From owner-freebsd-smp Tue Jun 30 12:19:07 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id MAA15981 for freebsd-smp-outgoing; Tue, 30 Jun 1998 12:19:07 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from critter.freebsd.dk (critter.freebsd.dk [195.8.133.1] (may be forged)) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA15971 for ; Tue, 30 Jun 1998 12:18:58 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.8.7/8.8.5) with ESMTP id VAA15174; Tue, 30 Jun 1998 21:13:34 +0200 (CEST) To: John Hay cc: freebsd-smp@FreeBSD.ORG Subject: Re: time problem? In-reply-to: Your message of "Sat, 30 Jun 1998 20:47:42 +0200." <199806301847.UAA19830@zibbi.mikom.csir.co.za> Date: Tue, 30 Jun 1998 21:13:33 +0200 Message-ID: <15172.899234013@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> >I followed Tony Li's advice and started ntpd again on the machine with >> >no local reference clocks and the machine has now been up for more than >> >24 hours with no jumps at all, so it looks like it is calling nanotime() >> >from within the sio interrupt that causes the problems. :-( >> >> yes, that may not be entirely safe to do according to Bruce. >> >> >So what are my options then? Can I somehow run the sio interrupts at >> >a lower priority? I would really prefer to have the pps signal on the >> >serial port, because the port is already used to read the serial stream >> >from the GPS. >> >> I don't know :-) > >Hmmm, so it seems that I'll have to get the soldering out and get the >pps signal hooked up to the printer port to see if that works better. :-) He's not too happy about that either... >What I do know is the way you calculate the second pararameter for >hardpps() in your printer port pps driver and the way I did it in >my patch, seems to make it very unstable. :-) yes, that parameter is a kludge for the hardclock() interpolation stuff which Dave Mills wrote for alphas. >> And it doesn't happen if you run !SMP with the TSC timecounter, right ? > >I'll have a look at it again. Please do, because the TSC doesn't require the interrupts to be disabled, so that would be a very valuable datapoint for us... -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." "ttyv0" -- What UNIX calls a $20K state-of-the-art, 3D, hi-res color terminal To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message