[Intel-gfx] [PATCH] Revert "drm/i915: Add option to support dynamic backlight via DPCD"

[David Weinehall]

On Wed, Jul 05, 2017 at 05:40:19PM +0300, David Weinehall wrote:
> On Wed, Jul 05, 2017 at 01:03:46PM +0300, Jani Nikula wrote:
> > On Tue, 04 Jul 2017, David Weinehall <david.weinehall at linux.intel.com> wrote:
> > > This reverts commit ae25eceab616d16a07bcaa434b84463d58a3bdc3.
> > >
> > > The introduction of dynamic backlight control causes
> > > Lenovo ThinkPad X1 Carbon 4th Gen to boot to a black screen;
> > > presumably the backlight is off.
> > 
> > Oh, I thought reverting just this commit was not enough for you, based
> > on the previous discussion in the patch thread. If this is enough, and
> > the DPCD backlight by default works for you, I might be inclined to go
> > with just defaulting to false here.
> 
> Yeah, that was indeed my experience, but I cannot reproduce that
> behaviour now; disabling DBC seems to be enough to make things work.
> It might be that I made some mistake in my earlier testing.
> 
> I retested now, with the following results:
> 
> Both DPCD & DBC enabled: black screen
> DBC enabled, DPCD disabled: works
> DBC disabled, DPCD enabled: works
> DBC disabled, DPCD disabled: works
> 
> Based on my understanding of the code, DPCD backlight is required for
> DBC to work, so it makes sense that disabling either of the two
> makes things work (hence I don't understand how I got my earlier
> result where just disabling DBC wasn't enough).
> 
> Anyhow, I'm fine with just changing the default to false.

I need to NAK my own revert. It turns out that while booting with
only i915.enable_dbc=0 gives backlight, the backlight will only work
until a suspend/resume cycle has taken place; after that the display
will come up blank, unless i915.enable_dpcd_backlight=0 is also
passed.

So I'll submit a new revert.


Kind regards, David