From: Andrew Lunn <and...@lunn.ch> Date: Mon, 17 Aug 2015 18:09:24 +0200
> This is RFC because i personally don't know if this is the best fix. > The patch restores previous behavior, while still keeping the bug fix. > > It is not obvious what is the correct source address for an IGMP > message when an interface has multiple addresses. IGMP messages are > sent either spontaneously, or as a result of a query. It could be > argued that when replying to a query, an address take from the same > subnet as the querier should be used. Doing this adds complexity for a > corner case which does not seem to effect people. In the spontaneous > case, there is no such hint, so an address has to be picked some other > way. Taking the highest scope address seems reasonable, and works for > me. Yes, unfortunately almost no guidance is given in this area in the various RFCs. Except that one recommended way to avoid forged IGMPs is to reject any IGMP that has a source address no on that interface's subnet. And if people actually do that, then the link scope address is the best address to use and that is what we do now if I understand things correctly. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
