adjtimex

BSD doesn't have an adjtimex syscall, so it's very unlikely for there to be a spinlock bug in the adjtimex syscall that doesn't exist.

OSX really needs adjtimex or ntp_adjtime for high-precision clock synchronization for PLL loops (unless there's another way).

If ntpd already issued the adjtimex(2) call, you may need to disable ntpd and reboot to be 100% safe.

Clock skew is tested, but the code paths in adjtimex() for handling a leap second are different.

NTP sends a special message to the kernel (using adjtimex), that boils down to: today you will insert a leap second.

That's a combination that teaches you very quickly about clocksources, ntp and adjtimex.

By using 'adjtimex' command, you can clear the problematic INS bit.

Not to mention that most people would want to adjust the clock speed for indefinite periods of time, which isn't what adjtimex is meant for.

I attempted to fix with adjtimex and the script in the linked question, but to no avail, in the end having to restart them all instead.

Our Linux instances running on Amazon EC2 had no issues since we are not running ntpd on these servers and adjtimex returns status as 64 (clock unsynchronized).

$./adjtimex --print | grep status status: 8209 8209's binary representation is like this.

(1) Directly through NTP (2) the adjtimex syscall, (3) the ntp_gettime call.