Date: Sat, 18 Nov 2006 19:19:02 +0100 From: Vincent Blondel <vincent@xtra-net.org> To: freebsd-stable@freebsd.org Cc: vincent@xtra-net.org, kris@obsecurity.org Subject: Re: kernel crash ... Message-ID: <1163873942.3315.12.camel@wbemfkaa.net.xtra-net.be> In-Reply-To: <20061117223140.GA13858@xor.obsecurity.org> References: <1163621364.85632.12.camel@wbemfkaa.net.xtra-net.be> <1163692748.2792.10.camel@wbemfkaa.net.xtra-net.be> <20061116161706.GB65054@xor.obsecurity.org> <1163695124.2792.16.camel@wbemfkaa.net.xtra-net.be> <20061116210151.GA68673@xor.obsecurity.org> <1163712849.8157.5.camel@wbemfkaa.net.xtra-net.be> <20061116214237.GA69412@xor.obsecurity.org> <1163780080.2697.6.camel@wbemfkaa.net.xtra-net.be> <20061117223140.GA13858@xor.obsecurity.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 2006-11-17 at 17:31 -0500, Kris Kennaway wrote: > On Fri, Nov 17, 2006 at 05:14:40PM +0100, Vincent Blondel wrote: > > > > I'm not sure about the cause though, does it happen every time you run > > > mailwrapper, or only under load? > > > > sendmail_enable is defined to "NONE" so I can suppose I do not use > > mailwrapper. > > > > How can I know it > > You are clearly running it, because of: > > > > > Unread portion of the kernel message buffer: > > > > x0, limit 0xfffff, type 0x1b > > > > = DPL 0, pres 1, def32 1, gran 1 > > > > processor eflags = interrupt enabled, resume, IOPL = 0 > > > > current process = 14294 (mailwrapper) > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > > If you don't know about mailwrapper running, that suggests you're not > using the system for email processing and it's the nightly admin > scripts which are sending mail automatically overnight. No I just forgot the existence of mailwrapper but as I said it in my last mail /etc/mail/mailer.conf points to /var/qmail/bin/sendmail on my machine because qmail is serving SMTP on my web server. Concerning nightly scripts, I receive them in my mailbox every day/week/month. > > I suspect that your mailwrapper executable may be corrupt. The kernel > is supposed to handle this, but you might have discovered a new and > better way to corrupt a binary to get it past the existing checks :-) Do not forget my original post concerned a kernel panic on process httpd so maybe problem does not come from mailwrapper ? > > Try to send an email to e.g. root on the system and see if it triggers > the crash. "echo Hello |mailx -s "Hello" root" works perfectly. > > If it does, then save a copy of the mailwrapper binary (i.e. don't > immediately "fix" it by reinstalling) and make it available so we can > see what is wrong with it. but I still get a new crash this morning at 09h20 AM and when I try to get kgdb running I get hundreds of lines with this ... kgdb: kvm_read: invalid address (0x0) ... > > Kris Vincent
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1163873942.3315.12.camel>