Comment 11 for bug 1082215

Revision history for this message
Tal Nevo (talnevo-i) wrote : Re: Writing on a certain USB HDD causes corruption

The following tar file (usb-debug.tgz) contains debugging obtained using the instructions in the link provided above.

I used two USB enclosures with the SATA hard drive, one that "works" and the one that exhibits the bug.
Testing was done with the last kernel installed (3.7.0-030700rc7).

I zeroed a large chunk at the beginning of the drive with the "good" USB interface and collected the data from /sys/kernel/debug/usb/usbmon/1u for two runs:

dd if=/dev/zero of=/dev/sdb bs=1M count=100
dd if=/dev/zero of=/dev/sdb sb=64K count=1000

I verified after each run that the drive was indeed zero by running hexdump.

I then repeated the same two dd tests with the "bad" USB (I zeroed the drive using the "good" interface in between). and also collected a hexdump of the drive (that shows the USBC headers that should not be there).

The bus traces are named bus1data*.txt and are marked "good" and "bad" depending on which USB interface was used.

I hope this will help.