Hey Janne, On Fri, 2003-08-22 at 00:40, Janne wrote: > OK, I figured out the second problem - with two fields per image, the > vertical size should be _half_ the 'full' image size.
That's the field size. :). The struct describes field configuration, not frame configuration. See the lavrec sources (well, you don't need to anymore, but if you'd want to double check, you could do that). > For the close problem, well, after running the app the first time, there > are a few messages, but nothing that would indicate a problem: > > LML33[0]: zoran_close, record pid=[7243] > LML33[0]: enable_jpg IDLE > LML33[0]: zoran_close done > LML33[0]: zoran_close, record pid=[7243] > LML33[0]: enable_jpg IDLE > LML33[0]: interrupts received: GIRQ0:10 queue_state=0/0/4/89 > LML33[0]: zoran_close done > > I run the app again, get the error, and look with dmesg - and there are > no new messages there. No error is logged, in other words. > > In a word: Huh? In that case, the driver is not spitting out the error. Something else is, apparently. Could you give the output of 'lsmod' or try with 'debug=5' (highest debug level, just for testing)? Oh, btw, you also said you dequeue all buffers to stop capturing. In case of MJPEG capture, you can also queue frame -1, which will auto-dequeue all frames in the driver and deintialize capture. This is only valid for MJPEG capture, though, not for v4l1 capture (where you have to dequeue all frames separately). V4l2 uses VIDIOC_STREAMOFF for this. Ronald -- Ronald Bultje <[EMAIL PROTECTED]> ------------------------------------------------------- This SF.net email is sponsored by: VM Ware With VMware you can run multiple operating systems on a single machine. WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the same time. Free trial click here:http://www.vmware.com/wl/offer/358/0 _______________________________________________ Mjpeg-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/mjpeg-users
