Re: Socket buffer sizes with autotuning

["Jerry Chu" <hkchu@google.com>]

On Tue, May 6, 2008 at 9:28 PM, David Miller <davem@davemloft.net> wrote:
> From: "Jerry Chu" <hkchu@google.com>
>
> Date: Tue, 6 May 2008 20:57:46 -0700
>
>
> > It may be getting a copy (e.g., when GSO is on?) hence losing all
>  > its connection to the original tp and any chance to have the pkt
>  > properly accounted for as host_infligh by TCP. The skb may also be
>  > cloned more than once (e.g., due to tcpdump)...
>
>  It only gets a copy if the SKB is already cloned and that clone is
>  alive somewhere (f.e. stuck in the device, a rare occurance by
>  the time we retransmit).

This is one of many things about skb that I still don't completely understand.
Why in tcp_transmit_skb() we'll have to pskb_copy() if skb_cloned()?
Can't we clone a skb mulitple times? Is it due to some special optimization
from skb->fclone stuff... that imposes this restriction?

>
>  It is a clone %99.999999 of the time.

When I turned on GSO but not TSO, I believe it's a copy %99.999999 of
the time. (I must confess I don't understand GSO code yet. I temporarily
ran out of stream after studying bunch of other code.)

Jerry

>