[PATCH V4] ARM: bcm281xx: Add timer driver

[csd@broadcom.com (Christian Daudt)]

On 13-02-11 07:39 PM, Stephen Warren wrote:
> On 02/11/2013 08:25 PM, Christian Daudt wrote:
>> On 13-02-11 07:00 PM, Stephen Warren wrote:
>>> On 02/11/2013 04:44 PM, Christian Daudt wrote:
>>>> This adds support for the Broadcom timer, used in the following SoCs:
>>>> BCM11130, BCM11140, BCM11351, BCM28145, BCM28155
>>>> diff --git a/drivers/clocksource/bcm_kona_timer.c
>>>> b/drivers/clocksource/bcm_kona_timer.c
>>>> +static const struct of_device_id bcm_timer_ids[] __initconst = {
>>>> +    {.compatible = "bcm,kona-timer"},
>>>> +    {},
>>>> +};
>>>> +
>>>> +static void __init kona_timers_init(void)
>>>> +{
>>>> +    struct device_node *node;
>>>> +    u32 freq;
>>>> +
>>>> +    node = of_find_matching_node(NULL, bcm_timer_ids);
>>>> +
>>>> +    if (!node)
>>>> +        panic("No timer");
>>> I assume this is for 3.10 now.
>> Oh. I missed 3.9 window ? Got busy with other stuff. Oh well, try again.
>>> Can you rework this to remove
>>> bcm_timer_ids[] and that of_find_matching_node() call, and replace it
>>> with e.g.:
>>>
>>> CLOCKSOURCE_OF_DECLARE(kona, "bcm,kona-timer", kona_timer_init);
>> where can I find an example ? I looked throught armsoc/for-next and the
>> code there has CLOCKSOURCE_OF_DECLARE but it still has
>> of_find_matching_node calls in the init fns - is it still in the process
>> of being removed ?
> It looks like Rob's patches that would allow you to eliminate the
> of_find_matching_node() call aren't checked in yet; I'm not sure if
> they'll make 3.9 or not. I has assumed so, but even if not, I imagine
> they will be available as a dependency branch for you to build on top of
> in 3.10. Those are:
>
> https://lkml.org/lkml/2013/2/7/417
>
> However, you could still convert to CLOCKSOURCE_OF_DECLARE right now;
> the patches to enable that should be in arm-soc's for-next; see
> drivers/clocksource/tegra20_timer.c there for example.
Ok, I've added the CLOCKSOURCE_OF_DECLARE and switched to 
clocksource_of_init in board_bcm, following what is in for-next. If this 
can't make it to 3.9 then I'll add the of_find_matching_node elimination 
in the next patch.

  Thanks,
    csd