[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-apmd
Subject:    Re: APM/ACPI merger?
From:       Craig Markwardt <craigm () pcasun3 ! gsfc ! nasa ! gov>
Date:       2000-02-08 1:40:21
[Download RAW message or body]


Greetings Andy--

I just read your message (Subject [PATCH] ACPI) about merging the APM
and ACPI drivers.  I am an occassional contributor to the APM
utilities so when I read the following I was a little bit surprised.

 > The only APM events that any driver is actually
 > handling are suspend and resume and
 > user/sys/critical are always handled as one.
 > So I believe I can safely map APM_XXX_SUSPEND to
 > PM_SUSPEND and APM_XXX_RESUME to PM_RESUME
 > and just discard all of the other APM events.
 > Or do you forsee the need to propagate any
 > additional APM events (eg. APM_LOW_BATTERY)?

This might bear mentioning to Stephen (Rothwell).  Doesn't look like
this message appeared on linux-kernel, the apmd list, or Rothwell's
bug tracking.  [ Is linux-laptop archived anywhere? ]  As for other
dependencies, I recall that PCMCIA depends on APM events in the kernel
modules, so perhaps David Hinds might be consulted too?

In any case, I can say that some *user* level programs depend on a few
of the "esoteric" events.  The APM_LOW_BATTERY event is a bad example,
because it's pretty useless, but the CAPABILITY_CHANGE and
POWER_STATUS_CHANGE events are very useful to detecting battery
changes, and when the AC power is unplugged.  The USER vs SYS types
could also be useful in determining policy, since SYS are time-out
events while USER are user-initiated.

Thanks for your consideration,

Craig

[prev in list] [next in list] [prev in thread] [next in thread]