On Tue, 21 Apr 2020 11:20:17 -0700
Tony Lindgren <tony@xxxxxxxxxxx> wrote:
> * H. Nikolaus Schaller <hns@xxxxxxxxxxxxx> [200421 18:14]:
> > > Am 21.04.2020 um 20:02 schrieb Tony Lindgren <tony@xxxxxxxxxxx>:
> > > This is 37xx though, maybe you have 35xx and there's some errata
> > > that we're not handling?
> >
> > No, it is dm3730 on three different units I have tried.
> >
> > > I'm only seeing "2.7. HDQTM/1-Wire® Communication Constraints"
> > > for external pull-up resitor in 34xx errata at [0].
> > >
> > > I wonder if wrong external pull could cause flakyeness after
> > > enabling the hdq module?
> >
> > I have checked and we have 10 kOhm pullup to 1.8 V and a 470 Ohm
> > series resistor.
>
> OK
>
> > > If nothing else helps, you could try to block idle for hdq
> > > module, but I have a feeling that's a workaround for something
> > > else.
> >
> > Well, what helps is reverting the patch and using the old driver
> > (which did work for several years). So I would not assume that
> > there is a hardware influence. It seems to be something the new
> > driver is doing differently.
>
> Well earlier hdq1w.c did not idle, now it does. If you just want
> to keep it enabled like earlier, you can just add something like:
>
> &hdqw1w {
> ti,no-idle;
> };
>
> > I need more time to understand and trace this issue on what it
> > depends... It may depend on the sequence some other modules are
> > loaded and what the user-space (udevd) is doing in the meantime.
>
> Yes would be good to understand what goes wrong here before we
> apply the ti,no-idle as that will block SoC deeper idle states.
>
hmm, he is testing without idling uarts, so I am a bit confused here,
the problem only seems to occur when more things are *active*.
Is something not handled in time.
Regards,
Andreas
