[Debian-eeepc-devel] Bug#603430: Bug#603430: Bug#603430: eeepc-acpi-scripts: Suspend to RAM takes several minutes since switch to acpi-support

[Axel Beckert]

Hi Ben,

thanks for having a look at that issue.

Ben Armstrong wrote:
> If acpi-support alone doesn't work for you with the Squeeze kernel, then
> this bug probably belongs on acpi-support instead.

I'm using the 2.6.36 from experimental currently. Will check with
2.6.32 from Squeeze/Sid.

> You just need to disable ACPI_SLEEP in /etc/default/acpi-support.
> You could have your WM bind some action (e.g. sudo pm-suspend
> --quirk-s3-bios) to the XF86Sleep key, perhaps. Or you could use
> gnome-power-manager or something similar.

Urgh. Sounds more like a hack than a solution...

> * A conflict with acpi-support meant not just that you would have to
> remove acpi-support, but actually *purge* it to avoid residual configs
> causing problems. This means it's impossible to have a general purpose
> system image that "just works" with all kinds of hardware, including Eee
> PCs.

Accepted.

Ben Armstrong wrote:
> On 13/11/10 10:19 PM, Ben Hutchings wrote:
> > Well, first Axel should check that the kernel itself does the right
> > thing:
> > 
> >     echo mem > /sys/power/state

JFYI:

$ cat /sys/power/state
mem disk

Will change it to mem only and check again.

> I have tested and re-tested with 1.1.11 on a model 701 and cannot
> reproduce your results.

Oh, ok. That gives hope! :-)

> I'm using 2.6.32-27 (current version in squeeze/sid). I have tested
> both with and without gnome-power-manager running. In both cases,
> pm-suspend is handling the suspension just fine with no delays.
> Also, "echo mem > /sys/power/state" as Ben suggested above works
> fine.

Ok, will check with 2.6.32 kernel.

> The ACPI implementation in the model 701 is pretty brain-dead and has
> had to be worked around in the kernel since many versions ago. If you
> generate enough of them at once*, it is still possible to have ACPI
> events "pile up" and not be handled until much later because the
> workaround is not perfect.

I see. Thanks for the explanation.

		Regards, Axel
-- 
 ,''`.  |  Axel Beckert <abe at debian.org>, http://people.debian.org/~abe/
: :' :  |  Debian Developer, ftp.ch.debian.org Admin
`. `'   |  1024D: F067 EA27 26B9 C3FC 1486  202E C09E 1D89 9593 0EDE
  `-    |  4096R: 2517 B724 C5F6 CA99 5329  6E61 2FF9 CD59 6126 16B5