| Message ID | 20210429133218.66971-1-u.kleine-koenig@pengutronix.de |
|---|---|
| State | Accepted |
| Headers | show |
| Series | pwm: tiecap: Drop .free() callback | expand |
Hello, On Thu, Apr 29, 2021 at 03:32:18PM +0200, Uwe Kleine-König wrote: > ecap_pwm_free is only called when a consumer releases the PWM (using > pwm_put() or pwm_free()). The consumer is expected to disable the PWM > before doing that. It's not clear if a warning about that is justified, but > if it is this is independent of the actual driver and can better be done in > the core. Also if there is a good reason it's wrong to disable the hardware > and so the call to pm_runtime_put_sync() should be dropped. Moreover there > is no matching pwm_runtime_get call and so the runtime usage counter might > become negative. For the record: The two people from TI (Vaibhav Bedia <vaibhav.bedia@ti.com>, Avinash Philip <avinashphilip@ti.com>) I addressed with that patch (because they were involved previously with this driver) didn't receive my mail. The addresses don't exist according to TI's mail server. I dropped them from To: for this reply. Best regards Uwe
diff --git a/drivers/pwm/pwm-tiecap.c b/drivers/pwm/pwm-tiecap.c index b9a17ab0c202..134f0db28870 100644 --- a/drivers/pwm/pwm-tiecap.c +++ b/drivers/pwm/pwm-tiecap.c @@ -168,16 +168,7 @@ static void ecap_pwm_disable(struct pwm_chip *chip, struct pwm_device *pwm) pm_runtime_put_sync(pc->chip.dev); } -static void ecap_pwm_free(struct pwm_chip *chip, struct pwm_device *pwm) -{ - if (pwm_is_enabled(pwm)) { - dev_warn(chip->dev, "Removing PWM device without disabling\n"); - pm_runtime_put_sync(chip->dev); - } -} - static const struct pwm_ops ecap_pwm_ops = { - .free = ecap_pwm_free, .config = ecap_pwm_config, .set_polarity = ecap_pwm_set_polarity, .enable = ecap_pwm_enable,
ecap_pwm_free is only called when a consumer releases the PWM (using pwm_put() or pwm_free()). The consumer is expected to disable the PWM before doing that. It's not clear if a warning about that is justified, but if it is this is independent of the actual driver and can better be done in the core. Also if there is a good reason it's wrong to disable the hardware and so the call to pm_runtime_put_sync() should be dropped. Moreover there is no matching pwm_runtime_get call and so the runtime usage counter might become negative. Fixes: 8e0cb05b3b75 ("pwm: pwm-tiecap: PWM driver support for ECAP APWM") Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> --- drivers/pwm/pwm-tiecap.c | 9 --------- 1 file changed, 9 deletions(-) base-commit: a6efb35019d00f483a0e5f188747723371d659fe