On Sun, Jan 27, 2002 at 04:30:48PM -0800, Guy Harris wrote: > I don't know what would break if the driver were to construct some kind > of VPI/VCI/AAL5 header at the beginning of the sk_buff and set "mac.raw" > to point to that;
Actually, I do. What would break would, at the minimum, be any program accessing LANE network interfaces with PF_PACKET sockets in raw mode... ...such as any program using libpcap, such as Ethereal. LANE devices are, well, emulated Ethernet devices, and supply a hardware type of ARPHRD_ETHER. This means that code that uses a PF_PACKET socket to send on or receive from that device will expect that received packets will start with an Ethernet header and that packets to be sent on the device should start with an Ethernet header. You'd have to introduce a new ARPHRD_ type for "raw ATM", and have the LANE code use that type when registering the network device (meaning it can't use "init_etherdev()"). Changing the ARPHRD_ type might well break other things as well (in fact, I wouldn't be surprised to hear that it did). So I don't think that would work partcularly well. - This is the TCPDUMP workers list. It is archived at http://www.tcpdump.org/lists/workers/index.html To unsubscribe use mailto:[EMAIL PROTECTED]?body=unsubscribe
