netdev - Re: Re: Re: limited network bandwidth with 3.2.x kernels

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: <1329845484.18384.20.camel@edumazet-laptop>
Date:	Tue, 21 Feb 2012 18:31:24 +0100
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	alekcejk@...glemail.com
Cc:	netdev@...r.kernel.org
Subject: Re: Re: Re: limited network bandwidth with 3.2.x kernels

Le mardi 21 février 2012 à 19:19 +0200, alekcejk@...glemail.com a
écrit :
> Download speed from ftp3.de.freebsd.org increases from 5 MB/s to 10 MB/s
> after I set "sysctl net.ipv4.tcp_rmem="4096 87380 4127616""" in 3.2.7 kernel.
> 
> --2012-02-21 18:56:36--  ftp://ftp3.de.freebsd.org/pub/FreeBSD/ISO-IMAGES-i386/8.2/FreeBSD-8.2-RELEASE-i386-dvd1.iso.xz
> 100%[======================================>] 2 066 424 512 10,3M/s  in 3m 17s
> 2012-02-21 18:59:56 (9,99 MB/s) - «/dev/null» saved [2066424512]
> 

Yep, this confirms the thing. You were at the limit of optimal receiver
window. Find a more remote server (80ms rtt for example) and you'll see
that your default tcp_rmem[2] was too small, even on 3.1.10 kernel.

> But maximum speed in 3.1.10 kernel (without this sysctl settings) still
> a bit more than in 3.2.7 - 11 MB/s instead of 10.3 MB/s.
> 
> --2012-02-21 18:56:36--  ftp://ftp3.de.freebsd.org/pub/FreeBSD/ISO-IMAGES-i386/8.2/FreeBSD-8.2-RELEASE-i386-dvd1.iso.xz
> 100%[======================================>] 2 066 424 512 11,0M/s   in 3m 30s   
> 2012-02-21 19:13:55 (9,39 MB/s) - «/dev/null» saved [2066424512]

Thats one measure, you'll have to make a lot of different measures to
make this significant, preferably in a lab to rule out different
external network conditions.




--
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