linux-kernel - Re: skbuff truesize incorrect.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1400860851.5367.210.camel@edumazet-glaptop2.roam.corp.google.com>
Date:	Fri, 23 May 2014 09:00:51 -0700
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Rick Jones <rick.jones2@...com>
Cc:	Bjørn Mork <bjorn@...k.no>,
	Jim Baxter <jim_baxter@...tor.com>,
	David Miller <davem@...emloft.net>, linux-usb@...r.kernel.org,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
	kamal@...onical.com, edumazet@...gle.com, mszeredi@...e.cz,
	fw@...len.de
Subject: Re: skbuff truesize incorrect.

On Fri, 2014-05-23 at 08:44 -0700, Rick Jones wrote:

> If you are measuring performance with the likes of netperf, you should 
> be able to get an idea of the performance effect from the change in 
> service demand (CPU consumed per unit of work) even if the maximum 
> throughput remains capped.

This wont be enough to truly get an idea of the gains this patch brings.

Add some random drops in the equation, and instead of dropping packets,
we gently add them in the out of order queue, gently coalesce them,
gently allow SACK enabled flows to recover thanks to fast retransmits,
with no latency effect (No expensive collapse/prunes)

# nstat -az|egrep "TCPRcvCoalesce|TCPOFOQueue|TCPRcvCollapsed|Prune"
TcpExtPruneCalled               118                0.0
TcpExtRcvPruned                 0                  0.0
TcpExtOfoPruned                 0                  0.0
TcpExtTCPRcvCollapsed           1009               0.0
TcpExtTCPRcvCoalesce            857806             0.0
TcpExtTCPOFOQueue               201246             0.0

Yep, we _might_ consume some cpu cycles in the perfect case where no
packet was dropped, but this is kind of an utopia.



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/