From owner-freebsd-hackers Mon Feb 16 14:52:42 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA25377 for freebsd-hackers-outgoing; Mon, 16 Feb 1998 14:52:42 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from alpo.whistle.com (alpo.whistle.com [207.76.204.38]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA25227 for ; Mon, 16 Feb 1998 14:52:08 -0800 (PST) (envelope-from julian@whistle.com) Received: (from daemon@localhost) by alpo.whistle.com (8.8.5/8.8.5) id OAA08626; Mon, 16 Feb 1998 14:42:12 -0800 (PST) Received: from UNKNOWN(), claiming to be "current1.whistle.com" via SMTP by alpo.whistle.com, id smtpd008623; Mon Feb 16 14:42:08 1998 Date: Mon, 16 Feb 1998 14:38:16 -0800 (PST) From: Julian Elischer To: Simon Shapiro cc: freebsd-hackers@FreeBSD.ORG Subject: Re: WatchDog Timer In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG look in wd.c (for example) for the code: #ifdef HW_WDOG if (wdog_tickler) (*wdog_tickler)(); #endif this is in the code to make a core dump and it tickles the watchdog each 1MB of core saved... we added it for our h/w watchdog at whistle. the existance of the function pointer is the entire support in freeBSD for a hardware watchdog, and took a whole 4 bytes :) we got tired of getting HALF of a core-dump :) under normal circumstances, the watchdog is tickled by a userland process. On Mon, 16 Feb 1998, Simon Shapiro wrote: > > On 16-Feb-98 Julian Elischer wrote: > > eh? > > what watchdog? > > we have a hook that is called by the kernel > > when it is doing long uninterrupted operations (e.g. dumping core) > > but it's just a hook. The user needs to supply their own hardware > > watchdog.. (and code to tickle it) > > > > julian > > If there is no driver for such hardware today, I'll write it this week. > This is really not the same as a kernel's internal watchdog.. What we are > talking about here is used to build a dead-man switch; In case of O/S > catastrophic failure, this device will reboot the machine, or in the case > of HA/FT, signal to the rest of the ``system'' that a Unix instance just > died. > > I simply do not want to write a driver that was already written. > What is the kernel's hook name/path? Maybe we can merge the two, or at > least relate to each other. As in ``If the kernel is busy with a long, > uninterrupted operation, we automatically mask off the dead-man switch''. > > ---------- > > > Sincerely Yours, > > Simon Shapiro > Shimon@Simon-Shapiro.ORG Voice: 503.708.7858 > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message