Date: Wed, 25 Mar 1998 11:45:35 -0800 From: David Greenman <dg@root.com> To: Pedro A M Vazquez <vazquez@IQM.Unicamp.BR> Cc: hackers@FreeBSD.ORG Subject: Re: ARP REQUEST question Message-ID: <199803251945.LAA12611@implode.root.com> In-Reply-To: Your message of "Wed, 25 Mar 1998 16:28:46 -0300." <199803251928.QAA04991@kalypso.iqm.unicamp.br>
next in thread | previous in thread | raw e-mail | index | archive | help
>David Greenman was saying that: >> >> >On Tue, 24 Mar 1998, David Greenman wrote: >> >> > pratap singh wrote: >> >> >I have a basic doubt. Every layer has a cehcksum being calculated >> >> >whereas the ARP frame does not have. Can anyone throw light on this >> >> >please. Is it because the ARP packets donot traverse the LAN boundary >> >> >and error rates in LAN environment are very low compared to the WAN >> >> >error rates???? >> >> All ethernet packets have a 32bit CRC, so the arps are protected by that. >> > >> >until it hits the first switch or router. Past that point the arp can be >> >garbaged any way you please, and the damage is undetectable. It's not an >> >end-to-end checksum. Do arps cross gateways and switches? in some places, >> >yes. >> >> Switches should be checking the CRC on inbound packets and discarding >> them if it is bad, so I don't see a problem. > >This is true for store and forward switches only, 3Com claims its >'Fast Forward' technology starts forwarding as soon as the destination >address has been received and the lookup completed, this would bypass CRC >checking. Yes, but in this case the CRC would also be passed on to the next device and will ultimately be checked at some point (and the packet discarded if it is bad). There is no practical difference between this and the treatment of IP through an IP router/switch. The only point when the ARP packet isn't protected by the ethernet CRC is when it gets to the final destination, and at that point only a problem with bad software or broken hardware could corrupt the packet. If we're arguing that IP checksums are better because they last longer in the software path, then that's silly because the first thing that happens to an IP packet is the checksum test. -DG David Greenman Core-team/Principal Architect, The FreeBSD Project To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199803251945.LAA12611>