00:10 <punit> Or cog is making the same assumption as weston? Though it does indicate that the issue is elsewhere..

00:19 <cockroach> could of course be that. also, I have to admit that I'm a bit lost, I'm very much stuck in the 'old ways' (Xorg etc.) :)

00:20 <cockroach> that said, cog *does* run without error when using the DRM backend, but only if I also add the swrast dri module. also, while it does seem to be running, I'm not seeing a picture anywhere, not sure where things are being rendered to...

02:18 <punit> If you're having to use swrast, that suggests that 3d /OpenGL is required. Sorry I have no idea what cog is or it's requirements

02:23 <cockroach> yeah I'll investigate further tomorrow. thanks already for the help!

03:36 <marex> punit: weston is picking the wrong drm device for you ?

03:37 <marex> weston should be using gles for compositing, so without etnaviv, your compositing performance will be lower, since it will be blitted on cpu

04:43 <punit> marex: Agreed about the performance impact - disabling DRM_ETNAVIV isn't a real solution. It is just a work around / debug step to confirm the symptoms of a problem

04:45 <punit> While investigating the conditions around planes assignment in weston it was noticed that in certain conditions the failure occurred because an object was being looked up in etnaviv while it was created by imx-drm

04:47 <punit> I am trying to understand how that happens and whether it's a bug or a configuration issue