From owner-freebsd-current Tue Sep 15 15:41:00 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA27308 for freebsd-current-outgoing; Tue, 15 Sep 1998 15:41:00 -0700 (PDT) (envelope-from owner-freebsd-current@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 PAA27275 for ; Tue, 15 Sep 1998 15:40:51 -0700 (PDT) (envelope-from julian@whistle.com) Received: (from daemon@localhost) by alpo.whistle.com (8.8.5/8.8.5) id PAA29869; Tue, 15 Sep 1998 15:22:03 -0700 (PDT) Received: from current1.whistle.com(207.76.205.22) via SMTP by alpo.whistle.com, id smtpdr29852; Tue Sep 15 22:22:02 1998 Date: Tue, 15 Sep 1998 15:21:58 -0700 (PDT) From: Julian Elischer To: Brian Feldman cc: Luoqi Chen , ben@rosengart.com, current@FreeBSD.ORG, bde@godzilla.zeta.org.au Subject: Re: soft updates panic In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG already done On Tue, 15 Sep 1998, Brian Feldman wrote: > I haven't heard back from you about this, but I suppose I shall take the > initiative. YES this does fix all the SoftUpdates instability I've had. I > can make -j4 world to my heart's content, with no crashes at all. Should > it be worth committing? And if it is the wrong thing to do for standard > ffs, maybe and #ifdef SOFTUPDATES #else #endif workaround? > > Cheers, > Brian Feldman > > On Fri, 11 Sep 1998, Luoqi Chen wrote: > > > > Another crash in make -j3 world last night, but no panic or core dump > > > this time. > > > > > > It's disturbing to me to see this this close to release. > > > > > > > > > Ben > > > > > > "You have your mind on computers, it seems." > > > > > I'm also able to reproduce this panic quite reliably. I tracked it down > > to the lost of a directory size increase. It may have something to do > > with Bruce's ffs_update() change in July (ffs_inode.c, -r1.43->1.44): > > many cases of the in-core update are no longer copied to the inode disk > > buffer. Please try the attached patch that restores the old behavior, and > > let me know if it helps. I was able to do a make -j8 buildworld myself, > > successfully up to the point that perl5 failed me 8( > > > > -lq > > > > Index: ffs_inode.c > > =================================================================== > > RCS file: /home/ncvs/src/sys/ufs/ffs/ffs_inode.c,v > > retrieving revision 1.46 > > diff -u -r1.46 ffs_inode.c > > --- ffs_inode.c 1998/07/04 20:45:38 1.46 > > +++ ffs_inode.c 1998/09/11 15:13:09 > > @@ -81,10 +81,12 @@ > > struct inode *ip; > > int error; > > > > - ufs_itimes(vp); > > ip = VTOI(vp); > > - if ((ip->i_flag & IN_MODIFIED) == 0 && waitfor != MNT_WAIT) > > + if (((ip->i_flag & > > + (IN_ACCESS | IN_CHANGE | IN_MODIFIED | IN_UPDATE)) == 0) && > > + (waitfor != MNT_WAIT)) > > return (0); > > + ufs_itimes(vp); > > ip->i_flag &= ~(IN_LAZYMOD | IN_MODIFIED); > > if (vp->v_mount->mnt_flag & MNT_RDONLY) > > return (0); > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > > with "unsubscribe freebsd-current" in the body of the message > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-current" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message