FreeLists / openbeosnetteam / [openbeosnetteam] Re: Status

[Waldemar Kornewald <wkornew@xxxxxxx>]

Philippe Houdoin wrote:
> I've made in the past a lame attempt to work on a bone_data like 
> structure which
> was supporting attributes to packets, and David Reid wrote also something to
> replace mbufs for his (dead?) Marrow project.

While the attribute-stuff sounded like a very nice concept our 
benchmark results
were not very promising. It's too slow for a real netstack concept.

Axel DÃrfler wrote:
> Since it seems we're currently getting a new idea about porting
> FreeBSD, I think that we need a module like that to base the whole
> stack upon.
> The implementation is allowed to be naive for now, but the API is
> important to get right.
> I think we should start with bone_data and change/simplify/extend
> whereever we see the need.

You must be crazy. This will cost a *lot* of time (IPv4+6, mobile IPv6, 
IPsec,
WiFi). OTOH, we'd have a nice and clean netstack. I'm not really against it.
:))

I doubt that BONE has a speed advantage over FreeBSD (probably it's even
slower). AFAIK, the main reason for bone_data was fragmentation, but this has
become less of a problem and IPv6 even disallows fragmentation (i.e.: the
sending host must know the route's MTU). I'd prefer a structure that is a
simple linked list of packet buffers (every packet is a buffer) instead of a
list of pointers to buffers and free() functions. But that's just IMHO and I'm
sure you have more experience in this area.

Bye,
Waldemar Kornewald