On Sat, 8 Sep 2012, Florian Wöhrl wrote: > Hi Alan, > > first I have to apologize - it's not kernel version 2.6.+ that's the > begin of the problems, it's 3.2.+ - included in Debian wheezy and Ubuntu > 12.04. > Up to kernel 2.6. it is currently working correct. > > I first thought it might be the change in the driver - because in the > past 04d8:000a was handled by cdc_acm, now the modalias shows ftdi_sio. No, that can't be the reason. The problem occurs before the driver is loaded. > But I had contact with Oliver Neukum, the maintainer of cdc_acm. He told > me to contact this mailing list as he thinks it is a problem in usbcore. I don't think the problem is in the core. Certainly those -71 errors in your usbmon trace are caused by hardware or firmware issues, not by software. > >> The problem seems to be in usbcore, because some descriptors cannot be > >> read (-32/-EPIPE). > > Actually this looks more like a hardware/firmware/driver problem. > > Maybe in the device, maybe in your Raspberry PI. > The problem doesn't change on different host system hardware but on the > kernel / usbcore version. > Firmware is on board - it's not uploaded by the driver - afai see. If you're certain that the problem is associated with the kernel version, you could try using git bisect to pinpoint exactly what change to the kernel caused the problem to start. > >> Has anybody an idea how to get this running? > > Does the device work when plugged into a regular desktop system? Have > > you tested it recently under a 2.4 kernel to make sure that combination > > still works? > > > > Alan Stern > > > yes, the device is working perfectly under Windows (7) and in Ubuntu > 10.04 with kernel 2.6.+. > The cdc_acm driver is loaded and a device /dev/ttyACM0 is created. > But starting with kernel 3.2.+ - as in Ubuntu 12.04 or Debian Wheezy - > the descriptors cannot be read. Do you mean that the descriptors cannot be read when running Ubuntu 12.04 on a desktop system? > @greg k-h: Thanks a lot, but it's not a power problem. > > Is there any further log or information that I can contribute? Can you provide a comparable usbmon trace from a kernel where the device works? Also, to make sure that nothing important has been omitted, please resend the earlier trace without any "grep" filtering. And do the same with the new trace. 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
