Re: PROBLEM: Linux kernel 2.6.31 IPv4 TCP fails to open huge amount of outgoing connections (unable to bind ... )

[Evgeniy Polyakov <zbr@ioremap.net>]

On Wed, Apr 21, 2010 at 07:46:39AM +0200, Eric Dumazet (eric.dumazet@gmail.com) wrote:
> 		if (atomic_read(&hashinfo->bsockets) > (high-low)+1) {
> 			spin_unlock(&head->lock);
> 			snum = smallest_rover;   // We select this, without checking for
> conflicts.
> 			goto have_snum;
> 		}
> 	}
> 
> 
> Then we goto to "have_snum" label
> 
> Then we realize (selected_IP, randomport) is already in use.
> End of first try.
> 
> We redo the thing 5 times, so we only look at 5 slots out of
> 32000-64000.

We only break out of the loop in above case when number of sockets is
already more than our range limit. If we would just try 5 random times
out of 1000 in Gaspar's case, we would not be able to select all 1000
sockets.

> Maybe the fix would need to check if there is a conflict before doing
> the "goto have_snum"

I believe this is a useful patch, but it addresses a different issue.
This path should not fire up when we bind to single address.

> diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
> index e0a3e35..0498daf 100644
> --- a/net/ipv4/inet_connection_sock.c
> +++ b/net/ipv4/inet_connection_sock.c
> @@ -120,9 +120,11 @@ again:
>  						smallest_size = tb->num_owners;
>  						smallest_rover = rover;
>  						if (atomic_read(&hashinfo->bsockets) > (high - low) + 1) {
> -							spin_unlock(&head->lock);
> -							snum = smallest_rover;
> -							goto have_snum;
> +							if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb))
> +								spin_unlock(&head->lock);
> +								snum = smallest_rover;
> +								goto have_snum;
> +							}
>  						}
>  					}
>  					goto next;
> 

-- 
	Evgeniy Polyakov