From owner-freebsd-bugs Sat Aug 18 11: 0:26 2001 Delivered-To: freebsd-bugs@hub.freebsd.org Received: from freefall.freebsd.org (freefall.freebsd.org [216.136.204.21]) by hub.freebsd.org (Postfix) with ESMTP id ED72437B415 for ; Sat, 18 Aug 2001 11:00:15 -0700 (PDT) (envelope-from gnats@FreeBSD.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.11.4/8.11.4) id f7II0Fh06989; Sat, 18 Aug 2001 11:00:15 -0700 (PDT) (envelope-from gnats) Date: Sat, 18 Aug 2001 11:00:15 -0700 (PDT) Message-Id: <200108181800.f7II0Fh06989@freefall.freebsd.org> To: freebsd-bugs@FreeBSD.org Cc: From: Andre Albsmeier Subject: Re: kern/29845: 4.4-PRERELEASE crashes under heavy net I/O Reply-To: Andre Albsmeier Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org The following reply was made to PR kern/29845; it has been noted by GNATS. From: Andre Albsmeier To: David Malone Cc: Andre Albsmeier , FreeBSD-gnats-submit@freebsd.org, Ian Dowse Subject: Re: kern/29845: 4.4-PRERELEASE crashes under heavy net I/O Date: Sat, 18 Aug 2001 19:55:37 +0200 On Sat, 18-Aug-2001 at 14:49:23 +0100, David Malone wrote: > On Sat, Aug 18, 2001 at 01:33:22PM +0200, Andre Albsmeier wrote: > > I have saved the crashdumps for further examination. As you can see, > > the box crashes in whichever process it wants... > > I'd guess that something is freeing an mbuf while it is still in > use. This would result in either a panic when the mbuf is corrupted > while in use or a double freeing of the mbuf. This could plausable > explain the panics you included trace backs for. > > I think Ian Dowse has some tools for examining the mbuf free lists > in kernel dumps. He did also have some patches for catching writes > to shared or free mbuf clusters, which might help figure out what's > going on here. As I said: I am glad to try anything. > The only thing that doesn't tally is that this is only effecting > your laptop and not all your machines. The first thing I thought of was a hardware problem. But the old version ran fine as does Win98 :-). But: I am using the Intel Etherexpress 100MBit PCMCIA card with the xe driver. The driver is somehow inefficient: When doing heavy net I/O with it, the load gets up to 4 and higher. It has always been like this. Maybe some changes with the mbuf handling and Warners recent pccard commits cause these problems under load now. Sometimes I can ftp the crashdumps to another machine, sometimes not. Hmm, I have the same box again at work. On monday I will swap the harddrives and see how this behaves... -Andre To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message