Re: ACPI Issues on Lifebook P7120 (with fixes)

[Grégoire Sutre <gregoire.sutre%labri.fr@localhost>]

Jared D. McNeill wrote:
> Looks mostly fine with me, although I would prefer if you could 
> integrate brightness handling with the display switching code in the 
> event that some obscure laptop actually implements that part of the spec.

Sure, I can try to integrate the two.

We would need a way to let the user choose whether (1) to let the BIOS 
perform display switching or (2) take over (_DOS call in attach funtion 
with arg 0x0).  Would a kernel configuration option be ok for this?

I forgot to mention that the laptop-specific drivers that listen to 
PMFE_DISPLAY_BRIGHTNESS_UP/DOWN events might end up processing 
brightness events twice if acpidisplay is enabled in the kernel.  This 
concerns sony_acpi.c and thinkpad_acpi.c.

I'm not sure how this can be properly fixed.  I would propose to only 
rely on ACPI notifications on output devices (as in acpidisplay), except 
for laptops that do not send notifications there (do you know of any?).

Grégoire