Messages in this thread | | | From | Eric Miao <> | Date | Mon, 1 Apr 2013 15:35:17 +0800 | Subject | Re: [PATCH 3/3] pwm: pxa: Remove clk_enabled field from struct pxa_pwm_chip |
| |
On Mon, Apr 1, 2013 at 3:32 PM, Axel Lin <axel.lin@ingics.com> wrote: > 2013/4/1 Eric Miao <eric.y.miao@gmail.com>: >> On Sun, Mar 31, 2013 at 11:04 PM, Axel Lin <axel.lin@ingics.com> wrote: >>> clk_enable/clk_disable maintain an enable_count, clk_prepare and clk_unprepare >>> also maintain a prepare_count. These APIs will do prepare/enable when the first >>> user calling these APIs, and do disable/unprepare when the corresponding counter >>> reach 0. Thus We don't need to maintain a clk_enabled counter here. >> >> The original intention is to keep a paired clk enable counter no matter >> how the user calls pwm_enable()/pwm_disable() in pair or not, if that's >> no longer the case. > > We don't need to worry that case: > In pwm core, both pwm_enable() and pwm_disable() will always check > PWMF_ENABLED flag.
That's good then, this part of the code was dated before the pwm core, looks like this has been carefully handled. Have my ack on this one then:
Acked-by: Eric Miao <eric.y.miao@gmail.com>
> > > /** > * pwm_enable() - start a PWM output toggling > * @pwm: PWM device > */ > int pwm_enable(struct pwm_device *pwm) > { > if (pwm && !test_and_set_bit(PWMF_ENABLED, &pwm->flags)) > return pwm->chip->ops->enable(pwm->chip, pwm); > > return pwm ? 0 : -EINVAL; > } > EXPORT_SYMBOL_GPL(pwm_enable); > > /** > * pwm_disable() - stop a PWM output toggling > * @pwm: PWM device > */ > void pwm_disable(struct pwm_device *pwm) > { > if (pwm && test_and_clear_bit(PWMF_ENABLED, &pwm->flags)) > pwm->chip->ops->disable(pwm->chip, pwm); > } > EXPORT_SYMBOL_GPL(pwm_disable);
| |