Re: [PATCH net-next-2.6] net: speedup udp receive path

[jamal <hadi@cyberus.ca>]

On Sat, 2010-05-01 at 13:42 +0200, Eric Dumazet wrote:

> But, whole point of epoll is to not change interest each time you get an
> event.
> 
> Without EV_PERSIST, you need two more syscalls per recvfrom()
> 
> epoll_wait()
>  epoll_ctl(REMOVE)
>  epoll_ctl(ADD)
>  recvfrom()
> 
> Even poll() would be faster in your case
> 
> poll(one fd)
> recvfrom()
> 

This is true - but my goal was/is to replicate the regression i was
seeing[1]. 
I will try with PERSIST next opportunity. If it gets better
then it is something that needs documentation in the doc Tom
promised ;->

> I always thought copybreak was borderline...
> It can help to reduce memory footprint (allocating 128 bytes instead of
> 2048/4096 bytes per frame), but with RPS, it would make sense to perform
> copybreak after RPS, not before.
> 
> Reducing memory footprint also means less changes on
> udp_memory_allocated /tcp_memory_allocate (memory reclaim logic)

Indeed, something that didnt cross my mind in the rush to test - it is
one of those things that need to be mentioned in some doc somewhere.
Tom, are you listening? ;->

cheers,
jamal

[1]i.e with this program rps was getting worse (it was much better
before say net-next of apr14) and that non-rps has been getting better
numbers since. The regression is real - but it is likely in another
subsystem.