Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

[Rafael J. Wysocki]

On Wednesday, June 26, 2013 07:31:00 PM Cork wrote:
> Hello. This is my first time sending a report like this, so please excuse me if I'm missing some vital information.
> 
> From updating the kernel in Arch linux to 3.9 the brightness controls stopped working on my Clevo P0150HM. I have tried the acpi_backlight=vendor boot parameter but it makes no difference.
> 
> While researching this problem I've found that /sys/class/backlight is empty in the versions where the buttons isn't working. Bisecting points to 00c43b9682507dc622c03172fde1032e2a216e9d [1] as the first commit showing the symptoms.
> 
> Attached is the output from journalctl -b _TRANSFORM=kernel for the two versions. Not sure what other useful information I can provide apart from that, so please direct me if there is something missing.
> 
> [1] https://github.com/torvalds/linux/commit/00c43b9682507dc622c03172fde1032e2a216e9d

Interesting.

Please send the output of acpidump from your machine.

Thanks,
Rafael


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

Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

[Rafael J. Wysocki]

On Wednesday, June 26, 2013 09:52:46 PM Rafael J. Wysocki wrote:
> On Wednesday, June 26, 2013 07:31:00 PM Cork wrote:
> > Hello. This is my first time sending a report like this, so please excuse me if I'm missing some vital information.
> > 
> > From updating the kernel in Arch linux to 3.9 the brightness controls stopped working on my Clevo P0150HM. I have tried the acpi_backlight=vendor boot parameter but it makes no difference.
> > 
> > While researching this problem I've found that /sys/class/backlight is empty in the versions where the buttons isn't working. Bisecting points to 00c43b9682507dc622c03172fde1032e2a216e9d [1] as the first commit showing the symptoms.
> > 
> > Attached is the output from journalctl -b _TRANSFORM=kernel for the two versions. Not sure what other useful information I can provide apart from that, so please direct me if there is something missing.
> > 
> > [1] https://github.com/torvalds/linux/commit/00c43b9682507dc622c03172fde1032e2a216e9d
> 
> Interesting.
> 
> Please send the output of acpidump from your machine.

Also please put initcall_debug into the kernel command line and send dmesg from
the working and not working kernels with that.

Thanks,
Rafael


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

Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

[Rafael J. Wysocki]

On Wednesday, June 26, 2013 08:38:49 PM Cork wrote:
> // Cork

OK, thanks.

The initialization ordering obviously changed in 3.9 and the ACPI video
driver fails to find your backlight device.

Can you please tell me what the output of

$ find /sys/devices/ -name LNXVIDEO\*

is on your system in 3.9 and 3.8?

Rafael


> 
> ----- Original Message -----
> > From: "Rafael J. Wysocki" <rjw@sisk.pl>
> > To: "Cork" <kernel@box.qzdev.se>
> > Cc: linux-acpi@vger.kernel.org, "Aaron Lu" <aaron.lu@intel.com>
> > Sent: Wednesday, 26 June, 2013 10:05:38 PM
> > Subject: Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"
> > 
> > On Wednesday, June 26, 2013 09:52:46 PM Rafael J. Wysocki wrote:
> > > On Wednesday, June 26, 2013 07:31:00 PM Cork wrote:
> > > > Hello. This is my first time sending a report like this, so
> > > > please excuse me if I'm missing some vital information.
> > > > 
> > > > From updating the kernel in Arch linux to 3.9 the brightness
> > > > controls stopped working on my Clevo P0150HM. I have tried the
> > > > acpi_backlight=vendor boot parameter but it makes no difference.
> > > > 
> > > > While researching this problem I've found that
> > > > /sys/class/backlight is empty in the versions where the buttons
> > > > isn't working. Bisecting points to
> > > > 00c43b9682507dc622c03172fde1032e2a216e9d [1] as the first commit
> > > > showing the symptoms.
> > > > 
> > > > Attached is the output from journalctl -b _TRANSFORM=kernel for
> > > > the two versions. Not sure what other useful information I can
> > > > provide apart from that, so please direct me if there is
> > > > something missing.
> > > > 
> > > > [1]
> > > > https://github.com/torvalds/linux/commit/00c43b9682507dc622c03172fde1032e2a216e9d
> > > 
> > > Interesting.
> > > 
> > > Please send the output of acpidump from your machine.
> > 
> > Also please put initcall_debug into the kernel command line and send
> > dmesg from
> > the working and not working kernels with that.
> > 
> > Thanks,
> > Rafael
> > 
> > 
> > --
> > I speak only for myself.
> > Rafael J. Wysocki, Intel Open Source Technology Center.
> > 
-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

[Aaron Lu]

On 06/27/2013 06:08 AM, Rafael J. Wysocki wrote:
> On Wednesday, June 26, 2013 08:38:49 PM Cork wrote:
>> // Cork
> 
> OK, thanks.
> 
> The initialization ordering obviously changed in 3.9 and the ACPI video
> driver fails to find your backlight device.
> 
> Can you please tell me what the output of
> 
> $ find /sys/devices/ -name LNXVIDEO\*
> 
> is on your system in 3.9 and 3.8?

This should be the same problem as bug 58091, where one of the reporter
is using a similar system Clevo_P150HMx. And I've checked the ACPI dump
of Cork's system, there is also a _ROM method under PCI0.

Cork,
Possible to test v3.10-rc6 or later kernels? There is a fix for your
problem, thanks.

Reference: https://bugzilla.kernel.org/show_bug.cgi?id=58091

-Aaron

> 
> Rafael
> 
> 
>>
>> ----- Original Message -----
>>> From: "Rafael J. Wysocki" <rjw@sisk.pl>
>>> To: "Cork" <kernel@box.qzdev.se>
>>> Cc: linux-acpi@vger.kernel.org, "Aaron Lu" <aaron.lu@intel.com>
>>> Sent: Wednesday, 26 June, 2013 10:05:38 PM
>>> Subject: Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"
>>>
>>> On Wednesday, June 26, 2013 09:52:46 PM Rafael J. Wysocki wrote:
>>>> On Wednesday, June 26, 2013 07:31:00 PM Cork wrote:
>>>>> Hello. This is my first time sending a report like this, so
>>>>> please excuse me if I'm missing some vital information.
>>>>>
>>>>> From updating the kernel in Arch linux to 3.9 the brightness
>>>>> controls stopped working on my Clevo P0150HM. I have tried the
>>>>> acpi_backlight=vendor boot parameter but it makes no difference.
>>>>>
>>>>> While researching this problem I've found that
>>>>> /sys/class/backlight is empty in the versions where the buttons
>>>>> isn't working. Bisecting points to
>>>>> 00c43b9682507dc622c03172fde1032e2a216e9d [1] as the first commit
>>>>> showing the symptoms.
>>>>>
>>>>> Attached is the output from journalctl -b _TRANSFORM=kernel for
>>>>> the two versions. Not sure what other useful information I can
>>>>> provide apart from that, so please direct me if there is
>>>>> something missing.
>>>>>
>>>>> [1]
>>>>> https://github.com/torvalds/linux/commit/00c43b9682507dc622c03172fde1032e2a216e9d
>>>>
>>>> Interesting.
>>>>
>>>> Please send the output of acpidump from your machine.
>>>
>>> Also please put initcall_debug into the kernel command line and send
>>> dmesg from
>>> the working and not working kernels with that.
>>>
>>> Thanks,
>>> Rafael
>>>
>>>
>>> --
>>> I speak only for myself.
>>> Rafael J. Wysocki, Intel Open Source Technology Center.
>>>

Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

[Cork]

Yes, 3.10.0-rc6 fixes the problem (sorry for not checking this before contacting you).

Rafael:
3.8:      /sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:18/LNXVIDEO:00
3.9:      /sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:18/LNXVIDEO:00
3.10-rc6: /sys/devices/LNXSYSTM:00/device:00/PNP0A08:00/device:19/LNXVIDEO:00

// Cork

----- Original Message -----
> From: "Aaron Lu" <aaron.lu@intel.com>
> To: "Rafael J. Wysocki" <rjw@sisk.pl>, "Cork" <kernel@box.qzdev.se>
> Cc: linux-acpi@vger.kernel.org
> Sent: Thursday, 27 June, 2013 3:38:49 AM
> Subject: Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"
> 
> On 06/27/2013 06:08 AM, Rafael J. Wysocki wrote:
> > On Wednesday, June 26, 2013 08:38:49 PM Cork wrote:
> >> // Cork
> > 
> > OK, thanks.
> > 
> > The initialization ordering obviously changed in 3.9 and the ACPI
> > video
> > driver fails to find your backlight device.
> > 
> > Can you please tell me what the output of
> > 
> > $ find /sys/devices/ -name LNXVIDEO\*
> > 
> > is on your system in 3.9 and 3.8?
> 
> This should be the same problem as bug 58091, where one of the
> reporter
> is using a similar system Clevo_P150HMx. And I've checked the ACPI
> dump
> of Cork's system, there is also a _ROM method under PCI0.
> 
> Cork,
> Possible to test v3.10-rc6 or later kernels? There is a fix for your
> problem, thanks.
> 
> Reference: https://bugzilla.kernel.org/show_bug.cgi?id=58091
> 
> -Aaron
> 
> > 
> > Rafael
> > 
> > 
> >>
> >> ----- Original Message -----
> >>> From: "Rafael J. Wysocki" <rjw@sisk.pl>
> >>> To: "Cork" <kernel@box.qzdev.se>
> >>> Cc: linux-acpi@vger.kernel.org, "Aaron Lu" <aaron.lu@intel.com>
> >>> Sent: Wednesday, 26 June, 2013 10:05:38 PM
> >>> Subject: Re: Sys backlight is empty since landing of "ACPI / PCI:
> >>> Make PCI root driver use struct acpi_scan_handler"
> >>>
> >>> On Wednesday, June 26, 2013 09:52:46 PM Rafael J. Wysocki wrote:
> >>>> On Wednesday, June 26, 2013 07:31:00 PM Cork wrote:
> >>>>> Hello. This is my first time sending a report like this, so
> >>>>> please excuse me if I'm missing some vital information.
> >>>>>
> >>>>> From updating the kernel in Arch linux to 3.9 the brightness
> >>>>> controls stopped working on my Clevo P0150HM. I have tried the
> >>>>> acpi_backlight=vendor boot parameter but it makes no
> >>>>> difference.
> >>>>>
> >>>>> While researching this problem I've found that
> >>>>> /sys/class/backlight is empty in the versions where the buttons
> >>>>> isn't working. Bisecting points to
> >>>>> 00c43b9682507dc622c03172fde1032e2a216e9d [1] as the first
> >>>>> commit
> >>>>> showing the symptoms.
> >>>>>
> >>>>> Attached is the output from journalctl -b _TRANSFORM=kernel for
> >>>>> the two versions. Not sure what other useful information I can
> >>>>> provide apart from that, so please direct me if there is
> >>>>> something missing.
> >>>>>
> >>>>> [1]
> >>>>> https://github.com/torvalds/linux/commit/00c43b9682507dc622c03172fde1032e2a216e9d
> >>>>
> >>>> Interesting.
> >>>>
> >>>> Please send the output of acpidump from your machine.
> >>>
> >>> Also please put initcall_debug into the kernel command line and
> >>> send
> >>> dmesg from
> >>> the working and not working kernels with that.
> >>>
> >>> Thanks,
> >>> Rafael
> >>>
> >>>
> >>> --
> >>> I speak only for myself.
> >>> Rafael J. Wysocki, Intel Open Source Technology Center.
> >>>
> 
> 

Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"

[Rafael J. Wysocki]

On Thursday, June 27, 2013 05:40:06 AM Cork wrote:
> Yes, 3.10.0-rc6 fixes the problem (sorry for not checking this before contacting you).

OK, great!

Thanks,
Rafael


> Rafael:
> 3.8:      /sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:18/LNXVIDEO:00
> 3.9:      /sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:18/LNXVIDEO:00
> 3.10-rc6: /sys/devices/LNXSYSTM:00/device:00/PNP0A08:00/device:19/LNXVIDEO:00
> 
> // Cork
> 
> ----- Original Message -----
> > From: "Aaron Lu" <aaron.lu@intel.com>
> > To: "Rafael J. Wysocki" <rjw@sisk.pl>, "Cork" <kernel@box.qzdev.se>
> > Cc: linux-acpi@vger.kernel.org
> > Sent: Thursday, 27 June, 2013 3:38:49 AM
> > Subject: Re: Sys backlight is empty since landing of "ACPI / PCI: Make PCI root driver use struct acpi_scan_handler"
> > 
> > On 06/27/2013 06:08 AM, Rafael J. Wysocki wrote:
> > > On Wednesday, June 26, 2013 08:38:49 PM Cork wrote:
> > >> // Cork
> > > 
> > > OK, thanks.
> > > 
> > > The initialization ordering obviously changed in 3.9 and the ACPI
> > > video
> > > driver fails to find your backlight device.
> > > 
> > > Can you please tell me what the output of
> > > 
> > > $ find /sys/devices/ -name LNXVIDEO\*
> > > 
> > > is on your system in 3.9 and 3.8?
> > 
> > This should be the same problem as bug 58091, where one of the
> > reporter
> > is using a similar system Clevo_P150HMx. And I've checked the ACPI
> > dump
> > of Cork's system, there is also a _ROM method under PCI0.
> > 
> > Cork,
> > Possible to test v3.10-rc6 or later kernels? There is a fix for your
> > problem, thanks.
> > 
> > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=58091
> > 
> > -Aaron
> > 
> > > 
> > > Rafael
> > > 
> > > 
> > >>
> > >> ----- Original Message -----
> > >>> From: "Rafael J. Wysocki" <rjw@sisk.pl>
> > >>> To: "Cork" <kernel@box.qzdev.se>
> > >>> Cc: linux-acpi@vger.kernel.org, "Aaron Lu" <aaron.lu@intel.com>
> > >>> Sent: Wednesday, 26 June, 2013 10:05:38 PM
> > >>> Subject: Re: Sys backlight is empty since landing of "ACPI / PCI:
> > >>> Make PCI root driver use struct acpi_scan_handler"
> > >>>
> > >>> On Wednesday, June 26, 2013 09:52:46 PM Rafael J. Wysocki wrote:
> > >>>> On Wednesday, June 26, 2013 07:31:00 PM Cork wrote:
> > >>>>> Hello. This is my first time sending a report like this, so
> > >>>>> please excuse me if I'm missing some vital information.
> > >>>>>
> > >>>>> From updating the kernel in Arch linux to 3.9 the brightness
> > >>>>> controls stopped working on my Clevo P0150HM. I have tried the
> > >>>>> acpi_backlight=vendor boot parameter but it makes no
> > >>>>> difference.
> > >>>>>
> > >>>>> While researching this problem I've found that
> > >>>>> /sys/class/backlight is empty in the versions where the buttons
> > >>>>> isn't working. Bisecting points to
> > >>>>> 00c43b9682507dc622c03172fde1032e2a216e9d [1] as the first
> > >>>>> commit
> > >>>>> showing the symptoms.
> > >>>>>
> > >>>>> Attached is the output from journalctl -b _TRANSFORM=kernel for
> > >>>>> the two versions. Not sure what other useful information I can
> > >>>>> provide apart from that, so please direct me if there is
> > >>>>> something missing.
> > >>>>>
> > >>>>> [1]
> > >>>>> https://github.com/torvalds/linux/commit/00c43b9682507dc622c03172fde1032e2a216e9d
> > >>>>
> > >>>> Interesting.
> > >>>>
> > >>>> Please send the output of acpidump from your machine.
> > >>>
> > >>> Also please put initcall_debug into the kernel command line and
> > >>> send
> > >>> dmesg from
> > >>> the working and not working kernels with that.
> > >>>
> > >>> Thanks,
> > >>> Rafael
> > >>>
> > >>>
> > >>> --
> > >>> I speak only for myself.
> > >>> Rafael J. Wysocki, Intel Open Source Technology Center.
> > >>>
> > 
> > 
-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.