On Monday, October 25th, 2021 at 8:11 PM, Tomasz Moń <desowin@xxxxxxxxx> wrote:
> If you have some major questions, then in my opinion it is best to
> open it as early as possible - just make sure to mark it as draft.
I've already opened a merge request. (Things ended up being easier than
I anticipated.) I think the dissector is usable enough; the main question
is how to integrate it with the wider USB dissector framework.
> Currently mass storage class dissector assumes Bulk-Only Transport. It
> should however check the interfaceProtocol value in usb_conv_info_t
> data and perform the dissection based on this. So the "shim dissector"
> you mention, should really be part of the existing mass storage
> dissector.
Well, I felt that was wrong to put in the BOT dissector, but on second
thoughts it's probably far better than what I did. At least as a one off.
I'd be willing to extend USB dissector selection to include more than
just interface class if that would be a welcome addition, since that's
really the problem I'm trying to paper over.