Date: Wed, 6 Feb 2002 21:17:34 +1100 (EST) From: Bruce Evans <bde@zeta.org.au> To: David Greenman <dg@root.com> Cc: <bugs@FreeBSD.ORG> Subject: Re: [legg@iastate.edu: Serious exception in FreeBSD found] Message-ID: <20020206210601.Y2405-100000@gamplex.bde.org> In-Reply-To: <20020205161313.B19463@nexus.root.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 5 Feb 2002, David Greenman wrote: > ----- Forwarded message from Tim Legg <legg@iastate.edu> ----- > ... > The tar file was being produced when the ttyv0 was being filled with > messages. Here is an example: > > Feb 5 17:43:56 pc047113 /kernal: microuptime() went backwards (1550895.778303 -> 1550895.-694593171) > Feb 5 17:43:56 pc047113 /kernal: microuptime() went backwards (1550895.839382 -> 1550895.832517) > Feb 5 17:43:56 pc047113 /kernal: microuptime() went backwards (1550895.291637 -> 1550895.274690) I fixed the one cause of negative tv_usec's recently (rev.1.111 of kern_tc.c and rev.1.105.2.7 of kern_clock.c), but this tends to make the spew of error messages worse if the timecounter actually goes backwards. Previously the timestamp comparisons were confused enough by the invalid tv_usec's to limit the spew. Since tv_usec went negative, we almost know that timecounter update was delayed for more than (1 - 1/hz) seconds. This can only happen if there are serious interrupt latency bugs elsewhere. > This made my /var/log/messages extremely huge very quickly. My > /var/log/messages was growing at a rate of approximately 70,000 lines per > minute. This is another example of why non-critical kernel printfs should be rate-limited. The "vm_fault: pager read error" regularly fills up my /var when I abuse nfs (run make install on the server). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020206210601.Y2405-100000>