Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

Windows Client v.0.2.0.0 Problems

DanT
2011-06-21
2014-09-22
1 2 3 .. 5 > >> (Page 1 of 5)
  • DanT
    DanT
    2011-06-21

    Curious if anyone is using this latest Windows Client on XP?  I just attempted to use it hoping it would fix blue-screen problems I was experiencing with v. 0.1.0.0 whenever I Detach a device on the Windows side using:  usbip -detach <port>
    Unfortunately..after updating the driver (and even tried uninstalling and reinstalling with latest driver), I now can not even get a listing of USB devices to work.  Here is my output. (of course a valid server IP was specified)
    usbip -debug -list 10.0.1.xx
    usbip dbg: usbip_network.c: 223 (tcp_connect ) trying 10.0.1.xx port 3240
    usbip dbg: usbip_network.c: 243 (tcp_connect ) connected to 10.0.1.xx:3240
    - 10.0.1.xx
    usbip err: usbip_network.c: 121 (usbip_recv_op_common) recv op_common, -1
    usbip err: usbip.c: 216 (query_exported_devices) recv op_common
    usbip err: usbip.c: 288 (show_exported_devices) query

    Is this a mismatch between the new v. 0.2.0.0 client code and what I am currently using on the Server? I don't think I am too far behind on the server version…but there is no indication on this forum (that I can see) that says if you upgrade to the latest windows client you need to at least have a server side version from a particular date or newer.  Any help would be appreciated!

     
  • DanT
    DanT
    2011-06-21

    I should have added that I working with 32 Bit Windows XP (not 64) and if I just replace the usbip.exe from the new v.0.2.0.0 back to the older v0.1.0.0…the usbip -list 10.0.1.xx works as expected. I will now test to see if I still Blue Screen upon detach to see if error was corrected in the new driver even though I have reverted back to the v.0.1.0.0 usbip executable.

     
  • DanT
    DanT
    2011-06-27

    Bump…nobody else having any problem with v. 0.2.0.0 (as I stated above)?  I am also experiencing XP blue-screen crashes whenever I detach a mouse or a keyboard by running the command usbip -detach <port> from XP.  Ironically, I also have a game controller that seems to detach just fine without any crashes.  Is anyone else experiencing similar issues when detaching devices?

     
  • Arjan Mels
    Arjan Mels
    2011-06-28

    No, not the same problems here… I assume you also used the 0.2.0.0 windows drivers?

    If it still remains I could try to debug using a virtual machne with XP, but I don't know yet when I will find the time.

     
  • DanT
    DanT
    2011-06-28

    Thanks for the reply! - please read on:
    Yes - 2 problems still remain and YES I currently have the 0.2.0.0 USB/IP Enumerator driver installed.
    1) I found that with both driver versions that whenever I detach a mouse or a keyboard using the usbip -detach <port> method, my windows XP system immediately blue screens with an "IRQL Not Less or Equal" problem.  If I detach a game controller, no blue screen takes place.  This blue screen is a big problem I would like to resolve as soon as possible.

    2) The usbip executable that came from downloading usbip_windows_v0.2.0.0_signed.zip performs in this manner:
    usbip -debug -list 10.0.1.xx
    usbip dbg: usbip_network.c: 223 (tcp_connect ) trying 10.0.1.xx port 3240
    usbip dbg: usbip_network.c: 243 (tcp_connect ) connected to 10.0.1.xx:3240
    - 10.0.1.xx
    usbip err: usbip_network.c: 121 (usbip_recv_op_common) recv op_common, -1
    usbip err: usbip.c: 216 (query_exported_devices) recv op_common
    usbip err: usbip.c: 288 (show_exported_devices) query

    Based upon your reply…it does not look like you tried 32 bit Windows XP when trying to reproduce the problem?  Can anyone who is currently using version 0.2.0.0 windows driver and usbip.exe on 32 bit XP comment as to whether they are experiencing similar issues as myself?

    Thanks!

     
  • DanT
    DanT
    2011-06-29

    Update - Tried on another 32 bit Windows XP system with the same results!  Both errors are present. XP systems are both fully patched (service pack 3).  Most concerning is the Blue Screens that occur when I detach keyboards and mice.  Anyone out there experience the same issues?  BTW I also notice that after a DETACH of a particular port, if I run the command usbip -port, the port I detached is still listed as a used port…why is this?  It does not seem correct.  So if I attach to and then detach a game controller repeatedly (the only device that does not blue screen XP), the port keeps incrementing and the command usbip -port will show all of the ports that were used in the past (but are no longer used). It is hard to get a sense for how many people out there are trying to make use of this project.  I know the Windows client is pretty new…and I am feeling like I am on the "bleeding edge".  Would love to hear that I am not alone…and that others are running into similar issues.  I would love to put this in front of a client for demonstration, but with the Blue Screens taking place upon detaching popular USB devices such as keyboards and mice…I am unable to. 

     
  • OliReadingUK
    OliReadingUK
    2011-07-03

    Hi all, I'm just having a go with the Windows version and not getting off the starting blocks…

    I have 2 machines, one's a virtual Win2008R2 the other a physical WinXP both with v0.2.0.0 installed. In each case, each can ping the other fine but fails connection when I try to list devices.

    >usbip -D -l andelain
    usbip dbg: usbip_network.c: 223 (tcp_connect ) trying 192.168.1.30 port 3240
    usbip dbg: usbip_network.c: 249 (tcp_connect ) andelain:3240, no destination to connect to
    usbip err:       usbip.c: 280 (show_exported_devices) - andelain failed

    …and the same in the other direction. Am I missing something? Do I need to run any sort of daemon app - I just installed the drivers on both ends (not rebooted etc),

    Any suggestion gratefully received - all seems like a neat system if it works,
    Oli

     
  • kinlau
    kinlau
    2011-07-04

    Are the both sides windows OS?
    USBIP couldnot support it.
    You should use linux OS as server.

     
  • OliReadingUK
    OliReadingUK
    2011-07-04

    Ah, thanks for clarifying that - it's a non-starter for my configuration then :(

    Oli

     
  • kinlau
    kinlau
    2011-07-05

    You're welcome.
    Now, I have many questions about usbip.
    I hope we can communicate about it as soon as possible.

     
  • Hi arjanmels,

    You can share the source code client windows GUI ?

    Thanks and Regards,

     
  • I'm experiencing the very same thing as cougardan describes with usbip_windows_v0.2.0.0_signed on WinXP x86 (kernel 2.6.39 on server side):

    usbip --debug -l 192.168.0.8
    usbip dbg: usbip_network.c: 223 (tcp_connect ) trying 192.168.0.8 port 3240
    usbip dbg: usbip_network.c: 243 (tcp_connect ) connected to 192.168.0.8:3240 - 192.168.0.8
    usbip err: usbip_network.c: 121 (usbip_recv_op_common) recv op_common, -1
    usbip err:       usbip.c: 216 (query_exported_devices) recv op_common
    usbip err:       usbip.c: 288 (show_exported_devices) query
    

    A workaround is indeed to replace usbip.exe with the one from usbip_windows_v0.1.0.0_signed.

     
  • For the records: I just tested usbip_windows_v0.2.0.0_signed on WinXP x64 and it definitely works there.

    And it is definitely also a brilliant pice of work (a fact that the little workaround needed for x86 as stated above cannot change)!

     
  • DanT
    DanT
    2011-07-15

    ogjelland - Thanks for verifying that you too experienced issues with x86.  I agree with you that usbip is a brilliant piece of work and it is apparent that much time has been spent on getting this far - I commend all contributors.  However, I do not agree with your statement "little workaround needed for x86 as stated above cannot change".  I am not sure what your meaning was here…perhaps just that you are overjoyed about the x64 functionality.  I very much think that changes need to be made in order to get x86 working appropriately.  The work-around of using the older usbip.exe is not really a viable work-around.  Although I can get past the errors, blue screen crashes still take place when disconnecting usb devices such as mice and keyboards.  I hope that some time will be dedicated to fixing x86 as many users are still using x86. One other item I have noticed is that PORTS never seem to be re-used after a detach takes place. I am not sure if that was by design or if this is another potential bug.  Anyhow, I would love to be able to use this project on both x86 and x64.

     
  • DanT
    DanT
    2011-07-27

    Another data point - Just fetched the source code that was under v0.2.0.0 tagged branch and built it in visual studio 2010 express.  Also rebuilt the driver.  tried out on Windows XP 32 bit virtual machine (VMware) and I am able to attach and detach mice and keyboard just fine.  If I try on a non-virtualized system (i.e. real physical box that is running WIndows XP 32 bit as its operating system), I still get the blue screen of death as previously reported. DRIVER_IRQL_NOT_LESS_THAN_OR_EQUAL.
    So..if any of the USBIP code gurus attempt to reproduce the problem by using a VM, you may not experience the problem.
    Hope this helps in narrowing the problem.  If anyone has any suggestions of what may be causing this, please let me know..I will also be attempting to poke around the source code to see if I can figure out a fix/work around to this problem.
    Thanks!

     
  • Bryan
    Bryan
    2011-08-01

    I am also facing the crashing issue on detaching HID devices with windows 7 as client. When i trace down the issue i got following sequence of calls:
    1.  Send comand  "usbip -d 1" from cmd
    2. Call goes to usbip_vbus_detach_device function which send
    IOCTL_USBVBUS_UNPLUG_HARDWARE to driver.
    3. call goes Bus_IoCtl and then bus_unplug_dev funtion in busenum.c
    which return successfully.
    4. call comes at Bus_PDO_PnP in buspdo.c
    5. call comes at IRP_MN_QUERY_DEVICE_RELATIONS 2-3 times


    6. then call comes at cancel_irp in bus_enum.c where pdodata is NULL
    (as pdo object has already deleted) which makes crash at "pdodata =
    (PPDO_DEVICE_DATA) pdo->DeviceExtension;"  

    For USB interrupt devices there is no continuous URB transfer and data transfers when ever there is any event happen in HID device. Bulk and Isochronous device doesn't have detaching issue. 

     
  • kinlau
    kinlau
    2011-08-02

    hi, cougardan.

    Would you tell me your usbip enviroment? Is client win7 x86 or x64? And server ?
    I have tested win7 x86 and linux 2.6.39. But it cannot work, some error as follows:

    usbip: failed to open usb.ids
    - 192.168.23.130
    usbip err: usbip_network.c: 121 (usbip_recv_op_common) recv op_common, -1
    usbip err:       usbip.c: 216 (query_exported_devices) recv op_common
    usbip err:       usbip.c: 288 (show_exported_devices) query

    Thanks!

     
  • DanT
    DanT
    2011-08-02

    Hi Kinlau,
    That is the same error I was getting (see top of thread). I have since fetched and built the source that I found here:
    http://usbip.svn.sourceforge.net/viewvc/usbip/windows/tag/usbip_windows-0.2.0.0/  using Visual Studio Express 2010.  It appears the recv op_common problem goes away - Apparently the pre-built client binary for usbip.exe x86 that you download is not a "valid" build.  However, even though the recv op_common problem goes away…there are still blue-screen of death issues when detaching simple devices such as keyboards and mice under x86 Windows.  (apparently my findings are that if x86 client is a virtual machine..the blue screen of death issue does not appear to happen).  Anyhow…to further answer your questions..my environment:
    Server: DaVinci (Embedded) Linux version 2.6.37  although my drivers/staging folder was grabbed from the latest git tree.
    Client: Windows XP Professional x86.
    Read this entire thread as I have also explained that if you replace the usbip.exe from the version 2.0 windows client download with the version 1.0 usbip.exe and install the windows 2.0 USB/IP Enumerator Driver…the recv op_common also goes away (although I do not recommend leaving things this way).
    Good luck and let me know if you hear of anyone fixing the blue screen of deaths that occur when detaching mice and keyboards.

     
  • kinlau
    kinlau
    2011-08-03

    Thank you for your answer, cougardan.

    Now, my first step is to connect device successfully, and second is to find the detaching problems.

    I hope we can make progress together. Good luck for you, too.

     
  • c
    c
    2011-08-04

    I can confirm the OP issue with v.0.2.0.0 on WinXP 32bit. Replacing with v. 0.1.0.0 does resolve it.

     
  • kinlau
    kinlau
    2011-08-05

    I used v0.1.0.0 on winxp x86, and usbipd 1.0.0 on server(linux 2.6.38 generic), but I still have above error.
    Who can help me to solve it?
    Thank you!

     
  • Arjan Mels
    Arjan Mels
    2011-08-07

    The recv op_common is a pretty generic error message if something goes wrong serverside. So please also include the server side debug log. (retrieve with the dmesg command).

    I will in the next couple of weeks try to double chekc the x32 driver with 0.2 and 0.1 usbip client release. It is unfortunate that the blue screen of death only seems to happen on physical systems and not on a VM session… This makes debugging for me lots harder.

    Regards,

    Arjan

     
  • Nicola
    Nicola
    2011-08-07

    I got rid of the blue screen on keyboard removal commenting out all the code in the following functions:
    pnp.c: complete_pending_read_irp
    pnp.c: complete_pending_irp
    But I don't really know what I've done
    Windows 7 64-bit and Windows XP 32-bit on VM

     
  • Bryan
    Bryan
    2011-08-11

    If i comment the function call of "complete_pending_irp"  inside "bus_unplug_dev"  function it works for me. It didn't crash while disconnecting HID device.
    Now PDODATA in cancel_irp call is not NULL. So the after cancel_irp IRP_SURPRISE_REMOVAL comes.
    And device detach successfully.

     
  • DanT
    DanT
    2011-08-11

    Thanks just4tech! I will also give that a try and see if that works for me.  Hopefully there is not many side effects to not making the "complete_pending_irp" call..but at this point if it fixes the Blue Screen of Death issues when detatching HID devices, I will certainly give it a go.
    Thanks for posting!

     
1 2 3 .. 5 > >> (Page 1 of 5)