[FS#655] ramips: mac address parsing

[LEDE Bugs]

The following task has a new comment added:

FS#655 - ramips: mac address parsing
User who did this - Mathias Kresin (mkresin)

----------
> and what magic do you think you will find in ifconfig output? my real mac address? 

As my first question was about the ifconfig output related to the original xiaomi mini firmware, I expect to get which mac address is used for the interfaces by the original firmware. But at least the LEDE ifconfig output proves that the wan interface uses a different mac address that the lan interface, in opposite to what you claimed in the intial bugreport comment.

I'm fine if you don't like to post the mac addresses of your device in public. But it shouldn't be that difficult to find mail address to send me the output via mail.

> does that 00:0C:43 looks like a valid prefix?!?

Yes it does. It is the Ralink prefix. For more details read the commit message of [[https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=9766cb709089fc460fda66c06a3da756af4a2930|9766cb709089fc460fda66c06a3da756af4a2930]].

What we have at the moment, is a MAC-Address with the Xiaomi Communications Co Ltd. prefix in the wireless eeprom and some apparently random data at the offset(s) for additional mac addresses. The 00:D0:59 prefix is assigned to AMBIT MICROSYSTEMS CORP., which doesn't seam to me related in any kind to Xiaomi. I'm not going to say that the prefix manipulation we are doing at the moment is correct, but using the not set additional mac address is wrong as well.

Therefore my point is still valid and it was correct to denial the reopen of the ticket.
----------

More information can be found at the following URL:
https://bugs.lede-project.org/index.php?do=details&task_id=655#comment2445