APMD-List:
Archives

  From: Jeffrey W. Baker <jwbaker@acm.org>
  To  : <craigm@milkyway.gsfc.nasa.gov>
  Date: Tue, 25 Apr 2000 17:08:47 -0700 (PDT)

Re: apmd not triggering AC online/offline

On Mon, 24 Apr 2000, Craig Markwardt wrote:

> 
> Jeffrey W. Baker writes:
>  > > It is conceivable that the BIOS just isn't sending those events.
>  > > Which would be wierd since SONY/Phoenix obviously know how to do APM
>  > > already: my 505G works fine.
>  > 
>  > Yes, and since the kernel is obviously aware that the power has been
>  > removed, I don't think it is a kernel problem.
> 
> Yes, but these could be two separate things.  Every time you do a cat
> /proc/apm the kernel module queries the BIOS for the power state.
> Power_change events are separately sent by the BIOS to the kernel.  In
> principle the kernel could be gobbling these up before they get out to
> apmd in userland.
> 
>  > > The ultimate test is to enable debugging in the kernel module, but
>  > > that requires a recompile.
>  > 
>  > What do I define to enable that?
> 
> I think you have to edit the source directly.  There's an #undef
> APM_DEBUG which needs to be changed to #define APM_DEBUG.  You will
> get a lot of kernel messages this way, so it's not good for general
> use.
> 
> But I agree, I think it's likely a BIOS problem.

I enabled debugging, and the kernel is not receiving the power status
events.  *sigh*.  Is it possible to setup apmd so that it periodically
polls for changed power status?  I launched it with "apmd  -c 1" but that
does not have any effect.  Wouldn't be difficult to modify the program to
do this, I can write a shell script with the equivalent functionality in a
few seconds.

Cheers,
Jeffrey

P.S.  The power change event does get triggered when apmd starts up.