Hi Boris,

Boris Barbour wrote:

> On a new, cheap Asus laptop ("Model X54L"), suspend to RAM does not
> work under some conditions.

Thanks for reporting it. Is this a regression? Does suspend-to-disk
(triggered with

echo disk >/sys/power/state

) work? If you have time to follow the instructions from
basic-pm-debugging.txt[1] to narrow down the cause, that would be very
useful.

Hope that helps,
Jonathan

[1] http://www.kernel.org/doc/Documentation/power/basic-pm-debugging.txt
also available from Documentation/power in the documentation dir of
the linux-doc-3.2 package



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120206044935.GA3407@burratino">http://lists.debian.org/20120206044935.GA3407@burratino

Thanks for the quick reply, Jonathan.

> Is this a regression?

Not really (the machine is new). I tried older debian 5.09 and 6.03 live cds.
With 6.03 suspending from a session worked under some conditions (lid) but not
all (menu).

> Does suspend-to-disk
> (triggered with
>
> echo disk >/sys/power/state

This doesn't work.

> ) work? If you have time to follow the instructions from
> basic-pm-debugging.txt[1] to narrow down the cause, that would be very
> useful.

Thanks for the pointer! For some reason I didn't find it while searching
around (we beginners don't know to search for "pm-" instead of "suspend").
Maybe a link could be added on the debian wiki at
http://wiki.debian.org/Suspend ? The "shutdown" method described in the
document seems to work. But it is a bit clunky and it wasn't immediately
obvious how to set this option permanently or whether it even applies to
suspend to RAM (more for the wiki...)

In the meantime I found the following site

http://thecodecentral.com/2011/01/18/fix-ubuntu-10-10-suspendhibernate-not-
working-bug

and shamelessly copy-pasted the suggested script. It works pretty well - the
only obvious problem at the moment is that waking from sleep requires a press
of the power button (lid-opening doesn't suffice). But at least there is no
freeze with data loss and reboot. Huge progress.

The working "shutdown" method suggests a bios issue and the copied script
indicates a usb problem. Follow up might therefore look at the usb settings in
the bios and investigate a bios upgrade (the changelog mumbles something about
usb).

The comments on the web site suggest that the solution has helped a lot of
people and with different machines.

Thanks and best wishes,

Boris




--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201202070157.35153.boris.barbour@ens.fr">http://lists.debian.org/201202070157.35153.boris.barbour@ens.fr

Hi again,

Boris Barbour wrote:

> The "shutdown" method described in the
> document seems to work. But it is a bit clunky

All right --- that's useful to know. Of course I was not suggesting
this as a workaround but as a diagnostic tool.

[...]
> In the meantime I found the following site
>
> http://thecodecentral.com/2011/01/18/fix-ubuntu-10-10-suspendhibernate-not-working-bug
>
> and shamelessly copy-pasted the suggested script. It works pretty well

Can you reproduce this result without that script? Like so:

modprobe -r ehci_hcd
modprobe -r xhci_hcd
echo disk >/sys/power/state
modprobe ehci_hcd
modprobe xhci_hcd

If so, could you also try unloading/loading only one of the two drivers to
see which one is the culprit?

Please attach "acpidump" output (so we can learn about the BIOS) and
full "dmesg" output from booting and successfully hibernating/resuming
in the "platform" mode with USB drivers unloaded (which will tell us
many things).

Thanks and sorry for the trouble,
Jonathan



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120207011635.GB2149@burratino">http://lists.debian.org/20120207011635.GB2149@burratino

# [1]
forwarded 658778 https://bugzilla.kernel.org/show_bug.cgi?id=37632
quit

Boris Barbour wrote:
> Hi Jonathan,

>> If so, could you also try unloading/loading only one of the two drivers to
>> see which one is the culprit?
>
> Removing only ehci_hcd, hibernation works perfectly. If only xhci_hcd is
> removed, the screen is blanked but stays on. A power cycle is necessary,
> although the computer then resumes OK.

Nicely done. Ok.

References:

https://bugzilla.kernel.org/show_bug.cgi?id=37632
https://bugzilla.redhat.com/show_bug.cgi?id=697150
http://thread.gmane.org/gmane.linux.kernel/1222803

Please add yourselves to the cc list for the upstream bug [1]. Further
work can be coordinated there.



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120213015907.GI2256@burratino">http://lists.debian.org/20120213015907.GI2256@burratino