Re: Brightness buttons broken on Toshiba Satellite Proc A100 in kernel 2.6.23 (Fedora8)

[Danny Baumann <dannybaumann@web.de>]

Hi,

> > > Ok, I played around with it some more. It seems that on my Laptop the
> > > firmware sends out notifications on Fn+F6/7 press.
> > > Kernel 2.6.23-rc2 worked fine, -rc3 is broken. Reverting ec.c in the
> > > current Fedora kernel to the version in 2.6.23-rc2 makes brightness
> > > control work again. After finding that, I tried the current HEAD version
> > > of ec.c - unfortunately to no avail.
> > > 2.6.23-rc3 doesn't show up the (unwanted) key events for Fn+F6/7 though
> > > - this behaviour was introduced in -rc4.
> > > 
> > > Is it possible to fix the problem with the amount of information
> > > provided? If no, what additional information can I provide?
> > Please check if attached patch changes situation.
> 
> Yes, it does. With that patch, everything works as intended.
> Gnome-power-manager doesn't seem to work properly with it, but that's a
> different story ;-)

Well, it is a different story. A fix for the gpm problem is at
http://bugzilla.kernel.org/show_bug.cgi?id=9277, but the underlying
problem is a different one that I need some help with:
The root problem is that I get two notifications per brightness hotkey
press. One is for video bus VGA, device LCD; the other one is for bus
GFX0, device LCD. The LCD device on GFX0 is of type 0x400 (internal
flat-panel), the one on VGA is of type 0x110 (external CRT - for
backwards compatibility). So I can somehow understand why Toshiba
duplicated the device in two busses, but the outcome of that is somewhat
inconvenient because it halves the amount of brightness levels I have
available ;-)

Now my question is: Can anything be done inside the kernel against this
kind of problem? I can hardly imagine that Toshiba is the only company
that duplicates the LCD device for backwards compatibility (as this is
even given in the example IDs in table B-3 in the ACPI spec). I thought
about that for some time, but didn't come up with a good idea so far.
Maybe an idea would be to prevent the inclusion of certain ACPI paths
into the namespace by a kernel command line parameter?

Thanks,

Danny