On Monday 14 December 2009, Martin Michlmayr wrote: > > So, to me this looks like a calculation error in rdate, probably in > > ntp.c. > > ntpdate uses adjtimex() to fix this but this is a Linux-specific > syscall so David is reluctant to use it in rdate. Maybe rdate can fix > the value manually somehow.
Are you sure? IIUC adjtimex() does gradual adjustment of the clock, but that is not what happens here. What's happening here, both in rdate and ntpdate is a one-shot major adjustment of the system time/date. I think the adjtimex() call is unrelated to the actual setting of the time by ntpdate. The real difference is in the values of the settimeofday() calls and how they are calculated. If I misunderstand and adjtimex() somehow really is the solution, then probably adjtime() could be used instead. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
