Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi all,

I've got Fedora 10 on Acer Aspire 1690 (1692 WLMi to be exact) and some time ago I noticed that pm-hibernate has the following problem:

pm-hibernate runs ok and cleanly shuts down the machine that at the very moment it is supposed to be hibernated it wakes up and resumes as if I had pressed the power button.

I compiled few kernel.org kernels and noticed this starts to occur between 2.6.23 and 2.6.24. 2.6.23 and all older versions work, 2.6.24 and new versions fail, including 2.6.28.

Does anyone have any suggestions what might be causing this? I mean, could I try some boot parameter or some simple patch before starting to bisect this?

FWIW, dmesg/lspci output can be found from my Fedora bug report at

https://bugzilla.redhat.com/show_bug.cgi?id=476430

Thanks!




      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Monday 05 January 2009, Daniel Qarras wrote:
> Hi all,
> 
> I've got Fedora 10 on Acer Aspire 1690 (1692 WLMi to be exact) and some time ago I noticed that pm-hibernate has the following problem:
> 
> pm-hibernate runs ok and cleanly shuts down the machine that at the very moment it is supposed to be hibernated it wakes up and resumes as if I had pressed the power button.
> 
> I compiled few kernel.org kernels and noticed this starts to occur between 2.6.23 and 2.6.24. 2.6.23 and all older versions work, 2.6.24 and new versions fail, including 2.6.28.
> 
> Does anyone have any suggestions what might be causing this? I mean, could I try some boot parameter or some simple patch before starting to bisect this?

Unfortunately, that's quite a few kernel versions back.  I'm afraid we won't be
able to resolve this without knowing which patch broke it for you.

Thanks,
Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!
 
> > I've got Fedora 10 on Acer Aspire 1690 (1692 WLMi
> > to be exact) and some time ago I noticed that pm-hibernate
> > has the following problem:
> > 
> > pm-hibernate runs ok and cleanly shuts down the
> > machine but at the very moment it is supposed to be
> > hibernated it wakes up and resumes as if I had pressed the
> > power button.
> > 
> > I compiled few kernel.org kernels and noticed this
> > starts to occur between 2.6.23 and 2.6.24. 2.6.23 and all
> > older versions work, 2.6.24 and newer versions fail, including
> > 2.6.28.
> 
> Unfortunately, that's quite a few kernel versions back.
>  I'm afraid we won't be
> able to resolve this without knowing which patch broke it
> for you.

I've now bisected this to this commit:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9cd9a0058dd35268b24fa16795a92c800f4086d4

If I revert that from 2.6.24 hibernation also works with it. With 2.6.28 the patch does not cleanly apply so I didn't test it.

I am able and willing to test any patches you might cook up.

Thanks!




      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Tuesday 06 January 2009, Daniel Qarras wrote:
> Hi!
>  
> > > I've got Fedora 10 on Acer Aspire 1690 (1692 WLMi
> > > to be exact) and some time ago I noticed that pm-hibernate
> > > has the following problem:
> > > 
> > > pm-hibernate runs ok and cleanly shuts down the
> > > machine but at the very moment it is supposed to be
> > > hibernated it wakes up and resumes as if I had pressed the
> > > power button.
> > > 
> > > I compiled few kernel.org kernels and noticed this
> > > starts to occur between 2.6.23 and 2.6.24. 2.6.23 and all
> > > older versions work, 2.6.24 and newer versions fail, including
> > > 2.6.28.
> > 
> > Unfortunately, that's quite a few kernel versions back.
> >  I'm afraid we won't be
> > able to resolve this without knowing which patch broke it
> > for you.
> 
> I've now bisected this to this commit:
> 
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9cd9a0058dd35268b24fa16795a92c800f4086d4
> 
> If I revert that from 2.6.24 hibernation also works with it. With 2.6.28 the patch does not cleanly apply so I didn't test it.
> 
> I am able and willing to test any patches you might cook up.

Have you tried with acpi_sleep=old_ordering in the kernel's command line?

Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!
 
> > I've now bisected this to this commit:
> >
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9cd9a0058dd35268b24fa16795a92c800f4086d4
> 
> Have you tried with acpi_sleep=old_ordering in the
> kernel's command line?

I hadn't done that yet, never heard about such parameter. Well, tested with Fedora 10 kernel and 2.6.28, it does not fix the issue.

Cheers!




      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Tuesday 06 January 2009, Daniel Qarras wrote:
> Hi!
>  
> > > I've now bisected this to this commit:
> > >
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9cd9a0058dd35268b24fa16795a92c800f4086d4
> > 
> > Have you tried with acpi_sleep=old_ordering in the
> > kernel's command line?
> 
> I hadn't done that yet, never heard about such parameter. Well, tested with Fedora 10 kernel and 2.6.28, it does not fix the issue.

Thanks for verification.

Apparently, one of devices is set up to wake up the system and it does that
right after it's been put to sleep.  I don't think it's a PCI device, because
2.6.28 prevents PCI devices from doing this, unless they are intentionally
set up for system wake-up.

Can you send me the output of /proc/acpi/wakeup, please?

Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!

> Apparently, one of devices is set up to wake up the system
> and it does that right after it's been put to sleep.  I don't think
> it's a PCI device, because 2.6.28 prevents PCI devices from doing this,
> unless they are intentionally set up for system wake-up.
> 
> Can you send me the output of /proc/acpi/wakeup, please?

Sure, please see below:

localhost:~> cat /proc/acpi/wakeup 
Device	S-state	  Status   Sysfs node
RP01	  S4	 disabled  pci:0000:00:1c.0
RP02	  S4	 disabled  pci:0000:00:1c.1
RP04	  S4	 disabled  
USB1	  S3	 disabled  pci:0000:00:1d.0
USB2	  S3	 disabled  pci:0000:00:1d.1
USB3	  S3	 disabled  pci:0000:00:1d.2
USB4	  S3	 disabled  pci:0000:00:1d.3
USB7	  S3	 disabled  pci:0000:00:1d.7
LANC	  S4	 disabled  pci:0000:06:08.0
MODM	  S3	 disabled  pci:0000:00:1e.3
localhost:~> 

Thanks!



      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Wednesday 07 January 2009, Daniel Qarras wrote:
> Hi!
> 
> > Apparently, one of devices is set up to wake up the system
> > and it does that right after it's been put to sleep.  I don't think
> > it's a PCI device, because 2.6.28 prevents PCI devices from doing this,
> > unless they are intentionally set up for system wake-up.
> > 
> > Can you send me the output of /proc/acpi/wakeup, please?
> 
> Sure, please see below:
> 
> localhost:~> cat /proc/acpi/wakeup 
> Device	S-state	  Status   Sysfs node
> RP01	  S4	 disabled  pci:0000:00:1c.0
> RP02	  S4	 disabled  pci:0000:00:1c.1
> RP04	  S4	 disabled  
> USB1	  S3	 disabled  pci:0000:00:1d.0
> USB2	  S3	 disabled  pci:0000:00:1d.1
> USB3	  S3	 disabled  pci:0000:00:1d.2
> USB4	  S3	 disabled  pci:0000:00:1d.3
> USB7	  S3	 disabled  pci:0000:00:1d.7
> LANC	  S4	 disabled  pci:0000:06:08.0
> MODM	  S3	 disabled  pci:0000:00:1e.3
> localhost:~> 

OK, thanks.

Can you try to unload the USB controller drivers before hibernation and see
what happens?

Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!
 
> > localhost:~> cat /proc/acpi/wakeup 
> > Device	S-state	  Status   Sysfs node
> > RP01	  S4	 disabled  pci:0000:00:1c.0
> > RP02	  S4	 disabled  pci:0000:00:1c.1
> > RP04	  S4	 disabled  
> > USB1	  S3	 disabled  pci:0000:00:1d.0
> > USB2	  S3	 disabled  pci:0000:00:1d.1
> > USB3	  S3	 disabled  pci:0000:00:1d.2
> > USB4	  S3	 disabled  pci:0000:00:1d.3
> > USB7	  S3	 disabled  pci:0000:00:1d.7
> > LANC	  S4	 disabled  pci:0000:06:08.0
> > MODM	  S3	 disabled  pci:0000:00:1e.3
> > localhost:~> 
> 
> OK, thanks.
> 
> Can you try to unload the USB controller drivers before
> hibernation and see what happens?

Tested with 2.6.28 with all USB related modules unloaded and, as if by magic, hibernation worked! :)

Cheers!




      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Wednesday 07 January 2009, Daniel Qarras wrote:
> Hi!
>  
> > > localhost:~> cat /proc/acpi/wakeup 
> > > Device	S-state	  Status   Sysfs node
> > > RP01	  S4	 disabled  pci:0000:00:1c.0
> > > RP02	  S4	 disabled  pci:0000:00:1c.1
> > > RP04	  S4	 disabled  
> > > USB1	  S3	 disabled  pci:0000:00:1d.0
> > > USB2	  S3	 disabled  pci:0000:00:1d.1
> > > USB3	  S3	 disabled  pci:0000:00:1d.2
> > > USB4	  S3	 disabled  pci:0000:00:1d.3
> > > USB7	  S3	 disabled  pci:0000:00:1d.7
> > > LANC	  S4	 disabled  pci:0000:06:08.0
> > > MODM	  S3	 disabled  pci:0000:00:1e.3
> > > localhost:~> 
> > 
> > OK, thanks.
> > 
> > Can you try to unload the USB controller drivers before
> > hibernation and see what happens?
> 
> Tested with 2.6.28 with all USB related modules unloaded and, as if by magic, hibernation worked! :)

IIRC, it should be sufficient to unload ehci-hcd.

It is a chipset issue rather than anything else.

Please send the output of lspci -vv

Thanks,
Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!
 
> > Tested with 2.6.28 with all USB related modules
> > unloaded and, as if by magic, hibernation worked! :)
> 
> IIRC, it should be sufficient to unload ehci-hcd.
> 
> It is a chipset issue rather than anything else.
> 
> Please send the output of lspci -vv

Please find lspci -vvv output from my Fedora bug report:

https://bugzilla.redhat.com/attachment.cgi?id=326874

Thanks!




      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Wednesday 07 January 2009, Daniel Qarras wrote:
> Hi!
>  
> > > Tested with 2.6.28 with all USB related modules
> > > unloaded and, as if by magic, hibernation worked! :)
> > 
> > IIRC, it should be sufficient to unload ehci-hcd.
> > 
> > It is a chipset issue rather than anything else.
> > 
> > Please send the output of lspci -vv
> 
> Please find lspci -vvv output from my Fedora bug report:
> 
> https://bugzilla.redhat.com/attachment.cgi?id=326874

OK, please check if it is sufficient to unload ehci-hcd and/or uhci-hcd before
hibernation.

Thanks,
Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!

> > Please find lspci -vvv output from my Fedora bug report:
> > 
> > https://bugzilla.redhat.com/attachment.cgi?id=326874
> 
> OK, please check if it is sufficient to unload ehci-hcd
> and/or uhci-hcd before hibernation.

It is sufficient to unload ehci-hcd. Unloading uhci-hcd does not help.

Thanks!



      

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Rafael J. Wysocki]

On Thursday 08 January 2009, Daniel Qarras wrote:
> Hi!
> 
> > > Please find lspci -vvv output from my Fedora bug report:
> > > 
> > > https://bugzilla.redhat.com/attachment.cgi?id=326874
> > 
> > OK, please check if it is sufficient to unload ehci-hcd
> > and/or uhci-hcd before hibernation.
> 
> It is sufficient to unload ehci-hcd. Unloading uhci-hcd does not help.

Thanks for verifying.

I think we'll need to use a blacklist to work around this issue, but I need to
discuss this with other people.

Please ping me if I don't get back to you for some time.

Thanks,
Rafael

Re: Hibernation breaks between 2.6.23 and 2.6.24 on Acer Aspire 1690

[Daniel Qarras]

Hi!
 
> > It is sufficient to unload ehci-hcd. Unloading
> > uhci-hcd does not help.
> 
> Thanks for verifying.
> 
> I think we'll need to use a blacklist to work around
> this issue, but I need to discuss this with other people.
> 
> Please ping me if I don't get back to you for some time.

Ok, will do, thanks!