From owner-freebsd-current Tue Oct 17 04:53:10 1995 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id EAA23562 for current-outgoing; Tue, 17 Oct 1995 04:53:10 -0700 Received: from bunyip.cc.uq.oz.au (pp@bunyip.cc.uq.oz.au [130.102.2.1]) by freefall.freebsd.org (8.6.12/8.6.6) with SMTP id EAA23557 for ; Tue, 17 Oct 1995 04:53:07 -0700 Received: from cc.uq.oz.au by bunyip.cc.uq.oz.au id <02456-0@bunyip.cc.uq.oz.au>; Tue, 17 Oct 1995 19:55:37 +1000 Received: from orion.devetir.qld.gov.au by pandora.devetir.qld.gov.au (8.6.10/DEVETIR-E0.3a) with ESMTP id SAA24875; Tue, 17 Oct 1995 18:55:02 +1000 Received: by orion.devetir.qld.gov.au (8.6.10/DEVETIR-0.3) id SAA18528; Tue, 17 Oct 1995 18:51:17 +1000 Date: Tue, 17 Oct 1995 18:51:17 +1000 From: Stephen McKay Message-Id: <199510170851.SAA18528@orion.devetir.qld.gov.au> To: David Greenman cc: freebsd-current@freebsd.org, syssgm@devetir.qld.gov.au Subject: Re: /kernel: biodone: buffer already done Sender: owner-current@freebsd.org Precedence: bulk On Mon, 2 Oct 1995 10:10:15 GMT, David Greenman explained: >>My memory starved test box has thrown up this: >> >>Oct 2 15:10:26 stupid /kernel: biodone: buffer already done >>Oct 2 15:10:27 stupid /kernel: biodone: buffer already done > > This is a "known" problem. It's apparantly caused by inadequate vnode >locking in the NFS code and the result is a race condition which then causes >the above. > >> UID PID PPID CPU PRI NI VSZ RSS WCHAN STAT TT TIME COMMAND >> 0 149 148 109 -18 5 300 12 objtrm DN p0 0:00.22 mv /tmp/_depend148 .depend > ^^^^^^ > >Yep, that's another symptom of the bug. We'll hopefully get this fixed soon. Just in case it throws any light on the matter, I've now got this hung process: UID PID PPID CPU PRI NI VSZ RSS WCHAN STAT TT TIME COMMAND 1056 16443 167 46 -18 0 144 24 vmopar D+ p0 0:00.43 tail /ctm/LOG and, yes, /ctm is NFS mounted. This hang came after the usual pair of "biodone: buffer already done" messages. The 'make world' is still going though, and writes to /ctm/LOG are still working. I'm running -current from 1995-10-15 21:39:29 (src-cur 1082). Stephen.