suspend-to-disk: fan always on and misc kernel errors

DISCUSSION Andrea Borgia 3 participants 7 messages Mar 27, 2005 – Apr 5, 2005 5 messages not in the archive ACPI

Share & tools

Andrea Borgia

author

2005-03-27 17:42 UTC #0 lore raw

Hello.

Long time after first trying suspend-to-disk, I gave it a spin again with 2.6.11.6, pristine save for two powernow/timer patches.

There are some rough corners, but it appears to be mostly working, including X. The glitches are:

1) scheduling while atomic (see kernel log)
2) script terminates with error and spews "command not found" problems only in the post-wakeup section, even though full pathnames are used (see script)
3) fan is always on, even though load approaches unity after a while

I'm attaching the suspend script I am using and the kernel log and I would like to hear from other people who have similar problems on a similar hardware: my laptop is an Acer 1357LMi, mobo chipset is VIA KM400 and vga is an ATI Mobility M9+. DSDT and other info available on request.

BR,
Andrea.

Signature · 1 line

Homepage: http://andrea.borgia.bo.it     /    Amateur radio: IZ4FHT

#!/bin/bash
set -x

# Modules to unload/reload (do not use radeonfb!)
RELOAD="psmouse mousedev acerhk"
# Subsystems to stop/start
RESTART="hotplug powernowd gpm"

# Switch to console to reduce problems
/usr/bin/chvt 1

# Stopping subsystems
for S in $RESTART
do

	/etc/init.d/$S stop

done

# Unloading modules
for M in $RELOAD
do
� � � � if /bin/grep -q "^$M[[:blank:]]" /proc/modules � � � � then
� � � � � � � � /sbin/modprobe -r "$M"
� � � � fi
done

# Zzzzz
sync
echo disk > /sys/power/state
#sync

# Reloading modules...
for M in $RELOAD
do
� � � � if /bin/grep -q "^$M[[:blank:]]" /proc/modules � � � � then
� � � � � � � � /sbin/modprobe "$M"
� � � � fi
done

# Restarting subsystems...
for S in $RESTART
do
/etc/init.d/$S start
done

A message that is not in the archive (sent off-list or not yet received): 4246F06D.40005-iEixELS/[email protected]

Stefan Seyfried

2005-03-28 20:50 UTC #2 lore raw

Quoted 3 lines · Andrea Borgia

2) script terminates with error and spews "command not found" problems only in the post-wakeup section, even though full pathnames are used (see script)

you have some funny special characters in there, no wonder bash barfs on that.

Quoted 17 lines

# Unloading modules
for M in $RELOAD
do
? ? ? ? if /bin/grep -q "^$M[[:blank:]]" /proc/modules ? ? ? ? then
? ? ? ? ? ? ? ? /sbin/modprobe -r "$M"
? ? ? ? fi
done
# Reloading modules...
for M in $RELOAD
do
? ? ? ? if /bin/grep -q "^$M[[:blank:]]" /proc/modules ? ? ? ? then
? ? ? ? ? ? ? ? /sbin/modprobe "$M"
? ? ? ? fi
done

Signature · 9 lines

Stefan Seyfried



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

A message that is not in the archive (sent off-list or not yet received): 20050328205027.GB28568-l0tNAEGuAhhzZ8+rp42Dbp9+tswZ0GTaehPwdyo5hKaELgA04lAiVw@public.gmane.org

Andrea Borgia

author

2005-03-29 18:50 UTC #4 lore raw

Quoted 3 lines · Stefan Seyfried

Quoted 1 line
2) script terminates with error and spews "command not found" problems only in the post-wakeup section, even though full pathnames are used (see script)
you have some funny special characters in there, no wonder bash barfs on that.

Thanks, I have no idea how they got in the script, that partially fixed
it: no more "command not found" errors, but the kernel messages about "scheduling while atomic" are still there and the script still aborts with SIG11 after resuming from suspend.

Anyhow, I found out that the fan issue is related to powernowd not being restarted because of the script failure.

Any clues about that? I tried both platform and shutdown for disk, but the only difference is that platform reboots immediately instead of powering down. Also, writing "4b" to acpi/sleep is the same as "disk" to power/state.

TIA,
Andrea.

Signature · 9 lines

Homepage: http://andrea.borgia.bo.it     /    Amateur radio: IZ4FHT



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

A message that is not in the archive (sent off-list or not yet received): 4249A359.9060301-iEixELS/[email protected]

Stefan Seyfried

2005-04-01 22:19 UTC #6 lore raw

Quoted 6 lines · Andrea Borgia

Stefan Seyfried wrote:
Thanks, I have no idea how they got in the script, that partially fixed
it: no more "command not found" errors, but the kernel messages about "scheduling while atomic" are still there and the script still aborts with SIG11 after resuming from suspend.

| Mar 27 18:49:57 localhost kernel: note: swsusp-script[3902] exited with preempt_count 1

disable preempt.

Quoted 5 lines

Anyhow, I found out that the fan issue is related to powernowd not being restarted because of the script failure.
Any clues about that? I tried both platform and shutdown for disk, but the only difference is that platform reboots immediately instead of

interesting, but another bug...

Quoted 2 lines

powering down. Also, writing "4b" to acpi/sleep is the same as "disk" to power/state.

known, s4bios is not working for a long time now, we should remove it. "echo 4 > /proc/acpi/sleep" is supposed to do the same thing as "echo disk > /sys/power/state" on an acpi system.

Signature · 9 lines

Stefan Seyfried



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

A message that is not in the archive (sent off-list or not yet received): 20050401221940.GC1868-l0tNAEGuAhhzZ8+rp42Dbp9+tswZ0GTaehPwdyo5hKaELgA04lAiVw@public.gmane.org

Andrea Borgia

author

2005-04-03 11:05 UTC #8 lore raw

Quoted 2 lines · Stefan Seyfried

| Mar 27 18:49:57 localhost kernel: note: swsusp-script[3902] exited with preempt_count 1 disable preempt.

Thanks, that did it: now it suspends/resumes with no errors reported and the script runs to completion (including restarting powernowd and stopping the fan soon afterwards).

Quoted 3 lines

known, s4bios is not working for a long time now, we should remove it. "echo 4 > /proc/acpi/sleep" is supposed to do the same thing as "echo disk > /sys/power/state" on an acpi system.

Does it make any difference which interface is used (/sys/power vs /proc/acpi) ?

Thanks again,
Andrea.

Signature · 9 lines

Homepage: http://andrea.borgia.bo.it     /    Amateur radio: IZ4FHT



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

Nigel Cunningham

2005-04-04 08:03 UTC #9 lore raw

In reply to a message not in the archive

Hi.

Quoted 11 lines · Stefan Seyfried

Quoted 6 lines · Andrea Borgia
Stefan Seyfried wrote:
Thanks, I have no idea how they got in the script, that partially fixed
it: no more "command not found" errors, but the kernel messages about "scheduling while atomic" are still there and the script still aborts with SIG11 after resuming from suspend.
| Mar 27 18:49:57 localhost kernel: note: swsusp-script[3902] exited with preempt_count 1
disable preempt.

Isn't that just dealing with the symptom, rather than the problem though? Presumably something a preempt_enable() or such like is missing somewhere.

Regards,

Nigel

Signature · 14 lines

Nigel Cunningham
Software Engineer, Canberra, Australia
http://www.cyclades.com
Bus: +61 (2) 6291 9554; Hme: +61 (2) 6292 8028;  Mob: +61 (417) 100 574

Maintainer of Suspend2 Kernel Patches http://suspend2.net



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

A message that is not in the archive (sent off-list or not yet received): 1112601782.3757.9.camel-r49W/1Cwd2ff0s6lnCXPX/[email protected]

Stefan Seyfried

2005-04-05 20:39 UTC #11 lore raw

Quoted 7 lines · Nigel Cunningham

Hi.
Quoted 1 line · Stefan Seyfried
disable preempt.
Isn't that just dealing with the symptom, rather than the problem

yes, of course. It is a workaround, not a fix.

Quoted 2 lines

though? Presumably something a preempt_enable() or such like is missing somewhere.

maybe the swsusp code has to be "preempt-audited" :-) Unfortunately i cannot do it.

Signature · 9 lines

Stefan Seyfried



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click