ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Alan Stern]

Rui:

Why does CONFIG_ACPI_VIDEO depend on BACKLIGHT_CLASS_DEVICE?  Doesn't
ACPI_VIDEO apply to any sort of video adapter, including plain old VGA
adapters attached to a non-LCD device?

Alan Stern

Re: ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Zhang Rui]

On Thu, 2010-11-11 at 06:02 +0800, Alan Stern wrote:
> Rui:
> 
> Why does CONFIG_ACPI_VIDEO depend on BACKLIGHT_CLASS_DEVICE?  Doesn't
> ACPI_VIDEO apply to any sort of video adapter, including plain old VGA
> adapters attached to a non-LCD device?
> 
There is an ACPI device node (an ACPI video bus device) for the VGA
controller and another ACPI device node (an ACPI video output device)
for the LCD display.
And ACPI_VIDEO knows how to switch the backlight of this integrated
LCD. :)

thanks,
rui

Re: ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Alan Stern]

On Fri, 12 Nov 2010, Zhang Rui wrote:

> On Thu, 2010-11-11 at 06:02 +0800, Alan Stern wrote:
> > Rui:
> > 
> > Why does CONFIG_ACPI_VIDEO depend on BACKLIGHT_CLASS_DEVICE?  Doesn't
> > ACPI_VIDEO apply to any sort of video adapter, including plain old VGA
> > adapters attached to a non-LCD device?
> > 
> There is an ACPI device node (an ACPI video bus device) for the VGA
> controller and another ACPI device node (an ACPI video output device)
> for the LCD display.
> And ACPI_VIDEO knows how to switch the backlight of this integrated
> LCD. :)

But my computer doesn't have an LCD display.  It's a desktop, not a
laptop.  So why do I need BACKLIGHT_CLASS_DEVICE to be enabled?

Are you saying I don't need ACPI_VIDEO if the machine doesn't have an
integrated LCD display?  That doesn't agree with the Kconfig help
message.

Alan Stern

Re: ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Matthew Garrett]

On Thu, Nov 11, 2010 at 10:00:03PM -0500, Alan Stern wrote:

> But my computer doesn't have an LCD display.  It's a desktop, not a
> laptop.  So why do I need BACKLIGHT_CLASS_DEVICE to be enabled?

Well, in the near future ddc-ci will be hooked up and you'll have 
backlight control over VGA if your monitor supports it, so even if we 
fix this up now you'll just end up getting it back in the near future 
because we can't tell that at build-time...

> Are you saying I don't need ACPI_VIDEO if the machine doesn't have an
> integrated LCD display?  That doesn't agree with the Kconfig help
> message.

But to that, yes. The only reason to use ACPI_VIDEO is to get EDID from 
firmware (which isn't relevant on desktops) or to control backlights. 
Every other theoretical feature it has is useless in the real world.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Alan Stern]

On Fri, 12 Nov 2010, Matthew Garrett wrote:

> On Thu, Nov 11, 2010 at 10:00:03PM -0500, Alan Stern wrote:
> 
> > But my computer doesn't have an LCD display.  It's a desktop, not a
> > laptop.  So why do I need BACKLIGHT_CLASS_DEVICE to be enabled?
> 
> Well, in the near future ddc-ci will be hooked up and you'll have 
> backlight control over VGA if your monitor supports it, so even if we 
> fix this up now you'll just end up getting it back in the near future 
> because we can't tell that at build-time...
> 
> > Are you saying I don't need ACPI_VIDEO if the machine doesn't have an
> > integrated LCD display?  That doesn't agree with the Kconfig help
> > message.
> 
> But to that, yes. The only reason to use ACPI_VIDEO is to get EDID from 
> firmware (which isn't relevant on desktops) or to control backlights. 
> Every other theoretical feature it has is useless in the real world.

Thanks for the information.  Why does the help text for ACPI_VIDEO 
mention defining the POST device or setting up a video output if the 
driver doesn't do these things?

Alan Stern

Re: ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Matthew Garrett]

On Fri, Nov 12, 2010 at 04:26:15PM -0500, Alan Stern wrote:

> Thanks for the information.  Why does the help text for ACPI_VIDEO 
> mention defining the POST device or setting up a video output if the 
> driver doesn't do these things?

The help text is documenting what the spec was intended to allow, rather 
than what it tends to in the real world...

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: ACPI_VIDEO and BACKLIGHT_CLASS_DEVICE

[Zhang Rui]

On Sat, 2010-11-13 at 05:36 +0800, Matthew Garrett wrote:
> On Fri, Nov 12, 2010 at 04:26:15PM -0500, Alan Stern wrote:
> 
> > Thanks for the information.  Why does the help text for ACPI_VIDEO 
> > mention defining the POST device or setting up a video output if the 
> > driver doesn't do these things?
> 
> The help text is documenting what the spec was intended to allow, rather 
> than what it tends to in the real world...
> 
At least the video driver is trying to do so right now, e.g. create
proc/sysfs I/F for output device switching, although they doesn't work
actually.
And I was preparing a patch to removing this stuff, together with an
update of the ACPI_VIDEO help text.

thanks,
rui