Date: Tue, 31 Oct 1995 13:21:08 -0500 From: "Garrett A. Wollman" <wollman@lcs.mit.edu> To: "Justin T. Gibbs" <gibbs@freefall.FreeBSD.org> Cc: "Garrett A. Wollman" <wollman@lcs.mit.edu>, current@FreeBSD.org, wollman@FreeBSD.org Subject: Re: Time problems Message-ID: <9510311821.AA29851@halloran-eldar.lcs.mit.edu> In-Reply-To: <199510311805.KAA08628@aslan.cdrom.com> References: <9510311747.AA29686@halloran-eldar.lcs.mit.edu> <199510311805.KAA08628@aslan.cdrom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Tue, 31 Oct 1995 10:05:22 -0800, "Justin T. Gibbs" <gibbs@freefall.FreeBSD.org> said: > yields poor results, or simply to fix up the time occasionally? I > thought the main reason for using this approach was to remove the > overhead of going out to the 8254, No. The main reason for using this approach is to ensure that `time' is incremented at the right rate. Previously, if a clock interrupt was late, the system would make no attempt to correct for that fact. Now it does. (Yes, that's a fringe case, but it can be a significant source of jitter under the old mechanism.) I have considered a number of mechanisms which might automatically detect when this situation has occurred, but I still hold out hope that the underlying cause for the mis-diagnosis might be fixed and the extra overhead will not be necessary. Even under the `old' way the cycle counter was still being used for microtime(), and was still giving incorrect results on your machine. Now it's just a bit more obvious about it. You should have complained six months ago that your CPU speed was not being correctly diagnosed. > Precision and overhead right? No, just precision. The overhead is actually slightly increased, since before we never checked to see how much time really elapsed between timer ticks; now we do. > Why is it that xntpd doesn't either > complain about my large offset (I started it just after using ntpdate), > or keep my time in sync? xntpd won't complain about a large offset at any time other than startup, and your clock is off by so much that it exceeds the bounds of the correction mechanism (and may very well have xntpd completely confused). -GAWollman -- Garrett A. Wollman | Shashish is simple, it's discreet, it's brief. ... wollman@lcs.mit.edu | Shashish is the bonding of hearts in spite of distance. Opinions not those of| It is a bond more powerful than absence. We like people MIT, LCS, ANA, or NSA| who like Shashish. - Claude McKenzie + Florent Vollant
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9510311821.AA29851>