[PATCH] clk: mvebu: Do not gate ge0/1 and runit clocks on Kirkwood

[Sebastian Hesselbarth]

On 01/27/2013 04:28 PM, Andrew Lunn wrote:
>> Except for loosing the MAC address, I disagree. From a "clock gating"
>> point-of-view all kirkwoods (except that crippled one on keymile board)
>> are the same. Even if there is no/one/two ethernet _jacks_ on a specific
>> board you still want to disable all _unused_ ethernet modules inside the
>> SoC.
>>
>> We should find a way to retain the MAC address and in the meantime we
>> can add a "always prepare ge clocks" workaround in
>> kirkwood_clk_legacy_init().
>
> I've been thinking the same, store the MAC address before turning the
> clock off and restore it when enabling the clock. We have a need for
> special clocks on kirkwood anyway for turning off the sata and pcie
> PHYs. So adding special clocks for ethernet is not too big a deal.

Andrew,

if Simon confirms that gbe is really loosing its MAC address when
clocks get gated, I still think that we should rely on other mechanisms
to get a valid MAC address.

Have a look at other DT ethernet drivers, they use local-mac-address
property to pass MAC address from boot loader. IMHO that is what we
should exploit on mv643xx_eth, too.

If the property is set on boot, we can store it in some private data.
If it is not set because some legacy u-boot without DT, we can rely
on what is in the MAC address registers and just call modular eth
on those platforms broken. If the easy fix is to update u-boot or
compile it as built-in we shouldn't care at all. Without proper DT
in u-boot you will always have a board specific kernel anyway.

Sebastian