On Mon, 7 Dec 2015, Per Hurtig wrote:
> This patch implements the RTO restart modification (RTOR). When data is
> ACKed, and the RTO timer is restarted, the time elapsed since the last
> outstanding segment was transmitted is subtracted from the calculated RTO
> value. This way, the RTO timer will expire after exactly RTO seconds, and
> not RTO + RTT [+ delACK] seconds.
>
> This patch also implements a new sysctl (tcp_timer_restart) that is used
> to control the timer restart behavior.
>
> Signed-off-by: Per Hurtig <[email protected]>
> ---
> Documentation/networking/ip-sysctl.txt | 12 ++++++++++++
> include/net/tcp.h | 4 ++++
> net/ipv4/sysctl_net_ipv4.c | 10 ++++++++++
> net/ipv4/tcp_input.c | 24 ++++++++++++++++++++++++
> 4 files changed, 50 insertions(+)
>
> diff --git a/Documentation/networking/ip-sysctl.txt
> b/Documentation/networking/ip-sysctl.txt
> index 2ea4c45..4094128 100644
> --- a/Documentation/networking/ip-sysctl.txt
> +++ b/Documentation/networking/ip-sysctl.txt
> @@ -591,6 +591,18 @@ tcp_syn_retries - INTEGER
> with the current initial RTO of 1second. With this the final timeout
> for an active TCP connection attempt will happen after 127seconds.
>
> +tcp_timer_restart - INTEGER
> + Controls how the RTO and PTO timers are restarted (RTOR and TLPR).
> + If set (per timer or combined) the timers are restarted with
> + respect to the earliest outstanding segment, to not extend tail loss
> + latency unnecessarily.
> + Possible values:
> + 0 disables RTOR and TLPR.
> + 1 enables RTOR.
> + 2 enables TLPR.
> + 3 enables RTOR and TLPR.
> + Default: 3
> +
> tcp_timestamps - BOOLEAN
> Enable timestamps as defined in RFC1323.
>
[...snip...]
> diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
> index fdd88c3..66e0425 100644
> --- a/net/ipv4/tcp_input.c
> +++ b/net/ipv4/tcp_input.c
[...snip...]
> /* Restart timer after forward progress on connection.
> * RFC2988 recommends to restart timer to now+rto.
> */
> @@ -3027,6 +3040,17 @@ void tcp_rearm_rto(struct sock *sk)
> */
> if (delta > 0)
> rto = delta;
> + } else if (icsk->icsk_pending == ICSK_TIME_RETRANS &&
> + (sysctl_tcp_timer_restart == 1 ||
> + sysctl_tcp_timer_restart == 3) &&
Use a bit operation here instead? Also I think that this sysctl would
benefit from named constants rather than use of literals (similar
comment applies to the other patch too).
> + (tp->packets_out + tcp_unsent_pkts(sk) <
> + TCP_RTORESTART_THRESH)) {
> + struct sk_buff *skb = tcp_write_queue_head(sk);
> + const u32 rto_time_stamp = tcp_skb_timestamp(skb);
> + s32 delta = (s32)(tcp_time_stamp - rto_time_stamp);
> +
> + if (delta > 0 && rto > delta)
> + rto -= delta;
> }
> inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, rto,
> TCP_RTO_MAX);
--
i.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
