From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Willy Tarreau , Eric Dumazet , "David S. Miller" Subject: [ 56/61] tcp: dont abort splice() after small transfers Date: Tue, 12 Feb 2013 12:35:16 -0800 Message-Id: <20130212203425.409719868@linuxfoundation.org> In-Reply-To: <20130212203417.890993903@linuxfoundation.org> References: <20130212203417.890993903@linuxfoundation.org> Sender: linux-kernel-owner@vger.kernel.org List-ID: 3.7-stable review patch. If anyone has any objections, please let me know. ------------------ From: Willy Tarreau [ Upstream commit 02275a2ee7c0ea475b6f4a6428f5df592bc9d30b ] TCP coalescing added a regression in splice(socket->pipe) performance, for some workloads because of the way tcp_read_sock() is implemented. The reason for this is the break when (offset + 1 != skb->len). As we released the socket lock, this condition is possible if TCP stack added a fragment to the skb, which can happen with TCP coalescing. So let's go back to the beginning of the loop when this happens, to give a chance to splice more frags per system call. Doing so fixes the issue and makes GRO 10% faster than LRO on CPU-bound splice() workloads instead of the opposite. Signed-off-by: Willy Tarreau Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman --- net/ipv4/tcp.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1490,15 +1490,19 @@ int tcp_read_sock(struct sock *sk, read_ copied += used; offset += used; } - /* - * If recv_actor drops the lock (e.g. TCP splice + /* If recv_actor drops the lock (e.g. TCP splice * receive) the skb pointer might be invalid when * getting here: tcp_collapse might have deleted it * while aggregating skbs from the socket queue. */ - skb = tcp_recv_skb(sk, seq-1, &offset); - if (!skb || (offset+1 != skb->len)) + skb = tcp_recv_skb(sk, seq - 1, &offset); + if (!skb) break; + /* TCP coalescing might have appended data to the skb. + * Try to splice more frags + */ + if (offset + 1 != skb->len) + continue; } if (tcp_hdr(skb)->fin) { sk_eat_skb(sk, skb, false);