Part Number: TDA4VM
Hello All,
We are planning on using OpenGL on our TDA4Mid platform. We did find this E2E question regarding how to enable Wayland for the Processor-RTOS: Can I use QT & wayland to render VisionApp result?
We were able to follow those instructions with a slight modification for the Processor RTOS v8.02 release. The location of the app_cfg_mcu2_0.h file has moved to the vision_apps/platform/j721e/rtos/common directory. After that and modifying the device tree file, we were able to run the OpenGL demo apps in /usr/bin/SGX/demos/Wayland/ and they displayed on the monitor. So far so good.
However, we are looking to do something more complicated. We would like to be able to run a simple OpenVX graph on mcu2_0 (the default which controls the display port output) at system boot, but later, when Linux is fully up and running, switch the control of the display to Linux (maybe using Wayland, maybe having the GPU render to the framebuffer).
It is not clear how something like that might be done.
One option might be to have the mcu2_0 control the display (as is the default now), but modify the code so that instead of a compile-time option, there is a run-time option to send a message to mcu2_0 to relinquish control of the display hardware. On the Linux side, we would also need to enable those disabled nodes in the device tree, and then somehow re-probe for the hardware that is now available to Linux.
Has anyone else done something like this? Is there a demo or something we could look at?
