Re: [PATCH] mm,page_alloc: Serialize out_of_memory() and allocation stall messages.

Tetsuo Handa wrote: > While warn_alloc() messages are completely unreadable, what we should note > are that > > (a) out_of_memory() => oom_kill_process() => dump_header() => show_mem() => > printk() > got stuck at console_unlock() despite this is schedulable context. > > -- > 2180

Re: [PATCH] mm,page_alloc: Serialize out_of_memory() and allocation stall messages.

Tetsuo Handa wrote: > Michal Hocko wrote: > > Hell no! I've tried to be patient with you but it seems that is just > > pointless waste of time. Such an approach is absolutely not acceptable. > > You are adding an additional lock dependency into the picture. Say that > > there is somebody stuck in w

Re: [PATCH] mm,page_alloc: Serialize out_of_memory() and allocation stall messages.

Michal Hocko wrote: > On Thu 19-10-17 19:51:02, Tetsuo Handa wrote: > > The printk() flooding problem caused by concurrent warn_alloc() calls was > > already pointed out by me, and there are reports of soft lockups caused by > > warn_alloc(). But this problem is left unhandled because Michal does n

Re: [PATCH] mm,page_alloc: Serialize out_of_memory() and allocation stall messages.

On Thu 19-10-17 19:51:02, Tetsuo Handa wrote: > The printk() flooding problem caused by concurrent warn_alloc() calls was > already pointed out by me, and there are reports of soft lockups caused by > warn_alloc(). But this problem is left unhandled because Michal does not > like serialization from

[PATCH] mm,page_alloc: Serialize out_of_memory() and allocation stall messages.

The printk() flooding problem caused by concurrent warn_alloc() calls was already pointed out by me, and there are reports of soft lockups caused by warn_alloc(). But this problem is left unhandled because Michal does not like serialization from allocation path because he is worrying about unexpect