[mythtv-users] ACPI sleep problems

[Greg Woods]

On Tue, 2005-09-20 at 12:27 +0100, Steve Hill wrote:

> /proc/acpi/sleep seems to be the old 
> 2.4 kernel interface, whereas /sys/power/state is the 2.6 kernel 
> interface.

I tried doing the "echo -n mem > /sys/power/state" on my laptop. It does
go to sleep; the blinking lights show it is in a suspend-to-RAM state.
But when I try to wake it up, it hangs. This is pretty much the same as
"echo 3 > /proc/acpi/sleep";  I have to power it off and reboot to get
it back. It looks like maybe this could be made to work; I have FC4 with
a 2.6 kernel. Maybe that hibernate script is what I need; I'll try it
out. Thanks to everyone for the pointers on this. Lack of suspend has
been by far the largest pain in the butt to running Linux on a laptop;
it's the one thing that Windows XP does that I wish Linux could do. 

This is wandering a bit off topic for Myth, getting suspend to work on a
laptop, but more to the point is that if I can make it work on my
laptop, perhaps I could make it work on my slave backend/frontend which
would be directly useful to my Myth setup.

>   Whilest suspending using ACPI seems to just power my machine 
> off completely, if I disable ACPI and use APM instead it sends the machine 
> to sleep but won't wake up again.

My machine only supports ACPI. Most new hardware will be ACPI-only. I
think the IBM Thinkpads may be one of the few exceptions left on the
market.

--Greg