[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CADVnQynNtha0Vc=oG=qiH94tAmWuQHjMcU-DFNudQdV_53+jeA@mail.gmail.com>
Date: Sun, 21 Jul 2013 22:45:13 -0400
From: Neal Cardwell <ncardwell@...gle.com>
To: Yuchung Cheng <ycheng@...gle.com>
Cc: David Miller <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Nandita Dukkipati <nanditad@...gle.com>,
Netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH 1/4 net-next] tcp: consolidate SYNACK RTT sampling
On Sun, Jul 21, 2013 at 1:39 AM, Yuchung Cheng <ycheng@...gle.com> wrote:
> This patch series improve RTT sampling in three ways:
> 1. Sample RTT during fast recovery and reordering events.
> 2. Favor ack-based RTT to timestamps because of broken TS ECR fields
> 3. Consolidate the RTT measurement logic.
>
> The first patch consolidates SYNACK and other RTT measurement to use a
> central function tcp_ack_update_rtt(). A (small) bonus is now SYNACK
> RTT measurement happens after PAWS check, potentially reducing the
> impact of RTO seeding on bad TCP timestamps values.
>
> Signed-off-by: Yuchung Cheng <ycheng@...gle.com>
The idea seems very nice to me, but I think a line or two got dropped
in a rebase. :-)
In the previous version of this patch you moved the TFO-specific call
to tcp_synack_rtt_meas() out to a spot in tcp_rcv_state_process()
where both TFO and non-TFO passive connections would hit it. In this
version of the patch it removes the existing calls to
tcp_synack_rtt_meas() but does not add this new call to replace them.
(And it does not seem to be added back in a later patch in the
series.)
Other than that, looks great.
neal
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists