Re: capturing packets with CRC errors

[Hannes Gredler <hannes () juniper net>]

most ISPs run MPLS today for packet transport in their backbones ....
so it may not be always the case that every transit router does check
the IP checksum. 


however the ingress and egress routers of the MPLS cloud should do at least;

/hannes

On Sun, Dec 15, 2002 at 06:22:19PM -0500, Michael Ang wrote:
| Someone also pointed out to me offlist that packets with IP checksum errors 
| should be dropped by any intermediate routers (although that may not 
| necessarily happen in practice).  So an IP checksum mismatch should only be 
| seen if the corruption happens between the last router and the sniffer.
| 
|       - Mike.
| 
| Guy Harris wrote:
| >On Thu, Dec 12, 2002 at 08:55:40PM -0500, Michael Ang wrote:
| >
| >>I of course meant checksum instead of CRC for IP and TCP.  Looking at 
| >>the code for tcpdump it seems that those checksums are passed intact 
| >>through pcap as they appear on the wire (i.e. packets are not rejected 
| >>based on IP or TCP checksum errors).
| >
| >
| >Correct.  libpcap pays no attention whatsoever to the IP or TCP (or UDP,
| >or SCTP, or ...) checksum - in link layers where the checksum happens to
| >be supplied (PPP in some cases, and some LANs on some versions of some
| >OSes, I think), it doesn't even pay attention to that.
| >
| >Tcpdump pays attention to them only by checking the IP, TCP, and UDP
| >checksums, if present, reporting whether they're valid.  (It does so for
| >IP headers only if the entire IP header was captured, and does so for
| >TCP and UDP only if the the entire packet was captured and the packet
| >isn't fragmented at the IP layer.)

-
This is the TCPDUMP workers list. It is archived at
http://www.tcpdump.org/lists/workers/index.html
To unsubscribe use mailto:tcpdump-workers-request () tcpdump org?body=unsubscribe