RE: kernel panic [linux 2.2.19-7] on UDP scan CP4.1-SP5

From: Yanek Korff (yanek@cigital.com)
Date: 11/14/01 

Message-ID: <4BC10D47D7ACD3119FA800104B1F8836013A4F12@exchange.cigital.com>
From: Yanek Korff <yanek@cigital.com>
To: 'Olaf Kirch' <okir@caldera.de>
Subject: RE: kernel panic [linux 2.2.19-7] on UDP scan CP4.1-SP5
Date: Wed, 14 Nov 2001 12:34:46 -0500

> -----Original Message-----
> From: Olaf Kirch [mailto:okir@caldera.de]
> Sent: Wednesday, November 14, 2001 11:40 AM
> To: Yanek Korff
> Cc: 'ed.rolison@power.alstom.com'; vuln-dev@securityfocus.com
> Subject: Re: kernel panic [linux 2.2.19-7] on UDP scan CP4.1-SP5
>
>
> On Wed, Nov 14, 2001 at 11:27:48AM -0500, Yanek Korff wrote:
> > Unfortunately, I don't think this is the case. If a table
> were being filled
> > up, I'd expect the FW to stay up for some period of time
> before eventually
> > crashing. Here are some relevant facts:
> >
> > 1. Linux FW crashes -immediately- before it has the
> opportunity to log a udp
> > packet with tcpdump
> > 2. Scans complete successfully against NT 4.0 and Solaris-x86
>
> There was a problem (kernel lockup) with certain types of UDP
> packets a few months ago (it could be though that happened
> only for locally generated packets). All vendors released fixes
> for these. Could be the scan checks for this vul. Check your
> vendor's security page for details.

Would not the OS itself crash without the FW kernel module loaded whena UDP
scan was initiated? When the machine is running without the FW active, it
stays up fine. I am running the latest updated kernel (source RPM) from
RedHat in the 2.2.x kernel sequence.

I've tried the -T Paranoid switch; the system crashes with the VERY FIRST
UDP packet, regardless of which port it's sent to. I subsequently
re-enabled icmp, as a "before last" implied rule... And I see this:
Initiating UDP Scan against (64.80.176.11)
12:43:34.168842 nmap_source.58153 > fw_under_test.973: udp 0
12:43:34.274503 fw_under_test > nmap_source: icmp: 64.80.176.11 udp port 973
unreachable

And that's the last packet I get from the machine.

If I run nslookup on nmap_source, set my server to fw_under_test, and
attempt to resolve something (even though fw_under_test is not running a
nameserver), the fw_under_test does not crash. It merely replies with udp
port unreachable and stays up.

Ideas?

-Yanek.

Relevant Pages

  • Re: jumbograms (& em) & nfs a no go
    ... >> Implies the sending host is not honoring the MTU restriction when ... In more modern UNIX systems, it's a kernel thread, ... Is this maybe UDP? ... make sure that the packet goes through the UDP fragmentation ...
    (freebsd-current)
  • kernel 2.6.18-92.1.6 produces errors when using nfs and nis
    ... kernel 2.6.18-53.1.21.el5, with no change to any conf files when I switch ... The problem is that when the nfs service starts I get the following error ... 100000 2 tcp 111 portmapper ... 100000 2 udp 111 portmapper ...
    (Linux-Kernel)
  • Re: Send-Q on UDP socket growing steadily - why?
    ... Send-Q on a moderately active UDP socket keeps growing steadily until it reaches ~128K at which point socket writes start failing. ... The application in question is standard ntpd from Fedora 7, kernel is the latest available for the distro, that is ... "I don't know how to send these packets" mode forever. ...
    (Linux-Kernel)
  • Identifying Kernel 2.4.x based Linux machines using UDP
    ... Identifying Kernel 2.4.x based Linux machines using UDP ... Linux Kernel 2.4.x has a bug with the UDP implementation which allows ... Combined with another fingerprinting method using ICMP this time ...
    (Bugtraq)
  • Re: Send-Q on UDP socket growing steadily - why?
    ... Send-Q on a moderately active UDP socket keeps growing steadily until it ... The application in question is standard ntpd from Fedora 7, kernel is ... and drops packets. ...
    (Linux-Kernel)