Alan Stern wrote:

                The byte in question should be set to
0 before it is transmitted.  Second, if this is a matter of sending a runt
USB packet, why not make it a 0-length packet instead of a 1-byte packet?

Set to zero or any other constant value, yes. Not a very high bandwidth covert channel ... fixing it would in some cases require an additional memory allocation though.

As for why not a ZLP, there is hardware that doesn't handle
them very well ... and drivers too.  Safer to avoid them as
a general policy.

FWIW I do have an updated "usbnet" that talks CDC Ether;
it's hardly any incremental code.  That resolves some bugs
beyond the one Ben noticed (including performance problems
at high speed -- usbnet rx/tx queue length > 1, strange
MTU rules that prevent bridging, and more).

- Dave




------------------------------------------------------- This SF.net email is sponsored by: The Definitive IT and Networking Event. Be There! NetWorld+Interop Las Vegas 2003 -- Register today! http://ads.sourceforge.net/cgi-bin/redirect.pl?keyn0001en _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to