On Wed, Jun 18, 2014 at 04:53:14PM +0200, Jiri Kosina wrote:
> On Wed, 18 Jun 2014, Paul E. McKenney wrote:
>
> > > > > - both RCU stall detector and 'echo l > sysrq-trigger' can (and we've
> > > > > seen it happening for real) cause a complete, undebuggable, silent
> > > > > hang
> > > > > of machine (deadlock in NMI context)
> > > >
> > > > I could easily add an option to RCU to allow people to tell it not to
> > > > use NMIs to dump the stack. Would that help?
> > >
> > > Well, that would make unfortunately the information provided by RCU stall
> > > detector rather useless ... workqueue-based stack dumping is very
> > > unlikely
> > > to point its finger to the real offender, as it'd be coming way too late.
> >
> > I would not use workqueues, but rather have the CPU detecting the
> > stall grovel through the other CPUs' stacks, which is what I do now for
> > architectures that don't support NMI-based stack dumps. Would that be
> > a reasonable approach?
>
> That would indeed solve lockups induced by RCU stall detector (and we
> should convert sysrq stack dumping code to use the same mechanism
> afterwards).
>
> But then, the kernel is still polluted by quite a few instances of
>
> WARN_ON(in_nmi())
>
> BUG_IN(in_nmi())
>
> if (in_nmi())
> printk(....)
>
> which need to be fixed separately afterwards anyway.
True enough!
Thanx, Paul
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
