Date: Mon, 8 Sep 1997 08:10:10 +0200 From: j@uriah.heep.sax.de (J Wunsch) To: hackers@FreeBSD.ORG Cc: ywliu1@tao.sinanet.com.tw (Hoffmann Yen-Wei Liu) Subject: Re: Kernel clock runs inaccurately Message-ID: <19970908081010.DS04250@uriah.heep.sax.de> In-Reply-To: <199709080124.JAA14004@tao.sinanet.com.tw>; from Hoffmann Yen-Wei Liu on Sep 8, 1997 09:24:11 %2B0800 References: <199709080124.JAA14004@tao.sinanet.com.tw>
next in thread | previous in thread | raw e-mail | index | archive | help
As Hoffmann Yen-Wei Liu wrote: > I just got myself a new AMD k6-200, and I found out a major problem : > it runs about 30 seconds faster per day. So two days later, it runs > 2 minutes faster. However, the CMOS clock runs around 2 seconds slower > per day. This is acceptable to me. The master clock on my machine at work is much worse. I was contemplating to hack xntpd so it could use the RTC (CMOS) clock as a refclock, but unfortunately, didn't get very far. Mainly, i don't grok the xnptd refclock implementations, and how to add a new one. An alternative solution would be to hack a small userland daemon that wakes up every hour, reads the RTC, and calls adjtime(2) to compensate for the drift. I didn't want to do it this way because it's more of a hack than the xntpd solution would be, but i might go this route. All the people i've been asking for help with xntpd haven't been any help for me, they only attempted to explain me that i don't wanna do what i do want. (They are probably living in somewhat of an ivory tower with good NTP refclocks readily available on a cheap Internet, something that is not my situation, sitting behind dialup lines everywhere.) -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970908081010.DS04250>