Hey Simon, preparing here a new release of libremesh.org firmware, i was revising our dnsmasq conf, and gladly found out that this workaround: http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2013q4/007687.html is no longer needed! apparently, dnsmasq now handles this "interesting" setup correctly and sends a unicast OFFER/ACK over the correct interface
i assume you finally applied this patch http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2013q4/007770.html and i must say, i'm awfully ashamed of never having tested (or replied) your original attached patch in time :( :( for what it's worth, i can now confirm that dnsmasq works correctly and just to add some "bisect datapoints" for anyone interested: using openwrt/lede opkg package: dnsmasq-dhcpv6 - 2.71-5 the bug was still present but with dnsmasq-dhcpv6 - 2.76-6 everything works as expected didn't try anything in the middle, i'm more than happy with 2.76 :) thanks again simon for your neverending help, cheers! gui _______________________________________________ Dnsmasq-discuss mailing list Dnsmasq-discuss@lists.thekelleys.org.uk http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss