This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TUSB7340 EVM fails to detect root hub in windows 7

Other Parts Discussed in Thread: TUSB7340

We have found an odd situation.

We have a embedded single core computer with the Intel® Celeron® M Processor ULV 722. When we connect the TUSB7340 EVM and install the latest driver the system will intermittently gives the following error:

Hardware Id Missing Windows cannot identify the device plugged into the Texas Instruments XHCI Controller. For assistance, contact the hardware manufacturer.

There is nothing connected to the board TUSB7340 board ports, the xHCI controller fails to identify the root hub and all the interfaces are not operational. Looking in device manager, the controller is there but the root hub is gone.

The setup:

  • Windows 7 SP1 Ultimate
  • CPU with x1 slot interface.
  • Driver package Texas_Instruments_xHCI_Driver_v1.16.3_WHQL
There are some things that do allow the driver to function.
  1. If we swap out the single core CPU with the Core 2 Duo version (same chipset), the driver failures disappear. If we disable one core the problem re-appears.
  2. If we only use USB keyboard and mouse (connected to the CPU USB 2.0 controller) instead of PS/2 keyboard and mouse , the driver failures disappear.
  3. If we disable the i8042prt driver in windows 7 (with the PS/2 keyboard attached ) the problems disappear.
  4. If we uninstall the driver and re-install the driver after the failure occurs the root hub is detected and everything works fine.
So we thought that was interesting enough to post.
  • Hello, 

                    Could you please verify if you are using the latest driver? you can find it at (http://www.ti.com/litv/zip/sllc423g ), we just want to confirm that. Also please try on a fresh system without older drivers installed. We want to avoid any possibly that a old installed drivers can cause any kind of conflicts

     

                    I’m wondering if you are able to see this behavior using different EVMs boards?

     

                    Regards,

                    Roberto.

  • I have seen the issue on the xHCI_Driver_v1.16.3_WHQL and the V1.12.18 driver. This is obviously a wide range for the Driver. 

    We have done many re-installations, I was pretty sure that the latest driver also had failures but I will verify it again. 

    Tony M. 

  • We were able to capture the issue again with the latest driver.

    Windows 7 SP1 , 32 bit install. The windows installation had a password that needs to be entered for start-up, in case this has relevance for the driver.

    This time it was a little hard to catch , failing only 4 times on 10 power cycles.

    We only have the one reference board. We have other expansion boards that have this same TI USB controller on them. We have seen the failure on these boards as well. We developed these expansion boards so for pure testing purposes they are not considered a reputable source, I imagine.

    After you made your suggestion I had to make sure that there were no oem*.inf files lying around from prior install, so I re-installed the drive again and only installed the latest driver package.

    The issue is probably small statistically. I was hoping that that error may "ring a bell" on the off chance an actual TI driver developer actually sees this thread.

    Tony M.

  • Along with the relationship of PS/2 keyboards being in the system we found that if we modify the registry key of the io8042prt driver we can make the issue disappear.

    The article on this option is here http://technet.microsoft.com/en-us/library/cc739722(v=ws.10).aspx 

    If we change the following register keys to 0x400 (the minimum suggested by Microsoft) the issue disappears:  HKLM\SYSTEM\CurrentControlSet\services\i8042prt\Parameters\PollingIterations and PollingIterationsMaximum  .

    So the big question is how this driver and the USB driver stack interrelate to cause an issue with enumeration of the root hub, but since our problem is solved, I guess it doesn't matter.

    There will be few systems shipping with PS/2 keyboards except in the embedded realm.

    Tony