Hello Kalev, On Monday, July 28, 2008 at 2:42:17 +0300, Kalev Soikonen wrote: > Clock adjustment at the time of boot is of rather dubious value. > With no synchronization source, writing the RTC can only increase > its entropy. 100% agreed. > This patch: > - compensate for drift when setting the system clock from rtc. Excellent principle of course. The mainline does just this since hwclock 2.7 in May 1999, and the results are excellent. Your patch has a little problem: the additional calculations increase the time elapsed between the read of the hardware clock tick and the final settimeofday(). This delay is not compensated. On my test machine this introduces an error of 1.6 ms in normal mode, and 6 ms in --debug mode. > Another worthwhile option would be --quick. This would do away with > syncing to tick boundary, trading precision for faster operation. Name it --fast for compatibility. Alain. -- To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
