LCD hardware fading

[Richard Hughes]

On Fri, 2006-06-30 at 12:04 -0400, David Zeuthen wrote:
> On Sun, 2006-01-29 at 13:16 -0500, David Zeuthen wrote:
> > On Sun, 2006-01-29 at 12:09 -0500, David Zeuthen wrote:
> > > On Sun, 2006-01-29 at 17:03 +0000, Richard Hughes wrote:
> > > > How do you disable the fading? I was under the impression that this was
> > > > done by hardware attached to the lcd backlight, rather than the kernel
> > > > doing anything remotely clever.
> > > 
> > > For IBM just look at the "cleverness" in ibm-acpi.c:brightness_write() -
> > > the patch should be simple to write. Hopefully upstream will take it...
> > 
> > FYI I've submitted a patch for this on LKML
> > 
> >  http://lkml.org/lkml/2006/1/29/96/index.html
> 
> So the patch was rejected because some hardware only works with changing
> the value one at a time...

Yes, I saw this this morning.

> Richard, remind me why this is a problem? It seems that if we're fading
> from g-p-m _anyway_ we still get the same effect (e.g. we take only one
> step at a time). Is it perhaps because you are not waiting for the
> return value of the previous SetLCDBrightness() call before invoking the
> next?

I'm not sure, although I probably agree with you that logically there
should be no problem.

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=173382 details the
problem.

Richard.