Hey methinks, my v4l2 experience is limited. But thats not stoppng me. I’m writing a daemon as we speak. Unfortunately, my C is a little rusty, so its going to take a while to debug. But I’m more than happy to share my (at the moment, somewhat messy) code. I’ve been pouring over the V4L2 API pages. At the moment, I’m getting the daemon to initialize buffers. There are a few different data acquisition methods, and I’m trying to make the daemon robust enough to use all of them. But for the most part, web cameras use a jpeg protocol. You can use read() methods to v4l2 devices, which may be acceptable. But since there is constant data acquisition, its probably better to use pointer type structures. If you have any questions at all, give me an email, or reply and ask me to email. Depending on how much reading you’ve done, I can help you out with some of the grunt work of parsing through all those verbose specification documents.
I’d love to work on code with you if you’re willing. I could really use the help of experience.