I moved the call to event_remove_handler to immediately after xennet_stop as suggested and that looks to have fixed the issue. Thanks!
On Mon, Feb 8, 2016 at 11:51 PM, Manuel Bouyer <bou...@antioche.eu.org> wrote: > On Mon, Feb 08, 2016 at 04:34:20PM -0800, Rohan Desai wrote: > > Thinking about this more, I don't understand this explanation. Won't the > > interrupt level of the cpu still be IPL_NET while detach is sleeping? > > No the spl is lowered while sleeping > > > BTW, > > in my crash detach is waiting on a cv in softint_disestablish() > > If softint_disestablish() can sleep that may be more problematic. > In this case event_remove_handler() should be called before > softint_disestablish(). > > -- > Manuel Bouyer <bou...@antioche.eu.org> > NetBSD: 26 ans d'experience feront toujours la difference > -- >
