Re: [PATCH 8/9] RFC: rtc: hctosys: support rtc_read_timeval() for high precision clocks

[John Stultz <john.stultz@linaro.org>]

On 06/14/2013 09:52 AM, Alexander Holler wrote:
> Some RTCs do provide a higher precision than seconds. Add support for them
> by trying rtc_read_timeval() before using rtc_read_time() to get the time
> in the hctosys mechanism.
>
>
[snip]
> +	rc = rtc_read_timeval(rtc, &tv);
> +	if (rc || (!tv.tv_sec && !tv.tv_usec)) {
> +		rc = rtc_read_time(rtc, &now);
> +		if (unlikely(rc)) {
> +			dev_err(rtc->dev.parent,
> +				"rtc core: error reading time from RTC: %d\n",
> +				rc);
> +			return;
> +		}
> +		rtc_tm_to_time(&now, &ts.tv_sec);
> +		ts.tv_nsec = NSEC_PER_SEC >> 1;
> +	} else {
> +		rtc_time_to_tm(tv.tv_sec, &now);
> +		ts.tv_sec = tv.tv_sec;
> +		ts.tv_nsec = tv.tv_usec*NSEC_PER_USEC;


Yea, this sort of fallback logic should be centralized in the RTC layer 
rather then in the individual users.

Might be easiest to modify the rtc_read_timeval() interface to try the 
ops->read_timeval() operation and do the fallback to rtc_read_time() 
internally.

thanks
-john