Luca et al:
I noticed a problem (for me, anyway) between 2.2.98 and 2.2.99
in pbuf.c. I typically run with decoders disabled without problems.
However, after to cvs update, I began getting a syslog full of:
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 90
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 590
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 1494
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 87
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 1514
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 78
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 74
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 740
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 1514
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 830
Jan 27 17:45:01 host ntop[25558]: **WARNING** dequeuePacket: caplen 68
!= len 1434
I thought it was related to stuff I'm playing with, so I did a
completely fresh checkout; however, the voluminous errors persisted.
So, I followed the logic as I understood it and I offer the following
patch:
--- ntop/pbuf.c.orig 2004-01-26 14:12:01.000000000 -0500
+++ ntop/pbuf.c 2004-01-27 18:13:49.000000000 -0500
@@ -2248,7 +2248,7 @@
2. MAX_PACKET_LEN should probably be removed
3. all the functions must check that they are not going beyond
packet boundaries
*/
- if (h.caplen != h.len) {
+ if ((h.caplen != h.len) && (myGlobals.enablePacketDecoding)) {
traceEvent (CONST_TRACE_WARNING, "dequeuePacket: caplen %d !=
len %d\n",
h.caplen, h.len);
}
It fixes the problem on my system and should be correct, if I
understand the logic correctly.
ken
_______________________________________________
Ntop-dev mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev
