Re: [PATCH] mm,oom: remove shortcuts for SIGKILL and PF_EXITING cases

2016-02-22 Thread David Rientjes
On Sun, 21 Feb 2016, Tetsuo Handa wrote: > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index ae8b81c..390ec2c 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -1253,16 +1253,6 @@ static void mem_cgroup_out_of_memory(struct mem_cgroup > *memcg, gfp_t gfp_mask, > >

Re: [PATCH] mm,oom: remove shortcuts for SIGKILL and PF_EXITING cases

2016-02-22 Thread David Rientjes
On Sun, 21 Feb 2016, Tetsuo Handa wrote: > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index ae8b81c..390ec2c 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -1253,16 +1253,6 @@ static void mem_cgroup_out_of_memory(struct mem_cgroup > *memcg, gfp_t gfp_mask, > >

[PATCH] mm,oom: remove shortcuts for SIGKILL and PF_EXITING cases

2016-02-21 Thread Tetsuo Handa
Currently, sending SIGKILL to all user processes sharing the same memory is omitted by three locations. But they should be removed due to possible OOM livelock sequence shown below. About setting TIF_MEMIE on current->mm && fatal_signal_pending(current) at out_of_memory(): There are two thread

[PATCH] mm,oom: remove shortcuts for SIGKILL and PF_EXITING cases

2016-02-21 Thread Tetsuo Handa
Currently, sending SIGKILL to all user processes sharing the same memory is omitted by three locations. But they should be removed due to possible OOM livelock sequence shown below. About setting TIF_MEMIE on current->mm && fatal_signal_pending(current) at out_of_memory(): There are two thread