On Fri, 11 Sep 2009, Toan Pham wrote: > Hi, > > I develop a usb device using CDC ACM device driver and a Linux distro > (LFS) that runs on DMP Ebox-3300L2 (ref. > http://www.compactpc.com.tw/ebox-3300.htm). > what I am experiencing is occasional usb port resets of the internal > usb device, usb device that is built into the computer and has > multiple interfaces such as Microphone/Audio/USBHID, when user > application starts communicating with my CDC device. Although usb > port resets occur, I notice that internal usb device does not > re-enumerate like a normal port reset (Ref Kernel log below). In what way is the re-enumeration not normal? It looks normal to me. > In some > rare occasions, I find that my CDC device also reset and redetected > due to strange events on the BUS. Not surprising, since your log is full of error messages caused by the CDC device. In fact, there's only one place in the log where ep1in on that device completed a transfer without an error! > Also, port reset only happens when > there is a USBHID driver attaches to internal usb device interface. That's because the usbhid driver requests the resets. So naturally without the driver there are no resets. > Everything works fine if i blacklisted USBHID driver. It seems the > problem is not in my CDC device's hardware or software because I > tested the device on multiple machines before. If there's nothing wrong with the CDC device then why does it generate all those error messages in the log? > I am suspecting USB OHCD host controller or internal USB device is > causing usb port resets. > I am not sure where the root cause is, and hope that someone would > help me on this issue. This might be easier to debug if you used a kernel that was more recent than 2.6.21. Alan Stern -- 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
