asus-ul30 team mailing list archive

[Tero Mononen <tero@xxxxxxxxxx>]

Hi

What kind of suspend do you use and does those work reliable? I compiled 
this Bucks version of nvidia_g210m_acpi and using it. Kernel does not 
puke anymore, but there is still occasional freezing (~20 %).

I'm using just hibernate-ram-command and which uses sysfs-ram.

In the log of hibernate are these lines:
Starting suspend at Thu Apr 1 21:14:27 EEST 2010
hibernate-ram: [01] Executing CheckLastResume ...
hibernate-ram: [01] Executing CheckRunlevel ...
hibernate-ram: [01] Executing LockFileGet ...
hibernate-ram: [01] Executing NewKernelFileCheck ...
hibernate-ram: [10] Executing EnsureSysfsPowerStateCapable ...
hibernate-ram: [11] Executing XHacksSuspendHook1 ...
hibernate-ram: [59] Executing RemountXFSBootRO ...
hibernate-ram: [89] Executing SaveKernelModprobe ...
hibernate-ram: [91] Executing ModulesUnloadBlacklist ...
hibernate-ram: [91] Executing ModulesUnloadBlacklist ...
hibernate-ram: [95] Executing XHacksSuspendHook2 ...
hibernate-ram: [98] Executing CheckRunlevel ...
hibernate-ram: [99] Executing DoSysfsPowerStateSuspend ...
hibernate-ram: Activating sysfs power state mem ...
<BLANK SCREEN + "_"-CHAR & FREEZING>

Does somebody have figured out some tricks to make hibernation 100% 
reliable? How about hibernation to disk?



Tero Mononen


Buck wrote:

>  Hi,

> I have an UL30VT-A1. I was happy to find nvidia-g210m-acpi, but I 
found that it was causing problems across suspend/resume cycles.
> I made some simple changes that seem to make it work better: 1) catch 
post suspend event so that module doesn't need to be
> unloaded/reloaded after suspend 2) unregister power notifier on 
unload, so that power events don't cause kernel panic after module
> is unloaded.

>  Hope someone else can benefit.  :)
>
>  Modified code is attached.   Maybe the deb can be updated accordingly?
>
> -- Buck

>

--
Tero Mononen
tero@xxxxxxxxxx
Skype: Corrafig | Tel: +358 45 654 3210