diff options
| author | Eric Dumazet <edumazet@google.com> | 2025-02-07 18:28:29 +0300 |
|---|---|---|
| committer | Paolo Abeni <pabeni@redhat.com> | 2025-02-11 15:08:00 +0300 |
| commit | 54a378f43425085d0684679d99735696b69165bc (patch) | |
| tree | 637db57f07abee77412a570f3e2028101c8b266c /net/ipv4/tcp_timer.c | |
| parent | 48b69b4c7e5d74ad66e5214ae8cbdae0b9ea154c (diff) | |
| download | linux-54a378f43425085d0684679d99735696b69165bc.tar.xz | |
tcp: add the ability to control max RTO
Currently, TCP stack uses a constant (120 seconds)
to limit the RTO value exponential growth.
Some applications want to set a lower value.
Add TCP_RTO_MAX_MS socket option to set a value (in ms)
between 1 and 120 seconds.
It is discouraged to change the socket rto max on a live
socket, as it might lead to unexpected disconnects.
Following patch is adding a netns sysctl to control the
default value at socket creation time.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Diffstat (limited to 'net/ipv4/tcp_timer.c')
| -rw-r--r-- | net/ipv4/tcp_timer.c | 17 |
1 files changed, 9 insertions, 8 deletions
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 6472f560f653..c0e601e4f39c 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -109,7 +109,7 @@ static int tcp_out_of_resources(struct sock *sk, bool do_reset) /* If peer does not open window for long time, or did not transmit * anything for long time, penalize it. */ - if ((s32)(tcp_jiffies32 - tp->lsndtime) > 2*TCP_RTO_MAX || !do_reset) + if ((s32)(tcp_jiffies32 - tp->lsndtime) > 2*tcp_rto_max(sk) || !do_reset) shift++; /* If some dubious ICMP arrived, penalize even more. */ @@ -189,12 +189,12 @@ static unsigned int tcp_model_timeout(struct sock *sk, { unsigned int linear_backoff_thresh, timeout; - linear_backoff_thresh = ilog2(TCP_RTO_MAX / rto_base); + linear_backoff_thresh = ilog2(tcp_rto_max(sk) / rto_base); if (boundary <= linear_backoff_thresh) timeout = ((2 << boundary) - 1) * rto_base; else timeout = ((2 << linear_backoff_thresh) - 1) * rto_base + - (boundary - linear_backoff_thresh) * TCP_RTO_MAX; + (boundary - linear_backoff_thresh) * tcp_rto_max(sk); return jiffies_to_msecs(timeout); } /** @@ -268,7 +268,7 @@ static int tcp_write_timeout(struct sock *sk) retry_until = READ_ONCE(net->ipv4.sysctl_tcp_retries2); if (sock_flag(sk, SOCK_DEAD)) { - const bool alive = icsk->icsk_rto < TCP_RTO_MAX; + const bool alive = icsk->icsk_rto < tcp_rto_max(sk); retry_until = tcp_orphan_retries(sk, alive); do_reset = alive || @@ -416,7 +416,8 @@ static void tcp_probe_timer(struct sock *sk) } max_probes = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_retries2); if (sock_flag(sk, SOCK_DEAD)) { - const bool alive = inet_csk_rto_backoff(icsk, TCP_RTO_MAX) < TCP_RTO_MAX; + unsigned int rto_max = tcp_rto_max(sk); + const bool alive = inet_csk_rto_backoff(icsk, rto_max) < rto_max; max_probes = tcp_orphan_retries(sk, alive); if (!alive && icsk->icsk_backoff >= max_probes) @@ -492,7 +493,7 @@ static bool tcp_rtx_probe0_timed_out(const struct sock *sk, const struct inet_connection_sock *icsk = inet_csk(sk); u32 user_timeout = READ_ONCE(icsk->icsk_user_timeout); const struct tcp_sock *tp = tcp_sk(sk); - int timeout = TCP_RTO_MAX * 2; + int timeout = tcp_rto_max(sk) * 2; s32 rcv_delta; if (user_timeout) { @@ -665,7 +666,7 @@ out_reset_timer: icsk->icsk_backoff = 0; icsk->icsk_rto = clamp(__tcp_set_rto(tp), tcp_rto_min(sk), - TCP_RTO_MAX); + tcp_rto_max(sk)); } else if (sk->sk_state != TCP_SYN_SENT || tp->total_rto > READ_ONCE(net->ipv4.sysctl_tcp_syn_linear_timeouts)) { @@ -673,7 +674,7 @@ out_reset_timer: * activated. */ icsk->icsk_backoff++; - icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX); + icsk->icsk_rto = min(icsk->icsk_rto << 1, tcp_rto_max(sk)); } tcp_reset_xmit_timer(sk, ICSK_TIME_RETRANS, tcp_clamp_rto_to_user_timeout(sk), false); |