On Thu, 2012-11-29 at 00:17 +0000, David Woodhouse wrote: > On Wed, 2012-11-28 at 14:35 -0800, Daniel Gimpelevich wrote: > > The br2684 driver seems to expect that the MAC is already on the device > > before a virtual circuit is created, but it seems to leave the mechanism > > for doing so to the ATM driver. The only other working ATM driver in > > OpenWrt is for AR7, > > There's also solos-pci, as used on the Geos. > > I mostly use that with PPPoA, rather than PPPoEoA with the pointless MTU > breakage that that implies. But it does work fine with BR2684 with the > MAC address 00:00:01:00:00:00, which is what it ends up with. >
br2684 picks that address if there is no 'set' mac and if the lower level provides nothing. It is a valid address, Xerox. I note that br2684 also has an ndo_set_mac_address(), enabling ifconfig nas0 hw ether OO:UU:II:aa:bb:cc > Do you actually *need* to set the MAC address on the br2684 link? > If the adsl/atm headend decides to actually bridge out onto a real ethernet, then it would be a good idea to have a reasonably unique MAC. How common is this? I wouldn't think it's too common, but probably would arise at some point. > Also, why is the br2684ctl package automatically starting br2684ctl at > boot, in the current AA builds? That actually prevents me from bringing > up the device in PPPoA mode. Surely br2684ctl should only be started if > it's actually been *configured*, and then only when the corresponding > network link is brought up? > I do not know. I was just routing my way around what the br2684 package does and does not do at start-up. It may be appropriate to have it set the mac either from a config option 'macaddr' like the ethernet bridge, or use a sensible default such as the system serial/MAC, with or without some adjustment like an increment or bit flip. Conor _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
