Andrea,
I think the lost_ticks accounting in gettimeofday can be slightly wrong
in the presence of ntp, thus beating the purpose of ntp. Instead of
fixed size tick = 1000000/HZ we should be using tick+time_adjust_step
from update_wall_time_one_tick() in timer.c. This solution can be wrong
if there were many lost ticks, and it is slightly off because of
possible additional coorection from time_phase.
Also, I still don't understand how lost_ticks can be >1. What can
inhibit bh from running between two ticks?
Boris
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 23 2000 - 21:00:21 EST