[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1472488885.14381.260.camel@edumazet-glaptop3.roam.corp.google.com>
Date: Mon, 29 Aug 2016 09:41:25 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: Jia He <hejianet@...il.com>
Cc: netdev@...r.kernel.org, linux-sctp@...r.kernel.org,
linux-kernel@...r.kernel.org, davem@...emloft.net,
Alexey Kuznetsov <kuznet@....inr.ac.ru>,
James Morris <jmorris@...ei.org>,
Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
Patrick McHardy <kaber@...sh.net>,
Vlad Yasevich <vyasevich@...il.com>,
Neil Horman <nhorman@...driver.com>,
Steffen Klassert <steffen.klassert@...unet.com>,
Herbert Xu <herbert@...dor.apana.org.au>
Subject: Re: [RFC PATCH 1/6] proc: Reduce cache miss in
{snmp,netstat}_seq_show
On Tue, 2016-08-30 at 00:03 +0800, Jia He wrote:
> This patch exchanges the two loop for collecting the percpu statistics
> data. This can aggregate the data by going through all the items of each
> cpu sequentially. In snmp_seq_show, just use one buff copy to dislay the
> Udp and UdpLite data because they are the same.
This is obviously not true.
On my laptop it seems it handled no UdpLite frames, but got plenty of
Udp ones.
$ grep Udp /proc/net/snmp
Udp: InDatagrams NoPorts InErrors OutDatagrams RcvbufErrors SndbufErrors InCsumErrors
Udp: 1555889 108318 0 3740780 0 0 0
UdpLite: InDatagrams NoPorts InErrors OutDatagrams RcvbufErrors SndbufErrors InCsumErrors
UdpLite: 0 0 0 0 0 0 0
Powered by blists - more mailing lists