Re: @Falmari - Can anyone explain?
The problem was actually caused by hitherto unknown bugs in Linux's HyperV drivers, which were fixed by Microsoft, and accepted by the mainstream Linux maintainers.
The underlying problem was apparently fixed, but as a belt and braces addition it looks like the possible value of errno for clock_gettime() was extended to include ESTALE which, now that the problem is fixed, presumably should not be encountered by an application.
That's not substantially changed the system interface for user applications.
Biting the hand that feeds IT
