dhcpcd-discuss

Good morning,

I am not sure whether this is related to dhcpcd at all. I have killed
dhcpcd and rely purely on the RAs in the network and kernel
autoconfiguration on 5.4.5 and 5.4.6.

The behaviour is identical. I posted about it to netdev@vger, however
did not yet get a reply.

Best regards,

Nico

Roy Marples <roy@xxxxxxxxxxxx> writes:

> On 01/01/2020 13:52, Nico Schottelius wrote:
>>
>> Update:
>>
>> On 5.4.7-arch1-1 the address also stays tentative for quite some time
>> (minutes?) but eventually transits into regular state:
>>
>> 2: wlp0s20f3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
>>      link/ether 24:ee:9a:54:c3:bf brd ff:ff:ff:ff:ff:ff
>>      inet6 2a0a:e5c0:10:bee:5fe8:6d2:b19a:3e8f/64 scope global dynamic mngtmpaddr noprefixroute
>>         valid_lft 86357sec preferred_lft 14357sec
>>      inet6 fe80::e00d:528:9af3:e20a/64 scope link
>>         valid_lft forever preferred_lft forever
>>
>> I wonder if something is broken with the DAD mechanism on these
>> kernels. I'll try to capture the MLD/icmp6 messages next time I connect
>> to that network.
>
> I still can't replicate this very easily, but does this patch fix it?
> https://roy.marples.name/cgit/dhcpcd.git/commit/?h=dhcpcd-8&id=114870290a8d3d696bc4049c32eef3eed03d6070
>
> Roy


--
Modern, affordable, Swiss Virtual Machines. Visit www.datacenterlight.ch