2018-07-13 00:50:38

by Jon Maxwell

Subject: [net-next, 3/3] tcp: convert icsk_user_timeout from jiffies to msecs

The final patch in this series. Create the tcp_clamp_rto_to_user_timeout()
helper routine. To calculate the correct rto, so that the TCP_USER_TIMEOUT
socket option is more accurate. Taking suggestions and feedback into account
from Eric Dumazet, Neal Cardwell and David Laight. Due to the 1st commit we
can avoid the msecs_to_jiffies() and jiffies_to_msecs() dance.

Signed-off-by: Jon Maxwell <[email protected]>
---
net/ipv4/tcp_timer.c | 19 ++++++++++++++++++-
1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 8ab8c9645294..d4d4b30a6bec 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -36,6 +36,22 @@ u32 tcp_retransmit_stamp(const struct sock *sk)
return start_ts;
}

+static u32 tcp_clamp_rto_to_user_timeout(const struct sock *sk)
+{
+ struct inet_connection_sock *icsk = inet_csk(sk);
+ u32 elapsed, start_ts;
+
+ start_ts = tcp_retransmit_stamp(sk);
+ if (!icsk->icsk_user_timeout || !start_ts)
+ return icsk->icsk_rto;
+ elapsed = tcp_time_stamp(tcp_sk(sk)) - start_ts;
+ if (elapsed >= icsk->icsk_user_timeout)
+ return 1; /* user timeout has passed; fire ASAP */
+ else
+ return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(icsk->icsk_user_timeout -
+ elapsed));
+}
+
/**
* tcp_write_err() - close socket and save error info
* @sk: The socket the error has appeared on.
@@ -544,7 +560,8 @@ void tcp_retransmit_timer(struct sock *sk)
/* Use normal (exponential) backoff */
icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX);
}
- inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, icsk->icsk_rto, TCP_RTO_MAX);
+ inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
+ tcp_clamp_rto_to_user_timeout(sk), TCP_RTO_MAX);
if (retransmits_timed_out(sk, net->ipv4.sysctl_tcp_retries1 + 1, 0))
__sk_dst_reset(sk);

--
2.13.6

2018-07-16 16:01:25

by Eric Dumazet

Subject: Re: [net-next, 3/3] tcp: convert icsk_user_timeout from jiffies to msecs

On 07/12/2018 05:49 PM, Jon Maxwell wrote:
> The final patch in this series. Create the tcp_clamp_rto_to_user_timeout()
> helper routine. To calculate the correct rto, so that the TCP_USER_TIMEOUT
> socket option is more accurate. Taking suggestions and feedback into account
> from Eric Dumazet, Neal Cardwell and David Laight. Due to the 1st commit we
> can avoid the msecs_to_jiffies() and jiffies_to_msecs() dance.
>
> Signed-off-by: Jon Maxwell <[email protected]>
> ---
> net/ipv4/tcp_timer.c | 19 ++++++++++++++++++-
> 1 file changed, 18 insertions(+), 1 deletion(-)
>
> diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
> index 8ab8c9645294..d4d4b30a6bec 100644
> --- a/net/ipv4/tcp_timer.c
> +++ b/net/ipv4/tcp_timer.c
> @@ -36,6 +36,22 @@ u32 tcp_retransmit_stamp(const struct sock *sk)
> return start_ts;
> }
>
> +static u32 tcp_clamp_rto_to_user_timeout(const struct sock *sk)
> +{
> + struct inet_connection_sock *icsk = inet_csk(sk);
> + u32 elapsed, start_ts;
> +
> + start_ts = tcp_retransmit_stamp(sk);
> + if (!icsk->icsk_user_timeout || !start_ts)
> + return icsk->icsk_rto;
> + elapsed = tcp_time_stamp(tcp_sk(sk)) - start_ts;
> + if (elapsed >= icsk->icsk_user_timeout)
> + return 1; /* user timeout has passed; fire ASAP */
> + else
> + return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(icsk->icsk_user_timeout -
> + elapsed));

if (elapsed >= icsk->icsk_user_timeout)
return 1; /* user timeout has passed; fire ASAP */

return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(icsk->icsk_user_timeout - elapsed));

Thanks !

2018-07-16 23:19:07

by Jon Maxwell

Subject: Re: [net-next, 3/3] tcp: convert icsk_user_timeout from jiffies to msecs

Sorry Eric, my bad. I'll do another version with the indentation fixed and
reformatting of min_t() and resubmit.

On Tue, Jul 17, 2018 at 2:00 AM, Eric Dumazet <[email protected]> wrote:
>
>
> On 07/12/2018 05:49 PM, Jon Maxwell wrote:
>> The final patch in this series. Create the tcp_clamp_rto_to_user_timeout()
>> helper routine. To calculate the correct rto, so that the TCP_USER_TIMEOUT
>> socket option is more accurate. Taking suggestions and feedback into account
>> from Eric Dumazet, Neal Cardwell and David Laight. Due to the 1st commit we
>> can avoid the msecs_to_jiffies() and jiffies_to_msecs() dance.
>>
>> Signed-off-by: Jon Maxwell <[email protected]>
>> ---
>> net/ipv4/tcp_timer.c | 19 ++++++++++++++++++-
>> 1 file changed, 18 insertions(+), 1 deletion(-)
>>
>> diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
>> index 8ab8c9645294..d4d4b30a6bec 100644
>> --- a/net/ipv4/tcp_timer.c
>> +++ b/net/ipv4/tcp_timer.c
>> @@ -36,6 +36,22 @@ u32 tcp_retransmit_stamp(const struct sock *sk)
>> return start_ts;
>> }
>>
>> +static u32 tcp_clamp_rto_to_user_timeout(const struct sock *sk)
>> +{
>> + struct inet_connection_sock *icsk = inet_csk(sk);
>> + u32 elapsed, start_ts;
>> +
>> + start_ts = tcp_retransmit_stamp(sk);
>> + if (!icsk->icsk_user_timeout || !start_ts)
>> + return icsk->icsk_rto;
>> + elapsed = tcp_time_stamp(tcp_sk(sk)) - start_ts;
>> + if (elapsed >= icsk->icsk_user_timeout)
>> + return 1; /* user timeout has passed; fire ASAP */
>> + else
>> + return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(icsk->icsk_user_timeout -
>> + elapsed));
>
> if (elapsed >= icsk->icsk_user_timeout)
> return 1; /* user timeout has passed; fire ASAP */
>
> return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(icsk->icsk_user_timeout - elapsed));
>
> Thanks !