Hi Denis,

Denis Laxalde wrote:

> I cannot hibernate my Lenovo T500 laptop, either from the Gnome power
> management menu or using pm-hibernate. In either cases, a two-lines
> message displays on the console before the computer shuts down:
>
> legacy_resume(): pnp_bus_resume+0x0/0x65 returns -19
> PM: Device 00:0a failed to thaw: error -19

Sorry for the slow response. That sounds like a bug indeed. Does
this only happen with suspend-to-disk or with suspend-to-ram, too?

Please attach full dmesg output from booting, suspending and resuming,
plus the version of kernel package you have installed and output of

/usr/share/bug/linux-image-$(uname -r)/script 3>&1

so we can learn a little about your configuration.



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110824231507.GA26932@elie.sbx02827.chicail.waypo rt.net">http://lists.debian.org/20110824231507.GA26932@elie.sbx02827.chicail.waypo rt.net

Hi Jonathan,

Thanks for the followup.

Jonathan Nieder a écrit :
> Sorry for the slow response. That sounds like a bug indeed. Does
> this only happen with suspend-to-disk or with suspend-to-ram, too?

It happens only with suspend-to-disk, suspend-to-ram works fine.

> Please attach full dmesg output from booting, suspending and resuming,
> plus the version of kernel package you have installed and output of
>
> /usr/share/bug/linux-image-$(uname -r)/script 3>&1
>
> so we can learn a little about your configuration.

See attached. Let me know if you need more information.

--
Denis

found 623514 linux-2.6/3.0.0-1
quit

Denis Laxalde wrote:

> See attached. Let me know if you need more information.

Thanks. Does unloading ("modprobe -r") the e1000e driver before
hibernating help?



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825180944.GE5537@elie.gateway.2wire.net">http ://lists.debian.org/20110825180944.GE5537@elie.gateway.2wire.net

Jonathan Nieder a écrit :
> Does unloading ("modprobe -r") the e1000e driver before
> hibernating help?

No. It does not make any noticeable difference.

--
Denis



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825143634.0b32f80c@mail.gmail.com">http://lists.debian.org/20110825143634.0b32f80c@mail.gmail.com

Denis Laxalde wrote:
> Jonathan Nieder a écrit :

>> Does unloading ("modprobe -r") the e1000e driver before
>> hibernating help?
>
> No. It does not make any noticeable difference.

Stupid of me --- I was misreading the log. Thanks for checking.

Now that I look more closely, I suspect the dmesg you attached is from
trying to resume after an unsuccessful hibernation. What would be
most useful is to get as much information as possible somehow from the
_unsuccessful_ hibernation, or even better, to trigger a successful
hibernation and find what causes the difference between the two.

Some good advice can be found at

http://www.kernel.org/doc/Documentation/power/basic-pm-debugging.txt



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825184713.GB9838@elie.gateway.2wire.net">http ://lists.debian.org/20110825184713.GB9838@elie.gateway.2wire.net

Jonathan Nieder a écrit :
> Now that I look more closely, I suspect the dmesg you attached is from
> trying to resume after an unsuccessful hibernation. What would be
> most useful is to get as much information as possible somehow from the
> _unsuccessful_ hibernation, or even better, to trigger a successful
> hibernation and find what causes the difference between the two.
>
> Some good advice can be found at
>
> http://www.kernel.org/doc/Documentation/power/basic-pm-debugging.txt

Ok. None of 'reboot', 'platform' or 'shutdown' hibernation modes worked
so I went through all steps in 1a "Tests modes of hibernation"
successfully. I then tried step 1b "Testing minimal configuration" (I
append init=/bin/bash to the line starting with 'linux' in grub2 prompt,
is that ok?). There none of the hibernation modes initially worked. I
got a "bash: echo: write error: No such device" message. Then I typed
'swapon -a' (as suggested in dmesg) and invoked hibernation again. Then
the machine halted or rebooted (depending on the mode) normally but I'm
not sure what was supposed to happen here. (I mean, I did not expect to
get back to the init shell I was on, since I left the grub cmd line
untouched upon restart...) Any idea?

--
Denis



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825162108.62d88665@mail.gmail.com">http://lists.debian.org/20110825162108.62d88665@mail.gmail.com

Denis Laxalde wrote:

> Then
> the machine halted or rebooted (depending on the mode) normally but I'm
> not sure what was supposed to happen here. (I mean, I did not expect to
> get back to the init shell I was on, since I left the grub cmd line
> untouched upon restart...)

It was supposed to get back to the init shell (at least, it does so for
me). Alas.

Looking back at the messages you see on the console:

legacy_resume(): pnp_bus_resume+0x0/0x65 returns -19
PM: Device 00:0a failed to thaw: error -19

In the dmesg you sent, we can see which device that is:

tpm_tis 00:0a: 1.2 TPM (device-id 0x1020, rev-id 6)
tpm_tis 00:0a: Intel iTPM workaround enabled

Unloading the tpm_tis and tpm modules before hibernating might be
worth a try. There have also been some suspend-to-ram related tpm
fixes upstream recently, meaning a test of v3.1-rc1 or later has a
chance of fixing this (though I don't find it likely).



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825204808.GA2303@elie.gateway.2wire.net">http ://lists.debian.org/20110825204808.GA2303@elie.gateway.2wire.net

Jonathan Nieder a écrit :
> > Then
> > the machine halted or rebooted (depending on the mode) normally but I'm
> > not sure what was supposed to happen here. (I mean, I did not expect to
> > get back to the init shell I was on, since I left the grub cmd line
> > untouched upon restart...)
>
> It was supposed to get back to the init shell (at least, it does so for
> me). Alas.

Ok. So this test fails and according to the documentation the culprit
should be a statically compiled driver. I guess I could then try to
rebuild the kernel with more drivers as modules, but where to start?

> Looking back at the messages you see on the console:
>
> legacy_resume(): pnp_bus_resume+0x0/0x65 returns -19
> PM: Device 00:0a failed to thaw: error -19
>
> In the dmesg you sent, we can see which device that is:
>
> tpm_tis 00:0a: 1.2 TPM (device-id 0x1020, rev-id 6)
> tpm_tis 00:0a: Intel iTPM workaround enabled
>
> Unloading the tpm_tis and tpm modules before hibernating might be
> worth a try.

It does not help either.

--
Denis




--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825171607.38507677@mail.gmail.com">http://lists.debian.org/20110825171607.38507677@mail.gmail.com

Denis Laxalde wrote:

> Ok. So this test fails and according to the documentation the culprit
> should be a statically compiled driver. I guess I could then try to
> rebuild the kernel with more drivers as modules, but where to start?

Thanks! In modern times, initramfs-tools lets udev load some modules,
such as the disk and network driver, before calling init. So
disabling the tpm module entirely to rule out that cause could be a
comfort. "make localmodconfig" is supposed to create a configuration
with just the currently loaded modules enabled, which could be useful
for coming up with a good starting point.

Another thing to try would be booting into an even more minimal
configuration, without loading any modules or mounting the root
filesystem, by passing the parameter "break=top" on the kernel command
line[1]. It would presumably be necessary to load the appropriate
disk driver ("modprobe ahci") and enable swap ("swapon /dev/sda5") to
get useful results.

[1] see initramfs-tools(8), http://wiki.debian.org/InitramfsDebug
[2] http://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-kernel-org-package



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110825214005.GA18713@elie.gateway.2wire.net">htt p://lists.debian.org/20110825214005.GA18713@elie.gateway.2wire.net

Jonathan Nieder a écrit :
> Another thing to try would be booting into an even more minimal
> configuration, without loading any modules or mounting the root
> filesystem, by passing the parameter "break=top" on the kernel command
> line[1]. It would presumably be necessary to load the appropriate
> disk driver ("modprobe ahci") and enable swap ("swapon /dev/sda5") to
> get useful results.

I do not manage to enable swap because there's no /dev/sda*. How can I
get those?

Thanks.

--
Denis



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110826122216.2a44ee20@mail.gmail.com">http://lists.debian.org/20110826122216.2a44ee20@mail.gmail.com