On 03/24/2010 03:54 PM, Mike Frysinger wrote: > On Wednesday 24 March 2010 02:24:41 Thomas Chou wrote: > >> +static int ethoc_set_mac_address(struct eth_device *dev, void *addr) >> +{ >> + struct ethoc *priv = (void *)dev->priv; >> + u8 *mac = (u8 *) addr; >> + >> + ethoc_write(priv, MAC_ADDR0, (mac[2]<< 24) | (mac[3]<< 16) | >> + (mac[4]<< 8) | (mac[5]<< 0)); >> + ethoc_write(priv, MAC_ADDR1, (mac[0]<< 8) | (mac[1]<< 0)); >> + >> + return 0; >> +} >> + >> +static int ethoc_init(struct eth_device *dev, bd_t * bd) >> +{ >> + eth_getenv_enetaddr("ethaddr", dev->enetaddr); >> + ethoc_set_mac_address(dev, dev->enetaddr); >> > yikes ! drivers shouldnt be poking the env -- the common net layer handles > syncing the env and the eth_device structure for you. so the code should > simply be: > ethoc_set_mac_address(dev); > > and that function works off of dev->enetaddr. or pass it dev->enetaddr and > have it work off the specified buffer. your choice. > -mike > Hi Mike,
Thank you. Please review the update. Best regards, Thomas _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot