Double button events?

Hi!

I've just gotten ACPI suspend-to-mem to work on my Dell Inspiron 8200, using kernel 2.6.11. I'm actually very happy about it, because I've been stuck on 2.4 until yesterday :-D

One confusing thing is a "double event" from acpi when I press the suspend button (Fn+Esc).

 [Wed Apr 13 12:49:19 2005] received event "button/sleep SBTN 00000080
000000c6"
[Wed Apr 13 12:49:19 2005] notifying client 9917[1000:1000]
[Wed Apr 13 12:49:19 2005] executing action
"/etc/acpi/actions/sleepbtn.sh button/sleep SBTN 00000080 000000c6"[Wed
Apr 13 12:49:19 2005] BEGIN HANDLER MESSAGES
[Wed Apr 13 13:49:25 2005] END HANDLER MESSAGES
[Wed Apr 13 13:49:25 2005] action exited with status 0
[Wed Apr 13 13:49:25 2005] completed event "button/sleep SBTN 00000080
000000c6"
[Wed Apr 13 13:49:25 2005] received event "button/sleep SBTN 00000080
000000c7"
[Wed Apr 13 13:49:25 2005] notifying client 9917[1000:1000]
[Wed Apr 13 13:49:25 2005] executing action
"/etc/acpi/actions/sleepbtn.sh button/sleep SBTN 00000080 000000c7"[Wed
Apr 13 13:49:25 2005] BEGIN HANDLER MESSAGES
[Wed Apr 13 13:49:25 2005] END HANDLER MESSAGES
[Wed Apr 13 13:49:25 2005] action exited with status 0
[Wed Apr 13 13:49:25 2005] completed event "button/sleep SBTN 00000080

000000c7"

The result is a suspend immediately after resuming. I've solved it temporarily by putting:

seqno=$4

parity=`(echo "ibase=16"; echo -n $seqno |\
           tr '[a-z]' '[A-Z]'; echo " % 2")|bc`

if [ $parity = 0 ]; then

# Do it!
echo mem > /sys/power/state
fi

in my sleepbtn.sh, but it's such a hack :-) Something must be doing something wrong, right? Is this a kernel bug, BIOS bug or what? I'd love to see a better solution.

/Mikael

Plus ça change, plus c'est la même chose

begin:vcard
fn:Mikael Nilsson
n:Nilsson;Mikael
org:Royal Institute of technology;KMR group, NADA adr;dom:;;;Stockholm email;internet:[email protected] tel;work:+46 8 7906896
x-mozilla-html:FALSE
url:http://kmr.nada.kth.se
version:2.1
end:vcard