On Thu, 6 Oct 2011, loody wrote:
> 1. when below source will be executed? it seems the context will
> always not be filled in.
> ( I temporarily remove that and program runs well.)
> if (req->status == 0 && req->context)
> ((fsg_routine_t) (req->context))(fsg);
This is in ep0_complete(), right? Therefore it will be executed each
time a transfer on ep0 completes. However, req->context will always be
NULL unless you enable the CBI or CB protocol by setting the
transport=CBI or transport=CB module parameter -- see the
USB_CBI_ADSC_REQUEST case in class_setup_req().
> 2. in following source code, it is possible process will fall in ***
> loop if the state in start_transfer be EMPTY, when queue failed,
> right?
> if that really happen, who or when the process will leave ***?
>
> start_transfer(fsg, fsg->bulk_out, bh->outreq,
> &bh->outreq_busy, &bh->state);
>
> /* We will drain the buffer in software, which means we
> * can reuse it for the next filling. No need to advance
> * next_buffhd_to_fill. */
>
> /* Wait for the CBW to arrive */
> **** while (bh->state != BUF_STATE_FULL) {
> rc = sleep_thread(fsg);
> if (rc)
> return rc;
This is in get_next_command(), right? What do you mean by "queue
failed"?
If the host doesn't send a command, the loop will end when
sleep_thread() returns -EINTR. This can happen for a few reasons: The
gadget could be disconnected, it could be reset, or the module could be
unloaded.
> 3. in http://www.mail-archive.com/linux-usb-devel@xxxxxxxxxxxxxxxxxxxxx/msg15529.html
> , it mentioned about Scsi-backed Storage Gadget, I cannot find it in
> the gadget file folder.Is that
> finally realized?
No, and I doubt that it will ever be written.
Alan Stern
--
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
