Re: [RFC 0/9] mce recovery for Sandy Bridge server

[Ingo Molnar <mingo@elte.hu>]

* Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:

> On Tue, 2011-05-24 at 14:30 -0700, Linus Torvalds wrote:
> > On Tue, May 24, 2011 at 2:24 PM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
> > >
> > > Right, so you can't do things like that from NMI context, but what perf
> > > can do is raise a self-IPI and continue from IRQ context (question for
> > > the HW folks, can there be cycles between the NMI iret and IRQ assert
> > > from whatever context was before the NMI hit?)
> > 
> > Of course there can be - the code where the NMI hit may have
> > interrupts disabled.
> 
> D'0h yes of course!..
> 
> Then I can't immediately see a way to stop a thread dead in its tracks.
> Userspace yes, but not kernel-space.

It's a best-effort thing: run proper kernel code (hardirq context) as soon as 
possible - but obviously not sooner than that! :-)

Thanks,

	Ingo