[9fans] unique MAC address allocation?

[9fans] unique MAC address allocation?

[fgergo]

Not really plan9 specific, but
1. does anybody know if and how MAC addresses are allocated to network
card manufacturers?
2. do manufacturers guarantee the uniqueness of MAC addresses they build?
3. is it still true (was it ever true), that if I bought 2 NIC's,
they'll have different MAC addresses?
4. is it the responsibility of the OS to override the default MAC
address if another card with the same (possibly overriden) MAC address
was present on the LAN?

 thanks: gergo

Re: [9fans] unique MAC address allocation?

[McLone]

On Wed, 30 Mar 2005 12:25:27 +0200, fgergo@gmail.com <fgergo@gmail.com> wrote:
> 1. does anybody know if and how MAC addresses are allocated to network
> card manufacturers?
trere was some comitee that allocates to manufacturers different MAC prefixes.
> 2. do manufacturers guarantee the uniqueness of MAC addresses they build?
true
> 3. is it still true (was it ever true), that if I bought 2 NIC's,
> they'll have different MAC addresses?
true [ for not very old ethernet cards]
> 4. is it the responsibility of the OS to override the default MAC
> address if another card with the same (possibly overriden) MAC address
> was present on the LAN?
As far as i know, driver must allow this, and this process isn't automatic.
Only modern drivers allow MAC override [read: win32, nvidia nforce,
and wireless :-]

--
wbr,                        |\      _,,,---,,_           dog bless ya!
`                       Zzz /,`.-'`'    -.  ;-;;,_
McLone at GMail dot com    |,4-  ) )-,_. ,\ (  `'-'
  net- and *BSD admin     '---''(_/--'  `-'\_)   ...sorry for translit

Re: [9fans] unique MAC address allocation?

[geoff]

Manufacturers can request (or perhaps buy) a largish block of
addresses from some standards organisation (I think it's
Bellcore/Telcordia/whatever-they're-called-this-week).  If they
exhaust that block, they can get another, and so on.

Manufacturers are supposed to ensure uniqueness of MAC addresses, but
they've been known to screw up (e.g., SGI), so yes, two NICs are
supposed to have different MAC addresses.  In practice, since we don't
run XNS, a MAC address only has to be unique within the Ethernet that
its card is plugged into (assuming no BOOTP through gateways).

Many cards, not just new ones, permit overriding the default MAC
address and the plan 9 drivers generally support this (with
ea=XXXXXXXXXXXX in the ether's plan9.ini entry).

Re: [9fans] unique MAC address allocation?

[Charles Forsyth]

>>Many cards, not just new ones, permit overriding the default MAC
>>address and the plan 9 drivers generally support this (with
>>ea=XXXXXXXXXXXX in the ether's plan9.ini entry).

indeed most cards require the address to be read from somewhere else (eg, EEPROM)
and explicitly programmed into the device proper, so that being
able to override it is not unusual.

Re: [9fans] unique MAC address allocation?

[fgergo]

On Wed, 30 Mar 2005 12:27:07 +0100, Charles Forsyth
<forsyth@terzarima.net> wrote:
> >>Many cards, not just new ones, permit overriding the default MAC
> >>address and the plan 9 drivers generally support this (with
> >>ea=XXXXXXXXXXXX in the ether's plan9.ini entry).
>
> indeed most cards require the address to be read from somewhere else (eg, EEPROM)
> and explicitly programmed into the device proper, so that being
> able to override it is not unusual.

I was probably not precise and the question was not well thought
through. What I wanted to ask was:
I did override the MAC address - for some obscure reason, I can't
remember - some 12 years ago on DOS via the packet driver interface,
but is it the OS's responsibility to detect and possibly deal with MAC
address collision or would the Ethernet layer take care of it?

Instead of googling for "MAC address" etc., I should have searched for
"Ethernet standard", days ago.
The pages I needed were http://standards.ieee.org/faqs/OUI.html and
http://standards.ieee.org/getieee802/802.3.html . Now I only have to
read thorugh the standard. Probably the subject is not worth it (~1500
pages?).

The OS can take care of the problem, if the driver supports MAC
address collision detection.

Re: [9fans] unique MAC address allocation?

[Artem Letko]

Ethernet assumes unique MACs. Save yourself some time and look up
Ethernet II encapsulation. Isn't that the one Plan9 uses anyway?


On Wed, 30 Mar 2005 15:00:57 +0200, fgergo@gmail.com <fgergo@gmail.com> wrote:
>
> I was probably not precise and the question was not well thought
> through. What I wanted to ask was:
> I did override the MAC address - for some obscure reason, I can't
> remember - some 12 years ago on DOS via the packet driver interface,
> but is it the OS's responsibility to detect and possibly deal with MAC
> address collision or would the Ethernet layer take care of it?
>
> Instead of googling for "MAC address" etc., I should have searched for
> "Ethernet standard", days ago.
> The pages I needed were http://standards.ieee.org/faqs/OUI.html and
> http://standards.ieee.org/getieee802/802.3.html . Now I only have to
> read thorugh the standard. Probably the subject is not worth it (~1500
> pages?).
>
> The OS can take care of the problem, if the driver supports MAC
> address collision detection.
>

Re: [9fans] unique MAC address allocation?

[geoff]

MAC addresses aren't supposed to collide; they're supposed to be
globally unique.  So it's manufacturer's responsibility to get this
right.  Users who override the MAC address do so at their own risk.

I'm not convinced that there's a genuine problem here, let alone a
problem worth solving.

Re: [9fans] unique MAC address allocation?

[Adrian Tritschler]

geoff@collyer.net wrote:
> MAC addresses aren't supposed to collide; they're supposed to be 
> globally unique.  So it's manufacturer's responsibility to get this 
> right.  Users who override the MAC address do so at their own risk.

Supposedly.

Unless they're the cheap taiwanese clones of NE2000 cards that my
previous employer bought about ten years ago.  The first couple worked,
since they were on different bits of cable, then everything went
pear-shaped.  50 *identical* reverse-engineered chinese copy cards anyone?

> I'm not convinced that there's a genuine problem here, let alone a 
> problem worth solving.

	Adrian

---------------------------------------------------------------
Adrian Tritschler                          mailto:ajft@ajft.org
Latitude 38°S, Longitude 145°E, Altitude 50m,      Shoe size 44
---------------------------------------------------------------

Re: [9fans] unique MAC address allocation?

[Michael Zappe]

fgergo@gmail.com wrote:

>Not really plan9 specific, but
>1. does anybody know if and how MAC addresses are allocated to network
>card manufacturers?
>2. do manufacturers guarantee the uniqueness of MAC addresses they build?
>3. is it still true (was it ever true), that if I bought 2 NIC's,
>they'll have different MAC addresses?
>4. is it the responsibility of the OS to override the default MAC
>address if another card with the same (possibly overriden) MAC address
>was present on the LAN?
>
> thanks: gergo
>
>
Well, since my company makes network cards, I can probably help you
here! ;-)  Each vendor registers one or more 3 octet OUIs  from the
IEEE, which is used as the first 3 octets of a MAC address.  The
remaining 3 are up to the vendor to use.  Theoretically, all MAC
addresses are supposed to be unique.  For one vendor code you only have
16 million addresses available, so there is the possibility of wrap
around with MAC addresses -- it's up to the vendor to comply.  (Once you
have a practical deployed base of 90%, you can request a new OUI.) For
all practical purposes, you can assume that MAC addresses are unique.

For this part here is a good tutorial:
http://standards.ieee.org/regauth/oui/tutorials/lanman.html

As to question 4, nope.  The OS has no good way of determining a new MAC
address to override to (The OS doesn't have an OUI, and then how do you
pick the serial number after that??), and it's not considered a real
problem because of the uniqueness of MACs.  Not to mention you don't
even want to think about how a switch could get confused by the whole
situation... :-)

    Mike

Re: [9fans] unique MAC address allocation?

[Derek Fawcus]

On Wed, Mar 30, 2005 at 11:00:08AM -0700, Michael Zappe wrote:
>
> As to question 4, nope.  The OS has no good way of determining a new MAC
> address to override to (The OS doesn't have an OUI, and then how do you
> pick the serial number after that??),

Pick a random number,  set the local allocation bit (second bit on wire)
and then test the address to see if anyone responds to it.

Didn't DECnet do something like that?

DF

> and it's not considered a real
> problem because of the uniqueness of MACs.  Not to mention you don't
> even want to think about how a switch could get confused by the whole
> situation... :-)
>
>     Mike

Re: [9fans] unique MAC address allocation?

[Michael Zappe]

Derek Fawcus wrote:

>On Wed, Mar 30, 2005 at 11:00:08AM -0700, Michael Zappe wrote:
>
>
>>As to question 4, nope.  The OS has no good way of determining a new MAC
>>address to override to (The OS doesn't have an OUI, and then how do you
>>pick the serial number after that??),
>>
>>
>
>Pick a random number,  set the local allocation bit (second bit on wire)
>and then test the address to see if anyone responds to it.
>
>Didn't DECnet do something like that?
>
>
>
Yeah, but you're not guaranteed a reply to the packet.  It's a distinct
possibility that if you pick an address and a protocol, the other
machines stack may disregard it, i.e. blackholing pings, etc.  There's
no good way to detect the collision.  Theoretically you could use LLC to
check, but I wouldn't count on that being implemented correctly, if at
all... :-)  If you're using locally administered addresses, you can
assume that valid NICs aren't going to have this address, and the chance
of a collision is low. (N in 16 million for N machines with the same
OUI.)  So I guess that would work! :-)

DECnet actually had a deterministic algorithm for assigning MAC
addresses.  You had an area number that was 6 bits, and a 10 bit node
number.  Then you appended this to AA:00:04:00 to get the MAC address.
Multiple interfaces on the same machine all used the same MAC.

>DF
>
>
>
>>and it's not considered a real
>>problem because of the uniqueness of MACs.  Not to mention you don't
>>even want to think about how a switch could get confused by the whole
>>situation... :-)
>>
>>    Mike
>>
>>

Re: [9fans] unique MAC address allocation?

[Martin C. Atkins]

On Wed, 30 Mar 2005 17:20:51 -0700 Michael Zappe <zapman@zappe.us> wrote:
> DECnet actually had a deterministic algorithm for assigning MAC
> addresses.  You had an area number that was 6 bits, and a 10 bit node
> number.  Then you appended this to AA:00:04:00 to get the MAC address.
> Multiple interfaces on the same machine all used the same MAC.

Anyone know how vmware allocates its MAC addresses? They appear to
own a OUI, but what about the other bits?

Martin

--
Martin C. Atkins			martin_ml@parvat.com
Parvat Infotech Private Limited		http://www.parvat.com{/,/martin}

Re: [9fans] unique MAC address allocation?

[Derek Fawcus]

On Thu, Mar 31, 2005 at 11:04:22AM +0530, Martin C. Atkins wrote:
> On Wed, 30 Mar 2005 17:20:51 -0700 Michael Zappe <zapman@zappe.us> wrote:
> > DECnet actually had a deterministic algorithm for assigning MAC
> > addresses.  You had an area number that was 6 bits, and a 10 bit node
> > number.  Then you appended this to AA:00:04:00 to get the MAC address.
> > Multiple interfaces on the same machine all used the same MAC.
>
> Anyone know how vmware allocates its MAC addresses? They appear to
> own a OUI, but what about the other bits?

I seem to recall they document it somewhere.  In their help pages or
online 'Knowledge Base'?

DF

Re: [9fans] unique MAC address allocation?

[boyd, rounin]

>> > You had an area number that was 6 bits, and a 10 bit node
>> > number.  Then you appended this to AA:00:04:00 to get the MAC address.

on reflection this sounds more accurate.  seems that had taken
out several blocks of vendor-ids and used AA:00:04:00 to say
'mangled address, includes DECnet'.
--
MGRS 31U DQ 52572 12604

Re: [9fans] unique MAC address allocation?

[boyd, rounin]

> Didn't DECnet do something like that?

it did.  as i recall, it smashed into the the ether addr's low bits
the machine's DECnet address for some vile hack so that
ether to DECnet address mapping would be 'easy' [on ULTRIX].

these all very vile and duplicate DECnet addresses would
ruin your whole day/week/month(s).

i knew there was something wrong with my machine but
could never figure it out until i stuck some printfs in the
kernel.  yup, duplicate DECnet addresses.  at this point
i surveyed the whole lab and made sure there were no dups.

as DECnet had such a small addres space (16 bits? 8
net, 12 host) there was 'hidden area' on net 62.
--
MGRS 31U DQ 52572 12604

Re: [9fans] unique MAC address allocation?

[boyd, rounin]

> as DECnet had such a small addres space (16 bits? 8
> net, 12 host) there was 'hidden area' on net 62.

arg, that should read 6 net, 10 host.
--
MGRS 31U DQ 52572 12604