From owner-freebsd-current Mon Mar 9 14:48:21 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA19651 for freebsd-current-outgoing; Mon, 9 Mar 1998 14:48:21 -0800 (PST) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from frmug.org (frmug-gw.frmug.org [193.56.58.252]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA19444 for ; Mon, 9 Mar 1998 14:48:01 -0800 (PST) (envelope-from roberto@keltia.freenix.fr) Received: (from uucp@localhost) by frmug.org (8.8.8/frmug-2.2/nospam) with UUCP id XAA09688 for current@FreeBSD.ORG; Mon, 9 Mar 1998 23:47:31 +0100 (CET) (envelope-from roberto@keltia.freenix.fr) Received: (from roberto@localhost) by keltia.freenix.fr (8.8.8/keltia-2.13/nospam) id XAA00358; Mon, 9 Mar 1998 23:38:34 +0100 (CET) (envelope-from roberto) Message-ID: <19980309233834.A297@keltia.freenix.fr> Date: Mon, 9 Mar 1998 23:38:34 +0100 From: Ollivier Robert To: current@FreeBSD.ORG Subject: Re: page fault while in kernel mode (-current from this morning && softupdates) Mail-Followup-To: current@FreeBSD.ORG References: <19980308230222.A9458@keltia.freenix.fr> <13409.889399148@critter.freebsd.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.90.4i In-Reply-To: <13409.889399148@critter.freebsd.dk>; from Poul-Henning Kamp on Mon, Mar 09, 1998 at 12:19:08AM +0100 X-Operating-System: FreeBSD 3.0-CURRENT ctm#4121 AMD-K6 MMX @ 225 MHz Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG According to Poul-Henning Kamp: > Considering that it mentions "allocindir" it would be obvious that it > involves files bigger than what we have in our source-tree. That > should also be the avenue to reproduction I think... Reproducting it is very easy :-) Finding the cause is not. One of the problem I have is that I can work only a kernel core dump, being unable to do remote debugging. The trace I get shows it stopped in _softdep_setup_allocindir_page when the real point of failure is the bcopy at the end of setup_allocindir_phase2. I don't have access to the latter's variables since it is a private function (I guess). I'll make it a public one and see if I can dig something out of this. Either one of the two pointers (newindirdep->ir_saveddata, newindirdep->ir_savebp->b_data) is bad or bp->b_bcount is too big (8192 which seems right for an indirect block on a 8k/1k FS). Just extending an existing file doen't seem to trigger it (just tried it). Using DDB and putting a break point in setup_allocindir_phrase2 shows that the function is called many times but I can't get the bug to manifest inside DDB . I'd appreciate if someone with more internal knowledge than I have could help :-) > >It happens for me in two cases: > >- rnews batch processing, > >- procmail delivering mail into mailboxes. I can see why for the latter case (procmail tries to append a message in one mailbox). The former could be .overview or history append... -- Ollivier ROBERT -=- FreeBSD: The Power to Serve! -=- roberto@keltia.freenix.fr FreeBSD keltia.freenix.fr 3.0-CURRENT #0: Sun Mar 1 18:50:39 CET 1998 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message