Re: [PATCH v3 2/4] backlight: Expose brightness curve type through sysfs

[Daniel Thompson <daniel.thompson@linaro.org>]

On Mon, Aug 19, 2019 at 11:50:49AM -0700, Matthias Kaehlcke wrote:
> Hi Daniel,
> 
> On Mon, Aug 19, 2019 at 11:02:41AM +0100, Daniel Thompson wrote:
> > On Fri, Aug 16, 2019 at 10:53:17AM -0700, Matthias Kaehlcke wrote:
> > > On Fri, Aug 16, 2019 at 04:54:18PM +0100, Daniel Thompson wrote:
> > > > On 07/08/2019 21:15, Matthias Kaehlcke wrote:
> > > > > On Tue, Jul 09, 2019 at 12:00:05PM -0700, Matthias Kaehlcke wrote:
> > > > > > Backlight brightness curves can have different shapes. The two main
> > > > > > types are linear and non-linear curves. The human eye doesn't
> > > > > > perceive linearly increasing/decreasing brightness as linear (see
> > > > > > also 88ba95bedb79 "backlight: pwm_bl: Compute brightness of LED
> > > > > > linearly to human eye"), hence many backlights use non-linear (often
> > > > > > logarithmic) brightness curves. The type of curve currently is opaque
> > > > > > to userspace, so userspace often uses more or less reliable heuristics
> > > > > > (like the number of brightness levels) to decide whether to treat a
> > > > > > backlight device as linear or non-linear.
> > > > > > 
> > > > > > Export the type of the brightness curve via the new sysfs attribute
> > > > > > 'scale'. The value of the attribute can be 'linear', 'non-linear' or
> > > > > > 'unknown'. For devices that don't provide information about the scale
> > > > > > of their brightness curve the value of the 'scale' attribute is 'unknown'.
> > > > > > 
> > > > > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
> > > > > 
> > > > > Daniel (et al): do you have any more comments on this patch/series or
> > > > > is it ready to land?
> > > > 
> > > > I decided to leave it for a long while for others to review since I'm still
> > > > a tiny bit uneasy about the linear/non-linear terminology.
> > > > 
> > > > However that's my only concern, its fairly minor and I've dragged by feet
> > > > for more then long enough, so:
> > > > Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
> > > 
> > > Thanks!
> > > 
> > > If you or someone else has another suggestion for the terminology that
> > > we can all agree on I'm happy to change it.
> > 
> > As you will see in my reply to Uwe. The term I tend to adopt when I want
> > to be precise about userspace behaviour is "perceptual" (e.g. that a
> > backlight can be mapped directly to a slider and it will feel right).
> > 
> > However that raises its own concerns: mostly about what is perceptual
> > enough.
> > 
> > Clear the automatic brightness curve support in the PWM driver is
> > perceptual.
> > 
> > To be honest I suspect that in most cases a true logarithmic curve (given a
> > sane exponent) would be perceptual enough. In other words it will feel
> > comfortable with a direct mapped slider and using it for animation
> > won't be too bad.
> > 
> > However when we get right down to it *that* is the information that is
> > actually most useful to userspace: explicit confirmation that the scale
> > can be mapped directly to a slider. I think it also aligned better with
> > Uwe's feedback (e.g. to start working towards having a preferred scale).
> 
> IIUC the conclusion is that there is no need for a string attribute
> because we only need to distinguish between 'perceptual' and
> 'non-perceptual'. If that is correct, do you have any preference for
> the attribute name ('perceptual_scale', 'perceptual', ...)?

More a summary than a conclusion! There is a reason I have left a bit or
space for others to comment on this over the last month (and a bit).

To be clear my Reviewed-by: means that I believe that the kernel is better
with "non-linear/linear/unknown" than without it and that I am comfortable
the API isn't likely to be a millstone for us.

Lee, Jingoo: Either of you care to offer $0.02


Daniel.