Re: [PATCH 0/6] Add nRF51 DETECT signal with test

[Peter Maydell]

On Thu, 27 Jul 2023 at 23:51, Chris Laplante <chris@laplante.io> wrote:
>
> Hi Peter,
>
> > > 2. I also have some implementations for pieces of CLOCK, namely the 
> > > HFCLKSTART/HFCLKSTOP events and HFCLKSTARTED event. Should I include that 
> > > in this patch series, or would you prefer it in a separate series? It is 
> > > unrelated to DETECT and POWER.
> >
> >
> > If you think they're ready to go in, and it doesn't
> > make the series more than about 12-15 patches long,
> > you can put them on the end of the series. If the
> > patchset is starting to get a bit big then it might
> > be easier to get the POWER/DETECT parts reviewed
> > first.
>
> Unrelated question regarding the CLOCK module. Should I model the startup 
> times for the various crystal oscillators? Or should I just assume they start 
> instantly for simplicity? External xtal startup time is 750-800 us. Internal 
> RC startup time is 4-5 us. I've already modeled the delay for the external 
> xtal, but just wondering if its worth the extra code.

We typically just have that sort of thing start instantly,
unless there's some specific guest workload that falls
over if you don't model the startup delay. Usually
modelling the delay is unnecessary complexity.

thanks
-- PMM