[ltp] [PATCH] New ibm-acpi patches for extended thermal sensor functionality

[Alex Deucher]

On 9/16/06, Henrique de Moraes Holschuh <hmh@hmh.eng.br> wrote:
> On Sat, 16 Sep 2006, Stefan Schmidt wrote:
> > Do you think about taking over the ibm-acpi maintainership? It seema
> > as Borislav has no more time for this code.
>
> I need to learn a lot more about writing code for the kernel before I'd even
> consider asking Borislav to take over ibm-acpi.
>
> I can certainly help with ibm-acpi, though.  And even if Borislav has little
> time for the code, as long as patches sent upstream still get accepted (and
> someone must be accepting them, 2.6.18 has silently added experimental
> support for enabling/disabling WAN cards, without documenting it).
>
> > It would be good to fetch all the patches flying around on this list
> > and the wiki and submit them into main. Here are my small list with
>
> Sure, one thing at a time.  Fan control goes next, but since there *are*
> firmware bugs at play with fan control, that one needs more thought.
>
> Which firmware bugs?  Well, the one I know for sure is that my T43 boots in
> auto mode, but reports it is on manual level 7 until you give it a boot to
> the head (like a S3 suspend).  Now imagine we wanted to handle persistence
> across sleep, and stored current state trusting it to be correct?
>
> > 1. Get rid of the experimental status of some features. See my patch
> > on this ml.
>
> But keep ecdump experimental.  Reading EC registers can have side-effects,
> some registers are likely to be reset-on-read.  The EC has some
> functionality that behaves this way, like HDAPS keyboard and mouse activity
> detection, it is only sane to expect that stuff like that might be in the EC
> register space too.
>
> Writing blindly to EC space is *really* dangerous, too.
>
> > 3. Integrate fan control. Patches on the wiki.
>
> They need some polishing yet, IMHO.  And a signed-off from the first author.
>
> > 4. Get rid of the of the docking feature and push the genric docking
> > driver in main instead.
>
> I can't help with that, unfortunately.  I don't own a dock.  But I believe
> the generic docking driver IS in mainline already?

We can probably just remove the docking hooks from ibm_acpi since the
generic code will handle it.

Alex

>
> --
>   "One disk to rule them all, One disk to find them. One disk to bring
>   them all and in the darkness grind them. In the Land of Redmond
>   where the shadows lie." -- The Silicon Valley Tarot
>   Henrique Holschuh