netdev - Re: e1000e 2.6.30.1 massive packet loss

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-next>] [day] [month] [year] [list]

Message-ID: <4A5EDEF7.7090303@cosmosbay.com>
Date:	Thu, 16 Jul 2009 10:04:07 +0200
From:	Eric Dumazet <dada1@...mosbay.com>
To:	Caleb Cushing <xenoterracide@...il.com>
CC:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Linux Netdev List <netdev@...r.kernel.org>
Subject: Re: e1000e 2.6.30.1 massive packet loss

Caleb Cushing a écrit :
> So from 2.6.30 to .1 I noticed that via an icmp test (mtr) 20-50%
> packet loss web browsing became impractical (downgrading fixed the
> issue). I'm wondering if the change to the driver (w/ msi) is the
> cause? also wondering if this is known and will be fixed in .2 I can
> provide more information if necessary.
> 

Added netdev to get more attention to this.

What do you call 20-50% packet loss ?
(Losses to icmp flood itself, or tcp losses ?
Is the machine answering to icmp flood, or is it the origin ?

Are you saying packet loss was not occurring with 2.6.30 ?

Could you describe your methodology ?

You could give us results of "cat /proc/interrupts" from both
versions, because it sounds as an interrupt affinity problem.

# cat /proc/interrupts
# tc -s -d qdisc
# ifconfig -a
--
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