Date: Thu, 19 Feb 1998 09:57:03 -0500 From: Norman C Rice <nrice@emu.sourcee.com> To: Nadav Eiron <nadav@barcode.co.il> Cc: jgrosch@superior.mooseriver.com, questions@FreeBSD.ORG Subject: Re: hostid equals Ethernet address or IP address ? Message-ID: <19980219095703.59598@emu.sourcee.com> In-Reply-To: <34EC3AA6.562C@barcode.co.il>; from Nadav Eiron on Thu, Feb 19, 1998 at 03:59:02PM %2B0200 References: <19980219004112.26662@mooseriver.com> <19980219084843.13569@emu.sourcee.com> <34EC3AA6.562C@barcode.co.il>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Feb 19, 1998 at 03:59:02PM +0200, Nadav Eiron wrote: > Norman C Rice wrote: > > > > On Thu, Feb 19, 1998 at 12:41:12AM -0800, Josef Grosch wrote: > [snip] > > > > While I am not sure about supporting documents, I am pretty sure that > > sethostid(3) is referring to the IP address, not the MAC address. My > > reasoning is that IP addresses are 32-bits while MAC address are > > 64-bits (the first 32-bits comprises the OUI which uniquely identifies > > the vendor). My other thought is that if sethostid() actually rewrote > > the MAC address (assuming it was implemented in FLASH, EEPROM, or some > > other reprogrammable device) is that there would be no guarantee of > > uniqueness. Reprogramming the MAC address would ``break'' many vendors > > cards due to the steps they take to prevent this, e.g., embedded > > CRC-32s, checksums, inverted MAC addresses, etc.. > > As stated in the man page, the function is deprecated anyhow, but just > to comment on Norman Rice's answer: MAC addresses are 48 bit long, not > 64 bit long, and there are indeed cases where they are reprogrammed. The Duh. You're right -- 6 bytes is definitely 48 bits. Thanks for making the correction. I should also point out that my ``generous'' byte-to-bit conversion on the 3 byte OUI was also incorrect; I should have reported that the OUI is the first 24 bits of the MAC address. > most notable of these is DECnet Phase IV where the MAC address is > reprogrammed to be the DECnet address (with some additions) as soon as > DECnet is up and running. Most NICs will let you reprogram their MAC > address temporarily, if you so wish. While most NICs use reprogrammable logic, most of them do use proprietary techniques to ensure that the end user cannot easily change the MAC address. If the card manufacturer provides the update algorithm, then this issue is moot. However, it is not a temporary change as the logic is non-volatile; the old MAC address would have to be explicitly reprogrammed. I would like to point out that FLASH and EEPROM devices ``wear out'' after 10,000 to 1,000,000 programming cycles -- actually their write cycle times become so excessive they no longer work with the surrounding circuitry. I find it hard to believe that NIC manufacturers would allow end users to freely change the MACs, especially the OUI portion. There would be no guarantee of uniqueness and I am sure that a new sport, ``MAC spoofing'', would soon be taken up by the crackers. > > > > > > > > > > > > Josef > > > > > > -- > > > Josef Grosch | Another day closer to a | FreeBSD 2.2.5 > > > jgrosch@MooseRiver.com | Micro$oft free world | UNIX for the masses > > -- > > Regards, > > Norman C. Rice, Jr. > > Nadav -- Regards, Norman C. Rice, Jr. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980219095703.59598>