From owner-freebsd-current Thu Oct 3 10:56:14 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id KAA11543 for current-outgoing; Thu, 3 Oct 1996 10:56:14 -0700 (PDT) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id KAA11538; Thu, 3 Oct 1996 10:56:12 -0700 (PDT) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id KAA06374; Thu, 3 Oct 1996 10:53:50 -0700 From: Terry Lambert Message-Id: <199610031753.KAA06374@phaeton.artisoft.com> Subject: Re: HELP!! kernel deadlock found.. To: dfr@render.com (Doug Rabson) Date: Thu, 3 Oct 1996 10:53:50 -0700 (MST) Cc: julian@freefall.freebsd.org, current@freefall.freebsd.org In-Reply-To: from "Doug Rabson" at Oct 3, 96 11:09:13 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > Take the following 3 processes: > > > > proc N, with a lock on file / (inode 2) > > wchan of that inode, waitstring of "ufslk2" > > is waiting for inode for /mnt in the root filesystem (inode M) > > > > proc N+1 with a lock on the inode M (/mnt in root filesystem) > > is waiting for inode for / (inode 2) in the mounted filesystem /mnt > > it is showing "uihget" as a waitstring. > > > > proc N+2 with a lock on inode 2 of the mnt filesystem (/ of that filesystem) > > is waiting for the inode for / and is showing "ufslk2" as a waitstring. > > > > It is my suspicion that process N+2 may be trying to unmount /mnt. > > It looks as if N+2 is attempting to perform a pathname lookup while it has > a locked vnode. I'm glad Doug jumped in here. 8-). I think the lookup might be the result of the NFS export handling. Is this system an NFS server? Do you remember what the exports list looked like? I think the current handling of the NFS stuff is a bit bogus. I'd like to eventually see an exportfs mechanism used instea, and take the string list mechanism out of the kernel entirely. The idea that I have to enable an FS for NFS export irks me a bit. All FS's, by their very nature as FS's, should be intrinsically consumable by all VFS consumers. NFS is a VFS consumer, just like the system calls are a VFS consumer. Just like a decent LKM system or kernel threading system would be a VFS consumer. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.