RFC3542 support on DragonFly BSD.

[Dashu Huang]

Hi Hasso
Another question, in your email, you said that "There is just no point
to preserve the code for compatibility with RFC2292", do you mean all
the RFC2292 related codes that have been obsolete by RFC3542 should be
removed from the OS (for example inet6_option_space,
inet6_rthdr_add()), or you just mean we can remove RH0?
As I know, in current FreeBSD OS those old RFC2292 related API such as
"inet6_option_space(), inet6_rthdr_add()" are preserved.

Thanks!
Huang Dashu

On Sun, Jul 27, 2008 at 3:24 AM, Hasso Tepper <hasso at estpak.ee> wrote:
> Dashu Huang wrote:
>> Hi All:
>>
>> This patch is for the Google summer code project of "RFC3542 support
>> on DragonFlyBSD", for more detail please see the document "The Design
>> and implementation of RFC3542 support on DragonFly BSD.pdf" in
>> http://leaf.dragonflybsd.org/~dhuang2/rfc3542/.
>
> Good work. I'm sure there are guys making more comments and more detailed
> review, but at first I have to kick myself not catching it earlier in the
> progress - routing header type 0 is really obsolete - see RFC5095 3:
> "IPv6 implementations are no longer required to implement RH0 in any
> way.". Therefore most of code handling RH0 can be just removed. There is
> just no point to preserve the code for compatibility with RFC2292.
>
> It does not mean that inet6_rth_*() functions could be removed - these
> have to stay as stubs (as they were AFAIK) and wait for code to handle
> type2 headers. You should be also careful not breaking whole world via
> removing members from structures, but the code handling these has to go
> (as most of rthdr.c in the kernel and related code in the ping6 for
> example).
>
>
> PS. Note that in vacation and not able to participate in discussion much,
> but I hope there are others who have opinion regarding the project :).
>
> --
> Hasso Tepper
>