[RFC weston 11/14] compositor-drm: Use actual drm_plane for cursor_plane

[Daniel Stone]

Hi,

> On 28 May 2015, at 10:49 pm, Derek Foreman <derekf at osg.samsung.com> wrote:
>> On 21/05/15 02:29 AM, Daniel Stone wrote:
>> Track the cursor plane through a real drm_plane, including pointing it
>> to the cursor plane we discover through universal planes, if applicable.
>> 
>> @@ -1535,7 +1541,8 @@ drm_assign_planes(struct weston_output *output_base)
>>                          &ev->transform.boundingbox);
>> 
>>        if (next_plane == primary ||
>> -            next_plane == &output->cursor_plane.base) {
>> +            (output->cursor_plane &&
>> +             next_plane == &output->cursor_plane->base)) {
>>            /* cursor plane involves a copy */
>>            ev->psf_flags = 0;
>>        } else {
>> @@ -1592,7 +1599,6 @@ drm_output_destroy(struct weston_output *output_base)
>>    }
>> 
>>    weston_plane_release(&output->fb_plane);
>> -    weston_plane_release(&output->cursor_plane.base);
> 
> Perhaps a totally naive question, but this release has disappeared and
> not been moved anywhere else?  Did we just leak something?

Subtle rather than naïve. Rather than explain here, I'll try to clarify through the medium of documentation in v2.

Cheers,
Daniel