[Bug 1914] Make timestamp implementation choice available for the CPU port

[bugzilla-daemon at rtems.org]

https://www.rtems.org/bugzilla/show_bug.cgi?id=1914

--- Comment #4 from Joel Sherrill <joel.sherrill at oarcorp.com> 2011-09-16 08:00:18 CDT ---
(In reply to comment #3)
> If we use __RTEMS_USE_TICKS_FOR_STATISTICS__ do we avoid the timestamps
> completely?  I think this would be an option for systems with very limited
> resources.

No.  If you use ticks for statistics, then you are changing the math for
updating and reporting statistics.  This primarily changes the cost for the 
statistics on a dispatch (around line 121). [1] There are two big differences
in the two cases:  data representation and calling the nanoseconds since tick
handler in the BSP (only not ticks).

You still have uptime and current time which are maintained as timestamps.

SCORE_EXTERN Timestamp_Control _TOD_Now;
SCORE_EXTERN Timestamp_Control _TOD_Uptime;

So the ticks statistics will speed up dispatching but won't completely
eliminate the use of timestamps.

[1] TOD_Get_uptime in ticks path NOT _TOD_Get_uptime.  It doesn't look to me
like the ticks path is right at the moment.

-- 
Configure bugmail: https://www.rtems.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.