Enhance VxWatchDogTimer overflow check

[kevin-f-ortega]

There are two overflow checks here https://github.com/fprime-community/fprime-vxworks/blob/33333635b6fd937ddc60a5d4be356e0f7ccf8cad/VxWorks/Svc/VxWatchDogTimer/VxWatchDogTimer.cpp#L38-L42. The checks use U64, but it's preferred not to use U64. Update the overflow check logic to do this instead

For https://github.com/fprime-community/fprime-vxworks/blob/devel/VxWorks/Svc/VxWatchDogTimer/VxWatchDogTimer.cpp#L39 do

FW_ASSERT( (U32::max() - milliseconds) >= usecsToMilliseconds);

And similar with the seconds in https://github.com/fprime-community/fprime-vxworks/blob/devel/VxWorks/Svc/VxWatchDogTimer/VxWatchDogTimer.cpp#L42

Remove https://github.com/fprime-community/fprime-vxworks/blob/devel/VxWorks/Svc/VxWatchDogTimer/VxWatchDogTimer.cpp#L38 and https://github.com/fprime-community/fprime-vxworks/blob/devel/VxWorks/Svc/VxWatchDogTimer/VxWatchDogTimer.cpp#L41C1-L41C96