Hi Kumar,

On Tue, Aug 30, 2011 at 7:19 AM, Kumar Gala galak@kernel.crashing.org wrote:

With the following commit I now see:

commit 7616e7850804c7c69e0a22c179dfcba9e8f3f587 Author: Simon Glass sjg@chromium.org Date:   Mon Jun 13 16:13:10 2011 -0700

Add Ethernet hardware MAC address framework to usbnet

Built-in Ethernet adapters support setting the mac address by means of a   ethaddr environment variable for each interface (ethaddr, eth1addr, eth2addr).

This adds similar support to the USB network side, using the names   usbethaddr, usbeth1addr, etc. They are kept separate since we don't want   a USB device taking the MAC address of a built-in device or vice versa.

Signed-off-by: Simon Glass sjg@chromium.org   Tested-by: Eric Bénard eric@eukrea.com

messages like:

Warning: failed to set MAC address

However, I don't see how this ever works properly as we do:

if (eth_write_hwaddr(dev, NULL, eth_number))

this ends up calling:

eth_getenv_enetaddr_by_index(base_name, eth_number, env_enetaddr)

but base_name was passed in as NULL.

So we lookup env strings like:

<NULL>1addr <NULL>2addr <NULL>3addr

etc., seems wrong.

Yes you are right. I dropped the handling of NULL from eth_getenv_enetaddr_by_index() in patch version 5 (due to review request) but did not add "eth" back into the eth.c call. This then carried on through the following patches. I will send a patch to fix this.

Regards, Simon

• k