Thank you Christophe! This is exactly the same which I was looking for!

After further analyze I got answer to several my questions, including
responsibility of Cyg_RealTimeClock class.

But there is one more question, where I'm having little difficulty.
That is: after h/w tick interrupt - first ISR of that interrupt then
DSR is called. And as you told:

> The tick value is incremented from the DSR:
>
> void Cyg_RealTimeClock::dsr(cyg_vector vector, cyg_ucount32 count, 
> CYG_ADDRWORD data)
>    rtc->tick( count );

and in tick( cyg_uint32 ticks ) function there is loop
(\kernel\current\src\common\clock.cxx) :

while( ticks-- )
{
...
}

where *ticks* is the value provided by call from DSR. So, my confusion
is what is value of this *ticks* and why we need decrement it in
*while* loop?

There is comment in source code:

// Increment the counter in a loop so we process
// each tick separately. This is easier than trying
// to cope with a range of increments.

but, couldn't get any idea.


2009/7/19 Christophe Coutand <[email protected]>:
> I believe what you are looking for is located in:
>
> kernel\current\src\common\clock.cxx
>
> The clock is created here and attached to the HW interrupt:
>
> Cyg_RealTimeClock::Cyg_RealTimeClock()
>    : Cyg_Clock(rtc_resolution),
>      interrupt(CYGNUM_HAL_INTERRUPT_RTC,
>                CYGNUM_KERNEL_COUNTERS_CLOCK_ISR_PRIORITY,
>                (CYG_ADDRWORD)this, isr, dsr)
> {
>    CYG_REPORT_FUNCTION();
>    HAL_CLOCK_INITIALIZE( CYGNUM_KERNEL_COUNTERS_RTC_PERIOD);
>    interrupt.attach();
>    interrupt.unmask_interrupt(CYGNUM_HAL_INTERRUPT_RTC);
>    Cyg_Clock::real_time_clock = this;
> }
>
> The tick value is incremented from the DSR:
>
> void Cyg_RealTimeClock::dsr(cyg_vector vector, cyg_ucount32 count, 
> CYG_ADDRWORD data)
>    rtc->tick( count );
>
> The period of the RTC clock is configurable from your .ecc file.
>
> Christophe
>

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Reply via email to