diff mbox

[REGRESSION,bisect] cxgb4 port failure with TSO traffic after commit 10d3be569243def8("tcp-tso: do not split TSO packets at retransmit time")

Message ID 1467034072.6850.152.camel@edumazet-glaptop3.roam.corp.google.com
State RFC, archived
Delegated to: David Miller
Headers show

Commit Message

Eric Dumazet June 27, 2016, 1:27 p.m. UTC 
On Mon, 2016-06-27 at 09:12 -0400, Neal Cardwell wrote:
> On Mon, Jun 27, 2016 at 8:46 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > Patch would be :
> >
> > diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> > index 8bd9911fdd16..3587efe22864 100644
> > --- a/net/ipv4/tcp_output.c
> > +++ b/net/ipv4/tcp_output.c
> > @@ -2784,6 +2784,10 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
> >                 segs = tp->snd_cwnd - tcp_packets_in_flight(tp);
> >                 if (segs <= 0)
> >                         return;
> > +               /* In case tcp_shift_skb_data() have aggregated large skbs,
> > +                * we need to make sure not sending too big TSO packets.
> > +                */
> > +               segs = min_t(int, segs, tp->gso_segs);
> >
> >                 if (fwd_rexmitting) {
> >  begin_fwd:
> 
> Nice catch, Eric. What do you think about using tcp_tso_autosize()
> instead of tp->gso_segs? The goal would be to get autosized skbs in
> this case, instead of 64KByte skbs. In addition to helping this corner
> case of SACK reneging, this might also help things in general, since
> if we are retransmitting packets then the cwnd and hence pacing rate
> and hence autosized skb length might be smaller now than they were
> when the packets were first sent. Just a thought.

Excellent idea ;)

Here is the v2 patch.

Comments

Neal Cardwell June 27, 2016, 1:36 p.m. UTC | #1 
On Mon, Jun 27, 2016 at 9:27 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Excellent idea ;)
>
> Here is the v2 patch.
>
> diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> index 8bd9911fdd16..e00e972c4e6a 100644
> --- a/net/ipv4/tcp_output.c
> +++ b/net/ipv4/tcp_output.c
> @@ -2751,7 +2751,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
>         struct tcp_sock *tp = tcp_sk(sk);
>         struct sk_buff *skb;
>         struct sk_buff *hole = NULL;
> -       u32 last_lost;
> +       u32 max_segs, last_lost;
>         int mib_idx;
>         int fwd_rexmitting = 0;
>
> @@ -2771,6 +2771,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
>                 last_lost = tp->snd_una;
>         }
>
> +       max_segs = tcp_tso_autosize(sk, tcp_current_mss(sk));
>         tcp_for_write_queue_from(skb, sk) {
>                 __u8 sacked = TCP_SKB_CB(skb)->sacked;
>                 int segs;
> @@ -2784,6 +2785,10 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
>                 segs = tp->snd_cwnd - tcp_packets_in_flight(tp);
>                 if (segs <= 0)
>                         return;
> +               /* In case tcp_shift_skb_data() have aggregated large skbs,
> +                * we need to make sure not sending too big TSO packets.
> +                */
> +               segs = min_t(int, segs, max_segs);
>
>                 if (fwd_rexmitting) {
>  begin_fwd:
>

Looks great to me. Thanks, Eric!

neal
Arjun Vynipadath June 27, 2016, 2:13 p.m. UTC | #2 
On Monday, June 06/27/16, 2016 at 19:06:05 +0530, Neal Cardwell wrote:
> On Mon, Jun 27, 2016 at 9:27 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > Excellent idea ;)
> >
> > Here is the v2 patch.
> >
> > diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> > index 8bd9911fdd16..e00e972c4e6a 100644
> > --- a/net/ipv4/tcp_output.c
> > +++ b/net/ipv4/tcp_output.c
> > @@ -2751,7 +2751,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
> >         struct tcp_sock *tp = tcp_sk(sk);
> >         struct sk_buff *skb;
> >         struct sk_buff *hole = NULL;
> > -       u32 last_lost;
> > +       u32 max_segs, last_lost;
> >         int mib_idx;
> >         int fwd_rexmitting = 0;
> >
> > @@ -2771,6 +2771,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
> >                 last_lost = tp->snd_una;
> >         }
> >
> > +       max_segs = tcp_tso_autosize(sk, tcp_current_mss(sk));
> >         tcp_for_write_queue_from(skb, sk) {
> >                 __u8 sacked = TCP_SKB_CB(skb)->sacked;
> >                 int segs;
> > @@ -2784,6 +2785,10 @@ void tcp_xmit_retransmit_queue(struct sock *sk)
> >                 segs = tp->snd_cwnd - tcp_packets_in_flight(tp);
> >                 if (segs <= 0)
> >                         return;
> > +               /* In case tcp_shift_skb_data() have aggregated large skbs,
> > +                * we need to make sure not sending too big TSO packets.
> > +                */
> > +               segs = min_t(int, segs, max_segs);
> >
> >                 if (fwd_rexmitting) {
> >  begin_fwd:
> >
> 
> Looks great to me. Thanks, Eric!
> 
> neal

This patch works fine. Now I am not seeing any packets with len > 64K. Thanks Eric for your support.

Arjun.
diff mbox

Patch

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 8bd9911fdd16..e00e972c4e6a 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2751,7 +2751,7 @@  void tcp_xmit_retransmit_queue(struct sock *sk)
 	struct tcp_sock *tp = tcp_sk(sk);
 	struct sk_buff *skb;
 	struct sk_buff *hole = NULL;
-	u32 last_lost;
+	u32 max_segs, last_lost;
 	int mib_idx;
 	int fwd_rexmitting = 0;
 
@@ -2771,6 +2771,7 @@  void tcp_xmit_retransmit_queue(struct sock *sk)
 		last_lost = tp->snd_una;
 	}
 
+	max_segs = tcp_tso_autosize(sk, tcp_current_mss(sk));
 	tcp_for_write_queue_from(skb, sk) {
 		__u8 sacked = TCP_SKB_CB(skb)->sacked;
 		int segs;
@@ -2784,6 +2785,10 @@  void tcp_xmit_retransmit_queue(struct sock *sk)
 		segs = tp->snd_cwnd - tcp_packets_in_flight(tp);
 		if (segs <= 0)
 			return;
+		/* In case tcp_shift_skb_data() have aggregated large skbs,
+		 * we need to make sure not sending too big TSO packets.
+		 */
+		segs = min_t(int, segs, max_segs);
 
 		if (fwd_rexmitting) {
 begin_fwd: