Linux-Activists - NET Channel digest. 94-9-20-1:26

Linux Activists (linux-activists@niksula.hut.fi)
Thu, 20 Oct 1994 15:07:07 +0100

Topics:
Re: tcpdump seg fault?

----------------------------------------------------------------------

From: longyear@netcom.com (Al Longyear)
Subject: Re: tcpdump seg fault?
Date: Wed, 19 Oct 1994 10:49:53 -0700

Olaf Erb <erb@insu1.etec.uni-karlsruhe.de> writes:

>Just wondering.. I'm still getting this since early 1.1 (I think)..:

You should still get it with 1.1.54! This bug has not been fixed in that
version either.

># > tcpdump
>tcpdump: listening on sl0
>
>Stack values give me
>
>00132998 t _dev_ifsioc
>0019eb3c d _slip0_dev

The tcpdump program will not work on the slip device any more than it will
work on the PPP device. I reported a similar problem to this group earlier
when someone complained about the same trap with PPP.

The problem is that the dev_ifsioc procedure does a blind call to put the
device into promiscous mode. The PPP and SLIP devices do not have an entry
for this procedure and it is zero.

The logic in the ioctl processing needs to test to see if the device may be
allowed to go into this mode before it accepts the request and sets the
device flag for promiscous. If there is no procedure, the ioctl operation
should not be allowed.

An alternative would be to put the entries for PPP and SLIP into the table.
However, there is no way that the driver can report back that "it can't do
this" to the ioctl processing so that it may return the error to tracedmp so
that tracedmp may report the error to the user and the user will know that
you can't run tracedmp with SLIP or PPP.

The best that this procedure could do is to turn off the flag bit in the
device flags. PLIP and possibly the AX-25 logic would possibly have the same
problem.

-- 
Al Longyear               longyear@netcom.com            longyear@sii.com
The public pgp 2.6 key is available by fingering longyear@netcom.com.

------------------------------

End of NET Digest ***************** -------