OS_TaskDelay implementation wrt errors not consistent

[skliper]

Is your feature request related to a problem? Please describe.
In the RTEMS implementation if OS_Milli2Ticks fails (over max int) it'll wait 10 ticks, although comment is somewhat misleading since 10 isn't a minimum... passing in 0 would also cause no wait:

osal/src/os/rtems/src/os-impl-tasks.c

Lines 212 to 237 in cb95655

	int32 OS_TaskDelay_Impl(uint32 milli_second)
	{
	int tick_count;
	int32 return_code;
	return_code = OS_Milli2Ticks(milli_second, &tick_count);
	if (return_code != OS_SUCCESS)
	{
	/*
	* always want to do some form of delay, because if
	* this function becomes a no-op then this might create a
	* tight loop that doesn't ever yield the CPU - effectively
	* locking the system in an RTOS environment.
	*/
	tick_count = 10;
	}
	/*
	** Always successful ( from RTEMS docs )
	*/
	rtems_task_wake_after((rtems_interval)tick_count);
	return (return_code);
	} /* end OS_TaskDelay_Impl */

VxWorks doesn't wait a minimum if OS_Milli2Ticks fails, just returns an error:

osal/src/os/vxworks/src/os-impl-tasks.c

Lines 325 to 343 in cb95655

	int32 OS_TaskDelay_Impl(uint32 milli_second)
	{
	/* msecs rounded to the closest system tick count */
	int sys_ticks;
	/* Convert to ticks if possible */
	if (OS_Milli2Ticks(milli_second, &sys_ticks) != OS_SUCCESS)
	{
	return OS_ERROR;
	}
	/* if successful, the execution of task will pend here until delay finishes */
	if (taskDelay(sys_ticks) != OK)
	{
	return OS_ERROR;
	}
	return OS_SUCCESS;
	} /* end OS_TaskDelay_Impl */

POSIX just converts to timespec and uses clock_nanosleep.

Describe the solution you'd like
Technically OS_TaskDelay should be functionally tested with a 0 and max value input, not that I'd ever want to wait that long (although RTEMS and VxWorks should return quickly due to the OS_Milli2Ticks error response). It would be nice if behavior was consistent, although for 0 it looks to me like none of them would yield. If we really want a yield in all cases then it needs an update.

Describe alternatives you've considered
None.

Additional context
From functional test scrub.

Requester Info
Jacob Hageman - NASA/GSFC