[U-Boot] [PATCH 2/2] mvgbe: fix network device indices

Wolfgang Denk wd at denx.de
Sat Nov 5 16:06:24 CET 2011


Dear Albert ARIBAUD,

In message <4EB54978.5020301 at aribaud.net> you wrote:
>
> > What would be the result?  A bord that comes up with a new MAC address
> > each time you reset it?
> 
> No -- the goal of the randomization code was, is, and will be to allow 
> the board to use the network when no correct MAC address can be found 
> anywhere (env vars, EEPROM, e-fuses, whatever). When a correct address 

And if this is the case, then the board will come up with a new MAC
address each time you reset it, right?

> is available, that address will be used. Typically, this happens when 
> the board has not been provisioned yet, at a point where the MAC address 
> it uses is not relevant yet.

I've done provisioning stuff a couple of times before, and I'm just
doing is again.  Random MAC addresses are a broken concept, and
anybody who considers using it should reassess his concepts.

Where is the real MAC address coming from, and how does it get
assigned to this specific board?  And how do you make sure not to make
mistakes when all you see is some board with a random MAC address?

[The systems I know usually either have the MAC address pre-programmed
in some storage device on the board, or printed on a barcode label, so
you can use ca barcode reader in combination with "askenv" and very
little U-Boot scripting as part of your production test /
initialization procedures.]

> 1. This code would only be available to kirkwood-based boards anyway.

That doe snot make things any better.

> 2. Although the code incorrectly describes it as "private", the random 
> address is actually a locally administered address (bit 1 of first octet 
> is set), which eliminates the risk of clashing against any 'normal' 
> (universally administered) address; and its last three octets are 
> randomized in order to limit the risk of clashing against other locally 
> administered addresses if we're unlucky enough to have any on the 
> network segment.

I consider the whole approach broken and object against it.

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Perfection is reached, not when there is no longer anything  to  add,
but when there is no longer anything to take away.
                                           - Antoine de Saint-Exupery


More information about the U-Boot mailing list