On Tue, Nov 11, 2014 at 05:18:56PM +0900, Ryota Ozaki wrote:
> + callout_halt(&in6m->in6m_timer_ch, softnet_lock);
> callout_destroy(&in6m->in6m_timer_ch);
Certainly a valid change - IMHO all callout_destroy() should have an
explicit callout_halt() unless we can be 100% sure no callout is in flight
for other reasons (which probably warrants a comment at the callout_destroy
call).