2.6.25-rc3 (g7704a8b) - freezer for suspend broken?
Hi All
Unstable Debian here, most packages last upgraded about 2 months ago.
On a Powerbook5,8
With git sources pulled last Friday: The kernel compiled nicely - it
even booted (I don't take that for granted, with fresh sources and my
- at best - incomplete knowledge, especially when it comes to the new
kernel switches .. :) ..
Here's the .config:
http://www.geocities.com/wolfgangpfeiffer/config.2.6.25-rc3-g7704a8b.txt
.. but sadly, with 'CONFIG_SUSPEND_FREEZER=y' the machine can be set
into sleep mode (sleep LED blinking like I was used to it with
2.6.18).
But instead of waking up when pressing the power button, the
system/machine simply was shut down. Logs that I found on it so far in
/var/log - the excerpt below should be one created the moment I
pressed the power button to get the machine into sleep mode, or the
moment I tried to resume from sleep ... not being sure:
------------------------------------
Mar 1 00:22:25 debby1-6 pmcs-bdflush: External laptop-mode script found - exiti
ng.
Mar 1 00:22:25 debby1-6 pbbuttonsd: INFO: Script '/etc/power/pmcs-pbbuttonsd su
spend ac ram' lauched but exitcode is not null
Mar 1 00:22:26 debby1-6 /usr/sbin/gpm[4463]: oops() invoked from gpn.c(205)
Mar 1 00:22:26 debby1-6 /usr/sbin/gpm[4463]: /var/run/gpm.pid: No such file or directory
Mar 1 00:22:26 debby1-6 pmcs-bdflush: External laptop-mode script found - exiting.
Mar 1 00:22:26 debby1-6 pbbuttonsd: INFO: Script '/etc/power/pmcs-pbbuttonsd suspend ac ram' lauched but exitcode is not null
Jan 1 01:01:16 debby1-6 syslogd 1.5.0#1: restart.
Jan 1 01:01:16 debby1-6 kernel: klogd 1.5.0#1, log source = /proc/kmsg started.
Jan 1 01:01:16 debby1-6 kernel: [ 0.000000] Crash kernel location must be 0x2000000
Jan 1 01:01:16 debby1-6 kernel: [ 0.000000] Reserving 0MB of memory at 32MB for crashkernel (System RAM: 1024MB)
Jan 1 01:01:16 debby1-6 kernel: [ 0.000000] Using PowerMac machine description
Jan 1 01:01:16 debby1-6 kernel: [ 0.000000] Total memory = 1024MB; using 2048kB for hash table (at cfe00000)
------------------------------------
The "Jan 1" snippet above for the time of the next reboot, as it seems
... :)
As for the gpm part: gpm is often shut down ... seems to be related to
experimental X packages (about 2 months old) .. IIRC gpm crashes at
times when I try to clone the machine LCD to some external VGA ... But
I didn't try to mirror the screen in the instance above ...
And this is how /etc/power/pmcs-pbbuttonsd is looking here:
----------------------------------------------
#!/bin/sh
#
# This script is invoked by pbbuttonsd to configure the system for a
# given power level. The script gets three arguments:
# The first and the the third argument belong together. The first
# argument is a command and the third a corresponding argument. Not
# each of the commands have a corresponding argument. The following
# table shows possible combinations:
#
# $1 $3
# 'powersave' | power policies transfered to slave
# 'custom' | scripts
# 'performance' |
# 'suspend' 'ram' prepare for suspend to RAM (sleep)
# 'disk' prepare for suspend to disk
# 'shutdown' prepare for system shutdown
# 'resume' 'ram' after wakeup from suspend to RAM
# 'disk' theoretically, not used yet
# 'emergency' battery is critically low -> shutdown.
# 'shutdown' user initiated a system shutdown.
# 'cover-open' 'open' cover has been opened
# 'cover-close' 'close' cover has been closed
#
# 'cover-open' and 'cover-closed' will only be called if no other
# suspend script is going to be called or sleep is not supported
# on this machine. The argument for this command is a future
# investment.
#
# The second argument contains the current powersource of the laptop
# 'ac'
# 'battery'
#
# The commands 'emergency' and 'shutdown' are handled directly. All
# other commands will be transfered to the slave scripts which hopefully
# will do the work.
PATH=/bin:/sbin:/usr/bin:/usr/sbin
# Logging is done by the pbbuttonsd daemon.
case "$1" in
emergency)
shutdown -h now "Low battery - system will go down now!"
;;
shutdown)
shutdown -h now "User requested shutdown - system will go down now!"
;;
*)
cd `dirname $0`
PATH=$PATH:$PWD
[ -d ${1}.d ] && run-parts --arg="$1" --arg="$2" --arg="$3" ${1}.d
run-parts --arg="$1" --arg="$2" --arg="$3" event.d
;;
esac
----------------------------------------------
So before I try to find out my mistakes (broken or even missing
userland packages that might be required for 2.6.25, etc. etc. ..):
*** 1:
Anyone out there who got sleep working with 2.6.25, on this machine:
$ cat /proc/cpuinfo
processor : 0
cpu : 7447A, altivec supported
clock : 833.333000MHz
revision : 0.5 (pvr 8003 0105)
bogomips : 16.57
timebase : 8320000
platform : PowerMac
machine : PowerBook5,8
motherboard : PowerBook5,8 MacRISC3 Power Macintosh
detected as : 287 (PowerBook G4 15")
pmac flags : 00000019
L2 cache : 512K unified
pmac-generation : NewWorld
*** 2:
Should pbbuttonsd 0.7.9-2 still work with that fresh kernel?
Because with 2.6.18-rc4 (some git kernel vers.) it worked like a
charm, .i.e. no problems with sleep mode here, so far, with this 2.6.18
and pbbuttonsd installed)
Or are there different tools - other than pbbuttonsd - necessary for
that kernel?
I know there is uswsusp, but this tool seems to require an initrd
driven kernel. I don't have an initrd image installed, and if possible
I really would like to avoid such a kernel ...
###################################################
Notes ( Please ignore them if it's not related to the suspend issue
from above ..):
###################################################
*** 1:
fn-key settings seem to be changed:
With the old kernel I had to press fn-F(x) to change sound, brightness
etc. - now the same can be done by pressing F(x) without the fn key.
Vice-versa on mc previously I only had to press the F(x) keys to
access its menu - now I have to add an fn press for the same outcome
...
*** 2:
Hibernation broke my swap partition, as it seems: I did a test as
recommended in Documentation/power/basic-pm-debugging.txt, like (excerpt):
#############################################
# echo reboot > /sys/power/disk
# echo disk > /sys/power/state
and the system should create a hibernation image, reboot, resume and get back to
the command prompt where you have started the transition. If that happens,
hibernation is most likely to work correctly.
############################################
But the above zeroed my swap partition. On reboot, IIRC, I read
"unable to find swap-space signature", IIRC ... and nothing was
resumed. I just got the usual prompt to log in, IIRC ...
###############################
Any hints welcome ...
Thanks in anticipation
Best Regards
--
Wolfgang
http://heelsbroke.blogspot.com/
http://keyserver.mine.nu/pks/lookup?search=0xE3037113&fingerprint=on
Reply to: