> > 2) modern networks almost never produce out-of-order fragments for packets
> > that are eventually accepted (this is a premise--needs to be tested).
>
> NFS produces zillions of them
UDP only currently, right? How does NFS product out-of-order fragments on
local networks on a regular basis? Is this all NFS implementations, or
just linux.
> > 3) workstations and server that do no fragment reassembly or at least no
> > out-of-order fragment reassembly would have simpler (and
> > better-performing) network code.
>
> Nope, its not on the fast path
But the code would still be simpler, no?
> > 4) if fragment reassembly (out-of-order or not) needs to be done, it
> > should be done by border routers (where people are increasingly willing to
> > do computationally intensive stuff like filter and reassemble fragments).
>
> We support this for internet stuff. However most fragmented packets are
> local
This is the part i could use some help understanding. As i understand it,
fragmentation only happens at routing when two links have different MTUs.
Is this covered by what you mean by 'local'?
todd
todd@unm.edu
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/