Re: [LTP] [PATCH 3/3] rtc02: loosen the compare precision with few seconds

[Cyril Hrubis <chrubis@suse.cz>]

Hi!
> That possibly has time elapse between twice operations, especially
> on VM we can't guarantee the time precisely equal, let's lose a few
> seconds to make the test happy:
> 
>   tst_test.c:1433: TINFO: Timeout per run is 0h 10m 00s
>   rtc02.c:66: TINFO: To set RTC date/time is: 2020-10-09 13:23:30
>   rtc02.c:80: TINFO: read RTC date/time is: 2020-10-09 13:23:31
>   rtc02.c:83: TFAIL: RTC SET TEST
> 
> Signed-off-by: Li Wang <liwang@redhat.com>
> Cc: Eirik Fuller <efuller@redhat.com>
> ---
>  testcases/kernel/device-drivers/rtc/rtc02.c | 46 +++++++++++++++++++--
>  1 file changed, 42 insertions(+), 4 deletions(-)
> 
> diff --git a/testcases/kernel/device-drivers/rtc/rtc02.c b/testcases/kernel/device-drivers/rtc/rtc02.c
> index 6198a5d5d..a008971d5 100644
> --- a/testcases/kernel/device-drivers/rtc/rtc02.c
> +++ b/testcases/kernel/device-drivers/rtc/rtc02.c
> @@ -41,10 +41,48 @@ static char *rtctime_to_str(struct rtc_time *tm)
>  
>  static int rtc_tm_cmp(struct rtc_time *set_tm, struct rtc_time *read_tm)
>  {
> -	return !((set_tm->tm_sec == read_tm->tm_sec)
> -		&& (set_tm->tm_min == read_tm->tm_min)
> -		&& (set_tm->tm_hour == read_tm->tm_hour)
> -		&& (set_tm->tm_mday == read_tm->tm_mday)
> +	int delta = read_tm->tm_sec - set_tm->tm_sec;
> +
> +	/*
> +	 * To handle the normal and special situations:
> +	 * 1#
> +	 *       set_tm:  2022-04-28 13:00:50
> +	 *       read_tm: 2022-04-28 13:00:50
> +	 * 2#
> +	 *       set_tm:  2022-04-28 13:00:50
> +	 *       read_tm: 2022-04-28 13:00:51
> +	 * 3#
> +	 *       set_tm:  2022-04-28 13:00:59
> +	 *       read_tm: 2022-04-28 13:01:00
> +	 * 4#
> +	 *       set_tm:  2022-04-28 13:59:59
> +	 *       read_tm: 2022-04-28 14:00:00
> +	 *
> +	 * Note: as we have avoided testing around the zero
> +	 * clock, so it's impossible to hit situation 5#
> +	 *       set_tm:  2022-04-28 23:59:59
> +	 *       read_tm: 2022-04-29 00:00:00
> +	 */
> +
> +	/* 1~3 */
> +	if (set_tm->tm_hour == read_tm->tm_hour) {
> +		if (set_tm->tm_min == read_tm->tm_min - 1)
> +			delta += 60;
> +		else if (set_tm->tm_min != read_tm->tm_min)
> +			return 1;
> +	}
> +
> +	/* 4 */
> +	if ((set_tm->tm_hour == read_tm->tm_hour - 1) &&
> +			(set_tm->tm_min == read_tm->tm_min + 59))
> +		delta += 60;
> +	else if ((set_tm->tm_hour != read_tm->tm_hour))
> +		return 1;
> +
> +	if (delta < 0 || delta > 3)
> +		return 1;
> +
> +	return !((set_tm->tm_mday == read_tm->tm_mday)
>  		&& (set_tm->tm_mon == read_tm->tm_mon)
>  		&& (set_tm->tm_year == read_tm->tm_year));

Wouldn't it be easier to convert both dates into a 64bit timestamp and
compare the timestamps?

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp