Hang on S3 resume with Acer TravelMate 222X

Hang on S3 resume with Acer TravelMate 222X

[Todd Pytel]

I've got a rather old Acer TravelMate 222X for which I would *really*
like to get proper S3 support. Right now, doing an 

echo -n "mem" > /sys/power/state

suspends the machine to RAM OK. But when it tries to resume, it shows a
bit of disk activity and then hard locks - nothing on the screen, no
keyboard input is accepted, and it doesn't respond to pings. This is
using 2.6.17, entirely from text mode with no framebuffer. I've tried
various combinations of acpi_sleep parameters and noapic/lapic with no
visible effect.

I've been reading through many of the ACPI docs on the net, and have
fixed my DSDT and added it to the kernel according to those
instructions. This had no visible effect on the resume behavior. Most of
the fixes were pretty obvious, but one I wasn't sure about - perhaps
it's related to my problem. I received the error "Method local variable
is not initialized (Local0)" for the statement "Store (Local0,
\_SB.PCI0.PCI1.OBLN.PMES" in the following code block:

Method (_L0B, 0, NotSerialized)
{
  Store (0x0B, PX80)
  And (\_SB.PCI0.PCI1.OZ68.PWRS, 0xFC, \_SB.PCI0.PCI1.OZ68.PWRS)
  Or (\_SB.PCI0.PCI1.OZ68.PMES, 0x80, \_SB.PCI0.PCI1.OZ68.PMES)
  Store (Local0, \_SB.PCI0.PCI1.OBLN.PMES)
  Notify (\_SB.SLPB, 0x02)
}

Well, yeah... it's not initialized. I assumed that the second Store
statement was just a sloppy way to zero out the variable. So I replaced
"Local0" with "0x00". That compiles without error, but I don't know if
it makes sense - is there a way that Local0 can be used as something
like a global variable in this specification? 

Other than that, I'm not sure what information is helpful to fix this
problem, but I'm *more* than happy to do whatever diagnostics will shed
better light on the problem. I really want to get this working! There's
nothing of obvious interest in my logs - the machine doesn't get far
enough when it tries to resume to write anything to syslog, and nothing
during bootup looks particularly suspicious. But maybe I just don't what
I should be looking for. 

Thanks for any help you can provide,
Todd

Re: Hang on S3 resume with Acer TravelMate 222X

[Todd Pytel]

Please? Anybody? Just some Google fodder or pointers on how to generate
better debugging information for the list would be greatly appreciated.

--Todd

On Thu, 2006-08-31 at 18:29 -0500, Todd Pytel wrote:
> I've got a rather old Acer TravelMate 222X for which I would *really*
> like to get proper S3 support. Right now, doing an 
> 
> echo -n "mem" > /sys/power/state
> 
> suspends the machine to RAM OK. But when it tries to resume, it shows a
> bit of disk activity and then hard locks - nothing on the screen, no
> keyboard input is accepted, and it doesn't respond to pings. This is
> using 2.6.17, entirely from text mode with no framebuffer. I've tried
> various combinations of acpi_sleep parameters and noapic/lapic with no
> visible effect.
> 
> I've been reading through many of the ACPI docs on the net, and have
> fixed my DSDT and added it to the kernel according to those
> instructions. This had no visible effect on the resume behavior. Most of
> the fixes were pretty obvious, but one I wasn't sure about - perhaps
> it's related to my problem. I received the error "Method local variable
> is not initialized (Local0)" for the statement "Store (Local0,
> \_SB.PCI0.PCI1.OBLN.PMES" in the following code block:
> 
> Method (_L0B, 0, NotSerialized)
> {
>   Store (0x0B, PX80)
>   And (\_SB.PCI0.PCI1.OZ68.PWRS, 0xFC, \_SB.PCI0.PCI1.OZ68.PWRS)
>   Or (\_SB.PCI0.PCI1.OZ68.PMES, 0x80, \_SB.PCI0.PCI1.OZ68.PMES)
>   Store (Local0, \_SB.PCI0.PCI1.OBLN.PMES)
>   Notify (\_SB.SLPB, 0x02)
> }
> 
> Well, yeah... it's not initialized. I assumed that the second Store
> statement was just a sloppy way to zero out the variable. So I replaced
> "Local0" with "0x00". That compiles without error, but I don't know if
> it makes sense - is there a way that Local0 can be used as something
> like a global variable in this specification? 
> 
> Other than that, I'm not sure what information is helpful to fix this
> problem, but I'm *more* than happy to do whatever diagnostics will shed
> better light on the problem. I really want to get this working! There's
> nothing of obvious interest in my logs - the machine doesn't get far
> enough when it tries to resume to write anything to syslog, and nothing
> during bootup looks particularly suspicious. But maybe I just don't what
> I should be looking for. 
> 
> Thanks for any help you can provide,
> Todd

Re: Hang on S3 resume with Acer TravelMate 222X

[Ben B]

Todd Pytel <tppytel@sophrosune.org> uttered the following thing:
> Please? Anybody? Just some Google fodder or pointers on how to generate
> better debugging information for the list would be greatly appreciated.

Try unloading various modules before suspend. On many machines, at least
the usb host drivers need to be unloaded.

Video's another possibility. Look at vbetool, it might be able to
restore video and get things on track again.

Also, do your suspend from a shell script, and have this script log
debugging info after resume (remembering to sync in case you have to
hard boot).

Ben

Re: Hang on S3 resume with Acer TravelMate 222X

[Yu Luming]

On Tuesday 05 September 2006 10:52, Todd Pytel wrote:
> Please? Anybody? Just some Google fodder or pointers on how to generate
>  better debugging information for the list would be greatly appreciated.
Not sure if this can help you. http://acpi.sourceforge.net/

>  > I've got a rather old Acer TravelMate 222X for which I would *really*
>  > like to get proper S3 support. Right now, doing an
>  >
>  > echo -n "mem" > /sys/power/state
>  >
>  > suspends the machine to RAM OK. But when it tries to resume, it shows a
>  > bit of disk activity and then hard locks - nothing on the screen, no
>  > keyboard input is accepted, and it doesn't respond to pings. This is
>  > using 2.6.17, entirely from text mode with no framebuffer. I've tried
>  > various combinations of acpi_sleep parameters and noapic/lapic with no
>  > visible effect.
Does windows work?
If you think this is a ACPI bug , please file it on bugzilla.kernel.org in 
ACPI category.

>  >
>  > I've been reading through many of the ACPI docs on the net, and have
>  > fixed my DSDT and added it to the kernel according to those
>  > instructions. This had no visible effect on the resume behavior. Most of
>  > the fixes were pretty obvious, but one I wasn't sure about - perhaps
>  > it's related to my problem. I received the error "Method local variable
>  > is not initialized (Local0)" for the statement "Store (Local0,
>  > \_SB.PCI0.PCI1.OBLN.PMES" in the following code block:
>  >
>  > Method (_L0B, 0, NotSerialized)
>  > {
>  >   Store (0x0B, PX80)
>  >   And (\_SB.PCI0.PCI1.OZ68.PWRS, 0xFC, \_SB.PCI0.PCI1.OZ68.PWRS)
>  >   Or (\_SB.PCI0.PCI1.OZ68.PMES, 0x80, \_SB.PCI0.PCI1.OZ68.PMES)
>  >   Store (Local0, \_SB.PCI0.PCI1.OBLN.PMES)
>  >   Notify (\_SB.SLPB, 0x02)
>  > }
>  >
>  > Well, yeah... it's not initialized. I assumed that the second Store
>  > statement was just a sloppy way to zero out the variable. So I replaced
>  > "Local0" with "0x00". That compiles without error, but I don't know if
>  > it makes sense - is there a way that Local0 can be used as something
>  > like a global variable in this specification?
>  >
>  > Other than that, I'm not sure what information is helpful to fix this
>  > problem, but I'm *more* than happy to do whatever diagnostics will shed
>  > better light on the problem. I really want to get this working! There's
>  > nothing of obvious interest in my logs - the machine doesn't get far
>  > enough when it tries to resume to write anything to syslog, and nothing
>  > during bootup looks particularly suspicious. But maybe I just don't what
>  > I should be looking for.

The first step, please take a close look at Documentation/power/video.txt


-- 
Thanks,
Luming
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Hang on S3 resume with Acer TravelMate 222X

[Todd Pytel]

Ben,

On Tue, 2006-09-05 at 08:16 +0200, Ben B wrote:
> Try unloading various modules before suspend. On many machines, at least
> the usb host drivers need to be unloaded.

OK, I'll give that a try. I don't have the machine with me to check at
the moment.

> Video's another possibility. Look at vbetool, it might be able to
> restore video and get things on track again.

Hmm... I read about vbetool in video.txt, but maybe I was
misunderstanding. From what I read, vbetool's restore function only
works after the resume gets back to userland. Since my machine is
hard-locking after the resume, I was assuming that vbetool would never
even get a chance to do its thing. Is it possible for the machine to get
back to running user processes and *then* hard lock?

> Also, do your suspend from a shell script, and have this script log
> debugging info after resume (remembering to sync in case you have to
> hard boot).

Similar questions as those about vbetool. But I'll give it a shot. 

Thanks for the suggestions.

--Todd

Re: Hang on S3 resume with Acer TravelMate 222X

[Todd Pytel]

On Tue, 2006-09-05 at 14:28 +0800, Yu Luming wrote:
> Not sure if this can help you. http://acpi.sourceforge.net/

I've checked out the site quite a bit - that's where I got a lot of the
DSDT info. But I didn't see anything specific to hard-locks on resume.

> Does windows work?

Of course. (Sadly...)

> If you think this is a ACPI bug , please file it on bugzilla.kernel.org in 
> ACPI category.

Well, I don't really know enough to say whether it's an ACPI bug or not.
>From what I've read, ACPI often needs a lot of troubleshooting because
the hardware and DSDT's don't follow the standard as strictly as
Linux-ACPI itself. If the regulars here think is a genuine Linux-ACPI
bug, then I can certainly file it.

> The first step, please take a close look at Documentation/power/video.txt

I've read it. But from what I understood, the various video tricks only
help after the kernel has resumed far enough to run processes. I'm not
getting that far. Maybe I'm not understanding...

Thank you for your suggestions.

--Todd