On Tue, Feb 12, 2019 at 03:28:50PM +0100, Uwe Kleine-König wrote:
> On Tue, Feb 12, 2019 at 10:49:27AM +0100, Mathieu Othacehe wrote:
> > Add support for hi3559v100-shub-pwm and hisilicon,hi3559v100-pwm
> > platforms. They require a special quirk: pwm has to be enabled again
> > to force duty_cycle refresh.
> >
> > Signed-off-by: Mathieu Othacehe <m.othacehe@xxxxxxxxx>
> > ---
> > drivers/pwm/pwm-hibvt.c | 26 +++++++++++++++++++++++---
> > 1 file changed, 23 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/pwm/pwm-hibvt.c b/drivers/pwm/pwm-hibvt.c
> > index 27c107e78d59..bf33aa24433c 100644
> > --- a/drivers/pwm/pwm-hibvt.c
> > +++ b/drivers/pwm/pwm-hibvt.c
> > @@ -49,6 +49,7 @@ struct hibvt_pwm_chip {
> > struct clk *clk;
> > void __iomem *base;
> > struct reset_control *rstc;
> > + bool quirk_force_enable;
> > };
> >
> > struct hibvt_pwm_soc {
> > @@ -56,6 +57,7 @@ struct hibvt_pwm_soc {
> > };
> >
> > static const struct hibvt_pwm_soc pwm_soc[2] = {
> > + { .num_pwms = 2 },
> > { .num_pwms = 4 },
> > { .num_pwms = 8 },
>
> The members of this struct are used as of-data (in struct
> of_device_id::data below). When looking at the usage:
>
> { .compatible = "hisilicon,hi3516cv300-pwm", .data = &pwm_soc[1] },
> { .compatible = "hisilicon,hi3519v100-pwm", .data = &pwm_soc[2] },
> { .compatible = "hisilicon,hi3559v100-shub-pwm", .data = &pwm_soc[2] },
> { .compatible = "hisilicon,hi3559v100-pwm", .data = &pwm_soc[0] },
>
> this isn't exactly easy to understand. I would prefer to do it as
> follows:
>
> static const struct hibvt_pwm_soc hi3516cv300_soc_info = {
> .num_pwms = 2,
> };
> ...
>
> static const struct of_device_id hibvt_pwm_of_match[] = {
> { .compatible = "hisilicon,hi3516cv300-pwm", .data = &hi3516cv300_soc_info },
> ...
> };
>
> Then you could also add a member to hibvt_pwm_soc to signal if that
> force_enable quirk is necessary and would not need to use
> of_device_is_compatible() to determine this. The result is that you have
> a description of all relevant differences in a single place.
>
> @Thierry: Also this is yet another driver instance where a num-pwms
> property would simplify the driver because up to before this patch this
> was the only difference between the different variants.
We don't need the num-pwms in device tree if it can be derived from the
compatible string.
Thierry
Attachment:
signature.asc
Description: PGP signature
