Jonathan Stone <jonathan@dsg.stanford.edu> writes:

> >This seems like a philosophical point as much as a practical one, and
> >I can imagine [but will not make] an argument that lo0 is the
> >"natural" local transport and that PF_LOCAL is redundant, and that
> >your benchmarking usage is not the case we should be accomodating.
> 
> I wouldn't care to see anyone make that argument, given Jason's stated
> rationale was *specifically* for performance relative to Linux on
> certain [again, we seem to agree such benchmarks are bogus?].

Right, I had veered from the original stated purpose and into the
abstract mental territory of "starting from scratch, if I want to make
a connection to myself, what do I do?" Of course, we're not starting
from scratch, so we've got PF_LOCAL. I think the argument I
hypothesized would have more appeal with those people ("young kids" or
something) who have only ever known IP networking, and whose answer to
"How do I make a local connection?" would be IP-centric rather than
popping up a layer.

> I'm still not aware of any real-world case where lo0 throughput is a
> significant bottleneck?  Speed up something that's well under 10% of
> your application, by some 10%[*]? Not a big win.

I'm not either, so I agree here.

        - Nathan