RTEMS4.7 PPC clock(was Re: IRQ latency, mvme5500 v1.3 )

[Kate Feng]

Sergei Organov wrote:

> Kate Feng <feng1 at bnl.gov> writes:
>
> > I  agree with what Till Straumann suggested for the IRQ handling
> > and prioritization of the RTEMS-MVME5500 BSP.  Thanks again.
> >
> > However, for the PPC shared file
> > c/src/lib/libcpu/powerpc/mpc6xx/clock/c_clock.c
> > in RTEMS4.7CVS,  can one  please replace the following:
> >
> [...]
> > with the same implementation as RTEMS4.6.0 shown below
> > until a better algorithm proposed ?
> > -------------------------------------------------------------
> > void clockIsr(void *unused)
> > {
> >   PPC_Set_decrementer( Clock_Decrementer_value );
> >
> >   Clock_driver_ticks += 1;
> >
> >   rtems_clock_tick();
> > }
> >
> > --------------------------------------------------------------------
> > The reason ?
> >
> > First, I do not think it's a good idea to disable interrupts
> > around the "asm volatile" code either.
>
> Why? There is absolutely no harm in disabling of interrupts for such a
> tiny period of time.

OK, I agree if it is needed  Could one  put in the interrupt  disable
around the assembly code in the 4.7 CVS  so that the inaccurance
is  not so  noticable, please ?  0.2 second out of 500 second is way
too much, which  at least I did'nt notice  in  4.6.0 yet.

Thanks,
Kate