Hi Krzysztof, On Thursday 04 September 2014 14:23:36 Krzysztof Borowczyk wrote: > Hello, > > I’ve recently noticed a problem with Modecom Venus and A4Tech PK-333E > webcams. Both can be put into a “bad state” in which they refuse to do > anything until they’re reconnected to the usb port. The test case is > simple: > > gst-launch-1.0 v4l2src ! videoconvert ! autovideosink > Setting pipeline to PAUSED ... > Pipeline is live and does not need PREROLL ... > Setting pipeline to PLAYING ... > New clock: GstSystemClock > eKilled > > gst-launch-1.0 v4l2src ! videoconvert ! autovideosink > Setting pipeline to PAUSED ... > Pipeline is live and does not need PREROLL ... > Setting pipeline to PLAYING ... > New clock: GstSystemClock > ERROR: from element /GstPipeline:pipeline0/GstV4l2Src:v4l2src0: Could not > read from resource. Additional debug info: > gstv4l2bufferpool.c(994): gst_v4l2_buffer_pool_poll (): > /GstPipeline:pipeline0/GstV4l2Src:v4l2src0: poll error 1: Invalid argument > (22) > Execution ended after 0:00:00.046823946 > Setting pipeline to PAUSED ... > Setting pipeline to READY ... > Setting pipeline to NULL ... > Freeing pipeline ... > > The GStreamer process has to be killed with the -9: > kill -9 `pidof gst-launch-1.0` > > The dmesg log shows this: > [88000.804362] usb 3-3: new high-speed USB device number 4 using xhci_hcd > [88000.864107] usb 3-3: New USB device found, idVendor=0ac8, idProduct=3460 > [88000.864113] usb 3-3: New USB device strings: Mfr=1, Product=2, > SerialNumber=0 > [88000.864116] usb 3-3: Product: Venus USB2.0 Camera > [88000.864118] usb 3-3: Manufacturer: Vimicro Corp. > [88000.865088] uvcvideo: Found UVC 1.00 device Venus USB2.0 Camera > (0ac8:3460) > [88000.866783] input: Venus USB2.0 Camera as > /devices/pci0000:00/0000:00:14.0/usb3/3-3/3-3:1.0/input/input16 > [88024.007404] uvcvideo: Failed to query (GET_DEF) UVC control 2 on unit 2: > -110 (exp. 2). > [88024.307041] uvcvideo: Failed to query (GET_DEF) UVC control 2 on unit 2: > -110 (exp. 2). > [88025.837884] uvcvideo: Failed to set UVC probe control : -32 (exp. 26). > [88030.830854] uvcvideo: Failed to set UVC probe control : -110 (exp. 26). > [88035.824615] uvcvideo: Failed to set UVC probe control : -110 (exp. 26). > [88040.818399] uvcvideo: Failed to set UVC probe control : -110 (exp. 26). > [88045.812171] uvcvideo: Failed to set UVC probe control : -110 (exp. 26). > [88050.805941] uvcvideo: Failed to set UVC probe control : -110 (exp. 26). > [88088.249967] xhci_hcd 0000:00:14.0: Signal while waiting for configure > endpoint command > [88088.250000] usb 3-3: Not enough bandwidth for altsetting 0 > [88090.907927] xhci_hcd 0000:00:14.0: ERROR Transfer event for disabled > endpoint or incorrect stream ring > [88090.907940] xhci_hcd 0000:00:14.0: @0000000115c5f460 00000000 00000000 > 0c000000 03058000 > [88091.021791] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with > disabled ep ffff88006b0fa540 > [88091.021797] xhci_hcd 0000:00:14.0: Trying to add endpoint 0x82 without > dropping it. > [88091.021802] usb 3-3: Not enough bandwidth for altsetting 7 > [88091.021805] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with > disabled ep ffff88006b0fa540 This looks like a low-level problem. Either the webcam firmware crashed, or the XCHI USB controller get in a bad state. What kernel version are you running ? -- Regards, Laurent Pinchart -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
