[PATCH] drm: allow changing DPMS mode for non-CRTC displays

[Lofstedt, Marta]

> -----Original Message-----
> From: Chris Wilson [mailto:chris at chris-wilson.co.uk]
> Sent: Thursday, December 1, 2016 4:15 PM
> To: Lofstedt, Marta <marta.lofstedt at intel.com>
> Cc: dri-devel at lists.freedesktop.org
> Subject: Re: [PATCH] drm: allow changing DPMS mode for non-CRTC displays
> 
> On Thu, Dec 01, 2016 at 03:30:01PM +0200, Marta Lofstedt wrote:
> > For non-CRTC displays,
> 
> A connector not attached to a CRTC works. It is the active connectors that
> were broken.
> 
> > the drm_atomic_helper_connector_dpms
> > will always set the connector back the old DPMS state before
> > returning. This makes it impossible to change DPMS state.
> >
> > fixes: 0853695c3ba46f97dfc0b5885f7b7e640ca212dd
> >
> > Signed-off-by: Marta Lofstedt <marta.lofstedt at intel.com>
> > ---
> >  drivers/gpu/drm/drm_atomic_helper.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_atomic_helper.c
> > b/drivers/gpu/drm/drm_atomic_helper.c
> > index 494680c..6a5acb9 100644
> > --- a/drivers/gpu/drm/drm_atomic_helper.c
> > +++ b/drivers/gpu/drm/drm_atomic_helper.c
> > @@ -2885,8 +2885,8 @@ int drm_atomic_helper_connector_dpms(struct
> > drm_connector *connector,
> >  fail:
> >  	if (ret == -EDEADLK)
> >  		goto backoff;
> > -
> > -	connector->dpms = old_mode;
> > +	if (ret != 0)
> > +		connector->dpms = old_mode;
> 
> Patch however is correct.
> 
> Fix the subject line to reflect the impact correctly, amend the fixes tag and
> add the appropriate CCs
> 
> Fixes: 0853695c3ba4 ("drm: Add reference counting to drm_atomic_state")
> Cc: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
> Cc: Eric Engestrom <eric.engestrom at imgtec.com>
> Cc: Sean Paul <seanpaul at chromium.org>
> Cc: dri-devel at lists.freedesktop.org
> Cc: <drm-intel-fixes at lists.freedesktop.org>
> 
> And mention if there was a testcase exercising the failure. If not please
> suggest one.


Thanks for the comments Chris.

About tests, I can't seem to find any test that is actually checking that we entered DPMS mode. The suggestion is to extend the kms_properties test, or maybe we should have a new dpms test. I CC Petri, so he can take part in such a discussion. 

/Marta

> -Chris
> 
> --
> Chris Wilson, Intel Open Source Technology Centre