tcp: tsq: restore minimal amount of queueing

After commit c9eeec26e32e ("tcp: TSQ can use a dynamic limit"), several
users reported throughput regressions, notably on mvneta and wifi
adapters.

802.11 AMPDU requires a fair amount of queueing to be effective.

This patch partially reverts the change done in tcp_write_xmit()
so that the minimal amount is sysctl_tcp_limit_output_bytes.

It also remove the use of this sysctl while building skb stored
in write queue, as TSO autosizing does the right thing anyway.

Users with well behaving NICS and correct qdisc (like sch_fq),
can then lower the default sysctl_tcp_limit_output_bytes value from
128KB to 8KB.

This new usage of sysctl_tcp_limit_output_bytes permits each driver
authors to check how their driver performs when/if the value is set
to a minimum of 4KB.

Normally, line rate for a single TCP flow should be possible,
but some drivers rely on timers to perform TX completion and
too long TX completion delays prevent reaching full throughput.

Fixes: c9eeec26e32e ("tcp: TSQ can use a dynamic limit")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Sujith Manoharan <sujith@msujith.org>
Reported-by: Arnaud Ebalard <arno@natisbad.org>
Tested-by: Sujith Manoharan <sujith@msujith.org>
Cc: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

2 files changed, 5 insertions, 7 deletions

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 8e8529d3c8c9..3dc0c6cf02a8 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -808,12 +808,6 @@ static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
                 xmit_size_goal = min_t(u32, gso_size,
                                        sk->sk_gso_max_size - 1 - hlen);
 
-                /* TSQ : try to have at least two segments in flight
-                 * (one in NIC TX ring, another in Qdisc)
-                 */
-                xmit_size_goal = min_t(u32, xmit_size_goal,
-                                       sysctl_tcp_limit_output_bytes >> 1);
-
                 xmit_size_goal = tcp_bound_to_half_wnd(tp, xmit_size_goal);
 
                 /* We try hard to avoid divides here */
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 672854664ff5..c5231d9b06d7 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1875,8 +1875,12 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
                  *  - better RTT estimation and ACK scheduling
                  *  - faster recovery
                  *  - high rates
+                 * Alas, some drivers / subsystems require a fair amount
+                 * of queued bytes to ensure line rate.
+                 * One example is wifi aggregation (802.11 AMPDU)
                  */
-                limit = max(skb->truesize, sk->sk_pacing_rate >> 10);
+                limit = max_t(unsigned int, sysctl_tcp_limit_output_bytes,
+                              sk->sk_pacing_rate >> 10);
 
                 if (atomic_read(&sk->sk_wmem_alloc) > limit) {
                         set_bit(TSQ_THROTTLED, &tp->tsq_flags);