T48G IPv6 SLAAC regression with V80

[bschmidt]

Hi,

we're currently testing T48G phones for a company wide replacement of our legacy Snom hardphones. We're pretty happy, but have hit an annoying regression regarding IPv6 in recent firmware versions.

I'll reply to this post with more dumps, but the short story looks like this.

Our test network is dualstacked with IPv4 DHCP and IPv6 SLAAC+stateless DHCPv6 (that means the address is assigned by SLAAC (via router advertisement), and the client/phone can request additional information using stateless DHCPv6 information request as described in RFC 3736 if necessary. This is signaled in the router advertisement using the O (Other Configuration) Flag.

We haven't tested DHCPv6 in this mix so far, but up to the latest 35.73.* versions this worked quite well. The phone got an IPv4 address from DHCP, an IPv6 address from SLAAC and was contacting dualstacked SIP servers via IPv6.

With all 35.80.* versions the phone does assign an IPv6 address from SLAAC (it is pingable and the Web GUI is reachable on that address), but the SIP part does not seem to accept it. The phone is only doing A DNS queries for the SIP servers (not AAAA), is connecting to dualstack SIP servers using IPv4 and cannot connect to IPv6 only servers at all. Additionally it is regularly sending out stateful DHCPv6 solicitations despite stateful DHCPv6 not being enabled in this network.

Can anyone from Yealink shed some light on this. SLAAC with stateless DHCPv6 would be our preferred operating mode.

Note that I am not sure whether 35.73.* did everything correctly here, as we didn't test IPv6-only operation with stateless DHCPv6 so far. But dualstacked with SLAAC did work.

Thanks,
Bernhard