Well, not really. Clock skew is tested, but the code paths in adjtimex() for handling a leap second are different. This is why the 2005 leap second crashed eve

Well, not really. Clock skew is tested, but the code paths in adjtimex() for handling a leap second are different. This is why the 2005 leap second crashed every Linux 2.4 kernel worldwide as I recall. Thankfully there's been a push toward leap smear (certainly on every system I run) which should obviate this type of problem entirely.