On 20.06.2012 16:20, Alan Stern wrote: > On Tue, 19 Jun 2012, Philipp Dreimann wrote: > >> Hello, >> >> my usb webcam using the uvcvideo module is not working since kernel >> 3.5-rc2. Kernel v3.4 is the last which is known to work, 3.5-rc3 is >> broken too. >> >> The webcam does not work if it is connected to an usb2 port. If the >> webcam is connected to an usb3 port it works. I compiled the uvcvideo >> module from kernel 3.4 together with 3.5-rc3 to rule out the driver. >> -- The behaviour did not change. (Thanks to Laurent Pinchart for the >> idea.) >> >> Here is a log using usb3: >> [ 1528.629506] usb 3-1: new high-speed USB device number 4 using xhci_hcd >> [ 1529.364396] uvcvideo: Found UVC 1.00 device <unnamed> (046d:0821) >> [ 1529.378344] input: UVC Camera (046d:0821) as >> /devices/pci0000:00/0000:00:1c.6/0000:0e:00.0/usb3/3-1/3-1:1.2/input/input16 >> -- Verifying whether the camera works -- successful. >> [ 1571.789894] usb 3-1: USB disconnect, device number 4 >> >> >> And the problematic one using usb2: >> [ 1636.434459] usb 2-1.2: new high-speed USB device number 4 using ehci_hcd >> [ 1637.237109] uvcvideo: Found UVC 1.00 device <unnamed> (046d:0821) >> [ 1637.250484] input: UVC Camera (046d:0821) as >> /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.2/input/input17 >> -- Verifying whether the camera works -- not successful. >> [ 1645.536061] uvcvideo: Failed to query (GET_DEF) UVC control 2 on >> unit 2: -71 (exp. 2). >> [ 1645.540292] uvcvideo: Failed to query (GET_DEF) UVC control 2 on >> unit 2: -71 (exp. 2). > >> The last line is being repeated a lot, the Failed to query lines >> appear every now and then.. > > -71 errors generally are caused by bad hardware. In this case, it may > be a bad cable connection between the device and the EHCI controller. > That would explain why the same device works okay when connected to > xHCI. Sorry for being late in the game. I didn't follow the thread here, only the one on alsa-devel. The core reason for this breakage was that the new USB audio streaming logic issued URBs to the device after altsetting 0 was selected to turn the stream off. The device didn't like this at all and signalled protocol errors in return. This is all fixed now and merged to 3.5-rc7. I'm just explaining this in case similar issue occur in the future (on other USB devices), which might then have a similar root cause. Daniel -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
