I'm seeing a funny very low probability anomaly when time-stamping
events using the pps com-port facility on -current/i386.  It looks
like every once in a long while the time-stamp indicates a time 10ms
(one clock-tick ?) early.

        http://www.wsrcc.com/wolfgang/ftp/pps-offset.ps

This spike only happens approximately 1 time out of 2k. Could this be
a race or lock-bug while some counter is rolling-over?

-wolfgang
-- 
       Wolfgang Rupprecht <wolfgang+gnus@dailyplanet.wsrcc.com>
		    http://www.wsrcc.com/wolfgang/
DGPS signals via the Internet  http://www.wsrcc.com/wolfgang/gps/dgps-ip.html