On Tue, 31 May 2011 linux@xxxxxxxxxx wrote: > On 05/31/2011 11:38 AM, Alan Stern wrote: > > Please configure your email client to wrap lines after 72 columns or > > so. > > Hopefully it's correct now. No, the lines are still very long. :-( > > Do you have any pointers to the discussion leading to that update or to > > the particular code change that fixed the problem? > > Sorry, I only have the package archive on this machine to hint at when new kernels where installed, along with notes collected while debugging this UPS issue last summer. I looked through the email archives. There are some interesting messages here: http://marc.info/?l=linux-usb&w=2&r=1&s=nut&q=b Look at the entries referring to Intel P55 chipset bugs and related messages. > > Actually it's a lot easier to read output from usbmon's text interface > > than wireshark output. See the instructions in > > Documentation/usb/usbmon.txt. > > cat .../3u data attached. I don't see anything out of the ordinary except that the UPS refused to accept a Set Idle request. (This may or may not be significant.) After the initial probing finished, there was no USB traffic for 20 seconds. Then the UPS was disconnected. Can you get a similar usbmon trace from an earlier kernel where everything worked correctly? Comparing the two traces might help. > > The reason for the disconnects is not obvious. If it was already > > figured out and fixed, it would be a big help to know what the solution > > was (and why it is no longer working). > > I also applied a patch to the nut driver (network ups tools) that was also required to make it work (summer 2010). This patch had to do with low-speed devices having a maximum packet size of 8 bytes. > > I tried to interpret the usbmon output to determine if descriptor requests, etc, where limited to 8 byte buffers. It does show a Malformed Packet at the end of Frame 2. (I have this textual tshark output too, but it's 625K file, first two frames below) Don't worry about that one bad packet. I don't think it has anything to do with the disconnections. > As an aside, I had this exact sequence of events with a FreeBSD server with this same UPS model: broken in 7.1, fixed with update to 7.3, broken again with recent update to 8.1. (Broken in same way, repeated disconnect/reconnect) > > > > > Alan Stern > > Thank you Alan! > > Please let me know if there's anyway for me to gather more specific details of the disconnect/reconnect activity. Comparison with a working system would be the best thing at this point. 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