[PATCH 2/8] drm: Don't update property values for atomic drivers

[Daniel Vetter]

On Sat, Aug 12, 2017 at 12:20 AM, Laurent Pinchart
<laurent.pinchart at ideasonboard.com> wrote:
> On Tuesday 25 Jul 2017 10:01:16 Daniel Vetter wrote:
>> diff --git a/drivers/gpu/drm/drm_mode_object.c
>> b/drivers/gpu/drm/drm_mode_object.c index da9a9adbcc98..92743a796bf0 100644
>> --- a/drivers/gpu/drm/drm_mode_object.c
>> +++ b/drivers/gpu/drm/drm_mode_object.c
>> @@ -233,6 +233,9 @@ int drm_object_property_set_value(struct drm_mode_object
>> *obj, {
>>       int i;
>>
>> +     WARN_ON(drm_drv_uses_atomic_modeset(property->dev) &&
>> +             !(property->flags & DRM_MODE_PROP_IMMUTABLE));
>
> It would have been nice to remove the calls to drm_object_property_set_value()
> for the dpms property from drivers before adding this :-/ Three drivers (rcar-
> du, shmobile and fsl-dcu) initialize the connector's DPMS property to OFF
> using this call (the default being ON).
>
> Following the DPMS code paths always give me a headache, so if you know by
> heart how I should replace the set property call, I'm all ears :-)

Remove, it doesn't do anything for kms drivers. The value these calls
update was never consulted when reading the property, instead
drm_atomic_connector_get_property() directly looks at
drm_connector->dpms. These calls where only needed for legacy modeset.
Also note that the default _ON is the right one, since DPMS state is
in addition to modeset state. dpms should actually be set to _ON when
you enable the display using a modeset call, atomic takes care of that
in drm_atomic_helper_update_legacy_modeset_state().

Aka we found a using this WARN_ON!

And if you think the above explanation should be somewhere in
kerneldoc, pls add :-)
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch