Re: [RESEND] [PATCH V1 2/2] rtc: da9063: RTC driver

From: Andrew Morton
Date: Wed May 07 2014 - 16:01:14 EST


On Wed, 7 May 2014 13:22:52 +0100 "Opensource [Steve Twiss]" <stwiss.opensource@xxxxxxxxxxx> wrote:

> Add the RTC driver for DA9063.

A few minor things:

>
> ...
>
> +static int da9063_rtc_stop_alarm(struct device *dev)
> +{
> + struct da9063_rtc *rtc = dev_get_drvdata(dev);
> + return regmap_update_bits(rtc->hw->regmap, DA9063_REG_ALARM_Y,
> + DA9063_ALARM_ON, 0);
> +}
> +
> +static int da9063_rtc_start_alarm(struct device *dev)
> +{
> + struct da9063_rtc *rtc = dev_get_drvdata(dev);
> + return regmap_update_bits(rtc->hw->regmap, DA9063_REG_ALARM_Y,
> + DA9063_ALARM_ON, DA9063_ALARM_ON);
> +}

It's conventional to put a newline between end-of-locals and
start-of-code. My version of checkpatch warns about this and soon
everyone's will.

> +static int da9063_rtc_read_time(struct device *dev, struct rtc_time *tm)
> +{
> + struct da9063_rtc *rtc = dev_get_drvdata(dev);
> + unsigned long tm_secs;
> + unsigned long al_secs;
> + u8 data[RTC_DATA_LEN];
> + int ret;
> +
> + ret = regmap_bulk_read(rtc->hw->regmap, DA9063_REG_COUNT_S,
> + data, RTC_DATA_LEN);
> + if (ret < 0) {
> + dev_err(dev, "Failed to read RTC time data: %d\n", ret);
> + return ret;
> + }
> +
> + if (!(data[RTC_SEC] & DA9063_RTC_READ)) {
> + dev_dbg(dev, "RTC not yet ready to be read by the host\n");
> + return -EINVAL;
> + }
> +
> + da9063_data_to_tm(data, tm);
> +
> + (void)rtc_tm_to_time(tm, &tm_secs);
> + (void)rtc_tm_to_time(&rtc->alarm_time, &al_secs);

The void casts are a bit useful - they tell the reader that the author
deliberately chose to ignore the return value. But kernel code
generally doesn't do this trick.


> + /* handle the rtc synchronisation delay */
> + if (rtc->rtc_sync == true && al_secs - tm_secs == 1)
> + (void)memcpy((void *)tm, (const void *)&rtc->alarm_time,
> + sizeof(struct rtc_time));

And all three casts here are unneeded.

And they're actually a bit harmful. If rtc->alarm_time has type `int'
then we really want the warning, but this cast will suppress it.

> + else
> + rtc->rtc_sync = false;
> +
> + return rtc_valid_tm(tm);
> +}


So how does this look?

--- a/drivers/rtc/rtc-da9063.c~rtc-da9063-rtc-driver-fix
+++ a/drivers/rtc/rtc-da9063.c
@@ -82,6 +82,7 @@ static void da9063_tm_to_data(struct rtc
static int da9063_rtc_stop_alarm(struct device *dev)
{
struct da9063_rtc *rtc = dev_get_drvdata(dev);
+
return regmap_update_bits(rtc->hw->regmap, DA9063_REG_ALARM_Y,
DA9063_ALARM_ON, 0);
}
@@ -89,6 +90,7 @@ static int da9063_rtc_stop_alarm(struct
static int da9063_rtc_start_alarm(struct device *dev)
{
struct da9063_rtc *rtc = dev_get_drvdata(dev);
+
return regmap_update_bits(rtc->hw->regmap, DA9063_REG_ALARM_Y,
DA9063_ALARM_ON, DA9063_ALARM_ON);
}
@@ -115,13 +117,12 @@ static int da9063_rtc_read_time(struct d

da9063_data_to_tm(data, tm);

- (void)rtc_tm_to_time(tm, &tm_secs);
- (void)rtc_tm_to_time(&rtc->alarm_time, &al_secs);
+ rtc_tm_to_time(tm, &tm_secs);
+ rtc_tm_to_time(&rtc->alarm_time, &al_secs);

/* handle the rtc synchronisation delay */
if (rtc->rtc_sync == true && al_secs - tm_secs == 1)
- (void)memcpy((void *)tm, (const void *)&rtc->alarm_time,
- sizeof(struct rtc_time));
+ memcpy(tm, &rtc->alarm_time, sizeof(struct rtc_time));
else
rtc->rtc_sync = false;

_

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/