Commit 3acf3ec3 authored by Lawrence Brakmo's avatar Lawrence Brakmo Committed by David S. Miller

tcp: Change txhash on every SYN and RTO retransmit

The current code changes txhash (flowlables) on every retransmitted
SYN/ACK, but only after the 2nd retransmitted SYN and only after
tcp_retries1 RTO retransmits.

With this patch:
1) txhash is changed with every SYN retransmits
2) txhash is changed with every RTO.

The result is that we can start re-routing around failed (or very
congested paths) as soon as possible. Otherwise application health
checks may fail and the connection may be terminated before we start
to change txhash.

v4: Removed sysctl, txhash is changed for all RTOs
v3: Removed text saying default value of sysctl is 0 (it is 100)
v2: Added sysctl documentation and cleaned code

Tested with packetdrill tests
Signed-off-by: default avatarLawrence Brakmo <brakmo@fb.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent fa5effe7
...@@ -192,6 +192,8 @@ static int tcp_write_timeout(struct sock *sk) ...@@ -192,6 +192,8 @@ static int tcp_write_timeout(struct sock *sk)
if (tp->syn_data && icsk->icsk_retransmits == 1) if (tp->syn_data && icsk->icsk_retransmits == 1)
NET_INC_STATS(sock_net(sk), NET_INC_STATS(sock_net(sk),
LINUX_MIB_TCPFASTOPENACTIVEFAIL); LINUX_MIB_TCPFASTOPENACTIVEFAIL);
} else if (!tp->syn_data && !tp->syn_fastopen) {
sk_rethink_txhash(sk);
} }
retry_until = icsk->icsk_syn_retries ? : net->ipv4.sysctl_tcp_syn_retries; retry_until = icsk->icsk_syn_retries ? : net->ipv4.sysctl_tcp_syn_retries;
syn_set = true; syn_set = true;
...@@ -213,6 +215,8 @@ static int tcp_write_timeout(struct sock *sk) ...@@ -213,6 +215,8 @@ static int tcp_write_timeout(struct sock *sk)
tcp_mtu_probing(icsk, sk); tcp_mtu_probing(icsk, sk);
dst_negative_advice(sk); dst_negative_advice(sk);
} else {
sk_rethink_txhash(sk);
} }
retry_until = net->ipv4.sysctl_tcp_retries2; retry_until = net->ipv4.sysctl_tcp_retries2;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment