Re: [RFC PATCH 13/50] ARM: at91: move at91rm9200 SoC to new at91clk implem

[boris brezillon]

Hello,

On 20/06/2013 08:52, Mike Turquette wrote:
> Quoting Boris BREZILLON (2013-06-07 08:11:03)
> > +static struct clk_lookup pioA_clk_lookup[] = {
> > +       CLKDEV_INIT(NULL, "pioA_clk", NULL),
> > +       CLKDEV_INIT(NULL, "pioA", NULL),
> > +};
> It would be great to get rid of this clkdev data from the kernel as
> well. Have you looked into encoding the aliases into your DT bindings?
I already switched all dt boards to dt clk definition (see patch 36 to 50).
These definition are kept for all non-dt boards (see in 
arch/arm/mach-at91/board-xxx.c).
This will be cleaned as soon as all non-dt boards are moved to dt.
> Or have you looked into using of_clk_get?
If I remember correctly, the clk_get function first tries to get the clk 
from dt binding.
Then if it fails it uses the clk_lookup info. Am I right ?
If so, I don't need to use of_clk_get in drivers calling clk_get with a 
non NULL device
parameter.

The only place where I use of_clk_get is in init timer driver 
(at91sam926x_time.c).
I first try to retrieve the clk from dt, then if it fails (or if dt is 
not supported),
I use clk_get with a NULL device param, which will lead to clk_lookup 
search.


Tell me if I misunderstood the API.
> Regards,
> Mike

Best Regards,
Boris
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/