Suppose a periodic kevent timer fires close to its deadline, so now -
kc->next is small. Then delta ends up being 1, and the next timer
deadline is set to (delta + 1) * kc->to, where kc->to is the timer
period. This means that the timer fires at half of the requested rate.
And, because the computation of delta rounds down, it seems that the
kn_data value is incorrect as well.
PR: 264131
Fixes: 7cb40543e964 ("filt_timerexpire: do not iterate over the interval")