Arch Linux

kludge

Member

Registered: 2008-08-03

Posts: 294

Y.A.L.F.A.P. -- Yet Another Laptop's eFfing ACPI Problems

greets!

so i finally got a battery for my thinkpad a21m.  visions of listening
to streaming talk radio... in the kitchen... *without* *rebooting*
danced through my head.

of course, now i want to get power management going.  after a long,
hairy goose-chase--the details of which i will spare you, dear
readers--i discovered that the "standard acpi modules" that udev loads
at boot up *should* give me the basics, a sleep button, a power button,
and battery and ac adapter events.  from there, i can go ahead and set
the rest up.  in fact, i surprised myself something fierce when i hit
fn+f4 and suddenly, my laptop just shutdown!  oh wait, no.  sleep
actually "just worked!"

except that something strange is going on here:

Oct 28 00:31:59 laura logger: SleepButton pressed: SLPB 00000001
Oct 28 00:32:01 laura logger: PowerButton pressed: PWRF 00000001

the buttons work fine...

Oct 28 00:32:15 laura logger: Running on Battery Power: AC 00000000
Oct 28 00:32:16 laura logger: Battery: BAT0 00000001

unplugging the ac adapter sends the appropriate events and "cat
/proc/acpi/ac_adapter/AC/state" gives:

state: off-line

and then...

Oct 28 00:33:34 laura logger: PowerButton pressed: PWRF 00000002

i press the power button just to watch the pretty events scroll by in
/var/log/everything.log and everything looks ok... except that what you
can't see in these listings is that in between 00:32:16 and 00:33:34, i
plugged the ac adapter back in and hit the sleep button (fn+f4) a
handful of times.  "cat /proc/acpi/ac_adapter/AC/state" now shows:

state: on-line

but the event never got sent.

this is a reliably repeatable, although the BAT0 event occasionally
happens twice in a row.  and not to long ago (while i was cooking dinner
and listening to a talk radio stream while running on the battery),
another BAT0 00000001 event showed up and everything started working as
expected.  i have no idea why or how... i was stirring the sauce at the
time.

there are no error messages; {ac,button,battery}.ko don't have debug
parameters, google gives me nothing.

i'm seriously considering hacking at my dsdt... but i'd really rather
not.  can anyone suggest next steps?

many thanks,
kludge

p.s.: when this is figgered out, i'll probably be back, whining about thinkpad-acpi.

p.p.s.: does *anyone* else out there rock the a21m?  google makes it seem like the forgotten step-child of the ibm thinkpad family.

---

[23:00:16]    dr_kludge | i want to invent an olfactory human-computer interface, integrate it into the web standards, then produce my own forked browser.
[23:00:32]    dr_kludge | can you guess what i'd call it?
[23:01:16]    dr_kludge | nosilla.
[23:01:32]    dr_kludge | i really should be going to bed.  i'm giggling madly about that.