On 2014-05-28 at 21:43:33 +0200, Chris Metcalf [off-list ref] wrote:

On 5/27/2014 8:04 AM, Tobias Klauser wrote:

quoted

Use is_valid_ether_addr() to check for a valid MAC address to set on the
device.  This will also check for the device address being multicast,
which would have been possible previously.

I don't think having a multicast address here is possible (i.e. as returned
by gxio_mpipe_link_enumerate_mac), and it confuses the actual issue, which
is handling the all-zeroes case in a simulator run.  I'd prefer to see
is_zero_ether_addr() instead.

Ok, I'll send an updated patch which only tests for
!is_zero_ether_addr()

quoted

Also use ether_addr_copy() instead of a manual memcpy() to set the
address.

Furthermore, get rid of a redundant assignment of dev->addr_len. This is
already set by ether_setup() which is called in tile_net_setup().

Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
---
 drivers/net/ethernet/tile/tilegx.c |   17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)

With the change above,

Acked-by: Chris Metcalf <redacted>

-- 
Chris Metcalf, Tilera Corp.
http://www.tilera.com