List,
I just tried the usb dissector with the example capture
usb_memory_stick.pcap which uses the (semi-)new dlt value from the
samples and dissection is quite broken.
Now, the usb dissector and the pcap format has been in flux so I am
not sure if the capture is just old and obsolete or if it is the
dissector that is just broken.
As far as i see the dissector and the capture file are seriously out
of sync regarding request/response matching, padding, detection of
transfer direction, etc, and no longer is able to dissect the SCSI
layer for memory sticks which the old dlt format and dissector
handled just fine.
I have a local fixed version of the dissector that fixes most of the
breakage in this capture file and dissects it almost properly.
I can not take usb traces myself so if someone could send me a capture
of usb when using a memory stick (==SCSI) traffic using the very
latest usbmon and linuxkernel i can test with i would be very
grateful.
best regards
ronnie s