Resume from standby not working from USB event

[Derek Broughton]

Daniel Eichhorn wrote:

> Hi everybody
> 
> I'm really getting desperate here. I turned an old Sony Vaio SR11K
> into a digital frame for watching TV, Movies, listening to music etc.
> I could manage a lot of problems so far, but I can't get the laptop to
> wake up from an USB event, such as pressing the Power Button on my USB
> remote. I adapted the hibernate configuration to send the laptop only
> to S1 mode, since it USB0 only seems to support this state:
>  > cat /proc/acpi/wakeup
> Device        S-state   Status   Sysfs node
> PWRB    S4    *enabled
> USB0    S1     enabled   pci:0000:00:07.2
> CRD0    S3     disabled  pci:0000:00:0c.0
> MODE    S3     disabled  pci:0000:00:0a.0
> 
> As you can see here I also managed to enabled USB0 to wakeup from S1
> mode. I changed a line in /etc/hibernate/sysfs-ram.conf to
> UseSysfsPowerState standby
> and now it really goes to S1 mode. After executing sudo hibernate-ram
> the screen gets black and the machine turns of.

I trust you're aware that S1 (sleep) does _not_ actually turn the machine
off.  It should turn your display and drives off.

> When I press now the Power button on the remote, the machine turns on
> (somewhat), the fan starts turning but there's only a blinking cursor
> on about the forth line. 

In the first place, I've seen many problems with using a power button
to "wake" from sleep - the result is often that the machine wakes - and
then goes into full shutdown mode!  Still, this doesn't seem to be what
yours is doing.

> The vaio doesn't react to keyboard input nor 
> remote buttons (also tried the (CTRL+) ALT-Fx) to switch the tty.

are you sure it doesn't really react to keyboard input?  If you left tty1
logged in as root, with "shutdown -r now" waiting for an <enter>, put the
system into standby, wake it, and then try ctrl-alt-F1 and <enter>, I
rather expect it to reboot (ie, your system really does wake, but it
doesn't reset your video).  Another thing to try is to switch to tty1,
_then_ sleep and resume (maybe even without X running).

If either of these methods is successful, then we know that it's just the
video - which is almost certainly fixable with a Video POST. 
Unfortunately, I haven't a clue where /etc/hibernate comes from (I use
either /etc/acpi or /etc/pm-utils) so I can't say what you'd need to do to
make it do that.

> So why doesn't the machine only doesn't wakeup when I try to activate
> from USB device (I also tried a mouse plugged to USB0: same thing).

I have to admit, it doesn't make a lot of sense to me, either.

> [   19.476171] ACPI: Looking for DSDT in initramfs... error, file /
> DSDT.aml not found.

That's an odd message I've never seen.  I don't think it's important - there
shouldn't _need_ to be a DSDT in initramfs, but I wonder if it called it an
error because there isn't one in your BIOS either - which means your ACPI
implementation is buggy.  Still everything else indicates that ACPI is
initializing properly.

> If I would have to guess, there must be a difference between the USB0
> device and the keyboard (maybe a IRQ) which lets the machine wake up
> completely with one and let it stay in zombie mode with the other...

Except that you see _some_ activity, which says to me that it's not a zombie
any more.
-- 
derek