[ltp] Thinkpad buttons, /dev/nvram, and return from S3 with ACPI

[Chun-Yu Shei]

Ari Pollak wrote:

> I've had ACPI working pretty much the same (fairly well, minus a few 
> features, like the loss of acpi interrupts after return from suspend) 
> since about 2.6.0-test9. I just built all of ACPI into modules and 
> load all of them at startup, as well as installed the acpi utility and 
> acpid daemon. Then i get a nice /proc/acpi tree with some fun 
> information. Executing "echo 3 > /proc/acpi/sleep" will put the 
> machine to sleep pretty quickly, and holding down the power button is 
> currently the only way to wake it up (there's a patch in the works to 
> relieve that). Suspending is the easy part, the problem is just that 
> certain drivers don't come back correctly all the time; psmouse seems 
> to have been fixed now, but I have sporadic problems with e1000 and 
> orinoco_cs.
> I never really tried APM on this laptop so I don't know how well it 
> works; I suppose I should, the battery life when using ACPI & cpudynd 
> isn't that great anyway (about 2 1/2 hours with the standard battery).

Heh, now this is a huge reason to run Windows :)  I can get 4 hours in 
Windows if I keep my CPU usage down, but in Linux with it fixed at 600 
MHz, I still can barely get 2 1/2 hours (with the 6 cell, not 9 cell 
battery).  Power management in Windows is far better than in Linux :-\  
Looking at the power usage ACPI provides, it looks like in Linux it 
never goes under 11-13W or so (if I remember correctly...APM only now).  
In Windows it can go down as low as 8-10W.

- Chun-Yu