[PATCH V2] watchdog: mpcore: Add DT probing support for ARM mpcore watchdog

[Marc Zyngier]

On 23/04/12 12:35, viresh kumar wrote:
> On Mon, Apr 23, 2012 at 2:06 PM, Marc Zyngier <marc.zyngier at arm.com> wrote:
>> Irk! Have you actually tested this?
> 
> Yes. But with incorrect base address. I passed timers address by mistake.

Ah, that explains why it worked. I suppose you do have additional
patches that fix the IRQ request bit?

>> The DT binding indicates:
>> - reg : Specify the base address and the size of the TWD watchdog
>>        register window.
>>
>> while all the offsets in smp_twd.h are expressed in bytes from the TWD
>> *timer* base. So you have to either fix these offsets (which breaks
>> potential users of the non-DT version of the driver), or correct the
>> base when using DT.
> 
> You are correct. Which one do you prefer:
> - Change bindings to pass timers base address

No. We already discussed this on LAK, and the outcome was the current
binding, so let's not change that again.

> - subtract 0x20 from base address for DT case

That's a possibility.

> - something else.

Given that no in-tree platform seem to be using this watchdog (at least
a quick grep didn't reveal anything), I'd be inclined to simply change
the offset in smp_twd.h and let them break.

	M.
-- 
Jazz is not dead. It just smells funny...