[ltp] Suspend woes and strange hardware quirks

[John Locke]

John Locke wrote:

> André Wyrwa wrote:
>
>> it seems to me that all of your three suspend/resume problems should be
>> solvable by compiling the according drivers as modules and unloading
>> them before suspend / reloading after. These problems (with also, usb
>> and network drivers) have been reported a lot here.
>>
>> So reconfigure your kernel, set these drivers to modules, recompile and
>> adjust your /etc/acpi/actions scripts to unload/reload the modules. I'm
>> sorry i can't give you detailed information on that for i'm not a
>> mandrake user and i only half adjusted my own scripts so far.
>>
> Okay. I did that, except for Alsa, and now my USB mouse is detected on 
> startup again. I'm still having a couple of weird quirks though.
>
> I think one of the issues is the Madwifi driver. If I unload the USB 
> modules, unload the madwifi modules, and power off the Alsa driver, 
> everything suspends and resumes perfectly--until I start using 
> networking again.
>
Okay. That was a dead end. Not a Madwifi problem--I get the exact same 
symptoms if I'm on a wired connection, with the madwifi drivers not loaded.

After resuming the machine from a suspend (either RAM or disk), 
everything works great until I start using any type of networking. 
Actually, it appears to take a LOT of networking to trigger the problem. 
If I fire up Thunderbird, or do a CVS update, it doesn't appear to cause 
any issues. But if I open up Evolution, I get an instant freeze of the 
networking, USB, and sound systems. Or if I start browsing in Firefox, 
hitting a few pages at once.

Again, what happens is my networking system appears to be still running 
and active, but no traffic goes through, either wired or wireless. The 
USB system stops working, and no sound comes out of my speakers. 
Sometimes I can suspend and resume again, and everything comes back 
up--until I start using the network extensively again. Other times, 
something blocks the machine from suspending. I was able to unload the 
Madwifi modules and successfully suspend, but when I tried to stop the 
USB system with the e1000 driver a few minutes ago, the entire system 
locked up.

Does anybody have a working ACPI suspend, with kernel 2.6.8.1, USB, 
Alsa, and Madwifi?

I'm using Mandrake 10.1 on a T41, but Brian Gerkey reported on the 
Madwifi list exactly the same issues on Gentoo with a 2.6.8 kernel on a 
T40 (he added that PCMCIA also freezes when this happens).


> The other issue I'm seeing now is this:
>
> Sep 29 09:45:36 hunter kernel: smapi: 32-bit protected mode SMAPI BIOS 
> found. :-)
> Sep 29 09:45:36 hunter kernel: smapi: SMAPI BIOS return codes differ!
>
> Not sure if this is the Thinkpad module or the ibm_acpi module.
>
> One other thing: I have it successfully suspending to RAM when I close 
> the lid. Is it possible to have it resume when I open it, like it does 
> in Windows? If so, how?
>
> Thanks,

--John