-next: no resume from suspend

-next: no resume from suspend

[Jiri Slaby]

Hi,

since friday's -next (the last known to be working is the last monday's)
I cannot resume from suspend. The last thing I see with
no_console_suspend is:
i915: No ACPI video bus found

But I used to see the message always, so this is no difference. Any idea
before I start bisecting?

It is x86_64, core 2 duo, intel g33 GPU, 6G RAM.

thanks,
-- 
js
suse labs

Re: -next: no resume from suspend

[Rafael J. Wysocki]

On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
> Hi,
> 
> since friday's -next (the last known to be working is the last monday's)
> I cannot resume from suspend. The last thing I see with
> no_console_suspend is:
> i915: No ACPI video bus found
> 
> But I used to see the message always, so this is no difference. Any idea
> before I start bisecting?

Well, what's the next message you normally see with no_console_suspend?

Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: -next: no resume from suspend

[Jiri Slaby]

On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
>> Hi,
>>
>> since friday's -next (the last known to be working is the last monday's)
>> I cannot resume from suspend. The last thing I see with
>> no_console_suspend is:
>> i915: No ACPI video bus found
>>
>> But I used to see the message always, so this is no difference. Any idea
>> before I start bisecting?
> 
> Well, what's the next message you normally see with no_console_suspend?

[drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

Adding DRM and intel guys.

-- 
js
suse labs

Re: -next: no resume from suspend

[Jani Nikula]

On Mon, 14 Jan 2013, Jiri Slaby <jslaby@suse.cz> wrote:
> On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
>> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
>>> Hi,
>>>
>>> since friday's -next (the last known to be working is the last monday's)
>>> I cannot resume from suspend. The last thing I see with
>>> no_console_suspend is:
>>> i915: No ACPI video bus found
>>>
>>> But I used to see the message always, so this is no difference. Any idea
>>> before I start bisecting?
>> 
>> Well, what's the next message you normally see with no_console_suspend?
>
> [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
>
> Adding DRM and intel guys.

Hi Jiri, the dmesgs for both cases with drm.debug=0xe module param might
give us some clues.

Does it work without no_console_suspend?


BR,
Jani.

Re: -next: no resume from suspend

[Jiri Slaby]

On 01/16/2013 10:20 AM, Jani Nikula wrote:
> Hi Jiri, the dmesgs for both cases with drm.debug=0xe module param might
> give us some clues.
> 
> Does it work without no_console_suspend?

Hi, no, it does not work either way. But what I found out with debug
enabled, that it's a completely different problem: in ACPI. The kernel
is spinning acpi_rs_get_aml_length. I'll report details without CCing
DRM people.

thanks,
-- 
js
suse labs

Re: -next: no resume from suspend

[Jiri Slaby]

On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
>> Hi,
>>
>> since friday's -next (the last known to be working is the last monday's)
>> I cannot resume from suspend. The last thing I see with
>> no_console_suspend is:
>> i915: No ACPI video bus found
>>
>> But I used to see the message always, so this is no difference. Any idea
>> before I start bisecting?
> 
> Well, what's the next message you normally see with no_console_suspend?

As I wrote in the other message, the kernel is spinning in
acpi_rs_get_aml_length. There are weird resources in the list. For
serial 00:05 there is resource with type and length equal to 0 and for
parport_pc 00:06 there is resource type 58196295 with length of
134284152. See:
acpi_rs_get_aml_length: ffff8801b9f7b380 type=3 len=16
acpi_rs_get_aml_length: ffff8801b9f7b390 type=0 len=0
acpi_rs_get_aml_length: BAD ffff8801b9f7b390 type=0 len=0
serial 00:05: activation failed
dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
PM: Device 00:05 failed to resume: error -5
acpi_rs_get_aml_length: ffff8801b9f7b680 type=3 len=24
acpi_rs_get_aml_length: ffff8801b9f7b698 type=58196295 len=134284152
parport_pc 00:06: activation failed
dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
PM: Device 00:06 failed to resume: error -5

if (!length)
  return error;

makes it resume again with the errors above...

This was 3.8.0-rc3-next-20130114. Going to test 20130116 as there are
these patches:
616979e memory-hotplug: remove sysfs file of node
a41a27d ACPI / PM: remove leading whitespace from #ifdef
ace2342 ACPI / PM: Consolidate suspend-specific and hibernate-specific code
27a5fb8 ACPI / PM: Move device power management functions to device_pm.c
717a5d1 ACPI / PM: Export power states of ACPI devices via sysfs
ee8f808 ACPI / PM: Common string representations of device power states
44fed9b ACPI / PM: More visible function for retrieving device power states
81ed027 ACPI / PM: Change the way power transitions to D3cold are
carried out
113542c ACPI: Use system level attribute of wakeup power resources
c6fdf80 ACPI: Take power resource initialization errors into account
b6568e5 ACPI / scan: Consolidate extraction of power resources lists
8a0a4d6 ACPI / scan: Remove unnecessary initialization of local variables
5c24aa4 ACPI / scan: Move power state initialization to a separate routine
59e8500 ACPI / PM: Take order attribute of wakeup power resources into
account
f369253 ACPI / PM: Take order attribute of power resources into account
3170b40 ACPI: Do not use device power states of power resources
17ece8c ACPI: Drop power resources driver
5e4cd24 ACPI / scan: Treat power resources in a special way
5b32520 ACPI / scan: More straightforward preparation of ACPI device objects
5a62e76 ACPI / PM: Rework the handling of devices depending on power
resources
5993c46 ACPI: update ej_event interface to take acpi_device
05404d8 ACPI / scan: Add second pass to acpi_bus_trim()
cecdb19 ACPI / scan: Change the implementation of acpi_bus_trim()
ae28179 ACPI / scan: Drop the second argument of acpi_bus_trim()
b17b537 ACPI / scan: Drop the second argument of acpi_device_unregister()
a412a11 ACPI / glue: Fix build with ACPI_GLUE_DEBUG set
abe9921 ACPI / scan: Fix check of device_attach() return value.
e92a79b ACPI thermal: remove unnecessary newline from exception message
2b16c77 ACPI sysfs: remove unnecessary newline from exception
f1f37cb ACPI video: remove unnecessary newline from error messages
b43e106 ACPICA: Cleanup table handler naming conflicts.
42f8fb7 ACPICA: Source restructuring: split large files into 8 new files.

thanks,
-- 
js
suse labs

Re: -next: no resume from suspend

[Jiri Slaby]

On 01/16/2013 11:43 AM, Jiri Slaby wrote:
> On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
>> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
>>> Hi,
>>>
>>> since friday's -next (the last known to be working is the last monday's)
>>> I cannot resume from suspend. The last thing I see with
>>> no_console_suspend is:
>>> i915: No ACPI video bus found
>>>
>>> But I used to see the message always, so this is no difference. Any idea
>>> before I start bisecting?
>>
>> Well, what's the next message you normally see with no_console_suspend?
> 
> As I wrote in the other message, the kernel is spinning in
> acpi_rs_get_aml_length. There are weird resources in the list. For
> serial 00:05 there is resource with type and length equal to 0 and for
> parport_pc 00:06 there is resource type 58196295 with length of
> 134284152. See:
> acpi_rs_get_aml_length: ffff8801b9f7b380 type=3 len=16
> acpi_rs_get_aml_length: ffff8801b9f7b390 type=0 len=0
> acpi_rs_get_aml_length: BAD ffff8801b9f7b390 type=0 len=0
> serial 00:05: activation failed
> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
> PM: Device 00:05 failed to resume: error -5
> acpi_rs_get_aml_length: ffff8801b9f7b680 type=3 len=24
> acpi_rs_get_aml_length: ffff8801b9f7b698 type=58196295 len=134284152
> parport_pc 00:06: activation failed
> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
> PM: Device 00:06 failed to resume: error -5
> 
> if (!length)
>   return error;
> 
> makes it resume again with the errors above...
> 
> This was 3.8.0-rc3-next-20130114. Going to test 20130116 as there are
> these patches:

It did not help...

> thanks,
-- 
js
suse labs

Re: -next: no resume from suspend

[Rafael J. Wysocki]

On Wednesday, January 16, 2013 11:43:15 AM Jiri Slaby wrote:
> On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
> > On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
> >> Hi,
> >>
> >> since friday's -next (the last known to be working is the last monday's)
> >> I cannot resume from suspend. The last thing I see with
> >> no_console_suspend is:
> >> i915: No ACPI video bus found
> >>
> >> But I used to see the message always, so this is no difference. Any idea
> >> before I start bisecting?
> > 
> > Well, what's the next message you normally see with no_console_suspend?
> 
> As I wrote in the other message, the kernel is spinning in
> acpi_rs_get_aml_length. There are weird resources in the list. For
> serial 00:05 there is resource with type and length equal to 0 and for
> parport_pc 00:06 there is resource type 58196295 with length of
> 134284152. See:
> acpi_rs_get_aml_length: ffff8801b9f7b380 type=3 len=16
> acpi_rs_get_aml_length: ffff8801b9f7b390 type=0 len=0
> acpi_rs_get_aml_length: BAD ffff8801b9f7b390 type=0 len=0
> serial 00:05: activation failed
> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
> PM: Device 00:05 failed to resume: error -5
> acpi_rs_get_aml_length: ffff8801b9f7b680 type=3 len=24
> acpi_rs_get_aml_length: ffff8801b9f7b698 type=58196295 len=134284152
> parport_pc 00:06: activation failed
> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
> PM: Device 00:06 failed to resume: error -5
> 
> if (!length)
>   return error;
> 
> makes it resume again with the errors above...
> 
> This was 3.8.0-rc3-next-20130114. Going to test 20130116 as there are
> these patches:
> 616979e memory-hotplug: remove sysfs file of node
> a41a27d ACPI / PM: remove leading whitespace from #ifdef
> ace2342 ACPI / PM: Consolidate suspend-specific and hibernate-specific code
> 27a5fb8 ACPI / PM: Move device power management functions to device_pm.c
> 717a5d1 ACPI / PM: Export power states of ACPI devices via sysfs
> ee8f808 ACPI / PM: Common string representations of device power states
> 44fed9b ACPI / PM: More visible function for retrieving device power states
> 81ed027 ACPI / PM: Change the way power transitions to D3cold are
> carried out
> 113542c ACPI: Use system level attribute of wakeup power resources
> c6fdf80 ACPI: Take power resource initialization errors into account
> b6568e5 ACPI / scan: Consolidate extraction of power resources lists
> 8a0a4d6 ACPI / scan: Remove unnecessary initialization of local variables
> 5c24aa4 ACPI / scan: Move power state initialization to a separate routine
> 59e8500 ACPI / PM: Take order attribute of wakeup power resources into
> account
> f369253 ACPI / PM: Take order attribute of power resources into account
> 3170b40 ACPI: Do not use device power states of power resources
> 17ece8c ACPI: Drop power resources driver
> 5e4cd24 ACPI / scan: Treat power resources in a special way
> 5b32520 ACPI / scan: More straightforward preparation of ACPI device objects
> 5a62e76 ACPI / PM: Rework the handling of devices depending on power
> resources
> 5993c46 ACPI: update ej_event interface to take acpi_device
> 05404d8 ACPI / scan: Add second pass to acpi_bus_trim()
> cecdb19 ACPI / scan: Change the implementation of acpi_bus_trim()
> ae28179 ACPI / scan: Drop the second argument of acpi_bus_trim()
> b17b537 ACPI / scan: Drop the second argument of acpi_device_unregister()
> a412a11 ACPI / glue: Fix build with ACPI_GLUE_DEBUG set
> abe9921 ACPI / scan: Fix check of device_attach() return value.
> e92a79b ACPI thermal: remove unnecessary newline from exception message
> 2b16c77 ACPI sysfs: remove unnecessary newline from exception
> f1f37cb ACPI video: remove unnecessary newline from error messages
> b43e106 ACPICA: Cleanup table handler naming conflicts.
> 42f8fb7 ACPICA: Source restructuring: split large files into 8 new files.

None of the ACPI patches above has anything to do with resources.

There are two things you can try.  First, revert all of the ACPICA patches
and see if that helps.  Second, if that doesn't help, try to revert things in
the PCI tree (alternatively, you can try the PCI tree alone).

I have tested suspend/resume with the ACPI (including ACPICA) patches applied,
but I haven't tested the PCI tree.

Thanks,
Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: -next: no resume from suspend

[Jiri Slaby]

On 01/16/2013 03:00 PM, Rafael J. Wysocki wrote:
> On Wednesday, January 16, 2013 11:43:15 AM Jiri Slaby wrote:
>> On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
>>> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
>>>> Hi,
>>>>
>>>> since friday's -next (the last known to be working is the last monday's)
>>>> I cannot resume from suspend. The last thing I see with
>>>> no_console_suspend is:
>>>> i915: No ACPI video bus found
>>>>
>>>> But I used to see the message always, so this is no difference. Any idea
>>>> before I start bisecting?
>>>
>>> Well, what's the next message you normally see with no_console_suspend?
>>
>> As I wrote in the other message, the kernel is spinning in
>> acpi_rs_get_aml_length. There are weird resources in the list. For
>> serial 00:05 there is resource with type and length equal to 0 and for
>> parport_pc 00:06 there is resource type 58196295 with length of
>> 134284152. See:
>> acpi_rs_get_aml_length: ffff8801b9f7b380 type=3 len=16
>> acpi_rs_get_aml_length: ffff8801b9f7b390 type=0 len=0
>> acpi_rs_get_aml_length: BAD ffff8801b9f7b390 type=0 len=0
>> serial 00:05: activation failed
>> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
>> PM: Device 00:05 failed to resume: error -5
>> acpi_rs_get_aml_length: ffff8801b9f7b680 type=3 len=24
>> acpi_rs_get_aml_length: ffff8801b9f7b698 type=58196295 len=134284152
>> parport_pc 00:06: activation failed
>> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
>> PM: Device 00:06 failed to resume: error -5
>>
>> if (!length)
>>   return error;
>>
>> makes it resume again with the errors above...
>>
>> This was 3.8.0-rc3-next-20130114. 
> 
> There are two things you can try.  First, revert all of the ACPICA patches
> and see if that helps.  Second, if that doesn't help, try to revert things in
> the PCI tree (alternatively, you can try the PCI tree alone).
> 
> I have tested suspend/resume with the ACPI (including ACPICA) patches applied,
> but I haven't tested the PCI tree.

I bisected that. It is caused by the following commit:
commit 2a3ab82714043580505634ff5b9a743ff4237376
Author: Bjorn Helgaas <bhelgaas@google.com>
Date:   Fri Jan 4 22:38:10 2013 +0000

    PNPACPI: Use _CRS buffer directly as _SRS template

Acpidump is here if you are interested:
http://www.fi.muni.cz/~xslaby/sklad/adump

thanks,
-- 
js
suse labs

Re: -next: no resume from suspend

[Bjorn Helgaas]

On Wed, Jan 16, 2013 at 3:48 PM, Jiri Slaby <jslaby@suse.cz> wrote:
> On 01/16/2013 03:00 PM, Rafael J. Wysocki wrote:
>> On Wednesday, January 16, 2013 11:43:15 AM Jiri Slaby wrote:
>>> On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
>>>> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
>>>>> Hi,
>>>>>
>>>>> since friday's -next (the last known to be working is the last monday's)
>>>>> I cannot resume from suspend. The last thing I see with
>>>>> no_console_suspend is:
>>>>> i915: No ACPI video bus found
>>>>>
>>>>> But I used to see the message always, so this is no difference. Any idea
>>>>> before I start bisecting?
>>>>
>>>> Well, what's the next message you normally see with no_console_suspend?
>>>
>>> As I wrote in the other message, the kernel is spinning in
>>> acpi_rs_get_aml_length. There are weird resources in the list. For
>>> serial 00:05 there is resource with type and length equal to 0 and for
>>> parport_pc 00:06 there is resource type 58196295 with length of
>>> 134284152. See:
>>> acpi_rs_get_aml_length: ffff8801b9f7b380 type=3 len=16
>>> acpi_rs_get_aml_length: ffff8801b9f7b390 type=0 len=0
>>> acpi_rs_get_aml_length: BAD ffff8801b9f7b390 type=0 len=0
>>> serial 00:05: activation failed
>>> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
>>> PM: Device 00:05 failed to resume: error -5
>>> acpi_rs_get_aml_length: ffff8801b9f7b680 type=3 len=24
>>> acpi_rs_get_aml_length: ffff8801b9f7b698 type=58196295 len=134284152
>>> parport_pc 00:06: activation failed
>>> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
>>> PM: Device 00:06 failed to resume: error -5
>>>
>>> if (!length)
>>>   return error;
>>>
>>> makes it resume again with the errors above...
>>>
>>> This was 3.8.0-rc3-next-20130114.
>>
>> There are two things you can try.  First, revert all of the ACPICA patches
>> and see if that helps.  Second, if that doesn't help, try to revert things in
>> the PCI tree (alternatively, you can try the PCI tree alone).
>>
>> I have tested suspend/resume with the ACPI (including ACPICA) patches applied,
>> but I haven't tested the PCI tree.
>
> I bisected that. It is caused by the following commit:
> commit 2a3ab82714043580505634ff5b9a743ff4237376
> Author: Bjorn Helgaas <bhelgaas@google.com>
> Date:   Fri Jan 4 22:38:10 2013 +0000
>
>     PNPACPI: Use _CRS buffer directly as _SRS template
>
> Acpidump is here if you are interested:
> http://www.fi.muni.cz/~xslaby/sklad/adump

Crap, sorry Jiri.

Rafael, I don't really care about that patch, so you can just drop it
unless you want to debug it.

Bjorn

Re: -next: no resume from suspend

[Rafael J. Wysocki]

On Wednesday, January 16, 2013 04:01:23 PM Bjorn Helgaas wrote:
> On Wed, Jan 16, 2013 at 3:48 PM, Jiri Slaby <jslaby@suse.cz> wrote:
> > On 01/16/2013 03:00 PM, Rafael J. Wysocki wrote:
> >> On Wednesday, January 16, 2013 11:43:15 AM Jiri Slaby wrote:
> >>> On 01/14/2013 01:56 PM, Rafael J. Wysocki wrote:
> >>>> On Monday, January 14, 2013 11:11:52 AM Jiri Slaby wrote:
> >>>>> Hi,
> >>>>>
> >>>>> since friday's -next (the last known to be working is the last monday's)
> >>>>> I cannot resume from suspend. The last thing I see with
> >>>>> no_console_suspend is:
> >>>>> i915: No ACPI video bus found
> >>>>>
> >>>>> But I used to see the message always, so this is no difference. Any idea
> >>>>> before I start bisecting?
> >>>>
> >>>> Well, what's the next message you normally see with no_console_suspend?
> >>>
> >>> As I wrote in the other message, the kernel is spinning in
> >>> acpi_rs_get_aml_length. There are weird resources in the list. For
> >>> serial 00:05 there is resource with type and length equal to 0 and for
> >>> parport_pc 00:06 there is resource type 58196295 with length of
> >>> 134284152. See:
> >>> acpi_rs_get_aml_length: ffff8801b9f7b380 type=3 len=16
> >>> acpi_rs_get_aml_length: ffff8801b9f7b390 type=0 len=0
> >>> acpi_rs_get_aml_length: BAD ffff8801b9f7b390 type=0 len=0
> >>> serial 00:05: activation failed
> >>> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
> >>> PM: Device 00:05 failed to resume: error -5
> >>> acpi_rs_get_aml_length: ffff8801b9f7b680 type=3 len=24
> >>> acpi_rs_get_aml_length: ffff8801b9f7b698 type=58196295 len=134284152
> >>> parport_pc 00:06: activation failed
> >>> dpm_run_callback(): pnp_bus_resume+0x0/0x80 returns -5
> >>> PM: Device 00:06 failed to resume: error -5
> >>>
> >>> if (!length)
> >>>   return error;
> >>>
> >>> makes it resume again with the errors above...
> >>>
> >>> This was 3.8.0-rc3-next-20130114.
> >>
> >> There are two things you can try.  First, revert all of the ACPICA patches
> >> and see if that helps.  Second, if that doesn't help, try to revert things in
> >> the PCI tree (alternatively, you can try the PCI tree alone).
> >>
> >> I have tested suspend/resume with the ACPI (including ACPICA) patches applied,
> >> but I haven't tested the PCI tree.
> >
> > I bisected that. It is caused by the following commit:
> > commit 2a3ab82714043580505634ff5b9a743ff4237376
> > Author: Bjorn Helgaas <bhelgaas@google.com>
> > Date:   Fri Jan 4 22:38:10 2013 +0000
> >
> >     PNPACPI: Use _CRS buffer directly as _SRS template
> >
> > Acpidump is here if you are interested:
> > http://www.fi.muni.cz/~xslaby/sklad/adump
> 
> Crap, sorry Jiri.
> 
> Rafael, I don't really care about that patch, so you can just drop it
> unless you want to debug it.

OK, dropping.

Thanks,
Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.