From mboxrd@z Thu Jan 1 00:00:00 1970 From: Neal Cardwell Subject: [PATCH net 2/3] tcp_bbr: reset full pipe detection on loss recovery undo Date: Thu, 7 Dec 2017 12:43:31 -0500 Message-ID: <20171207174332.17689-3-ncardwell@google.com> References: <20171207174332.17689-1-ncardwell@google.com> Cc: netdev@vger.kernel.org, Neal Cardwell To: David Miller Return-path: Received: from mail-yb0-f194.google.com ([209.85.213.194]:45919 "EHLO mail-yb0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752573AbdLGRnp (ORCPT ); Thu, 7 Dec 2017 12:43:45 -0500 Received: by mail-yb0-f194.google.com with SMTP id 184so3273303ybw.12 for ; Thu, 07 Dec 2017 09:43:45 -0800 (PST) In-Reply-To: <20171207174332.17689-1-ncardwell@google.com> Sender: netdev-owner@vger.kernel.org List-ID: Fix BBR so that upon notification of a loss recovery undo BBR resets the full pipe detection (STARTUP exit) state machine. Under high reordering, reordering events can be interpreted as loss. If the reordering and spurious loss estimates are high enough, this could previously cause BBR to spuriously estimate that the pipe is full. Since spurious loss recovery means that our overall sending will have slowed down spuriously, this commit gives a flow more time to probe robustly for bandwidth and decide the pipe is really full. Signed-off-by: Neal Cardwell Reviewed-by: Yuchung Cheng Acked-by: Soheil Hassas Yeganeh --- net/ipv4/tcp_bbr.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/ipv4/tcp_bbr.c b/net/ipv4/tcp_bbr.c index 3089c956b9f9..ab3ff14ea7f7 100644 --- a/net/ipv4/tcp_bbr.c +++ b/net/ipv4/tcp_bbr.c @@ -874,6 +874,10 @@ static u32 bbr_sndbuf_expand(struct sock *sk) */ static u32 bbr_undo_cwnd(struct sock *sk) { + struct bbr *bbr = inet_csk_ca(sk); + + bbr->full_bw = 0; /* spurious slow-down; reset full pipe detection */ + bbr->full_bw_cnt = 0; return tcp_sk(sk)->snd_cwnd; } -- 2.15.1.424.g9478a66081-goog