it seems it works! :D

From: sandr8 <sandr8(at)blu.it>
Date: Sun Jun 15 2003 - 10:28:03 EDT

it seems to work! everything is ok both with tcp and udp with or without options in tcp or ip. i tested it with tcp1.c, tcp2.c and udp1.c, udp2.c tcpdump -w and cmp -i 31 (i think the first 42 bytes are used by tcpdump for timestamping and so on, however skipping 31 bytes is already enough) said that they produce the same result as if linked with the old libnet.a this is not true with the previous patch i did submit... as i expected, they differ when there are pblocks such as DATA_H, IPO_H, TCPO_H, IP_DATA and TCPDATA! in fact i said i was only sure it would not break things that didn't need checksum (and that was not too good ;°) now it seems ok...

i beg your help with a question: is there any header that needs a very bastard pseudoheader in order to compute its checksum? for a very bastard pseudoheader i mean a pseudoheader that includes some data more than 1 layer below... i don't think there exist any but i can't say that for 100% sure...

being in doubt i tried to be as general as i could: in fact even in such a strange case, this code could work. it would be enough that between the strange bastard header and the lowest header frow which it needs informations there were no headers that want to be checksummed...

maybe the comments i put in the code ar more clear than i am when i speak about it... cheers
sandr8)

---

To unsubscribe, e-mail: libnet-unsubscribe@securityfocus.com For additional commands, e-mail: libnet-help@securityfocus.com Received on Sun Jun 15 10:33:55 2003