On 22/03/13 21:31, Alan Stern wrote:
On Fri, 22 Mar 2013, Roger James wrote:
Sorry Alan,
My bad. I stuck a binary file on pastebin by mistake.
It is a binary wireshark trace (I think that is pcap). Here it is via
dropbox
https://dl.dropbox.com/u/84613021/wireshark.log
I will make a pair of host linux and a guest windows traces that refer
to the same run. However I do not think they will show anything
different from the slightly unmatched pair I have already posted.
The packet to look at if you can load the file into wireshark is number
110 it looks pretty identical to the one in the failing windows log to me.
Yes.
Just as a matter of interest, Ubuntu 12.04 which has a 3.2.0-39 kernel
only ever sees one drive on the Garmin device whereas windows sees two
when it works.
You could try moving to a more recent kernel version. However, you're
wrong about Ubuntu seeing only one drive. The wireshark log shows that
it sees two, but the device says that the second drive does not contain
any media.
It is clear from the log that VirtualBox is messing you up. The resets
that the guest tries to do never make it out to the actual device.
Nor do the Set-Config and Set-Interface requests. This may explain the
differing reactions to the READ FORMAT CAPACITIES command.
The Garmin device isn't blameless either. After sending its 12-byte
response to that command, it is supposed to halt the bulk-in endpoint.
But it doesn't; it responds to the next transfer with a 0-length
packet and _then_ halts the endpoint. Things go downhill from there
and it stops responding, after which Windows gives up.
Alan Stern
Thanks again for the rapid response. There is not much point in me
trying to get this going on Linux as Garmin do not support downloading
maps via Linux.
Your analysis still leaves one thing I do not understand. Why does my
trace on the native windows machine not show the 12 byte header making
it up to the "user" level. The only thing I can think is that it is a
peculiarity of the Windows Vista usb implementation that is not present
in XP and windows 8.
I guess I can head back over to the VirtualBox forum now and see if
there is any way to configure the virtualbox usb driver to act in a more
transparent fashion :-(
Roger
--
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
