Re: BUG: using smp_processor_id() in preemptible arch_trigger_all_cpu_backtrace_handler

[Cyrill Gorcunov <gorcunov@gmail.com>]

On Fri, Nov 05, 2010 at 05:06:55PM +0100, Frederic Weisbecker wrote:
> 2010/11/1 Cyrill Gorcunov <gorcunov@gmail.com>:
...
> >  yup, this will do the trick for a while. In general I believe we might have
> > kind of NMI exclusive chain so we wouldn't need the 'case:'s.
> 
> Yeah.  And seperating NMIs from the rest of the DIE notifiers would
> probably improve
> performance of things like PMI handling.
> 

It will make it more straight and clean as minimum which is already
quite a benefit ;) There were patches floating from Don with notify
priorities which are good start. Though I didn't manage to look into
most of patches I've been CC'ed yet :(

> And I've always been confused with this "die" notifier semantic. We
> are not dying when
> we handle a counter overflow interrupt.
> The same applies to DIE_INT3, DIE_TRAP, DIE_DEBUG, ....
> 

As the comment says on top of the enum they are "Grossly misnamed" ;)
Though we could consider them not as "dying" but rather in terms of say
on-die signals (or something like that).

> But until then, as having a seperate notifier is quite a refactoring,
> we should enqueue Don's fix.
> Don, can you resend it with usual SOB and changelog?
> 
> Thanks.
> 
  Cyrill