The comments in cmos_read_alarm() and cmos_set_alarm() do not apply to
current code, replace them.
These comments were added in
commit fbb974ba693b ("rtc: cmos: Do not export alarm rtc_ops when we do not support alarms")
which introduced a separate struct rtc_class_ops, which was used when
the device did not support RTC alarms. The functions cmos_read_alarm()
and cmos_set_alarm() were called not only from the rtc_op struct, but
also explicitly, so they had to independently check for RTC alarm
support.
The separate rtc_class_ops structure was later removed in
commit 30f5bd537fdb ("rtc: cmos: remove cmos_rtc_ops_no_alarm")
but the comments remained and now are obsolete.
Signed-off-by: Mateusz Jończyk <[email protected]>
Cc: Alessandro Zummo <[email protected]>
Cc: Alexandre Belloni <[email protected]>
---
This patch applies cleanly on rtc-next.
drivers/rtc/rtc-cmos.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/rtc/rtc-cmos.c b/drivers/rtc/rtc-cmos.c
index 7c006c2b125f..cd82eff2630a 100644
--- a/drivers/rtc/rtc-cmos.c
+++ b/drivers/rtc/rtc-cmos.c
@@ -290,7 +290,7 @@ static int cmos_read_alarm(struct device *dev, struct rtc_wkalrm *t)
.time = &t->time,
};
- /* This not only a rtc_op, but also called directly */
+ /* Fail if the RTC alarm is not supported */
if (!is_valid_irq(cmos->irq))
return -EIO;
@@ -523,7 +523,7 @@ static int cmos_set_alarm(struct device *dev, struct rtc_wkalrm *t)
unsigned char rtc_control;
int ret;
- /* This not only a rtc_op, but also called directly */
+ /* Fail if the RTC alarm is not supported */
if (!is_valid_irq(cmos->irq))
return -EIO;
--
2.25.1
On 07/01/2022 19:30:29+0100, Mateusz Jończyk wrote:
> The comments in cmos_read_alarm() and cmos_set_alarm() do not apply to
> current code, replace them.
>
> These comments were added in
> commit fbb974ba693b ("rtc: cmos: Do not export alarm rtc_ops when we do not support alarms")
> which introduced a separate struct rtc_class_ops, which was used when
> the device did not support RTC alarms. The functions cmos_read_alarm()
> and cmos_set_alarm() were called not only from the rtc_op struct, but
> also explicitly, so they had to independently check for RTC alarm
> support.
Isn't cmos_read_alarm still called directly from cmos_check_wkalrm and
cmos_suspend ?
>
> The separate rtc_class_ops structure was later removed in
> commit 30f5bd537fdb ("rtc: cmos: remove cmos_rtc_ops_no_alarm")
> but the comments remained and now are obsolete.
>
> Signed-off-by: Mateusz Jończyk <[email protected]>
> Cc: Alessandro Zummo <[email protected]>
> Cc: Alexandre Belloni <[email protected]>
> ---
>
> This patch applies cleanly on rtc-next.
>
> drivers/rtc/rtc-cmos.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/rtc/rtc-cmos.c b/drivers/rtc/rtc-cmos.c
> index 7c006c2b125f..cd82eff2630a 100644
> --- a/drivers/rtc/rtc-cmos.c
> +++ b/drivers/rtc/rtc-cmos.c
> @@ -290,7 +290,7 @@ static int cmos_read_alarm(struct device *dev, struct rtc_wkalrm *t)
> .time = &t->time,
> };
>
> - /* This not only a rtc_op, but also called directly */
> + /* Fail if the RTC alarm is not supported */
> if (!is_valid_irq(cmos->irq))
> return -EIO;
>
> @@ -523,7 +523,7 @@ static int cmos_set_alarm(struct device *dev, struct rtc_wkalrm *t)
> unsigned char rtc_control;
> int ret;
>
> - /* This not only a rtc_op, but also called directly */
> + /* Fail if the RTC alarm is not supported */
> if (!is_valid_irq(cmos->irq))
> return -EIO;
>
> --
> 2.25.1
>
--
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com