[PATCH] ARM: imx: Explicitly pass the active level of the PHY reset

[Olof Johansson]

On Sun, Jan 5, 2014 at 6:24 PM, Shawn Guo <shawn.guo at linaro.org> wrote:
> On Sun, Jan 05, 2014 at 09:29:37AM -0800, Olof Johansson wrote:
>> Actually, Russell brought this to my attention (I should have looked
>> closer earlier). You guys have made an incompatible binding change
>> here. In -rc6? That's not cool, we're not doing those any more.
>
> Yes, the driver change hit mainline too quickly.
>
>>
>> What should happen is that the driver change should be reverted
>> instead. You can't rely on every device tree out there changing to
>> accommodate these kind of things any more.
>
> Fabio just sent a revert patch for it, and David has queued it.
>
> This is a typical example that we did not implement device tree
> bindings perfectly right in the first place, and we have to live with it
> forever.  If someday there is board design using an active high PHY
> reset, we basically have to introduce another property for it.

Yes, or another, more precise, binding name.
> But
> it does no harm to correct the level type of the existing property in
> device tree sources anyway.  So I will still queue Fabio's dts patch [1]
> for 3.14 - a different one from what you're dropping here.  The
> patch [1] is against my for-next branch while the dropped one is
> against mainline.

Sounds good. Thanks!


-Olof