• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 98 subscribers
  • Views 465 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Mass Storage device with apple OSX

Michiel van Leeuwen
Prodigy 190 points

Hello,

I'm working with a custom board based on the C6746. We need to transfer data from the attached SD card to a host computer over USB. I implemented the mass-storage code from the example code from starterware. It works fine using a windows host. However, using an apple with OSX, the host doesn't see the device as a disk.

I added printf's in USBDSCSICommand (in usbdmsc.c), to see what commands get in. This is what I get:

SCSI_TEST_UNIT_READY
SCSI_REQUEST_SENSE
SCSI_INQUIRY_CMD
SCSI_TEST_UNIT_READY
Allow usb access: 1
SCSI_REQUEST_SENSE
SCSI_INQUIRY_CMD
SCSI_TEST_UNIT_READY
SCSI default: 1e
SCSI_REQUEST_SENSE
SCSI_READ_CAPACITY
SCSI default: 5a

Then all transfers just stop.

I did change the code a bit, so the user can allow or disallow access to the SD card. That's the "allow usb access" log line.

Any idea what is going on? Do I need to implement command 0x5a? But that is (from what I find) only used for usb-attached floppy disk. So why is OSX sending that command?

Thanks,

Michiel

  • 0
    TI__Intellectual 1640 points

    Hi Michiel,

    From a first look at it  seems the host has sent  a MODE SENSE 10 command and we don't have a handler for it. 

    As with regard to "allow usb access" change I suspect you have made it as a response to TEST_UNIT_READY. Is that correct ?

    Is there a way you can get a USB bus trace ( using an USB analyser) or at least a log on what's happening on the host PC side? 

  • 0 in reply to Vineeth Balan
    Prodigy 190 points

    I made some progress: If I send a version in the INQUIRY response, it seems to work a lot better. I now have this:

        _mem4(&g_pucCommand[0]) = SCSI_INQ_PDT_SBC | (SCSI_INQ_RMB << 8) | (SCSI_INQ_VERSION_SPC2 << 16); // version SPC2 is needed for OSX support

    and in the .h:

    #define SCSI_INQ_VERSION_UNSPECIFIED 0x00
    #define SCSI_INQ_VERSION_SPC2        0x04
    #define SCSI_INQ_VERSION_SPC3        0x05
    #define SCSI_INQ_VERSION_SPC4        0x06

    This seems to be a solution. I hope it is the right solution. I will test and let you know.

    I this the usb access by returing 0 in USBDMSCStorageOpen, which is indeed called from SCSI_TEST_UNIT_READY.

  • 0 in reply to Michiel van Leeuwen
    TI__Intellectual 1640 points

    Hi Michiel ,

    Good to know that you have made progress.

    Which field of the SCSI INQUIRY response are you trying to modify? ISO version field ? Not all devices report/ fill this field ( hitachi drives report this and Disk on Key does not). Please let me know if you face any issues with this change. I would like to pull this change into the StarterWare stack if this works. 

  • 0 in reply to Vineeth Balan
    Prodigy 190 points

    Hi,

    I'm changing the version field, as described in http://www.t10.org/cgi-bin/ac.pl?t=f&f=spc4r36h.pdf on page 336 and 339.
    The modification is based on a look through the Darwin code: http://opensource.apple.com/source/IOSCSIArchitectureModelFamily/IOSCSIArchitectureModelFamily-1.1_35/IOSCSIBlockCommands/IOSCSIBlockCommandsDevice.cpp  function DetermineDeviceCharacteristics.

    It seems to me that sending a zero for the version (as the original starterware code) is OK. But then the OSC host goes into some mode, and doesn't understand that the device doesn't have MODE_SENSE_10 implemented. Sending an actual version sets the host in another mode, and now it sends MODE_SENSE_6, and all works.

    However, this is based on testing with just one mac, one device and one set of software. A bit more testing would be good. I haven't done any testing with other hosts yet.

    Michiel