Date: Mon, 27 Aug 2001 09:06:38 -0700 From: "Kevin Oberman" <oberman@es.net> To: Harry Putnam <reader@newsguy.com> Cc: freebsd-mobile@FreeBSD.ORG Subject: Re: Frustrating network problem - need diagnotic help Message-ID: <200108271606.f7RG6cR23974@ptavv.es.net> In-Reply-To: Your message of "Sun, 26 Aug 2001 22:37:04 PDT." <m11yly6p33.fsf@reader.newsguy.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> From: Harry Putnam <reader@newsguy.com> > Date: Sun, 26 Aug 2001 22:37:04 -0700 > Sender: owner-freebsd-mobile@FreeBSD.ORG > > "Kevin Oberman" <oberman@es.net> writes: > > > For all of you who don't have a copy of 802.3 handy, there are two > > special bits in a MAC address. They are the first two bits on the wire > > in the MAC, but because of the bit ordering on the wire, these are the > > last bits of the first byte of the MAC. > > How would you phrase this, speaking to a `network impaired' (dodo) > person, please? Or at least maybe a diagram of the posted MAC as it > relates to the items enumerated above. > > > The first bit is the "group address" bit. This means multicast or > > broadcast. (Broadcast is simply a special case of multicast.) It > > should NEVER be set in the source address, so the first octet must be > > even. > > Ditto > > > The second bit the "locally administered" bit, indicating that the MAC > > address is locally assigned and not always globally unique. Some > > protocols like DECnet IV and Xerox PUP make use of this, but it is > > pretty much unused these days. > > > > In any case, a MAC address with an odd first byte is clearly not legal > > as a source address. > > Well, sounding like a perfect lamer here, I'm not sure I follow what > this all means. Following bits, bytes, hexidecimal et al, confuses me > greatly. Can you offer a guess as to what this all means for the > posted MAC: 01:d4:ff:03:00:20? > > Sounds as if it is a patent dud, what with starting like `01'. Is > that true? What does one do to correct bogus recognition? > Or, when you say `source address' does that not apply to the address > revealed with `ifconfig ed0'? If you ever have really bad insomnia, 802.3 section "3.2.3 Address Fields" is the relevant information. This is getting a bit far from the mobile charter, but... 8:0:2b:3f:a5:15 is the MAC address of some ancient and forgotten DEC device. In binary that is: 00001000 00000000 00101011 00111111 10100101 00010101 When the data is placed on the Ethernet wire, it is transmitted LSB first, so it goes on the wire like this: 00010000 00000000 11010100 11111100 10100101 10101000 If you add the first bit in the stream, the hex representation turns into 9:0:2b:3f:a5:15 (This should NEVER be seen in the source address!) If you tweak the software to assign a MAC address to the something else that is locally administered (different from setting the MAC to the globally unique address of some older interface that failed), the hex would be: a:0:2b:3f:a5:15 (not that anyone does this any longer and there is no reason to assume that any part of the hardware MAC would be used if you are using a locally administered MAC). So a MAC address on a card of 01:d4:ff:03:00:20 is clearly not valid. Either the software is not reading the address correctly or the device is broken. (If some other OS (Windows?) sees the same MAC address, it's hardware. If the MAC is valid in that OS, it's software. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: oberman@es.net Phone: +1 510 486-8634 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200108271606.f7RG6cR23974>