On Wed, 18 Mar 2009, Dave Mielke wrote: > I'm using usbfs to support braille devices. I'm experiencing a problem with one > such device that I've never encountered before. When the device is opened for > the first time (after connecting the USB cable), it works fine. All operations > succeed as expected. When the device is closed and reopened, however, it no > longer responds (for example, to "get descriptor" requests), i.e. the ioctl()s > yield timeouts. Disconnecting/reconnecting the USB cable restores normal > operation, but only for one open. > > I know this problem has nothing to do with faulty support of autosuspend > because I've made sure that feature was disabled via writing "on" to > "power/level" and by writing "-1" to "power/autosuspend". > > One unusual thing I'm noticing is that the "release interface" which is done > just before the close of the device takes about five seconds to complete. From > strace: > > 08:59:17.106788 ioctl(10, USBDEVFS_RELEASEINTERFACE, 0xbfe8a58c) = 0 <5.001493> > 08:59:22.108524 close(10) = 0 <0.000032> > > The vendor assures me that reopening (without disconnect/reconect) the device > does work on Windows. > > Can any of you think of any possibilities that I should check? Most likely you have found the problem: The device doesn't like the Set-Interface request sent automatically by the kernel when the "release interface" occurs. In 2.6.29 the kernel won't send this request if it isn't needed. You might want to try 2.6.29-rc8 to see if it fixes the problem. > Is there any way I can get the kernel to log all of the low-level USB > operations involving the device? There are two things you can do: enable CONFIG_USB_DEBUG and collect a usbmon trace (see Documentation/usb/usbmon.txt). 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
