[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20160113.103825.475570791449704545.davem@davemloft.net>
Date: Wed, 13 Jan 2016 10:38:25 -0500 (EST)
From: David Miller <davem@...emloft.net>
To: edumazet@...gle.com
Cc: hans.westgaard.ry@...cle.com, hannes@...essinduktion.org,
David.Laight@...lab.com, kuznet@....inr.ac.ru, jmorris@...ei.org,
yoshfuji@...ux-ipv6.org, kaber@...sh.net, ast@...mgrid.com,
jiri@...lanox.com, daniel@...earbox.net, nicolas.dichtel@...nd.com,
ebiederm@...ssion.com, noureddine@...sta.com, jarod@...hat.com,
makita.toshiaki@....ntt.co.jp, ja@....bg, ying.xue@...driver.com,
kraig@...gle.com, mgorman@...hsingularity.net, edjee@...gle.com,
julia.lawall@...6.fr, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, haakon.bugge@...cle.com,
knut.omang@...cle.com, wei.lin.guay@...cle.com,
santosh.shilimkar@...cle.com, yuval.shaia@...cle.com
Subject: Re: [PATCH] net: add per device sg_max_frags for skb
From: Eric Dumazet <edumazet@...gle.com>
Date: Wed, 13 Jan 2016 06:19:11 -0800
> 2) TCP stack has coalescing support. write(2) or sendmsg(2) should
> append data into the last skb in write queue, and still use 32 KB
> frags.
Another way to get pathological SKBs is to do lots of tiny sendpage()
calls over discontiguous areas of the file.
Powered by blists - more mailing lists