Standby (S1) does not work

Standby (S1) does not work

[Michael Grigoriev]

Hi,

I am running 2.6.7-rc2-mm2 on an HP zt3000 series (centrino based)
notebook. The problem is I can't seem to get normal standby (S1) working.

When I echo "1" into /proc/acpi/sleep or "standby" into /sys/power/state,
the screen blanks for a second but comes right back (and yes, I tried
delaying it with a sleep, and using echo -n). The dmesg shows:

PM: Preparing system for suspend
Stopping tasks: ================|
Restarting tasks... done

Same thing happens under X and with a very minimal kernel from the console.

In possibly related news, I noticed this abnormality:

# cat /sys/power/state
standby mem disk

# cat /proc/acpi/sleep
S0 S3 S4 S4bios S5

Notice that proc does not even list S1!

I was able to get suspend to disk kind of working, so not *everything* is
broken :)

I'd be happy to provide more info about my system, etc, but I'd even
appreciate some pointers as per how I can debug this... I've done a bit of
kernel development in the past, but nothing related to pm/acpi, so I'm
kind of clueless here. For example, I compiled a kernel with acpi debug
info, but that did not seem to give me any more messages. Is there
something special I need to do to actually change the debug level?

Oh a couple more questions while I am here:

1. Sometimes my system goes into a state where even holding the power
button down for a while does not turn it off - I have to pop out the
battery to reboot it (my broken wireless driver sometimes causes this, and
so does trying to suspend to ram). Is this also an ACPI problem, or should
this be handled entirely in hardware?

2. What does S4bios mean? I am assuming it somehow has the BIOS restore
the RAM image rather than having to boot up the kernel to do it, so it's
faster? How can I actually use this mode? Are there any docs anywhere?

Thanks in advance!

-- 
Have fun,
Michael "mag" Grigoriev          The words I have, the spells I know,
mag-wxkiEGMmYARAfugRpC6u6w@public.gmane.org                     Were not enough to make you happy
http://www.luminal.org                        To make you smile again



-------------------------------------------------------
This SF.Net email is sponsored by the new InstallShield X.
>From Windows to Linux, servers to mobile, InstallShield X is the one
installation-authoring solution that does it all. Learn more and
evaluate today! http://www.installshield.com/Dev2Dev/0504

Re: Standby (S1) does not work

[Michael Grigoriev]

On Sun, 06 Jun 2004 15:05:10 +0000, Michael Grigoriev wrote:
> In possibly related news, I noticed this abnormality:
> 
> # cat /sys/power/state
> standby mem disk
> 
> # cat /proc/acpi/sleep
> S0 S3 S4 S4bios S5
> 
> Notice that proc does not even list S1!

Further research shows that S1 does in fact not exist. _S1 is not defined
in the DSDT. I thought S1 is like the most basic ACPI mode and everybody
would have it, no? Do you think that it is actually not supported by
hardware, or is it just that the didn't bother adding it to the DSDT?

Would it be a safe experiment to try throwing something like 

    Name (\_S1, Package (0x03)
    {
        0x02, 
        0x02, 
        0x00
    })

into the DSDT?

Has anybody tried anything like this? :)

Thanks in advance.

-- 
Have fun,                                As the moment elapsed
Michael "mag" Grigoriev                We walked in slowmotion
mag-wxkiEGMmYARAfugRpC6u6w@public.gmane.org                               Denying the tide
http://www.luminal.org                 We'll find our devotion



-------------------------------------------------------
This SF.Net email is sponsored by the new InstallShield X.
>From Windows to Linux, servers to mobile, InstallShield X is the one
installation-authoring solution that does it all. Learn more and
evaluate today! http://www.installshield.com/Dev2Dev/0504

Re: Re: Standby (S1) does not work

[Pavel Machek]

Hi!

> > In possibly related news, I noticed this abnormality:
> > 
> > # cat /sys/power/state
> > standby mem disk
> > 
> > # cat /proc/acpi/sleep
> > S0 S3 S4 S4bios S5
> > 
> > Notice that proc does not even list S1!
> 
> Further research shows that S1 does in fact not exist. _S1 is not defined
> in the DSDT. I thought S1 is like the most basic ACPI mode and everybody
> would have it, no? Do you think that it is actually not supported by
> hardware, or is it just that the didn't bother adding it to the
> DSDT?

S1 is basically useless, no wonder hw manufacturers do not support
it. It was workaround for old windows...
							Pavel


> Would it be a safe experiment to try throwing something like 
> 
>     Name (\_S1, Package (0x03)
>     {
>         0x02, 
>         0x02, 
>         0x00
>     })
> 
> into the DSDT?

You are not linkely to fry anything.

-- 
934a471f20d6580d5aad759bf0d97ddc


-------------------------------------------------------
This SF.Net email is sponsored by the new InstallShield X.
>From Windows to Linux, servers to mobile, InstallShield X is the one
installation-authoring solution that does it all. Learn more and
evaluate today! http://www.installshield.com/Dev2Dev/0504