HDAPS

Kay Sievers kay.sievers at vrfy.org
Mon Oct 2 12:10:49 PDT 2006


On Mon, 2006-10-02 at 19:52 +0100, Richard Hughes wrote:
> On Mon, 2006-10-02 at 13:34 +0100, Bastien Nocera wrote:
> > On Mon, 2006-10-02 at 12:35 +0100, Richard Hughes wrote:
> > <snip>
> > > IMO, the kernel is the place where over-temperature is actioned, so it
> > > probably makes sense for the "laptop-is-falling" code to be there too.
> > > But then userspace can do funcy things and set policy. Opinions
> > > welcomed.
> > 
> > Overheating isn't as time-critical as falling down. Falling down to
> > possible failure < 1 sec. Overheating to possible failure ~= minutes?
> > 
> > And there isn't much of a policy either. It should park the heads, and
> > tell userspace about it.
> 
> Yes, you are probably correct. I'll start a thread on LKML.

It's a bit more complicated, you need to detect movements like you see
in a car while driving a rough road, and disable the automatic head
parking. That's some kind of policy - not sure if you need to be able to
tweak that in the kernel, or communicate with userspace for that, but
it's at least not too trivial.

Kay



More information about the hal mailing list