Re: MASQ: failed TCP/UDP checksum?

Nigel Metheringham (
Mon, 04 Nov 1996 09:50:06 +0000

} [root@crasher2 log]# grep checksum messages | tail -5
} Nov 2 12:15:07 crasher2 kernel: MASQ: failed TCP/UDP checksum from

} I'm getting tons of those.

They mean that the packets running through the masquerade are corrupt when
they arrive at your machine. I added to code to do this after realising
that the previous incarnation of the code reworked the port numbers and ip
addresses in the packets and then recalced the checksum, potentially
validating packets that are corrupt.

Causes of these are:-
- Duff network cabling between you and the ip address(es) in the message.
For most people this would happen on the internal network.
- Duff PPP config on the external network - typically serial overruns.
- Duff network stacks at the other end of the link - unlikely.
- Duff routers somewhere - less likely but possible.

} It's happened ever since I started using IP masquearading, a few dozen
} kernels ago. They are very annoying (6 out of 10 syslog'd messages consist
} of that). Any ideas on how to get rid of them? ;-) Thanks. -rt

Check your PPP config is the most likely effective answer.

The messages are given with far too high a priority. Modding ip_masq.c
and changing the priority of the messages would be a good solution, but
merely hides the problem. I don't think that the IP counters are updated
correctly under these conditions so probably these messages are the only
way of warning you!


[   - Unix Applications Engineer ]
[ *Views expressed here are personal and not supported by PLAnet* ]
[ PLAnet Online : The White House          Tel : +44 113 251 6012 ]
[ Melbourne Street, Leeds LS2 7PS UK.      Fax : +44 113 2345656  ]
[[[ Welcome to Grace, arrived 01:37 BST, 18 Sept 1996, 5lb 15oz ]]]