BeagleBone questions (was Re: Towards an ARM system-building script)

[Ian Lepore]

On Tue, 2012-10-30 at 18:42 +0800, Alie Tan wrote:
> On Tue, Oct 30, 2012 at 6:40 AM, Ian Lepore
> <freebsd at damnhippie.dyndns.org>wrote:
> 
> > On Mon, 2012-10-29 at 09:14 -0700, Adrian Chadd wrote:
> > > On 29 October 2012 08:08, Alie Tan <alie at affle.com> wrote:
> > > > I have tried angstrom, ArchLinux and FreBSD. Only FreeBSD makes the
> > > > chip/board quite hot.
> > > >
> > > > Anyone have same experiece?
> > >
> > > Are we not halting the CPU during idle ticks on ARM?
> > >
> > > Does Linux have some extra power management code for ARM that we don't?
> >
> > I'm not sure all arm SoCs we support have the ability to do this well
> > (or at all).  I know for sure the atmel chips can do it, and we don't
> > use that abillity.  We're pretty poor on power management in general in
> > the atmel world.  I mean poor, like we turn on all on-chip devices early
> > in kernel init and leave them that way, even the ones we don't have
> > drivers for.  A comment in the code labels that a "temporary hack" but
> > it's been that way for 5 years.
> >
> > More directly on the OP's point, I wonder if the combo of the system not
> > fully booting and too much heat imply that something has gone into a
> > tight loop somewhere.
> >
> Is there any way to debug/detect this issue?
> I cant see any heavy processes from 'top'
> 

Well if the system is running well enough to use top, and it isn't
showing any process or kernel thread using a ton of cpu, then I think
that probably invalidates my theory.

One of my arm eval boards has an LCD interface that I don't use, and a
note with the system mentioned that there's a part on the board that
will heat up if you don't connect a display; they recommend unsoldering
it.  The part is some sort of choke or coil (it's silkscreened as
"L11").  On my board it gets warm, not really hot, so I haven't bothered
to remove it.

-- Ian