JackPosixProcessSync: Fix mutex owner on TimedWait() timeout.

[0EVSG]

Per POSIX definition, pthread_cond_timedwait() re-acquires the mutex when a timeout occurs and ETIMEDOUT is returned. In that case also mark the waiting thread as owner again, for consistency. Otherwise subsequent attempts to unlock the mutex will fail, leaving the mutex locked forever.

The return value of TimedWait() is still false on timeout, the same as for other platform's implementations.
But AFAICT, JackPosixProcessSync::TimedWait() is only used by my new FreeBSD OSS backend, see #943.

[0EVSG]

Same here, should I just merge this if nobody's got time for a review?

[falkTX]

Would be nice to confirm with some automated testing that we do not get regressions, but since as you say this is only used for BSD the breakge is more minimal and I can assume you tested it.
Feel free to merge yes

[0EVSG]

It's difficult to test for regression when the code isn't even in use ;-)

I only discovered this problem when testing #943 - the timed wait treated a timeout like an error. Which left the mutex locked, but with ownership unset, and thus prevented any further unlocking of the mutex. I'm pretty sure that behavior was wrong and this patch at least improved it. Thread synchronization has been robust since then, so I'm gonna merge.