mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Stephen Hemminger <shemminger@osdl.org>
To: linux-kernel@vger.kernel.org
Subject: Re: 2.6.12.3 network slowdown?
Date: Wed, 27 Jul 2005 10:55:06 -0700	[thread overview]
Message-ID: <20050727105506.78b82aaa@dxpl.pdx.osdl.net> (raw)
In-Reply-To: <42E7497B.5040202@highlandsun.com>

On Wed, 27 Jul 2005 01:44:43 -0700
Howard Chu <hyc@highlandsun.com> wrote:

> I just recently compiled the 2.6.12.3 kernel for my x86_64 machine
> (Asus A8V motherboard); was previously running a SuSE-compiled 2.6.8
> kernel (SuSE 9.2 distro). I'm now seeing extremely slow throughput on
> the onboard Yukon (Marvell) ethernet interface, but only in certain
> conditions. Going back to the 2.6.8 kernel shows no slowdown.
> 
> I have a few machines connected to a Linksys 5 port 10/100 hub. There
> is also a Linksys WRT54G wireless router on this hub. When doing a
> raw ftp transfer of a large (600MB) file (using a Windows XP client
> to do a GET) with both machines plugged into the hub, I see transfer
> rates only as high as 1MB/sec initially, which quickly degrades to
> ~200KB/sec over a span of 20-30 seconds. Going the opposite
> direction, a PUT runs at a steady 7.5MB/sec.
> 
> If I unplug the Windows client and just connect via the wireless 
> network, I see GETs run a steady 2.8MB/sec and PUTs run a steady 
> 3.2MB/sec, which is about the same as I saw with the 2.6.8 kernel, so 
> that all appears normal. (But it's still odd, that adding one hop
> like this avoids the throughput degradation.)
> 
> I don't see any collisions or any other errors on the ifconfig 
> statistics, just a very slow transmit rate. Does anyone recognize
> this symptom? Any suggestions on tunables to tweak, etc.? (Please cc:
> me directly when replying, thanks.)
> 

There a couple of possibilities, one is driver differences. SUSE ships
the SysKonnect vendor driver vs the older version in 2.6.12, and the
TCP congestion and TSO code has changed since 2.6.8.

What is the output of ethtool (ethtool -i and ethtool -k)?
Perhaps one driver has checksum offload on and the other off?
Could you get a tcpdump capture of a slow transfer?

  parent reply	other threads:[~2005-07-27 17:55 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1122452018.772579.63900@g49g2000cwa.googlegroups.com>
     [not found] ` <20050727082020.C73AC5F7CA@attila.bofh.it>
2005-07-27  8:44   ` Howard Chu
2005-07-27 10:30     ` Mihai Rusu
2005-07-27 13:00       ` Bernd Schubert
2005-07-27 17:55     ` Stephen Hemminger [this message]
2005-07-27 19:42       ` Howard Chu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20050727105506.78b82aaa@dxpl.pdx.osdl.net \
    --to=shemminger@osdl.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome