Received: from ATHENA-AS-WELL.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA19950; Sun, 6 Jun 93 07:40:56 EDT
Received: from ACHATES.MIT.EDU by Athena.MIT.EDU with SMTP
	id AA03815; Sun, 6 Jun 93 07:40:53 EDT
Received: by Achates.MIT.EDU (5.61) id AA19818; Sun, 6 Jun 93 07:40:50 -0400
Message-Id: <9306061140.AA19818@Achates.MIT.EDU>
To: Marc Horowitz <marc@MIT.EDU>
Cc: sipb-afsreq@Athena.MIT.EDU, jweiss@Athena.MIT.EDU, mhbraun@Athena.MIT.EDU
Subject: Re: opus 
In-Reply-To: Your message of "Sun, 06 Jun 1993 02:10:59 EDT."
             <9306060611.AA24946@bill-the-cat.MIT.EDU> 
Date: Sun, 06 Jun 1993 07:40:50 EDT
From: John Carr <jfc@Athena.MIT.EDU>


Make that 11 times.  It crashed again around 4:30 AM.  This means the
bufcache increase was not the cause, and it should be set back to 25%
(kvar -s bufcache -wl -v 10 /vmunix).

It is crashing in this code in ufs_gnode.c:

			if (gp->g_flag & GACC) {
				if (ta == &atime || ta == &btime)
					gp->g_atime = guniqtm;
				else
					gp->g_atime= *ta;
			}

ta is apparently NULL; it is supposed to be a pointer to the value to
set the inode access time to.

I'm still trying to figure out the exact call path that led to this
code.