4.1-rc7, thinkpad X60: platform mode hibernation does very wrong thing

4.1-rc7, thinkpad X60: platform mode hibernation does very wrong thing

[Pavel Machek]

Hi!

On thinkpad X60, if I hibernate in platform mode, it does not power
down, but I get black screen, moon icon blinking, and CPU fan running
(CPU probably too).

I can hold the power button down to force powerdown... if I notice
what is going on before the battery is empty.

Any ideas what could be wrong there?

platform mode hibernation works ok on thinkpad T40p, and shutdown mode
hibernation works ok on thinkpad X60 with the same kernel.

Best regards,
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Rafael J. Wysocki]

On Wednesday, June 17, 2015 08:19:37 PM linus wrote:
> there might be something wrong with driver, can you provide serial log with no_console_suspend? or using rtctrace
> 
> ---Original---
> From: "Pavel Machek"<pavel@ucw.cz>
> Date: 2015/06/17 20:15:28
> To: "rjw"<rjw@rjwysocki.net>;"lenb"<lenb@kernel.org>;"kernel list"<linux-kernel@vger.kernel.org>;"linux-acpi"<linux-acpi@vger.kernel.org>;"linux-pm"<linux-pm@vger.kernel.org>;
> Subject: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing
> 
> 
> Hi!
> 
> On thinkpad X60, if I hibernate in platform mode, it does not power
> down, but I get black screen, moon icon blinking, and CPU fan running
> (CPU probably too).
> 
> I can hold the power button down to force powerdown... if I notice
> what is going on before the battery is empty.
> 
> Any ideas what could be wrong there?

Not right away.

Have you tried any previous 4.1-rc kernels?

Rafael

Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Pavel Machek]

Hi!

> On Wednesday, June 17, 2015 08:19:37 PM linus wrote:
> > there might be something wrong with driver, can you provide serial log with no_console_suspend? or using rtctrace
> >

Who are you?

The hibernation image is stored successfully, that means drivers work
ok. Serial console... might be fun to set up, but might help, right.

> > Hi!
> > 
> > On thinkpad X60, if I hibernate in platform mode, it does not power
> > down, but I get black screen, moon icon blinking, and CPU fan running
> > (CPU probably too).
> > 
> > I can hold the power button down to force powerdown... if I notice
> > what is going on before the battery is empty.
> > 
> > Any ideas what could be wrong there?
> 
> Not right away.
> 
> Have you tried any previous 4.1-rc kernels?

Yes, and it was broken there, too. It is possible that 4.0 is broken,
too, I did not use hibernation for some time.

3.14 kernel works ok. I'm trying v3.19 now.

Best regards,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Rafael J. Wysocki]

On Wednesday, June 17, 2015 02:33:33 PM Pavel Machek wrote:
> Hi!
> 
> > On Wednesday, June 17, 2015 08:19:37 PM linus wrote:
> > > there might be something wrong with driver, can you provide serial log with no_console_suspend? or using rtctrace
> > >
> 
> Who are you?
> 
> The hibernation image is stored successfully, that means drivers work ok.

Well, not entirely necessarily.  We execute the ->poweroff callback for all
of them in the last step, maybe one of these hangs.

You can verify that by adding a reboot trigger after the dpm_suspend_end()
invocation in hibernation_platform_enter().  If it triggers 100% of the time,
drivers are OK.

> Serial console... might be fun to set up, but might help, right.
> 
> > > Hi!
> > > 
> > > On thinkpad X60, if I hibernate in platform mode, it does not power
> > > down, but I get black screen, moon icon blinking, and CPU fan running
> > > (CPU probably too).
> > > 
> > > I can hold the power button down to force powerdown... if I notice
> > > what is going on before the battery is empty.
> > > 
> > > Any ideas what could be wrong there?
> > 
> > Not right away.
> > 
> > Have you tried any previous 4.1-rc kernels?
> 
> Yes, and it was broken there, too. It is possible that 4.0 is broken,
> too, I did not use hibernation for some time.
> 
> 3.14 kernel works ok. I'm trying v3.19 now.

I see.

OK, it would be good to narrow down the time frame when it broke.

Thanks,
Rafael

Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Pavel Machek]

On Wed 2015-06-17 15:14:02, Rafael J. Wysocki wrote:
> On Wednesday, June 17, 2015 02:33:33 PM Pavel Machek wrote:
> > Hi!
> > 
> > > On Wednesday, June 17, 2015 08:19:37 PM linus wrote:
> > > > there might be something wrong with driver, can you provide serial log with no_console_suspend? or using rtctrace
> > > >
> > 
> > Who are you?
> > 
> > The hibernation image is stored successfully, that means drivers work ok.
> 
> Well, not entirely necessarily.  We execute the ->poweroff callback for all
> of them in the last step, maybe one of these hangs.

shutdown mode works ok, that's why I think driver problem is not too
likely.

> > Yes, and it was broken there, too. It is possible that 4.0 is broken,
> > too, I did not use hibernation for some time.
> > 
> > 3.14 kernel works ok. I'm trying v3.19 now.
> 
> I see.
> 
> OK, it would be good to narrow down the time frame when it broke.

3.19 is broken. Some kind of strange 3.17 works ok.

(On an additional note, x60 fails to boot when "hot" halting at

"ACPI: All ACPI Tables successfully acquired"

That was happening before, but it is really annoying for debugging. As
you do two reboots, machine gets hot enough, and you have to wait to
cool it down. Any ideas what to do there?)

I tried no_console_suspend=1, but it still suspended the console.
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Paul Bolle]

On Wed, 2015-06-17 at 15:16 +0200, Pavel Machek wrote:
> 3.19 is broken. Some kind of strange 3.17 works ok.
> 
> (On an additional note, x60 fails to boot when "hot" halting at
> 
> "ACPI: All ACPI Tables successfully acquired"
> 
> That was happening before, but it is really annoying for debugging. As
> you do two reboots, machine gets hot enough, and you have to wait to
> cool it down. Any ideas what to do there?)
> 
> I tried no_console_suspend=1, but it still suspended the console.

And this all reminded me of https://lkml.org/lkml/2015/3/18/133 . So I
think that's another system not covered by commit ab3be73fa7b4
("drm/i915: gen4: work around hang during hibernation").

Hope this helps,


Paul Bolle

da2bc1b9db3351addd293e5b82757efe1f77ed1d breaks hibernation was Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Pavel Machek]

Hi!

> And this all reminded me of https://lkml.org/lkml/2015/3/18/133 . So I
> think that's another system not covered by commit ab3be73fa7b4
> ("drm/i915: gen4: work around hang during hibernation").

Yes, another system broken by

commit da2bc1b9db3351addd293e5b82757efe1f77ed1d
Author: Imre Deak <imre.deak@intel.com>
Date:   Thu Oct 23 19:23:26 2014 +0300

    drm/i915: add poweroff_late handler

    The suspend_late handler saves some registers and powers off the
    device,
        so it doesn't have a big overhead. Calling it at S4
    poweroff_late time
        makes the power off handling identical to the S3 suspend and
    S4 freeze
        handling, so do this for consistency.
	

And no, adding another system to the "blacklist" is a not good idea,
because we know there are other broken systems out there. Can we
revert da2bc1b9db3351addd293e5b82757efe1f77ed1d, please? Consistency
is nice, but not having regressions is even nicer.
									Pavel
									
PS: Now I have real Linus and fake Linus on the Cc list. Or maybe two
real Linuses, its hard to tell.

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: da2bc1b9db3351addd293e5b82757efe1f77ed1d breaks hibernation was Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Pavel Machek]

Hi!

> > And this all reminded me of https://lkml.org/lkml/2015/3/18/133 . So I
> > think that's another system not covered by commit ab3be73fa7b4
> > ("drm/i915: gen4: work around hang during hibernation").
> 
> Yes, another system broken by
> 
> commit da2bc1b9db3351addd293e5b82757efe1f77ed1d
> Author: Imre Deak <imre.deak@intel.com>
> Date:   Thu Oct 23 19:23:26 2014 +0300

Imre, your patch already broken 3 machines, and now we discovered it
breakes another. Can you please revert it before it causes more
damage?

Thanks,
									Pavel


>     drm/i915: add poweroff_late handler
> 
>     The suspend_late handler saves some registers and powers off the
>     device,
>         so it doesn't have a big overhead. Calling it at S4
>     poweroff_late time
>         makes the power off handling identical to the S3 suspend and
>     S4 freeze
>         handling, so do this for consistency.
> 	
> 
> And no, adding another system to the "blacklist" is a not good idea,
> because we know there are other broken systems out there. Can we
> revert da2bc1b9db3351addd293e5b82757efe1f77ed1d, please? Consistency
> is nice, but not having regressions is even nicer.
> 									Pavel
> 									
> PS: Now I have real Linus and fake Linus on the Cc list. Or maybe two
> real Linuses, its hard to tell.
> 

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: da2bc1b9db3351addd293e5b82757efe1f77ed1d breaks hibernation was Re: 4.1-rc7, thinkpad X60: platform mode hibernation does very wrongthing

[Rafael J. Wysocki]

On Tuesday, June 30, 2015 11:44:55 AM Pavel Machek wrote:
> Hi!
> 
> > > And this all reminded me of https://lkml.org/lkml/2015/3/18/133 . So I
> > > think that's another system not covered by commit ab3be73fa7b4
> > > ("drm/i915: gen4: work around hang during hibernation").
> > 
> > Yes, another system broken by
> > 
> > commit da2bc1b9db3351addd293e5b82757efe1f77ed1d
> > Author: Imre Deak <imre.deak@intel.com>
> > Date:   Thu Oct 23 19:23:26 2014 +0300
> 
> Imre, your patch already broken 3 machines, and now we discovered it
> breakes another. Can you please revert it before it causes more
> damage?

It looks like Imre has just sent a fix for that.  Can you please test that one?

Rafael