Hello, I am looking at writing a v4l program that is tightly coupled to a device driver using the mmap interface. The programs will be closely linked to webcams so there are probably some issues that need to be addressed. The question I have (which to me doesn't seem to explained fully in the API documentation) is about the recommended way of reading frames from a camera. I have seen some differences between userland and kernel implementations. In some of webcam implementations, it looks like a frame is captured and processed when a VIDIOCMCAPTURE ioctl is called. The frame is then freed when a VIDIOCSYNC is called. Other drivers seem to capture and process frames on each. Can anyone point me to a 'correct' way of working with the mmap interface? Regards, Matthew -- +---[ Matthew Tippett ]-[ Ph:+61 416 006 047 ]-[ Fax:+61 414 000 694 ]---+ +---[ TICons ]-[ mtippett@xxxxxxxxxxxxx ]-[ http://www.ticons.com.au/ ]--+ +---[ http://commercialos.org/ - Commercial Open Source Devel. Info ]---+
