If the last calibration was less than 23 hours ago, hwclock [--set | --systohc] refuses to adjust the drift factor in /etc/adjtime but records the current time as last calibration time. Thus, a stale drift factor will persist forever on a machine that is booted more than once a day (quite common on laptops) with hwclock --systohc in a shutdown script. There is no warning unless run with --debug. Documentation says: "Every time you calibrate (set) the clock (using --set or --systohc ), hwclock recalculates the systematic drift rate based on how long it has been since the last calibration". Options to fix this: 1. Just print a warning when the drift factor was not adjusted and update the man page. 2. Reduce the required minimum time since the last calibration to a value that is more likely to be exeeded (maybe 3 hours). 3. Add an option to select the required minimum time. 4. Add a value to /etc/adjtime that records calibration amounts, making it possible to defer drift factor adjustments until the 23 hours are reached. If an extension to the /etc/adjtime format is acceptable, I'm willing to code option 4. Also, hwclock lacks an --initadjfile option to start over with a fresh /etc/adjtime (current workaround is to remove it by hand before invoking hwclock). - 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
