Alan Stern schrieb: > No, no. The host sent an 8-byte SETUP packet to ep0 followed by a > 24-byte data-OUT packet to ep0. Sorry for my non precise wording! OK, here is my problem. Where should I put the 24 bytes from ep0? For the bulk out ep I have a pending request in the queue where I could put incoming data and call the callback from the request. But I have not seen any ep0 request pending in the queue. I need a callback, a request or a buffer where I could put that packet. But I have no idea where the gadget infrastructure expects these packets. I really miss a simple beginners guide to the gadget framework where a simple data flow diagram give answers to these beginners questions. Is there maybe a sample implementation which I could look in? Many drivers look very complex cause of all the hardware access. Is there really no documentation for the gadget framework? Maybe it is easier to write a pair of host and gadget driver at my own which enable ethernet functionality? To run ip over can needs 2 days... Regards Klaus > >> But why wireshark could not see the two >> packets? > > Wireshark for USB doesn't see packets at all. It sees only URBs, which > are purely software entities. Each URB corresponds to multiple > packets. The URB you posted was a 24-byte control-OUT transfer, which > means the hardware will try to carry out three transactions: SETUP (8 > bytes), data-OUT (24 bytes), and data-IN (0 bytes). > > And if you want to be really technical about it, each transaction > consists of multiple packets: > > The SETUP transaction involves the host sending a SETUP packet > and the device replying with an ACK packet. > > The data-OUT transaction involves the host sending an OUT > packet and a DATA0 packet, followed by the device replying with > an ACK packet. (Or the device might reply with NAK, in which > case host will retry the OUT and DATA0 packets.) > > The data-IN transaction involves the host sending an IN packet, > the device replying with a 0-length DATA1 packet, and then the > host sending an ACK packet. > >> I could believe that my hardware do mysterious things and my >> driver also, but how could the tracer show the things in that way? >> >> I maybe have Monday the chance to look directly on the bus with a >> hardware tracer. Maybe this clears the fog a bit :-) > > That's a good thing to try. > > 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 > -- 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
