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.

Broken USB_DEV_MSC example

Hi,

I'm using Logic-PD EVM and the latest StarterWare_1_00_03_03

I have a disappointing performance with example USB_DEV_BULK.

So I tried another example USB_DEV_MSC.

However, it hang at very beginning, it gets stuck at line with the code below

    /* Enables MMU */
       CP15MMUEnable();

What's wrong with the code ?

Thanks,

~Duy-Ky

  • Duy-Ky,

    There is a known issue with USB host MSC and USB dev MSC examples that is filed in the bug filing system. We are making a new release shortly that should fix the issue.

    Regards,

    Rahul

  • Hi Rahul,

    Thanks goodness!

    I also realized the example does NOT USE DMA, but the example usb_dev_msc does

    So, please have an option DMA for usb_dev_bulk, like usb_dev_msc

    Thanks, Rahul

    ~Duy-Ky

  • Hi Rahul,

    Over weekend, I did try example usb_dev_msc, its data throughput is awesome, a lot faster than usb_dev_buld (about 250 times), owing to usage of DMA

    I had to remove all code for SCSI disk, ... and it works intermittent due to my lack knowledge of how CPPI DMA works,

    In addition, I have to remove section of MMU at start of the main code as it hangs there

    So, it good idea when the example usb_dev_bulk is fixed, it should have option of DMA, buld opertion is for large data anyway

    Thanks, Rahul

    ~Duy-Ky

  • Hi Rahul,

    I just want to update my status with the example usb_dev_msc

    1) I remove MMU at start of main() as it gets stuck there

    2) Replace disk SCSI operations by a RAM of 4 MB

    Below is ny test results for a transfer of 4 MB from host PC to the target board

    1) No DMA : always successful with 4 MB in about 7 seconds

    2) With DMA : occasionally success with 4 MB in 1 second (7 times faster), BUT most of the time it fails in reach the expected count 0x400000 (4 MB), it stop short at, say 0x3FE200, ...

    So, it appears to me something wrong with CPPI DMA

    I'm anxious in waiting for the fix.

    I'd appreciate if you could send me the new code once the team have it fixed

    Thanks,

    ~Duy-Ky

  • Hi Rahul,

    I keep testing USB_DEV_MSC while waiting for the fix

    Below are my latest test resul

    1) 4 MB data from PC to Device

    1.1) DMA : very rarely the device receive the full 4 MB in 1 second

    1.2) NO DMA : the device always receive full 4 MB in 7 seconds

    2) 4 MB from device to PC

    2.1) DMA : very rarely the PC receive full 4 MB

    2.2) NO DMA : the PC receive NONE from the device

    So, both DMA and NO_DMA are broken

    ~Duy-Ky

  • Hi TI employee,

    Any update on this problem ?

    It has 2 modes : DMA and NO_DMA

    In DMA mode, it's possible to do both direction, but intermitten

    But in NO_DMA mode, only one way from host to device is OK, but impossible from device to host PC

    Please confirm.

    Thanks,

    ~Duy-Ky