On 07.11.15 06:10, Bill Fischofer wrote:
A reportable statistic for this situation would be appropriate, but this 
certainly should not result in an abort.  Ticks can be lost for all sorts of 
reasons and should be considered unusual but hardly fatal conditions.  Aborts 
are reserved for serious internal errors or detected inconsistencies that 
compromise the integrity of the system.

It's not simple statistic, it's indicator that actual resolution in the system 
is worse
than set in the timer by the application and allows to set appropriate ... 
later.
It depends on load on system, jiffy, size of the handler and is only way to 
draw attention on it.
But abort is not correct, I agree. I'll send new patch with warn instead of 
abort.


On Fri, Nov 6, 2015 at 8:05 AM, Ivan Khoronzhuk <[email protected] 
<mailto:[email protected]>> wrote:

    On 31.10.15 00:56, Ivan Khoronzhuk wrote:

        No need to continue test example if tick is lost. It means that
        actual timer resolution is worse than expected. The default timer
        resolution is set 10ms, that is equal to jiffy on most systems.
        The default resolution is set bearing in mind that on a CPU runs
        maximum two threads that ideally fits in 10ms. But user can change
        it to be smaller, in case if CPU0 is isolated and handles the timer
        ticks. This patch saves user from incorrect choice by aborting
        the application.

        Signed-off-by: Ivan Khoronzhuk <[email protected] 
<mailto:[email protected]>>
        ---
           platform/linux-generic/odp_timer.c | 7 +++++++
           1 file changed, 7 insertions(+)

        diff --git a/platform/linux-generic/odp_timer.c 
b/platform/linux-generic/odp_timer.c
        index e8f0267..3728ac3 100644
        --- a/platform/linux-generic/odp_timer.c
        +++ b/platform/linux-generic/odp_timer.c
        @@ -632,7 +632,14 @@ static unsigned 
odp_timer_pool_expire(odp_timer_pool_t tpid, uint64_t tick)

           static void timer_notify(sigval_t sigval)
           {
        +       uint64_t overrun;
                 odp_timer_pool *tp = (odp_timer_pool *)sigval.sival_ptr;
        +
        +       overrun = timer_getoverrun(tp->timerid);
        +       if (overrun)
        +               ODP_ABORT("Error: need to increase resolution, lost tickets: 
%"
        +                         PRIu64"\n", overrun);


    Maybe here should be warn instead of abort?

        +
           #ifdef __ARM_ARCH
                 odp_timer *array = &tp->timers[0];
                 uint32_t i;


    --
    Regards,
    Ivan Khoronzhuk

    _______________________________________________
    lng-odp mailing list
    [email protected] <mailto:[email protected]>
    https://lists.linaro.org/mailman/listinfo/lng-odp



--
Regards,
Ivan Khoronzhuk
_______________________________________________
lng-odp mailing list
[email protected]
https://lists.linaro.org/mailman/listinfo/lng-odp

Reply via email to