Re: [PATCH net-next] tcp: introduce tcp_try_coalesce

[Eric Dumazet <eric.dumazet@gmail.com>]

On Mon, 2012-04-23 at 21:13 -0400, Neal Cardwell wrote:
> On Mon, Apr 23, 2012 at 1:11 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > From: Eric Dumazet <edumazet@google.com>
> >
> > commit c8628155ece3 (tcp: reduce out_of_order memory use) took care of
> > coalescing tcp segments provided by legacy devices (linear skbs)
> >
> > We extend this idea to fragged skbs, as their truesize can be heavy.
> >
> > ixgbe for example uses 256+1024+PAGE_SIZE/2 = 3328 bytes per segment.
> >
> > Use this coalescing strategy for receive queue too.
> >
> > This contributes to reduce number of tcp collapses, at minimal cost, and
> > reduces memory overhead and packets drops.
> 
> The mechanics look solid, but I'm a little concerned about the
> potential added overhead for the new case where tcp_try_coalesce()
> does a skb_copy_bits() for in-order data that it is coalescing at the
> end of the sk_receive_queue. Do you have any performance numbers for
> this case to help suggest whether this added copy is a concern?
> 
> neal

This never happens on connections where performance matters : skb head
can only contains one full mss segment.

This part is only used on wifi devices, where skb head is really fat.