Hi Wolfgang, Manuel, On Tuesday 20 July 2010 12:10:54 Wolfgang Grandegger wrote: > Hi Manuel, > > On 07/19/2010 02:55 PM, Manuel Lauss wrote: > > Modify au1000_eth to receive an ethernet address from platform data, > > or choose a random one. > > > > The default address is usually provided by the firmware; modify > > platform device registration to use it if the board code has not > > already overridden it. > > > > Cc: Wolfgang Grandegger <wg@xxxxxxx> > > Cc: Florian Fainelli <florian@xxxxxxxxxxx> > > Signed-off-by: Manuel Lauss <manuel.lauss@xxxxxxxxxxxxxx> > > --- > > v2: diffed against linus-git, on top of Wolfgang's patch > > > > "mips/alchemy: define eth platform devices in the correct order" > > This one should actually apply cleanly. > > > > IMHO a device driver should not call firmware-specific functions > > (be it MIPS-style prom_get_*(), OF properties or whatever) to > > get missing information. Instead this should be done by the > > platform code which sets up the device. This patch does just that. > > > > Compile-tested only. Florian, Wolfgang: could you please give this > > a try on your boards? If it works and you agree to it, I'll > > resubmit it also to linux-netdev. Thank you! (I don't have > > accessible au1000-eth hardware). > > I gave the patch a try. The kernel builds and runs fine. The eth's are > realized in the correct order and do work properly. Feel free to add my > "Tested-by: Wolfgang Grandegger <wg@xxxxxxx>". I will not be able to test it soon, so go ahead for merging it. Thanks! -- Florian
