Oh no, it was 6.3.2. Sorry.

> I've got a high priority thread which has strict timing requirements. That 
> thread is monitored by even higher priority thread which is to crash 
> application if first thread does not respond within the required time interval
> .
> Just few minutes again it crashed and first thread is calling ThreadCtl(
> _NTO_TCTL_IO, 0).
> I'm still using 6.3.0SP3, so I can't say what would be with latest release and
> , of course it has not necessarily spent all the time slice in call to 
> ThreadCtl() but could not by occasion some priority inversion point be there 
> in interior of ThreadCtl()?