Fw: ospf endianless bug

----- Original Message -----
From: "Kathleen Poulsen" <[email protected]>
To: "Valerio 'click' Genovese" <[email protected]>
Sent: Wednesday, May 12, 2004 12:17 AM
Subject: Re: ospf endianless bug


> this IS broken...as are several of the ospf functions.  i reported this
> awhile ago (maybe 6 months), but was unable to supply a fix at the time.
>
> there is also a problem in the ospf neighbor list, in HELLO packets.
> as i recall, at least one neighbor is required by libnet, but the RFC does
not
> require this (i may be recollecting the bug incorrectly, but i think that
> the description is correct).
> kathleen
>
> On Tue, May 11, 2004 at 10:48:27PM +0200, Valerio 'click' Genovese wrote:
> > hu?
> > it seems to be a bug in libnet_build_ospf.c , htonl() in s.addr field
> > are wrong, or not?
> > If I sniff packets generated with ospf_hello and ospf_lsa samples IP
> > adresses are printed (using ethereal and tcpdump) with wrong endianless
> >
> > let me know
> >
> > libnet 1.1.2.1
> > linux kernel 2.6.5
> >
> > regards
> > Valerio 'click' Genovese
[...]