On Tue, Jun 22, 2010 at 05:03:24PM -0400, Alan Stern wrote: > On Tue, 22 Jun 2010, Greg KH wrote: > > > On Tue, Jun 22, 2010 at 03:08:33PM -0400, Alan Stern wrote: > > > On Tue, 22 Jun 2010, stephane duverger wrote: > > > > > > > Hello, > > > > > > > > This is a patch that implements an USB EHCI Debug Device using the > > > > Gadget API. It does make use of "u_serial" to provide userland access. > > > > > > Specifically, it emulates a Netchip debug device. However it's not > > > clear to me why any emulation is needed at all. Why not just use a > > > regular serial gadget as a debug device? > > > > You would need to emulate the ehci debug port connection in the gadget, > > right? I didn't think the regular serial gadget did that, but I might > > be totally wrong. > > I don't understand what you mean. The connection from the debug port > to the gadget is a regular USB cable; it doesn't need to be emulated. > > I don't remember the details about how the Netchip debug device works. It's an electrical blob that does some type of DEBUG negotiation as a client device. > Does it simply copy bytes from one port over to the other? At any > rate, the EHCI spec doesn't say anything about how a debug device has > to work; it only says how the debug port itself works. That port is > theoretically capable of handling any high-speed device, subject to its > length limitation (no packet can be longer than 8 bytes). > > Is that limitation the reason a special gadget driver is needed? The EHCI controller has some type of handshake happening for it to determine that this really is a debug device, so that needs to be emulated in the gadget somehow. Because of that, I would assume we need some code to do so. thanks, greg k-h -- 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
