Hi Joe, On Thursday 08 January 2009, jmk wrote: > Hi Laurent, > > On 1/7/09 9:05 AM, "Laurent Pinchart" <[email protected]> wrote: > > Where did you get the source from ? I had a look at the > > http://hg.jannau.net/hdpvr/ repository, and v4l2-dev.c:261 tests for > > vdev->release being on NULL. This is quite unlikely, as the uvcvideo > > driver provides a release callback. > > > > Could you please try the latest uvcvideo sources available from the > > linuxtv.org/hg/~pinchartl/uvcvideo repository ? > > You are correct, the version I was running came from hg.jannau.net. I > performed a "rminstall" on that version and built and installed a checkout > from linuxtv.org/hg/~pinchart1/uvcvideo. I've attached the dmesg output > from the module load: > > --- > Linux video capture interface: v2.00 > uvcvideo: Found UVC 1.00 device <unnamed> (046d:0990) > WARNING: at /h/downloads/multimedia/myth/hd-pvr/uvcvideo/v4l/v4l2-dev.c:423 > video_register_device_index() > Pid: 1455, comm: modprobe Not tainted 2.6.24-gentoo-r8 #3 > [<f0fb8643>] video_register_device_index+0x4b3/0x500 [videodev] > [<f0fb869f>] video_register_device+0xf/0x20 [videodev] > [<f0ff4ad1>] uvc_simplify_fraction+0x6f1/0x790 [uvcvideo] > [<f0ff57bf>] uvc_delete+0xc4f/0x1eb0 [uvcvideo] > [<c02fa01a>] idr_get_empty_slot+0xea/0x270 > [<c02fa227>] ida_get_new_above+0x87/0x180 > [<c03e5d65>] usb_probe_interface+0x75/0x100
There's something seriously wrong with the top of the backtrace. The backtrace might just be corrupted, but given the warning I suspect a deeper issue. > [<c037f598>] driver_sysfs_add+0x58/0x80 > [<c037f6ce>] driver_probe_device+0x8e/0x190 > [<c04ccc3d>] klist_next+0x5d/0xa0 > [<c037f910>] __driver_attach+0x80/0x90 > [<c037eaaa>] bus_for_each_dev+0x3a/0x60 > [<c037f536>] driver_attach+0x16/0x20 > [<c037f890>] __driver_attach+0x0/0x90 > [<c037ee5b>] bus_add_driver+0x8b/0x1e0 > [<c03e5cf0>] usb_probe_interface+0x0/0x100 > [<c03e58e5>] usb_register_driver+0x95/0x110 > [<f0fcf06f>] init_module+0x6f/0x86 [uvcvideo] > [<c014a28b>] sys_init_module+0x16b/0x1930 > [<f0fe4000>] fw_dev_release+0x0/0x20 [firmware_class] > [<c010437e>] sysenter_past_esp+0x5f/0x85 > ======================= > usbcore: registered new interface driver uvcvideo > USB Video Class driver (v0.1.0) > --- > > Would seeing my v4l .config be useful? It's pretty minimal, enabling the > V4L module, V4L API1, IVTV, and UVC. I'm not sure if I'm missing something > critical or not. I don't think it would help. Could you add printk's in video_register_device_index to print vdev and vdev->release (if vdev is not NULL), as well as in uvc_register_video to print vdev and vdev->release just before the call to video_register_device ? Best regards, Laurent Pinchart _______________________________________________ Linux-uvc-devel mailing list [email protected] https://lists.berlios.de/mailman/listinfo/linux-uvc-devel
