Re: [PATCH] Revert mm/vmstat.c: fix vmstat_update() preemption BUG
From: Andrew Morton
Date: Wed Apr 18 2018 - 15:54:35 EST
On Wed, 18 Apr 2018 17:44:36 +0200 Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> wrote:
> On 2018-04-11 21:07:29 [+0200], To Tejun Heo wrote:
> > On 2018-04-11 16:42:21 [+0200], To Tejun Heo wrote:
> > > > > So is this perhaps related to the cpu hotplug that  mentions? e.g. is
> > > > > the cpu being hotplugged cpu 1, the worker started too early before
> > > > > stuff can be scheduled on the CPU, so it has to run on different than
> > > > > designated CPU?
> > > > >
> > > > >  https://marc.info/?l=linux-mm&m=152088260625433&w=2
> > > >
> > > > The report says that it happens when hotplug is attempted. Per-cpu
> > > > doesn't pin the cpu alive, so if the cpu goes down while a work item
> > > > is in flight or a work item is queued while a cpu is offline it'll end
> > > > up executing on some other cpu. So, if a piece of code doesn't want
> > > > that happening, it gotta interlock itself - ie. start queueing when
> > > > the cpu comes online and flush and prevent further queueing when its
> > > > cpu goes down.
> > >
> > > I missed that cpuhotplug part while reading it. So in that case, let me
> > > add a CPU-hotplug notifier which cancels that work. After all it is not
> > > need once the CPU is gone.
> > This already happens:
> > - vmstat_shepherd() does get_online_cpus() and within this block it does
> > queue_delayed_work_on(). So this has to wait until cpuhotplug
> > completed before it can schedule something and then it won't schedule
> > anything on the "off" CPU.
> > - The work item itself (vmstat_update()) schedules itself
> > (conditionally) again.
> > - vmstat_cpu_down_prep() is the down event and does
> > cancel_delayed_work_sync(). So it waits for the work-item to complete
> > and cancels it.
> > This looks all good to me.
> > > > Thanks.
(top-posting repaired, Please don't do that - how am I supposed to
reply to you while maintaining appropriate context?)
> any reason not to accept the revert?
That will make the warnings come back. Or was the hotplug issue
addressed by other means? If so, that fix should be referred to in