<html><body>
<p><tt>>So, yes, currently, IPoIB is broken in that DAD for new addresses is<br>
>not synchronized to the SM join. But, DAD for startup addresses is<br>
>OK due to the trick that is played with carrier. Your patch breaks<br>
>both equally :></tt><br>
<br>
As you descripted, DAD is not some mechanism to prevent duplicated address. The default DAD timer is 1HZ. Even the solicited node IB multicast join successfully before the carrier on, that doesn't mean we won't get any potentical duplicated link local addresses at all. For example, the NA returns back after 1HZ. <br>
<br>
Comparing to the IPoIB accessibility(let IPv4 working) with playing trick to carrier on to avoid IPv6 link local DAD in a small possibility, this patch is a better choice for switches with limited MGCs resouce today in a large cluster. Then IPoIB will have the same behavior as IPv4 and IPv6 over ethernet: interface will be up and running no matter whether there is any duplicated address or not.<br>
<br>
Thanks<br>
Shirley Ma</body></html>