tcp: tsq: relax tcp_small_queue_check() when rtx queue contains a single skb

In commit 75eefc6c59fd ("tcp: tsq: add a shortcut in tcp_small_queue_check()")
we allowed to send an skb regardless of TSQ limits being hit if rtx queue
was empty or had a single skb, in order to better fill the pipe
when/if TX completions were slow.

Then later, commit 75c119afe14f ("tcp: implement rb-tree based
retransmit queue") accidentally removed the special case for
one skb in rtx queue.

Stefan Wahren reported a regression in single TCP flow throughput
using a 100Mbit fec link, starting from commit 65466904b015 ("tcp: adjust
TSO packet sizes based on min_rtt"). This last commit only made the
regression more visible, because it locked the TCP flow on a particular
behavior where TSQ prevented two skbs being pushed downstream,
adding silences on the wire between each TSO packet.

Many thanks to Stefan for his invaluable help !

Fixes: 75c119afe14f ("tcp: implement rb-tree based retransmit queue")
Link: https://lore.kernel.org/netdev/7f31ddc8-9971-495e-a1f6-819df542e0af@gmx.net/
Reported-by: Stefan Wahren <wahrenst@gmx.net>
Tested-by: Stefan Wahren <wahrenst@gmx.net>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20231017124526.4060202-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index bbd85672fda70c8de259471311bf46fe5171c905..f0723460753c5d035022ce705a8ba533ce9ab2b8 100644 (file)
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2542,6 +2542,18 @@ static bool tcp_pacing_check(struct sock *sk)
        return true;
 }
 
+static bool tcp_rtx_queue_empty_or_single_skb(const struct sock *sk)
+{
+       const struct rb_node *node = sk->tcp_rtx_queue.rb_node;
+
+       /* No skb in the rtx queue. */
+       if (!node)
+               return true;
+
+       /* Only one skb in rtx queue. */
+       return !node->rb_left && !node->rb_right;
+}
+
 /* TCP Small Queues :
  * Control number of packets in qdisc/devices to two packets / or ~1 ms.
  * (These limits are doubled for retransmits)
@@ -2579,12 +2591,12 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb,
                limit += extra_bytes;
        }
        if (refcount_read(&sk->sk_wmem_alloc) > limit) {
-               /* Always send skb if rtx queue is empty.
+               /* Always send skb if rtx queue is empty or has one skb.
                 * No need to wait for TX completion to call us back,
                 * after softirq/tasklet schedule.
                 * This helps when TX completions are delayed too much.
                 */
-               if (tcp_rtx_queue_empty(sk))
+               if (tcp_rtx_queue_empty_or_single_skb(sk))
                        return false;
 
                set_bit(TSQ_THROTTLED, &sk->sk_tsq_flags);