Hello,
I am running Ubuntu Linux 12.04 LTS on the OMAP5432 EVM. I need to use this version because I am using V4L2 with an Omnivision OV8850 imager.
I need to have a very large, contiguous memory space in RAM as a buffer for high-speed streaming of up to 61 MIPI CSI-2 images at the imager's maximum frame rate of 24 fps. Based on the relevant documents, it appears that either bigphysarea or CMA (contiguous memory allocator) can do this, and that the large buffer can be mapped to user space with mmap(). Presumably I would "carve off" this memory at boot time and linux and my app would run in the balance of the memory.
- Does anyone know if ti-glsdk_6 supports CMA or bigphysarea?
- Has anyone done this before? If so, could they share some example code with me, please?
- Are there any other examples of something like this?
Another question: Does anyone have any documentation on the internals of the OV8850 V4L2 driver and yavta application? It would help me to have an overview.
Best regards,
Peter Gregson
