Hi Lucas / Marek,
On Wed, Aug 22, 2012 at 1:32 PM, Lucas Stach dev@lynxeye.de wrote:
Am Mittwoch, den 22.08.2012, 20:22 +0200 schrieb Marek Vasut:
Dear Lucas Stach,
If the environment doesn't provide a MAC address it's not an error to use the one provided by the eth adapter. Actually it should be the default case if the adapter exports it's own address during get_info().
I think we should warn if the addr in env doesn't match the one in the device. Otherwise, if the env doesn't contain ethaddr (or eth1addr, eth2addr ...), pull one from the device and inform user.
Agreed... we also do that.
Sadly it's not obvious from the diff context, but the warning if MAC in env does not match the one from the device is just above this hunk. Joe explained to me that the ueth core should always pull the address from the device and if this succeeds it always sets this address back if none is found in env. As this is the default case (and we are not actually changing the MAC of the device) there is no point in bothering the user with this info.
This is the whole point of this patch. That no one has seen this warning with the current code is simply because no usb eth driver had implemented the mac fetch at the right spot. ASIX now does this.
That's a good thing. It's not saying "Error:"... it's a warning. It's just letting the user know that the MAC is coming from the EEPROM.
Nak.
-Joe