There is code in _syspage_time() (called by clock_gettime() ) sepecifically
to check for clock interrupts disturbing the clock read. It loops until it
gets a stable reading.

So I'd say it's thread safe.

-ad

Dieter Bittner wrote:
> Hy
> 
> since some weeks I have a problem with getting the Realtimeclock values 
> by "clock_gettime". Sometimes I gets a timevalue back which is not a 
> correct one.
> 
> The system on which the software is running is a X86 plattform with a 
> dualcore prozessor with multiple threads. Cause of timing issues it 
> would be possible that
> 
> these calls ->
> 
> coBindCpu_m.vfuBindCpu(0);
> lRetVal_l = clock_gettime( CLOCK_REALTIME, &tActTimeSpec_l);
> coBindCpu_m.vfuRestoreCpu();
> 
> is done in parallel.
> 
> Is it possible that the function clock_gettime is not threadsafe or that 
> it has the same problem like fopen, fclose ... in a multithreaded 
> environment?
> 
> Best regards
> Dieter
> 
> _______________________________________________
> OSTech
> http://community.qnx.com/sf/go/post7700
>