[OLPC-AU] lid close while shutting down

Paul Fox pgf at laptop.org
Tue Feb 14 11:09:32 EST 2012

sridhar wrote:
 > On 14 February 2012 22:09, James Cameron <quozl at laptop.org> wrote:
 > > For example, because of an episode of damage that happened when a stack
 > > of laptops was left on and their lid sensors interacted, and because of
 > > the use of early prototypes with an incomplete heat spreader, no doubt a
 > > myth has developed that the thermal design of the laptop is bad.
 > Not a myth, just a concern and a need on my part to understand the
 > implementation better.
 > > You can verify from power logs whether the users have properly shutdown
 > > their laptops.  Please ensure this logging works well, as it will give
 > > you good data.  Obtain the logs whenever there is a report of laptop
 > > thermal issues.  Require that the logging is implemented as part of any
 > > remedial action.
 > Agreed. We are working on a means to more easily collect this
 > information from the field.
 > > I don't think a powered-on XO gets very hot at all if the lid is shut.
 > > It merely grows slightly warmer.
 > What about if it's also charging in an XOP rack?

in that case i think this part of james' reply is the pertinent part:
 > There is certainly possibility for damage if the laptop is ...
 > stacked with other laptops without being shutdown.

a powered-on laptop with a closed lid might get warm.  but it's not an
issue unless there are a stack (or a rack) of such laptops.  that's
why we've gone to some effort to ensure the system stays off when the
lid switch is triggered (even though that's a very convenient feature --
i can tell you it hurt to disable it.)

as for closing the lid after commencing shutdown?  once powerd
commences the shutdown sequence, it won't return to its main loop
before exiting or being killed, so there's little chance of it causing
a suspend due to the closed lid.  if the shutdown process got hung
somewhere, i suppose it's possible that powerd might be respawned
(i'll check on that), and run far enough to cause a suspend -- but i'm
skeptical.  (at that point the system was already in trouble, so
arguably suspending would be better than not, anyway.)

 paul fox, pgf at laptop.org

More information about the Devel mailing list