linux-mips
[Top] [All Lists]

Re: [PATCH] mips/alchemy: define eth platform devices in the correct

To: Manuel Lauss <manuel.lauss@googlemail.com>
Subject: Re: [PATCH] mips/alchemy: define eth platform devices in the correct order
From: Wolfgang Grandegger <wg@grandegger.com>
Date: Sat, 17 Jul 2010 19:17:52 +0200
Cc: linux-mips@linux-mips.org, Wolfgang Grandegger <wg@denx.de>
In-reply-to: <AANLkTikEjMarWjpMLQYCceRzmr7mL8RZp3-9MZ_tvKO8@mail.gmail.com>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <1279377528-3190-1-git-send-email-wg@grandegger.com> <AANLkTikEjMarWjpMLQYCceRzmr7mL8RZp3-9MZ_tvKO8@mail.gmail.com>
Sender: linux-mips-bounce@linux-mips.org
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.9) Gecko/20100430 Fedora/3.0.4-2.fc12 Thunderbird/3.0.4
On 07/17/2010 07:01 PM, Manuel Lauss wrote:
> Servus Wolfgang,
> 
> On Sat, Jul 17, 2010 at 4:38 PM, Wolfgang Grandegger <wg@grandegger.com> 
> wrote:
>> From: Wolfgang Grandegger <wg@denx.de>
>>
>> Currently, the eth devices are probed in the inverse order, first
>> au1xxx_eth1_device and then au1xxx_eth0_device. On the GPR board,
>> this makes trouble:
>>
>>  # ifconfig|grep HWaddr
>>  eth0      Link encap:Ethernet  HWaddr 00:50:C2:0C:30:01
>>  eth1      Link encap:Ethernet  HWaddr 66:22:01:80:38:10
>>
>> A bogous ethernet hwaddr is assigned to the first device and
>> au1xxx_eth0_device is mapped to eth1, which even does not work
>> properly. With this patch, the problems are gone:
>>
>>  # ifconfig|grep HWaddr
>>  eth0      Link encap:Ethernet  HWaddr 66:22:11:32:38:10
>>  eth1      Link encap:Ethernet  HWaddr 66:22:11:32:38:11
> 
> Interesting.  I don't disagree with the patch; what do you think about
> passing MAC address via platform_data?   I don't particularly like
> how the driver is trying to get a MAC address using the prom interface.

Well, I don't think it's a good idea. Each board should have a different
mac address and it's nomally stored somewhere in the boards non-volatile
storage during board bringup.

> I'll try to cook something up.

But not via platform data, please. Or have I missed something. With the
flat device tree (as used for PowerPC) the situation is different
because the boot-loader can fixup the MAC address before booting Linux.

Wolfgang.

<Prev in Thread] Current Thread [Next in Thread>