[LEDE-DEV] impressed by progress on odhcpd and procd and elsewhere.

[Hans Dedecker]

On Friday, 10 February 2017 09:11:17 CET Dave Taht wrote:
> Hans, jow, blogic, everybody...
> 
> I have been pulling lede-head stuff and starting to build lede again
> (for the first time in 3 years), but I'd given up on the ipv6
> issues[1].
> 
> After reviewing odhcpd/procd/etc latest git commits today, I'm impressed.
> 
> My intent is to return to testing a few things this weekend - using
> odhcpd as the default... if it is "ready enough"?
Refactoring was done to make the odhcpd implementation more stable and less 
error prone; still further work is required to improve its behavior. 
> 
> 0) More wifi ATF work at scale (adding ipv6 back in as a first class test)
> 0.1) cake profiling
> 
> 1) IPv6 & comcast
> 2) 6rd on sonic fiber
> 3) DHCPv6-PD from an internal router to edgerouters and to lede
> 4) Poking at multicast (uftpd flooding, primarily, but also babeld and mdns)
> 
> About the biggest request from the field seems to be supplying RA and
> dhcpv6 dns on a more real IPv6 address than fe80 to make android users
> happier (https://github.com/dtaht/dnsmasq-lede/issues/11).
My idea would be to first pass the ULA IPv6 address; next a GUA IPv6 address 
and finally the link local IPv6 address  as DNS server

Hans
> 
> Please let me know what I can best do to help, not be noisy, and stay
> out of the way, and yet be useful.
> 
> Thanks for jumping all over these issues!
> 
> 
> [1] And ranting. TL;DR - https://github.com/dtaht/libv6 I am painfully
> aware of how much ipv6 can suck.