Re: [lwip-users] performance

[Leon Woestenberg]

Hello Sam,

thanks for your insight.

Sam Jansen wrote:

> Leon Woestenberg wrote:
>
> > (I think the major part of what determines the performance on such
> > configurations is the device driver and processor plus bus, not so much
>
>
> But the original question was "does it perform well on a high speed 
> network" and I'll answer that with: if it's buffer size is large 
> enough, then yes. It wont get as much throughput as FreeBSD, but 
> you'll probably find the difference isn't so great. Of course, this is 
> all assuming perfect hardware and device drivers.
I do not have the means to test lwIP outside embedded environments much 
(due to time constraints, not hardware per se), but I really wonder 
whether the outgoing packet queueing (together with DMA) keep the 
Ethernet line full on a back-to-back basis.

lwIP hands off an outgoing packet to the device driver, returns to the 
application, which may in turn send a new packet. There is no queue 
pressure against the device driver and there is some loop latency in 
this thread of execution.

Ideal would be (I think) a queue from which the driver can read, an lwIP 
can write against, and have driver and application run in a concurrent 
(and asynchrone) fashion. I think the BSDs and Linux do this this way, I 
am not sure though.

Regards,

Leon.