BSP benchmark timer to be replaced with RTEMS counter support?

[Sebastian Huber]

On 23/03/18 12:47, Amaan Cheval wrote:
> The BSP porting guide[1] includes this warning when talking about the
> benchmark timer driver:
>
>> The Timer Driver is superfluous and should be replaced by the RTEMS
> counter support. Ask on the mailing list if you plan to write a Timer
> Driver.
>
> We also have this[2] open ticket about the same issue. It seems like this
> might be a part of the BSP's clock driver, exposed through the
> "_CPU_Counter_read" function. Is that right?
>
> The tests in ./testsuites/tmtests/ do call several benchmark_timer_*
> functions, so I imagine the warning means that we can implement these
> functions using the counter instead of reading the hardware timer by itself
> as the guide shows?

For new BSPs, please implement the CPU counter and use

c/src/lib/libbsp/shared/timercpucounter.c

for the timer driver.

>
> (For context: this is relevant to my GSoC plan for the x86_64 port and BSP;
> I'd like to architect it to be of a style which the RTEMS community deems
> ideal, despite other BSPs not quite being there yet, understandably.)
>
> [1] https://docs.rtems.org/branches/master/bsp-howto/timer.html
> [2] https://devel.rtems.org/ticket/2847
> _______________________________________________
> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel

-- 
Sebastian Huber, embedded brains GmbH

Address : Dornierstr. 4, D-82178 Puchheim, Germany
Phone   : +49 89 189 47 41-16
Fax     : +49 89 189 47 41-09
E-Mail  : sebastian.huber at embedded-brains.de
PGP     : Public key available on request.

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.