iio: adc: rzg2l: Use RUNTIME_PM_OPS() instead of SET_*
Subsystems (2)
Jonathan Cameron
maintainerPATCH 1/1 iio: adc: rzg2l: Use RUNTIME_PM_OPS() instead of SET_*
From: Jonathan Cameron <[email protected]>
Here we could use DEFINE_RUNTIME_DEV_PM_OPS() but that would have the side effect of providing suspend and resume support. That would be harmless but also of little purpose as this driver does very simplistic power management with synchronous power up and down around individual channel reads.
In general these new PM macros avoid the need to mark functions __maybe_unused, whilst still allowing the compiler to remove them if they are unused.
--- drivers/iio/adc/rzg2l_adc.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/iio/adc/rzg2l_adc.c +6 -6
@@ -547,7 +547,7 @@ static const struct of_device_id rzg2l_adc_match[] = { };MODULE_DEVICE_TABLE (of, rzg2l_adc_match); -static int __maybe_unused rzg2l_adc_pm_runtime_suspend (struct device *dev) +static int rzg2l_adc_pm_runtime_suspend (struct device *dev) {struct iio_dev *indio_dev =dev_get_drvdata (dev);struct rzg2l_adc *adc =iio_priv (indio_dev);
@@ -559,7 +559,7 @@ static int __maybe_unused rzg2l_adc_pm_runtime_suspend(struct device *dev)return 0 ; } -static int __maybe_unused rzg2l_adc_pm_runtime_resume (struct device *dev) +static int rzg2l_adc_pm_runtime_resume (struct device *dev) {struct iio_dev *indio_dev =dev_get_drvdata (dev);struct rzg2l_adc *adc =iio_priv (indio_dev);
@@ -581,9 +581,9 @@ static int __maybe_unused rzg2l_adc_pm_runtime_resume(struct device *dev) }static const struct dev_pm_ops rzg2l_adc_pm_ops = { -SET_RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, - rzg2l_adc_pm_runtime_resume, -NULL ) +RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, + rzg2l_adc_pm_runtime_resume, +NULL ) };static struct platform_driver rzg2l_adc_driver = {
@@ -591,7 +591,7 @@ static struct platform_driver rzg2l_adc_driver = { .driver = { .name = DRIVER_NAME, .of_match_table = rzg2l_adc_match, - .pm = &rzg2l_adc_pm_ops, + .pm = pm_ptr(&rzg2l_adc_pm_ops), }, };
Signature · 1 line
2.37.1
Quoted 12 lines · Jonathan Cameron
From: Jonathan Cameron <[email protected]>
Here we could use
DEFINE_RUNTIME_DEV_PM_OPS()but that would have the side effect of providing suspend and resume support. That would be harmless but also of little purpose as this driver does very simplistic power management with synchronous power up and down around individual channel reads.In general these new PM macros avoid the need to mark functions
__maybe_unused, whilst still allowing the compiler to remove them if they are unused.
...
drivers/iio/adc/rzg2l_adc.c static int __maybe_unused rzg2l_adc_pm_runtime_resume(struct device *dev) · line 587
Quoted 8 lines
static const struct
dev_pm_opsrzg2l_adc_pm_ops= {drivers/iio/adc/rzg2l_adc.c +3
-3-SET_RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, - rzg2l_adc_pm_runtime_resume, - NULL) +RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, + rzg2l_adc_pm_runtime_resume, + NULL) };
DEFINE_RUNTIME_DEV_PM_OPS() ?
Signature · 2 lines
With Best Regards, Andy Shevchenko
Jonathan Cameron
maintainerQuoted 25 lines · On Mon, 8 Aug 2022 11:34:23 +0200 Andy Shevchenko <[email protected]>…
Quoted 12 lines · Jonathan Cameron
From: Jonathan Cameron <[email protected]>
Here we could use
DEFINE_RUNTIME_DEV_PM_OPS()but that would have the side effect of providing suspend and resume support. That would be harmless but also of little purpose as this driver does very simplistic power management with synchronous power up and down around individual channel reads.In general these new PM macros avoid the need to mark functions
__maybe_unused, whilst still allowing the compiler to remove them if they are unused....
drivers/iio/adc/rzg2l_adc.c static int __maybe_unused rzg2l_adc_pm_runtime_resume(struct device *dev) · line 587
Quoted 8 lines
static const struct
dev_pm_opsrzg2l_adc_pm_ops= {drivers/iio/adc/rzg2l_adc.c +3
-3-SET_RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, - rzg2l_adc_pm_runtime_resume, - NULL) +RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, + rzg2l_adc_pm_runtime_resume, + NULL) };
Quoted 2 lines
DEFINE_RUNTIME_DEV_PM_OPS()?
Disagreeing with the patch description argument on why I didn't do that? The extra ops set will never have anything to do... Mostly harmless, but kind of gives the wrong impression of what is going on in this driver.
Jonathan
Quoted 25 lines · Jonathan Cameron
Quoted 23 lines · On Mon, 8 Aug 2022 11:34:23 +0200 Andy Shevchenko <[email protected]>…
Quoted 9 lines · Jonathan Cameron
Here we could use
DEFINE_RUNTIME_DEV_PM_OPS()but that would have the side effect of providing suspend and resume support. That would be harmless but also of little purpose as this driver does very simplistic power management with synchronous power up and down around individual channel reads.In general these new PM macros avoid the need to mark functions
__maybe_unused, whilst still allowing the compiler to remove them if they are unused....
Quoted 8 lines
static const struct
dev_pm_opsrzg2l_adc_pm_ops= {drivers/iio/adc/rzg2l_adc.c +3
-3-SET_RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, - rzg2l_adc_pm_runtime_resume, - NULL) +RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, + rzg2l_adc_pm_runtime_resume, + NULL) };
Quoted 6 lines
Quoted 2 lines
DEFINE_RUNTIME_DEV_PM_OPS()?Disagreeing with the patch description argument on why I didn't do that? The extra ops set will never have anything to do... Mostly harmless, but kind of gives the wrong impression of what is going on in this driver.
As per thread with Paul, this patch has no function change intentions, but also, if tested on hardware, enabling system sleep states shouldn't be harmful.
Signature · 2 lines
With Best Regards, Andy Shevchenko
Jonathan Cameron
maintainerQuoted 28 lines · On Sun, 14 Aug 2022 22:12:59 +0300 Andy Shevchenko <[email protected]>…
Quoted 27 lines · Jonathan Cameron
Quoted 25 lines · On Mon, 8 Aug 2022 11:34:23 +0200 Andy Shevchenko <[email protected]>…
Quoted 9 lines · Jonathan Cameron
Here we could use
DEFINE_RUNTIME_DEV_PM_OPS()but that would have the side effect of providing suspend and resume support. That would be harmless but also of little purpose as this driver does very simplistic power management with synchronous power up and down around individual channel reads.In general these new PM macros avoid the need to mark functions
__maybe_unused, whilst still allowing the compiler to remove them if they are unused....
drivers/iio/adc/rzg2l_adc.c static int __maybe_unused rzg2l_adc_pm_runtime_resume(struct device *dev) · line 587
Quoted 8 lines
static const struct
dev_pm_opsrzg2l_adc_pm_ops= {drivers/iio/adc/rzg2l_adc.c +3
-3-SET_RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, - rzg2l_adc_pm_runtime_resume, - NULL) +RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, + rzg2l_adc_pm_runtime_resume, + NULL) };
DEFINE_RUNTIME_DEV_PM_OPS()?
Quoted 9 lines
Quoted 4 lines
Disagreeing with the patch description argument on why I didn't do that? The extra ops set will never have anything to do... Mostly harmless, but kind of gives the wrong impression of what is going on in this driver.
As per thread with Paul, this patch has no function change intentions, but also, if tested on hardware, enabling system sleep states shouldn't be harmful.
This one is different from that case where there might be side effects. Here the suspend and resume are (I think) guaranteed to have nothing to do in all cases - because the driver does synchronous power up and power down in all paths. So in all cases we are already in runtime suspended state on a call to suspend.
Joanthan
Jonathan Cameron
maintainerQuoted 8 lines · On Sun, 7 Aug 2022 20:07:12 +0100 Jonathan Cameron <[email protected]> wrote:
From: Jonathan Cameron <[email protected]>
Here we could use
DEFINE_RUNTIME_DEV_PM_OPS()but that would have the side effect of providing suspend and resume support. That would be harmless but also of little purpose as this driver does very simplistic power management with synchronous power up and down around individual channel reads.
Quoted 7 lines
In general these new PM macros avoid the need to mark functions
__maybe_unused, whilst still allowing the compiler to remove them if they are unused.Signed-off-by Jonathan Cameron <[email protected]>Cc Biju Das <[email protected]>Cc Geert Uytterhoeven <[email protected]>
+CC Prabhakar on basis of recent DT binding patch might have hardware
to test changing this to DEFINE_RUNTIME_DEV_PM_OPS.
If not I'm tempted to just pick this one up on basis it does no harm and we can revisit later.
Thanks,
Jonathan
Quoted 48 lines
--- drivers/iio/adc/rzg2l_adc.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)drivers/iio/adc/rzg2l_adc.c +6
-6@@ -547,7 +547,7 @@ static const struct of_device_id rzg2l_adc_match[] = { };MODULE_DEVICE_TABLE (of, rzg2l_adc_match); -static int __maybe_unused rzg2l_adc_pm_runtime_suspend (struct device *dev) +static int rzg2l_adc_pm_runtime_suspend (struct device *dev) {struct iio_dev *indio_dev =dev_get_drvdata (dev);struct rzg2l_adc *adc =iio_priv (indio_dev);@@ -559,7 +559,7 @@ static int __maybe_unused rzg2l_adc_pm_runtime_suspend(struct device *dev)return 0 ; } -static int __maybe_unused rzg2l_adc_pm_runtime_resume (struct device *dev) +static int rzg2l_adc_pm_runtime_resume (struct device *dev) {struct iio_dev *indio_dev =dev_get_drvdata (dev);struct rzg2l_adc *adc =iio_priv (indio_dev);@@ -581,9 +581,9 @@ static int __maybe_unused rzg2l_adc_pm_runtime_resume(struct device *dev) }static const struct dev_pm_ops rzg2l_adc_pm_ops = { -SET_RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, - rzg2l_adc_pm_runtime_resume, -NULL ) +RUNTIME_PM_OPS (rzg2l_adc_pm_runtime_suspend, + rzg2l_adc_pm_runtime_resume, +NULL ) };static struct platform_driver rzg2l_adc_driver = {@@ -591,7 +591,7 @@ static struct platform_driver rzg2l_adc_driver = { .driver = { .name = DRIVER_NAME, .of_match_table = rzg2l_adc_match, - .pm = &rzg2l_adc_pm_ops, + .pm = pm_ptr(&rzg2l_adc_pm_ops), }, };