Hello, I am running mainline(ish) 3.16.1 on a custom i.MX6 board which is working very well including a basic rgb parallel display and fb support through mxc_lcdif etc which i brought forward in /drivers/video/mxc/.
I would like to get an accelerated graphics stack running and am looking for some advice/guidance on getting the vivante gpu up and running with a basic xfce or similar windowed environment.
Currently I am simply using qt 4x in qws mode and I would like to get accelleration working before I start looking at qt5.
I am using oe to generate my angstrom image, but the vivante kernel support doesn't seem to be available for mainline and there seems to be a myriad of drivers/files and modifications needed to get this going.
Specifically, I am looking for a list of 'requirements' that need to be met so that I can tackle them one at a time but haven't found any documentation outlining all of the requirements.
1. /drivers/gpu/drm requirements?
2. /drivers/gpu/ipu_v3 requirements?
3. vpu power configuration?
4. gpc gpu/vpu clocks?
5. vivante kernel drivers?
6. user-space drivers required?
7.device tree bindings required in my dts files?
8. other things that I probably forgot....
bonus question, since I currently am NOT using the gpu, how can I confirm that is is not using up extra power unnecessarily until I get this sorted out?
I know that this is a very board question/set of questions, so don't feel the need to answer everything but please chime in if you have any ideas or questions!
Thanks everyone
Brent
Brent,
For parallel support you can use the 'fsl,imx-parallel-display' property in mainline. No need to use the mxc_lcdif from FSL kernel.
Also, you could try using this 3.14 kernel instead:
Thank you.
That is good to know about the fsl,imx-parallel-display, that should help to clean up alot of my display configs in the meantime.
Has anyone used this?
https://github.com/xobs/novena-linux/commits/v3.17-rc5-novena-vivante
Hi,
I do not know of the requirements, I am not an expert in GPU/Vivante; but according to one of my peers it seems that we are already working on this.
Let me delve into this.
Best Regards,
Alejandro