diff options
| author | Ben Hutchings <bhutchings@solarflare.com> | 2012-07-30 16:11:42 +0000 |
|---|---|---|
| committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2012-10-02 09:47:04 -0700 |
| commit | 09c403dc7c8e73a4f9c553e4d84aef68af25ff65 (patch) | |
| tree | 0a6bfbdde7aa192140a8b7398e58d3f9f9a7b953 /net/ipv4/tcp_output.c | |
| parent | 8a15a4b44f4f3348db373f70103b25a1175ea25b (diff) | |
tcp: Apply device TSO segment limit earlier
[ Upstream commit 1485348d2424e1131ea42efc033cbd9366462b01 ]
Cache the device gso_max_segs in sock::sk_gso_max_segs and use it to
limit the size of TSO skbs. This avoids the need to fall back to
software GSO for local TCP senders.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diffstat (limited to 'net/ipv4/tcp_output.c')
| -rw-r--r-- | net/ipv4/tcp_output.c | 21 |
1 files changed, 12 insertions, 9 deletions
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index faf257b9415..e0b8bd197fc 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1310,21 +1310,21 @@ static void tcp_cwnd_validate(struct sock *sk) * when we would be allowed to send the split-due-to-Nagle skb fully. */ static unsigned int tcp_mss_split_point(struct sock *sk, struct sk_buff *skb, - unsigned int mss_now, unsigned int cwnd) + unsigned int mss_now, unsigned int max_segs) { struct tcp_sock *tp = tcp_sk(sk); - u32 needed, window, cwnd_len; + u32 needed, window, max_len; window = tcp_wnd_end(tp) - TCP_SKB_CB(skb)->seq; - cwnd_len = mss_now * cwnd; + max_len = mss_now * max_segs; - if (likely(cwnd_len <= window && skb != tcp_write_queue_tail(sk))) - return cwnd_len; + if (likely(max_len <= window && skb != tcp_write_queue_tail(sk))) + return max_len; needed = min(skb->len, window); - if (cwnd_len <= needed) - return cwnd_len; + if (max_len <= needed) + return max_len; return needed - needed % mss_now; } @@ -1551,7 +1551,8 @@ static int tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb) limit = min(send_win, cong_win); /* If a full-sized TSO skb can be sent, do it. */ - if (limit >= sk->sk_gso_max_size) + if (limit >= min_t(unsigned int, sk->sk_gso_max_size, + sk->sk_gso_max_segs * tp->mss_cache)) goto send_now; /* Middle in queue won't get any more data, full sendable already? */ @@ -1777,7 +1778,9 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle, limit = mss_now; if (tso_segs > 1 && !tcp_urg_mode(tp)) limit = tcp_mss_split_point(sk, skb, mss_now, - cwnd_quota); + min_t(unsigned int, + cwnd_quota, + sk->sk_gso_max_segs)); if (skb->len > limit && unlikely(tso_fragment(sk, skb, limit, mss_now, gfp))) |