[Please keep the bug address in cc.] On Mon, 2016-08-29 at 03:31 +0900, Hughe Chung wrote: > Hi Ben, > > I posted an article on my blog in Sep 2012, updated Nov 2013. > https://butterflyofdream.wordpress.com/2012/09/11/kernel-fails-to-load-atheros-ar8161-ethernet-controller-module/ > > It showed MAC address of AR8161 was changing at that time which I didn't > aware. So the network driver of AR8161 has gotten floating MAC address > from kernel 3.4.x to kernel 4.7.x. It wasn't officially present in version 3.4, though. You're comparing the version from compat-drivers (which Debian also included for some time) with the in-tree version. [...] > According to the dmesg log, alx's developer wrote random MAC address > generator. [...] > Aug 29 01:37:30 dv1-station1 kernel: [ 1.592921] alx 0000:02:00.0: Invalid permanent address programmed, using random one [...] That's a standard networking function used by many network drivers in the case where they can't find a valid permanent MAC address. Comparing the functions used to read the permanent MAC address in the the two driver versions, they look superficially quite different but seem to be functionally the same. The in-tree version of alx was first included in Debian's package version 3.10.1-1 and the version from compat-drivers was last included in version 3.10~rc7-1~exp1. Those are still available from <http://snapshot.debian.org/package/linux/> and it would be helpful if you could test them to see which of those can read the permanent MAC address. Ben. -- Ben Hutchings I say we take off; nuke the site from orbit. It's the only way to be sure.
Attachment:
signature.asc
Description: This is a digitally signed message part