RE: [net-next,v3] tcp: Improve setsockopt() TCP_USER_TIMEOUT accuracy

From: David Laight <David.Laight@ACULAB.COM>
To: 'Jon Maxwell' <jmaxwell37@gmail.com>,
	"davem@davemloft.net" <davem@davemloft.net>
Cc: "edumazet@google.com" <edumazet@google.com>,
	"eric.dumazet@gmail.com" <eric.dumazet@gmail.com>,
	"ncardwell@google.com" <ncardwell@google.com>,
	"kuznet@ms2.inr.ac.ru" <kuznet@ms2.inr.ac.ru>,
	"yoshfuji@linux-ipv6.org" <yoshfuji@linux-ipv6.org>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"jmaxwell@redhat.com" <jmaxwell@redhat.com>
Subject: RE: [net-next,v3] tcp: Improve setsockopt() TCP_USER_TIMEOUT accuracy
Date: Tue, 10 Jul 2018 09:17:27 +0000	[thread overview]
Message-ID: <5f34e43135ed4948bbc32849289aeb20@AcuMS.aculab.com> (raw)
In-Reply-To: <20180710065147.27647-1-jmaxwell37@gmail.com>

From: Jon Maxwell
> Sent: 10 July 2018 07:52
...
> +static __u32 tcp_clamp_rto_to_user_timeout(struct sock *sk)
> +{
> +	struct inet_connection_sock *icsk = inet_csk(sk);
> +	__u32 elapsed, user_timeout;
> +	u32 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;
> +	user_timeout = jiffies_to_msecs(icsk->icsk_user_timeout);
> +	if (elapsed >= user_timeout)
> +		return 1; /* user timeout has passed; fire ASAP */
> +	else
> +		return (icsk->icsk_rto < msecs_to_jiffies(user_timeout - elapsed)) ?
> +			icsk->icsk_rto : msecs_to_jiffies(user_timeout - elapsed);

You've not looked at what the above actually does - something like:
msecs_to_jiffies() probably rounds up, the old code rounded down
so dividing by HZ is probably right.

	int new_tmo = icsk->icsk_user_timeout - (tcp_time_stamp(tcp_sk(sk)) - start_ts)/HZ;

	if (new_tmo > (int) icsk->icsk_rto;
		return icsk->icsk_rto;
	return new_tmo <= 0 ? 1 : new_tmo;
}