09:57 <marex> cphealy: what is it ?

09:58 <marex> austriancoder: is LIBGL_ALWAYS_SOFTWARE=true supposed to force software rendering with etnaviv too ?

09:58 <marex> austriancoder: I was under the impression it should

09:58 <marex> flto: ^

09:58 <austriancoder> marex: yes

10:03 <marex> austriancoder: hmmmm, I wonder if passing that variable into qt5 doesn't work somehow

10:03 <marex> austriancoder: time to debug :)

10:38 <xexaxo1> marex: it should work with any driver. you can set LIBGL_DEBUG=verbose EGL_LOG_LEVEL=debug -> a confirmation will be printed on stderr/out

10:39 <xexaxo1> plus it should list/attempt only the swrast_dri.so module.

10:40 <xexaxo1> one potential gotcha - many qt/kde apps are spawned via fork/exec - where the parent may not have the variables set

11:05 <marex> xexaxo1: I think the ^ is the problem, yep

14:09 <sa7mfo> Any idea why drmModeGetResources() fails with "Operation not supported" when using etnaviv on an iMX6? I do not fully understand all the graphical layers

14:10 <sa7mfo> I have card0 and renderD128 in /dev/dri. Not sure what renderD128 actually is, but it seems to be the one that reports the error

14:11 <lynxeye> sa7mfo: so seems you only have the etnaviv GPU DRRM device

14:11 <lynxeye> on i.MX6 you also need the imx-drm device for display

14:13 <sa7mfo> thanks, seems like I miss DRM_IMX. I will resolve some dependencies and try

15:49 <sa7mfo> I have struggled with "imx-drm display-subsystem: failed to bind ldb (ops imx_ldb_ops): -2". Any idea what could cause this?

17:29 <sa7mfo> lynxeye: how did you figure out that imx-drm was missing? Am I missing nodes in /dev/dri? I'm asking because I have activated imx-drm, got the framebuffer to bind, but still get the same error

17:31 <austriancoder> sa7mfo: etnaviv is a renderonly device and you need a KMS device... and that's imx-drm for imx6

17:36 <sa7mfo> thank you

17:36 <austriancoder> sa7mfo: your device tree is okay?

17:47 <cphealy> sa7mfo: can you perhaps share a dump of "dmesg" in pastebin or some other online service like that? This may provide some valuable insight to your configuration.

17:59 <lynxeye> sa7mfo: If the framebuffer is there, the imx-drm DRM device should also be there. Now you just need to make sure that you application is actually using the right DRM device for display purposes.

19:53 <sa7mfo> it turned out that the kernel was based on the imx-tree. Use the etnaviv instead of vivande driver simply did not work. I rebased the work on mainline instead and the drm devices popped up nice :)

19:54 <sa7mfo> Thank you for your time