netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Sébastien Barré" <sebastien.barre@uclouvain.be>
To: David Miller <davem@davemloft.net>
Cc: "Sébastien Barré" <sebastien.barre@uclouvain.be>,
	netdev@vger.kernel.org,
	"Gregory Detal" <gregory.detal@uclouvain.be>,
	"Nandita Dukkipati" <nanditad@google.com>,
	"Yuchung Cheng" <ycheng@google.com>,
	"Neal Cardwell" <ncardwell@google.com>
Subject: [PATCH net-next v2] tcp: avoid reducing cwnd when ACK+DSACK is received
Date: Thu, 8 Jan 2015 13:20:09 +0100	[thread overview]
Message-ID: <1420719609-18638-1-git-send-email-sebastien.barre@uclouvain.be> (raw)

When the peer has delayed ack enabled, it may reply to a probe with an
ACK+D-SACK, with ack value set to tlp_high_seq. In the current code,
such ACK+DSACK will be missed and only at next, higher ack will the TLP
episode be considered done. Since the DSACK is not present anymore,
this will cost a cwnd reduction.

This patch ensures that this scenario does not cause a cwnd reduction, since
receiving an ACK+DSACK indicates that both the initial segment and the probe
have been received by the peer.

Cc: Gregory Detal <gregory.detal@uclouvain.be>
Cc: Nandita Dukkipati <nanditad@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Sébastien Barré <sebastien.barre@uclouvain.be>

---

Changes:
Applied Neal's comments:
-adapted commit first line
-moved logic to if condition, and removed is_tlp_dupack

Thanks Neal for those comments !

 net/ipv4/tcp_input.c | 30 +++++++++++++-----------------
 1 file changed, 13 insertions(+), 17 deletions(-)

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 075ab4d..cf63a29 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3363,29 +3363,25 @@ static void tcp_replace_ts_recent(struct tcp_sock *tp, u32 seq)
 static void tcp_process_tlp_ack(struct sock *sk, u32 ack, int flag)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
-	bool is_tlp_dupack = (ack == tp->tlp_high_seq) &&
-			     !(flag & (FLAG_SND_UNA_ADVANCED |
-				       FLAG_NOT_DUP | FLAG_DATA_SACKED));
 
 	/* Mark the end of TLP episode on receiving TLP dupack or when
 	 * ack is after tlp_high_seq.
+	 * With delayed acks, we may also get a regular ACK+DSACK, in which
+	 * case we don't want to reduce the cwnd either.
 	 */
-	if (is_tlp_dupack) {
+	if (((ack == tp->tlp_high_seq) &&
+	     !(flag & (FLAG_SND_UNA_ADVANCED |
+		       FLAG_NOT_DUP | FLAG_DATA_SACKED))) ||
+	    (!before(ack, tp->tlp_high_seq) && (flag & FLAG_DSACKING_ACK))) {
 		tp->tlp_high_seq = 0;
-		return;
-	}
-
-	if (after(ack, tp->tlp_high_seq)) {
+	} else if (after(ack, tp->tlp_high_seq)) {
 		tp->tlp_high_seq = 0;
-		/* Don't reduce cwnd if DSACK arrives for TLP retrans. */
-		if (!(flag & FLAG_DSACKING_ACK)) {
-			tcp_init_cwnd_reduction(sk);
-			tcp_set_ca_state(sk, TCP_CA_CWR);
-			tcp_end_cwnd_reduction(sk);
-			tcp_try_keep_open(sk);
-			NET_INC_STATS_BH(sock_net(sk),
-					 LINUX_MIB_TCPLOSSPROBERECOVERY);
-		}
+		tcp_init_cwnd_reduction(sk);
+		tcp_set_ca_state(sk, TCP_CA_CWR);
+		tcp_end_cwnd_reduction(sk);
+		tcp_try_keep_open(sk);
+		NET_INC_STATS_BH(sock_net(sk),
+				 LINUX_MIB_TCPLOSSPROBERECOVERY);
 	}
 }
 
-- 
tg: (44d84d7..) net-next/tlp-dsack-handling (depends on: net-next/master)

             reply	other threads:[~2015-01-08 12:24 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-01-08 12:20 Sébastien Barré [this message]
2015-01-08 15:07 ` [PATCH net-next v2] tcp: avoid reducing cwnd when ACK+DSACK is received Eric Dumazet
2015-01-08 15:24   ` Sébastien Barré
2015-01-08 15:43     ` Neal Cardwell
2015-01-08 16:00       ` Sébastien Barré
2015-01-08 15:59     ` Eric Dumazet
2015-01-08 15:19 ` Eric Dumazet
2015-01-08 15:39   ` Sébastien Barré
2015-01-08 15:49     ` Neal Cardwell
2015-01-08 15:52       ` Neal Cardwell
2015-01-08 16:25       ` Eric Dumazet
2015-01-08 17:17         ` Eric Dumazet
2015-01-08 17:27           ` Eric Dumazet
2015-01-09 19:43         ` Yuchung Cheng
2015-01-09 20:36           ` Neal Cardwell
2015-01-10 11:51           ` Sébastien Barré
2015-01-10 17:37             ` Eric Dumazet
2015-01-12 11:52           ` David Laight
2015-01-12 15:02             ` Neal Cardwell

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1420719609-18638-1-git-send-email-sebastien.barre@uclouvain.be \
    --to=sebastien.barre@uclouvain.be \
    --cc=davem@davemloft.net \
    --cc=gregory.detal@uclouvain.be \
    --cc=nanditad@google.com \
    --cc=ncardwell@google.com \
    --cc=netdev@vger.kernel.org \
    --cc=ycheng@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).