FS#24029 - [pm-utils intel] blank screen after suspend, also inputs dead
Attached to Project:
Arch Linux
Opened by mark (mmm) - Sunday, 01 May 2011, 18:24 GMT
Last edited by Allan McRae (Allan) - Saturday, 02 June 2012, 12:43 GMT
Opened by mark (mmm) - Sunday, 01 May 2011, 18:24 GMT
Last edited by Allan McRae (Allan) - Saturday, 02 June 2012, 12:43 GMT
|
Details
Description:
Dear all, at 29/Apr a started to experience broken suspend on my Archlinux. I'm on i686 netbook (intel_dri GMA945, Atom N270), running [testing]. First started on the 29th, my machine didn;t reboot for about a week, but suspends were all ok for/during the week time. Synopsis: suspends goes (probably) well, the machine suspends, on wake up though --i move mouse or press keyboard-- ..like it seems cpu goes for a while (5sec - some sound, LED goes on) then nothing. Screen remains dark for all the time. All inputs seem death, cpu activity probbably goes off too - no SysRq keys, ctr+alt+del nothing. I have to hard power-off. This happens every time. I tried: * resuming from [testing] * from [core] * resume from tty (not kde) * read the topic https://bbs.archlinux.org/viewtopic.php?id=116314 * and have the resume hook in mkinitcpio.conf * boot param intel_idle.max_cstate=0 * didnt do bios disable power setting as my bios doesnt allow that * change /etc/pm/config.d/module SLEEP_MODULE to kernel/uswsusp * suspend via pm-suspend; s2ram; echo mem > /sys/power/state; rtcwake -m mem --- all same * i even tried downgrading to 2011/04/14 & 2011/04/04 thanks to Arch Rollback Machine Additional info: I'm attaching various configs/logs. Now I'm clueless :( , please if onyone has an idea to try, let me know. Steps to reproduce: suspend on my machine and then wake Thank you, Mark 
dmesg
(51.5 KiB)
grub.cfg
(3.3 KiB)
lspci
(11.2 KiB)
mkinitcpio.conf
(2.9 KiB)
pm-suspend.log
(38.7 KiB)
Xorg.0.log
(26.3 KiB)
|
This task depends upon
Closed by Allan McRae (Allan)
Saturday, 02 June 2012, 12:43 GMT
Reason for closing: Fixed
Additional comments about closing: submitter reports fixed by magic!
Saturday, 02 June 2012, 12:43 GMT
Reason for closing: Fixed
Additional comments about closing: submitter reports fixed by magic!
and tried installing dmidecode & vbetool.
Tried with 2.6.38.{3,4}, 2.6.37 and 2.6.32-lts kernels.
$ls /etc/pm/*
/etc/pm/conf.d:
module
/etc/pm/power.d:
/etc/pm/sleep.d:
If your machine ever worked before, then revert to the packages from when it worked, and try to figure out which package update broke suspend.
Sorry that I cannot be of more help. There might be guides out there about how to debug suspend problems.
I'll check out the lkml as you suggested. Strange is the suspend seems to fail by default, but works if a hibernation has been done before (a hypothesis now:)
Anyways, thanks a lot!
I'll try a custom kernel with the patch provided.
I also confirmed a strange workaround:
-fresh boot -> resume from suspend not working
-once hibernated -> since then resume is OK (till next fresh start)
was still present in 3.0, will test latest 3.1 soon but i dont expect much, i dont remember how the lkml thread ended, but nothing crystalized there..