Re: [PATCH 4/10] cpu: deliver CPU_UP_CANCELED only to NOTIFY_OKed callbacks with CPU_UP_PREPARE

["Akinobu Mita" <akinobu.mita@gmail.com>]

> > >[...] However, it might break slab.
> > >If I am not mistaken, slab code initializes multiple objects in
> > >CPU_UP_PREPARE and relies on the CPU_UP_CANCELLED to destroy the
> > >objects which successfully got initialized before the some object's
> > >initialization went bad.
> >
> > My testing machine is ordinary dual core non numa box. So it might not
> > trigger the problem that you are warried about under heavy slab alloc
> > failure injection.
> >
> > At first glance I couln't find the problem in cpu hottplug code in slab.c
> > yet,
> > but found some memory leak path. (it doesn't break slab though)
>
> That's what I meant. I shouldn't have used the word "break" :-)
> In case of slab, freeing up of resources on an error during CPU_UP_PREPARE,
> is currently handled in CPU_UP_CANCELLED.

Now I perfectly understand your concern. The last memleak fix
patch did not cover for each cachep->array[cpu] in cache_chain.
So cpu hotplug error handling in slab becomes worse by this change.

> But, like you reasoned out, it makes more sense for such a subsystem
> to free up all the correctly allocated resources before sending a
> NOTIFY_BAD, rather than handling it in CPU_UP_CANCELLED. And slab
> needed that fix, which you've provided, before we send the notification
> to (nr_calls - 1) callers.
>
> So could you add this patch to series?

Sure, and I'll CC you on the slab change.