Ethereal-users: Re: [Ethereal-users] ethereal 0.10.12 + synclink / lapd / isis

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: clay wispell <clay.wispell@xxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2005 07:40:37 -0500
hello guy,

the link format from the omniber is really lapd (sonet section dcc). the driver from the card manufacturer does not support native lapd on layer 2; the card is currently setup as cisco-hdlc. packet-isis.c has been modified to register the isis protocol by name and all is well:


 1   0.000000      Network -> User         LAPD S, func=RR, N(R)=59
2 9.231336 Network -> User ISIS P2P HELLO, System-ID: 00e0.004f.af21
 3   9.991451      Network -> User         LAPD S, func=RR, N(R)=60
 4  10.431185      Network -> User         LAPD S, func=RR, N(R)=61
5 10.437877 [39|84:0f][80|00:00:00|00:00][00:00|00:00]00e0.004f.af21[1d] -> [39|84:0f][80|00:00:00|00:00][00:00|00:00]0000.0e64.925c[1d] COTP CC TPDU src-ref: 0x1401 dst-ref: 0x1401
 6  10.450576      Network -> User         LAPD S, func=RR, N(R)=62
 7  10.458710      Network -> User         LAPD S, func=RR, N(R)=63
8 10.467170 [39|84:0f][80|00:00:00|00:00][00:00|00:00]00e0.004f.af21[1d] -> [39|84:0f][80|00:00:00|00:00][00:00|00:00]0000.0e64.925c[1d] COTP AK TPDU (1) dst-ref: 0x1401 9 10.546747 [39|84:0f][80|00:00:00|00:00][00:00|00:00]00e0.004f.af21[1d] -> [39|84:0f][80|00:00:00|00:00][00:00|00:00]0000.0e64.925c[1d] ACSE
10  10.562910      Network -> User         LAPD S, func=RR, N(R)=64
11  10.631883      Network -> User         LAPD S, func=RR, N(R)=65
12 10.639348 [39|84:0f][80|00:00:00|00:00][00:00|00:00]00e0.004f.af21[1d] -> [39|84:0f][80|00:00:00|00:00][00:00|00:00]0000.0e64.925c[1d] COTP AK TPDU (2) dst-ref: 0x1401 13 10.899767 [39|84:0f][80|00:00:00|00:00][00:00|00:00]00e0.004f.af21[1d] -> [39|84:0f][80|00:00:00|00:00][00:00|00:00]0000.0e64.925c[1d] PRES Data transfer PDU
14  10.914565      Network -> User         LAPD S, func=RR, N(R)=66
15 11.033555 [39|84:0f][80|00:00:00|00:00][00:00|00:00]00e0.004f.af21[1d] -> [39|84:0f][80|00:00:00|00:00][00:00|00:00]0000.0e64.925c[1d] PRES Data transfer PDU
16  11.048487      Network -> User         LAPD S, func=RR, N(R)=67
17  13.989504      Network -> User         LAPD S, func=RR, N(R)=68
18  14.228583      Network -> User         ESIS IS HELLO
19  20.496730      Network -> User         LAPD S, func=RR, N(R)=69


thank you,

clay





Guy Harris wrote:

clay wispell wrote:

i am capturing and decoding data from an agilent omniber captured from a microgate synclink card. packet-lapd.c has been modified so that the incoming data is decoded as lapd (native cisco-hdlc on layer 2)


So what's the *actual* link-layer format on the wire? Is it really LAPD, or is it Cisco HDLC, or is it something else?

and so that any information packets are decoded based on the nlpid of the packet. clnp and esis work well (nlpid 0x81 and 0x82), however, isis (nlpid 0x83) causes a segmentation fault. it appears that the culprit is a null value returned here in packet-lapd.c

   isis_handle = find_dissector("isis");


The ISIS dissector doesn't register itself by name, so the "find_dissector()" call fails.

If you want to dissect the packet based on the NLPID, you should probably make "dissect_osi()" (in packet-osi.c) register itself by name, and call *that* dissector, so that it'll do all the work for you (handling not only ISIS, but CLNP and ESIS).

_______________________________________________
Ethereal-users mailing list
Ethereal-users@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-users