Hi,
I've been looking at some issues with some disconnection issues seen
when using Barry (http://barry.sourceforge.net) to communicate with some
BlackBerry devices.
The underlying problem seems to be that the BlackBerry device fails to
respond to a Test Unit Ready SCSI command within the timeout period.
This causes the usb-storage driver to reset the USB device when handling
the timeout in usb_stor_invoke_transport(), which causes the Barry
communications on other endpoints to fail.
A sample of the output from usb_storage with debugging enabled is:
[24924.910351] usb-storage: queuecommand called
[24924.910390] usb-storage: *** thread awakened.
[24924.910397] usb-storage: Command TEST_UNIT_READY (6 bytes)
[24924.910399] usb-storage: 00 00 00 00 00 00
[24924.910409] usb-storage: Bulk Command S 0x43425355 T 0x5a L 0 F 0 Trg
0 LUN 0 CL 6
[24924.910413] usb-storage: usb_stor_bulk_transfer_buf: xfer 31 bytes
[24924.910543] usb-storage: Status code 0; transferred 31/31
[24924.910546] usb-storage: -- transfer complete
[24924.910549] usb-storage: Bulk command transfer result=0
[24924.910551] usb-storage: Attempting to get CSW...
[24924.910554] usb-storage: usb_stor_bulk_transfer_buf: xfer 13 bytes
[24955.477542] usb-storage: command_abort called
[24955.477547] usb-storage: usb_stor_stop_transport called
[24955.477550] usb-storage: -- cancelling URB
[24955.477660] usb-storage: Status code -104; transferred 0/13
[24955.477663] usb-storage: -- transfer cancelled
[24955.477665] usb-storage: Bulk status result = 4
[24955.477668] usb-storage: -- command was aborted
[24955.477787] usb-storage: usb_stor_pre_reset
[24955.557453] usb 2-1.4: reset high speed USB device using ehci_hcd and
address 28
[24955.679226] usb-storage: usb_stor_post_reset
My question therefore is if it is necessary to always reset the USB
device when a usb-storage command times out?
If there are other alternatives to resetting the whole device then could
someone give me a few suggestions to point me in the right direction?
The BlackBerry continues to communicate correctly on the other endpoints
while the usb-storage command is pending and when I tried a nasty hack
of just commenting out the resetting of the USB device so that the SCSI
layer could retry the command then it at least seems like the BlackBerry
can recover, with usb-storage and Barry continuing to work correctly.
I'm not suggesting that this hack is anything like a solution, but was
curious to see if the hanging command could be recovered from.
Regards,
Toby
--
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
