Re: [PATCH] ACPI video: Be more liberal in validating _BQC behaviour

[Matthew Garrett <mjg59@srcf.ucam.org>]

On Mon, Feb 22, 2010 at 09:51:19AM +0800, Zhang Rui wrote:
> On Fri, 2010-02-19 at 21:54 +0800, Matthew Garrett wrote:
> > My system statically initialises the variable containing the current 
> > brightness to 100, but doesn't include 100 in the list of valid 
> > brightnesses.
> 
> sorry, I don't understand.
> does the video driver set the backlight to 100, which is not a valid
> value in the _BCL package?

No, the firmware does.

> >  Right now this causes us to stop believing _BQC. However, 
> > the enxt thing we do is set the brightness to maximum anyway - at this 
> > point _BQC will now return a correct value.
> 
> hmmm, could you attach the acpidump please?

The relevant bits are:

                    Name (BRIG, 0x64)
                    Method (_BQC, 0, Serialized)
                    {
                        Store (BRIG, Local0)
                        Return (Local0)
                    }
                    Method (_BCM, 1, Serialized)
                    {
			...
                        Store (Arg0, BRIG)

0x64 is an invalid value as far as _BCL goes. So, _BQC will give an 
invalid response until we set a value - after that, it'll be correct.

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