waits. > > > Once the OOM reaper reaped that task's memory (or gave up reaping it), > > > the OOM killer must not select that task again when oom_task_origin(task) > > > returned true. We need to select different victims until that task can > > > hit signal_pending() test or finish the iteration loop. > > > > > > Since oom_badness() is a function which returns score of the given thread > > > group with eligibility/livelock test, it is more natural and safer to let > > > oom_badness() return highest score when oom_task_origin(task) == true. > > > > > > This patch moves oom_task_origin() test from oom_scan_process_thread() to > > > after MMF_OOM_REAPED test inside oom_badness(), changes the callers to > > > receive the score using "unsigned long" variable, and eliminates > > > OOM_SCAN_SELECT path in the callers. > > > > I do not think this is the right approach. If the problem is real then > > the patch just papers over deficiency of the oom_task_origin which > > should be addressed instead. > > > > > Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> > > > > Nacked-by: Michal Hocko <mhocko@xxxxxxxx> > > > > Quoting from http://lkml.kernel.org/r/20160523075524.GG2278@xxxxxxxxxxxxxx : > > > Is it guaranteed that try_to_unuse() from swapoff is never blocked on memory > > > allocation (e.g. mmput(), wait_on_page_*()) ? > > > > It shouldn't. All the waiting should be killable. If not it is a bug and > > should be fixed. > > So, you think that we should replace mmput() with mmput_async(), lock_page() > with lock_page_killable(), wait_on_page_bit() with wait_on_page_bit_killable(), > mutex_lock() with mutex_lock_killable(), down_read() with down_read_killable() > and so on, don't you? Yes where appropriate. And even more importantly. First think whether you are trying to address a real problem. It doesn't make any sense to complicate the code for something that is simply not realistic. In this particular case we are talking about _root_ doing a potentially expensive operation and I am pretty sure no reasonable admin would do it under a high memory pressure. So all the cases you are trying to address are maybe completely unrealistic. -- Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>