[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1402331047.3645.379.camel@edumazet-glaptop2.roam.corp.google.com>
Date: Mon, 09 Jun 2014 09:24:07 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: Per Hurtig <per.hurtig@....se>
Cc: Nandita Dukkipati <nanditad@...gle.com>,
Netdev <netdev@...r.kernel.org>,
Anna Brunström <anna.brunstrom@....se>,
mohammad.rajiullah@....se, Neal Cardwell <ncardwell@...gle.com>,
Sergei Shtylyov <sergei.shtylyov@...entembedded.com>
Subject: Re: [PATCH] tcp: fixing TLP's FIN recovery
On Mon, 2014-06-09 at 09:15 -0700, Eric Dumazet wrote:
> On Mon, 2014-06-09 at 17:56 +0200, Per Hurtig wrote:
>
> >
> > Ok, I guess you mean that the retransmission was not fast enough? But
> > will the same not happen if the original FIN is not lost and triggers a
> > SACK (i.e., if the two last data segments are still lost)?
>
> Yes, I'd like to understand why Nandita specifically added the original
> test. In her tests, fast retransmit was not really effective.
>
> Running packetdrill in a separate container gives me these interesting
> counters :
>
> # nstat
> #kernel
> IpInReceives 4 0.0
> IpInDelivers 4 0.0
> IpOutRequests 5 0.0
> TcpPassiveOpens 1 0.0
> TcpInSegs 4 0.0
> TcpOutSegs 10 0.0
> TcpRetransSegs 2 0.0
> TcpExtTCPPureAcks 3 0.0
> TcpExtTCPSackRecovery 1 0.0
> TcpExtTCPFastRetrans 1 0.0
> TcpExtTCPLossProbes 1 0.0
> TcpExtTCPSackRecoveryFail 1 0.0
> TcpExtTCPSackShiftFallback 1 0.0
> TcpExtTCPRetransFail 4 0.0 <<<< HERE >>>
> TcpExtTCPOrigDataSent 9 0.0
> IpExtInOctets 184 0.0
> IpExtOutOctets 9212 0.0
> IpExtInNoECTPkts 4 0.0
>
> I guess we need to understand why the retransmit is in error.
>
> I am investigating.
>
Hmm... We hit this point... This is embarrassing I guess.
if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk))
return -EHOSTUNREACH; /* Routing failure or similar. */
--
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