From owner-freebsd-arch Tue Aug 20 16:14:45 2002 Delivered-To: freebsd-arch@freebsd.org Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 72C3637B400 for ; Tue, 20 Aug 2002 16:14:43 -0700 (PDT) Received: from mailman.zeta.org.au (mailman.zeta.org.au [203.26.10.16]) by mx1.FreeBSD.org (Postfix) with ESMTP id 3F45D43EA9 for ; Tue, 20 Aug 2002 16:14:41 -0700 (PDT) (envelope-from bde@zeta.org.au) Received: from bde.zeta.org.au (bde.zeta.org.au [203.2.228.102]) by mailman.zeta.org.au (8.9.3/8.8.7) with ESMTP id XAA30888; Tue, 20 Aug 2002 23:14:11 GMT Date: Wed, 21 Aug 2002 09:19:44 +1000 (EST) From: Bruce Evans X-X-Sender: bde@gamplex.bde.org To: Poul-Henning Kamp Cc: Alexander Leidinger , Subject: Re: timestamping kernel messages In-Reply-To: <24056.1029843322@critter.freebsd.dk> Message-ID: <20020821085414.O25287-100000@gamplex.bde.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-arch@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Tue, 20 Aug 2002, Poul-Henning Kamp wrote: > In message <20020820133215.0545063d.Alexander@Leidinger.net>, Alexander Leidinger wr > >what do you think about using timestamps in kernel messages? I think this would be useful for some cases (mainly when syslogd is blocked or not running). > I think we need to revamp the console/logging system in toto. > > I think we should have a pseudo-device for console/log use, which stores > thing in a circular buffer. > > Syslogd would then retreive stuff from that buffer, somewhat like what > it does today. > > For tty console usage, we should have a kernel thread which picks things out > of the buffer and prints it on the chosen console, and putting a timestamp > on it would be a Good Thing. This would only help for broken console drivers. Non-broken ones must be able to do output at almost any time, including in the middle of message buffer pointer update (which should be locked somehow), since they may be invoked then when the code is traced using ddb. I wouldn't want to do the buffering in a device driver. The current message buffer has never been locked correctly for writing despite it being very simple and low level. But read accesses are easy provided the reader doesn't try to lock the writer. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message