10:59 <ecrn> btw. probably you are already aware of this, but NXP provided vg_lite library for its GC355-containing MCUs, which surprisingly is provided as a full source code

10:59 <ecrn> could aid in the development

11:02 <ecrn> VG_LITE_RETURN_ERROR(push_state(&tls->t_context, 0x0A2B, source->stride | tiled_source));

11:02 <ecrn> VG_LITE_RETURN_ERROR(push_state(&tls->t_context, 0x0A2D, 0));

11:02 <ecrn> VG_LITE_RETURN_ERROR(push_state(&tls->t_context, 0x0A2F, source->width | (source->height << 16)));

11:02 <ecrn> and so on

15:48 <cockroach> punit: I saw your question about the "Failed to lookup GEM object" on imx6. I'm currently having the same issue and was wondering if you ever found a solution.

16:24 <austriancoder> ecrn: please do not proprietary source code snippets. I do not think the source code licence allows you do to so and we only allow clean room reverse engineering!

16:35 <ecrn> austriancoder: it is MIT licensed

16:51 <mntmn> oh, etnaviv on genode https://genodians.org/jws/2021-06-25-experimental-etnaviv-driver-port

17:28 <pcercuei> nice!

22:53 <punit> cockroach: Haven't found the solution but a work around - disabling DRM_ETNAVIV makes the problem go away

22:55 <cockroach> punit: oh, fun. how about performance? wouldn't disabling DRM hurt that?

22:56 <punit> The issue from my understanding so far is i.MX6 has two DRM devices (/dev/dri/card[01]) - the etnaviv driver for the 3d engine and the imx-drm for the display pipeline

22:58 <punit> And somewhere in the software stack (weston, libgbm, somewhere else) there is a mixup and the query ends up with the wrong device - hence the failed to lookup...

22:58 <cockroach> ah okay, so as long as I don't need the 3d engine I should be fine without DRM_ETNAVIV

22:58 <cockroach> thanks, I'll give it a try

22:58 <punit> Disabling the etnaviv driver means that the query ends up in the right driver and problem goes away

22:59 <punit> I haven't had the time to dig into what is causing the mixup - and whether it's a configuration issue (I am using Debian) or a bug that needs fixing

23:01 <cockroach> I'm using buildroot myself so there's not a whole lot of unintended configuration happening, plus it seems others are having the same problem

23:01 <cockroach> https://github.com/Igalia/cog/issues/144

23:02 <cockroach> might be worth documenting somewhere / looking into it

23:05 <punit> Thanks for the pointer - I had not seen that. It definitely would be worth documenting / reporting - the question is identifying the component to report it against

23:07 <punit> At the moment I don't think it's a kernel driver issue - weston and libgbm are my current top suspects

23:46 <cockroach> ah, speaking of things: on my system I'm not even running weston at the moment, trying to get cog to run directly with the drm backend. I suppose that means it's not weston's fault...