From owner-freebsd-questions Sat Feb 12 21:31:19 2000 Delivered-To: freebsd-questions@freebsd.org Received: from scientia.demon.co.uk (scientia.demon.co.uk [212.228.14.13]) by builder.freebsd.org (Postfix) with ESMTP id C51CE44A2; Sat, 12 Feb 2000 21:30:59 -0800 (PST) Received: from strontium.scientia.demon.co.uk ([192.168.91.36] ident=exim) by scientia.demon.co.uk with esmtp (Exim 3.12 #1) id 12JrVh-000GmF-00; Sun, 13 Feb 2000 05:23:53 +0000 Received: (from ben) by strontium.scientia.demon.co.uk (Exim 3.12 #1) id 12JrVh-0004WZ-00; Sun, 13 Feb 2000 05:23:53 +0000 Date: Sun, 13 Feb 2000 05:23:53 +0000 From: Ben Smithurst To: questions@freebsd.org, current@freebsd.org Subject: timed/adjtime() on -current Message-ID: <20000213052353.C16083@strontium.scientia.demon.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG [sorry about the crosspost - I'm not sure if this is me being a dumbass, or something wrong with adjtime() on current. adjtime() certainly behaves as I expect it to on stable.] I've recently updated one of my machines to -current, and now timed(8) seems to have stopped working. Is there some reason why timed(8) on -current wouldn't like the master being on -stable? (I can't see one.) Enabling tracing with timedc shows the corrections that timed thinks it should be making, and ktracing the timed process shows it is calling adjtime, but nothing seems to be changing the time. Hmm. timed is doing something very wierd... before starting timed, my clock seems to be fairly stable (gaining about 1ms per minute, but I can cope with that). As soon as I start timed, the clock starts gaining about 1ms per 2 seconds, which is a bit much. Actually, now I've written this, it's gaining about 5ms per second. Soon after I kill timed, the clock goes back to normal. So, I write a small program which calls adjtime itself... if I specify tv_sec=-4 and tv_usec=0, this seems to speed the clock up. tv_sec=4 seems to slow it down, in direct contradiction of the manpage and common sense: time on scientia.demon.co.uk is 4366 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4316 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4266 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4216 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4166 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4116 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4066 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 4016 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 3966 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 3916 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 3866 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 3816 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 3766 ms. behind time on platinum.scientia.demon.co.uk time on scientia.demon.co.uk is 3716 ms. behind time on platinum.scientia.demon.co.uk (platinum is the -current machine, scientia is the -stable timed master). This seems wierd... anyone know what's going on here? (It *is* 5 a.m., so I may be missing something obvious.) adjtime() seems to behave as I expect it to on -stable. If you want more information, just let me know. This is a fairly recent -current, cvsupped around 1 a.m. GMT saturday morning. -- Ben Smithurst / ben@scientia.demon.co.uk / PGP: 0x99392F7D To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message