Dear Heiko Schocher,
In message 4BB18E5F.6060705@denx.de you wrote:
initialize mac address with the value from "ethaddr", before doing some network commands. This is not in line with u-boot design principle "not to initalize not used devices", and maybe should go away, if there is a solution for passing the mac address to arm linux kernels.
Tested on the suen3 board.
Signed-off-by: Heiko Schocher hs@denx.de
posting this patch as a result of this discussion:
http://lists.denx.de/pipermail/u-boot/2010-March/069025.html
Hm... it seems I misunderstood you there :-(
You wrote (here: http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/76173/focus=76338):
| > know about these, of course). So if kirkwood_egiga is clean (in this | > respect), there is no need to change it. | | It ends in the same problem, as the fec_mxc.c driver has ...
At this point I thought that you were referring to the problem that the fec_mxc.c would under certain conditions ignore the "ethaddr" environment setting.
}
while (!eth_getenv_enetaddr(s, dev->enetaddr)) {
if (!eth_getenv_enetaddr(s, dev->enetaddr)) {
I think this change is actually not related to the other modifi- cations in this patch. Axctually, it is not even needed. OK, you can call it paranoid and a waste of time to re-check the setting of the environment variable after running setenv(), but then it's not a real bug either. In any case it is unrelated and should not be mixed into this commit.
dev->enetaddr[0] = 0x02; dev->enetaddr[1] = 0x50;@@ -688,6 +688,7 @@ int kirkwood_egiga_initialize(bd_t * bis) dev->enetaddr[5] = get_random_hex(); eth_setenv_enetaddr(s, dev->enetaddr); }
port_uc_addr_set(dkwgbe->regs, dev->enetaddr);
This is another issue - this is not a bug fix.
I will leave it up to others (especially Ben) to comment on this part.
Best regards,
Wolfgang Denk