Messages in this thread |  | | Date | Tue, 26 Nov 2013 13:32:13 -0800 | Subject | Re: [PATCH 2/5] pwm: kona: Introduce Kona PWM controller support | From | Tim Kryger <> |
| |
On Tue, Nov 26, 2013 at 1:45 AM, Thierry Reding <thierry.reding@gmail.com> wrote: > On Mon, Nov 25, 2013 at 05:38:44PM -0800, Tim Kryger wrote:
> Okay, that's fine then. Do you have a pointer to that documentation?
Unfortunately the documentation is not publicly available at the moment.
>> The hardware is always enabled and at best can be be configured to >> operate at zero duty. > > What good are the PWMOUT_ENABLE bits then? Is that really only used for > triggering updates? That's what another comment suggests, but if so, can > the comment in kona_pwmc_apply_settings() be extended to mention that?
The guidance from the documentation is to treat the enable bits as a trigger.
Sure I can expand the comment to try and make this more clear.
>> The settings in HW may have already been triggered and might not be >> what you want. >> >> For example: >> >> /sys/class/pwm/pwmchip0/pwm1# echo 1 > enable >> /sys/class/pwm/pwmchip0/pwm1# echo 5000 > period >> /sys/class/pwm/pwmchip0/pwm1# echo 2000 > duty_cycle >> /sys/class/pwm/pwmchip0/pwm1# echo 0 > enable >> /sys/class/pwm/pwmchip0/pwm1# echo 1 > enable > > I'm not exactly sure what this is supposed to demonstrate.
The computation has to be performed for every operation except the disable.
> All of that said, I wasn't able to dig up a normative policy and either > of your proposed alternatives do exist in the kernel, so I withdraw any > objections regarding MODULE_AUTHOR. Take your pick.
Thanks
|  |