On 2/11/19 8:06 PM, Uwe Kleine-König wrote: > Hello Fabrice, > > On Mon, Feb 11, 2019 at 05:12:02PM +0100, Fabrice Gasnier wrote: >> Add a device link between the PWM consumer and the PWM provider. This >> enforces the PWM user to get suspended before the PWM provider. It >> allows proper synchronization of suspend/resume sequences: the PWM user >> is responsible for properly stopping PWM, before the provider gets >> suspended: see [1]. Add the device link in: >> - pwm_get() >> - devm_pwm_get() >> - devm_of_pwm_get() >> as it requires a reference to the device for the PWM consumer. >> >> [1] https://lkml.org/lkml/2019/2/5/770 >> >> Suggested-by: Thierry Reding <thierry.reding@xxxxxxxxx> >> Signed-off-by: Fabrice Gasnier <fabrice.gasnier@xxxxxx> >> --- >> drivers/pwm/core.c | 13 +++++++++++-- >> 1 file changed, 11 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/pwm/core.c b/drivers/pwm/core.c >> index 1581f6a..2835e27 100644 >> --- a/drivers/pwm/core.c >> +++ b/drivers/pwm/core.c >> @@ -770,8 +770,13 @@ struct pwm_device *pwm_get(struct device *dev, const char *con_id) >> int err; >> >> /* look up via DT first */ >> - if (IS_ENABLED(CONFIG_OF) && dev && dev->of_node) >> - return of_pwm_get(dev->of_node, con_id); >> + if (IS_ENABLED(CONFIG_OF) && dev && dev->of_node) { >> + pwm = of_pwm_get(dev->of_node, con_id); >> + if (!IS_ERR(pwm)) >> + device_link_add(dev, pwm->chip->dev, >> + DL_FLAG_AUTOREMOVE_CONSUMER); >> + return pwm; >> + } >> >> /* >> * We look up the provider in the static table typically provided by >> @@ -851,6 +856,8 @@ struct pwm_device *pwm_get(struct device *dev, const char *con_id) >> pwm->args.period = chosen->period; >> pwm->args.polarity = chosen->polarity; >> >> + device_link_add(dev, pwm->chip->dev, DL_FLAG_AUTOREMOVE_CONSUMER); >> + >> return pwm; >> } >> EXPORT_SYMBOL_GPL(pwm_get); >> @@ -943,6 +950,8 @@ struct pwm_device *devm_of_pwm_get(struct device *dev, struct device_node *np, >> if (!IS_ERR(pwm)) { >> *ptr = pwm; >> devres_add(dev, ptr); >> + device_link_add(dev, pwm->chip->dev, >> + DL_FLAG_AUTOREMOVE_CONSUMER); > > IMHO it's surprising that devm_of_pwm_get() does more than of_pwm_get() > + devres stuff. I'd put device_link_add() into of_pwm_get(). Hi Uwe, I also agree with this. But I think this implies modifying the API for of_pwm_get(): /** * of_pwm_get() - request a PWM via the PWM framework + * @dev: device for PWM consumer * @np: device node to get the PWM from * @con_id: consumer name It seems there aren't much of_pwm_get() users currently. Does this look sensible ? Best regards, Fabrice > > Best regards > Uwe >
