On Sat, 2 Feb 2008, Robert Spitzenpfeil wrote:
> if I understand right you wanted me to make my system mimic this:
>
> Loop 3 times {
> Get-Max-LUN => STALL
> clear-halt(ep0)
> }
>
> I don't know how to hack usb-storage to run this very loop (lack of
> knowledge / understanding of the inner workings).
> Usually I just have to deal with perl/bash and not c/c++.
>
> the closest thing I could manage was to activate these subroutines
> within transport.c:
>
> if (result == -EPIPE) {
> /*
> * usb_stor_clear_halt(us, us->recv_bulk_pipe);
> * usb_stor_clear_halt(us, us->send_bulk_pipe);
> */
>
> /* sb_stor_clear_halt(us, us->send_ctrl_pipe); */
>
> usb_stor_clear_halt(us, us->recv_bulk_pipe); /* A */
> usb_stor_clear_halt(us, us->send_bulk_pipe); /* B */
> usb_stor_clear_halt(us, us->send_ctrl_pipe); /* C */
> }
>
>
> combinations of ABC: (order not changed, no multiple runs, A = A was
> executed, a = A was not executed):
>
> * abc : OK
> * Abc : FAILS
> * aBc : OK
> * abC : OK
> * aBC : OK
> * ABc : FAILS
> * AbC : FAILS
> * ABC : FAILS
>
> so for this order running A is the killer.
That's impressively thorough testing!
IMO this indicates we shouldn't issue any clear-halts at all unless the
device actually needs it. In general it's not a good idea to do a
clear-halt for an endpoint that isn't actually halted; devices are
prone to misinterpret the request.
And since the only device we know of that does need the clear-halts is
long obsolete, the simplest strategy is just to leave them out. That
ancient ZIP-100 drive can be accomodated by adding a US_FL_SINGLE_LUN
flag for it, since the Get-Max-LUN is never issued when that flag is
set.
Alan Stern
-
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
