Re: [PATCH] ACPI / video: Add HP EliteBook Revolve 810 to the blacklist

From: Aaron Lu
Date: Wed Feb 12 2014 - 20:09:58 EST


On 02/12/2014 05:34 PM, Mika Westerberg wrote:
> On Wed, Feb 12, 2014 at 10:31:59AM +0800, Aaron Lu wrote:
>> On 02/07/2014 06:25 PM, Mika Westerberg wrote:
>>> On Fri, Feb 07, 2014 at 09:29:58AM +0800, Aaron Lu wrote:
>>>> On 02/03/2014 06:59 PM, Mika Westerberg wrote:
>>>>> On HP EliteBook Revolve 810 the ACPI backlight device doesn't work as
>>>>> expected. For example when resuming from system sleep, it seems to lose
>>>>> backlight settings.
>>>>>
>>>>> Forcing Intel driver fixes the problem so add this machine the ACPI
>>>>> video detect blacklist.
>>>>
>>>> For reference's purpose, can you please file a bug to kernel bugzilla
>>>> under the ACPI/Power-Video category and attach the acpidump and dmesg
>>>> there? Thanks.
>>>
>>> Done, the bug is here:
>>>
>>> https://bugzilla.kernel.org/show_bug.cgi?id=70231
>>>
>>> I noticed that using only the acpi_video0 for tuning backlight works but if
>>> I use intel_backlight directly, on resume I get values from acpi_video0
>>> instead. I'm not really familiar how this is supposed to work, though so
>>> please let me know if I'm missing something.
>>
>> Thanks for the report, as communicated on the bug page, your system is
>> already blacklisted for win8 by commit:
>>
>> commit 2d4054d8422462cb2771fdb4eb1925df55d2b320
>> Author: Takashi Iwai <tiwai@xxxxxxx> Fri Dec 20 00:09:12 2013
>> Committer: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx> Fri Dec 20 22:50:56 2013
>>
>> ACPI: Blacklist Win8 OSI for some HP laptop 2013 models
>>
>> And the fact that acpi_video interface works made it not appropriate to
>> add your system into the video_detect_dmi_table. You can just use acpi_video
>> interface and everything should work fine.
>
> OK, thanks for the investigation Aaron.
>
> I have few questions, though.
>
> 1) Is it always right thing to do to use acpi_video0 over anything else?

Not necessarily I think. X will pick an interface to use and AFAIK, the X
driver for i915 favour acpi_video over the native one for some reason. But
of course you can override that in xorg.conf, I think this is the way you
have done, right?

> In my case that device first disapeared and then reappeared so I got kind

It is disappeared by some of my patches that didn't get into mainline :-)

> of confused about that. I'm going to tune my scripts to use acpi_video0 as
> you suggested.

It seems to be the guideline that kernel just provides functionality
without doing any decision. From this perspective, it seems to be a
correct thing for kernel to expose as many interfaces as possible as
long as they are not broken and leaving the decision on which interface
to use to user space. The fact that ACPI video driver will restore
backlight setting according to its own knowledge is not a right thing
to do IMO, how does it know that user space is using it? I would like
to remove that in video's resume callback but then I guess some user
would report regressions since I suppose the GUI has relied on this
behaviour and doesn't do anything on resume.

>
> 2) What should we do with this commit? It is already in mainline so I
> guess revert is the sane thing to do.

I agree.

>
> I should have asked/investigated this more before submitting the patch.
> Sorry about that.

Never mind, the handling of backlight is in a confusing state anyway.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/