Comment 193 for bug 1098216

(In reply to comment #66)
> (In reply to comment #65)
> > (In reply to comment #64)
> > > The upstream maintainers requested the first behavior
> >
> > Out of curiosity, who are the “upstream maintainers” (and why aren't they
> > commenting here, since we're on the kernel bugzilla)
>
> Here's the thread:
>
> http://www.spinics.net/lists/linux-acpi/msg42180.html
>
> And yeah, the term "upstream" was probably misused since we're on b.k.o ;-)

Thanks for the link. It's somehow weird that Henrique de Moraes Holschuh was not in the loop (I think he's subscribed to linux-acpi, but maybe not). He has a deep knowledge of Thinkpads.

Note that Matthew Garrett is wrong when he says:

> So the problem is that userspace is writing values that don't happen to
> be aligned with the values the hardware reacts to, and so nothing gets
> changed?

Here the value is not written by userspace but by the kernel. Not sure if it matters though.

>
> > > and probably aren't going
> > > to take the patch to quirk these machines with ascpi_osi="!Windows 2012".
> > > However, nothing prevents you from continuing to boot with this option to
> get
> > > the second behavior.
> >
> > Well, I can't test right now but are you sure I'll get that behavior with
> your
> > patches applied? On top of that, acpi_osi might have other side effects, so
> > it's not really a durable workaround I think.
>
> Actually you also have to pass video.brightness_switch_enabled=1.

Good point.

The thing is, every userspace daemon does this differently, and not everyone uses a userspace daemon (nor want to). And userspace daemon only works when it's running. All in all, deffering this to userspace seems to me like a bad workaround. I think the kernel should handle this as much as possible in order for people to get a consistent behavior.

I'm not too sure if Matthew Garrett, Ben Jencks and other people involved in the thread do read this bug, but I'm open to writing this on linux-acpi if needed (I'm not subscribed but I can get a copy of the initial mail if needed).