netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bart Alewijnse <scarfboy@gmail.com>
To: netdev@oss.sgi.com
Cc: linux-kernel@vger.kernel.org
Subject: Re: gigabit trouble
Date: Fri, 30 Jul 2004 00:41:03 +0200	[thread overview]
Message-ID: <b71082d80407291541f9d6f93@mail.gmail.com> (raw)
In-Reply-To: <20040729210401.A32456@electric-eye.fr.zoreil.com>

I was going to do an exhaustive test, but because my computer stopped
running completely, I'll reply to the one or two bits I can now.

On Thu, 29 Jul 2004 21:04:01 +0200, Francois Romieu
<romieu@fr.zoreil.com> wrote:
> Bart Alewijnse <scarfboy@gmail.com> :
> [...]
> > I run gentoo on both, which until yesterday was 2.6.7-ck5 (on both),
> > and currently run 2.6.7-mm6 (again, both), as I saw the suggestion
> > somewhere it had better support for the card - something about a new
> > net card inferface that's nicer to interrupts.
> 
> NAPI support for r8169 is available in recent -mm kernel and there is
> a small (though noticeable) optimization wrt to interrupt disabling.
Well, I noticed a max of 6500 interrupts/s on eth1 on both computers, with or
without napi - and that 6500 is also the figure of packets per second.
So I'm slightly dubious. (notice 6500 *1500 bytes is about 10MB/s)

> [...]
> > So, question one - how do I see the link speed under linux, and how,
> > if at all, do I control it?
> 
> ethtool
Thanks. That wasn't the problem - the line speed's a gbit.

> [...]
> > Disturbingly, in such a linux-to-linux speed test, my new computer
> > froze.As in, in text mode, have the screen freeze and apparently be
> > half written full of nonsense.
> 
> These messages would be welcome (pen/paper/serial line/image/log file
> or whatever).

No messages, no oops, no log messages that I noticed. 
It was video memory corruption in text mode.

As to the rest, I'll do it when I revive my computer. Right now, I'm
thinking the power supply may be dodgy. I'll see if I can get
a minimum to run off my even older 235Watt. But as a note,
with two cards, two cdroms and a hard drive less, it was still
making the noise. I think it still is now, but since no OS actually
boots completely right now, I can't say for sure.
I'll do a memtest, that makes sense.

--Bart

       reply	other threads:[~2004-07-29 22:41 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <b71082d8040729094537e59a11@mail.gmail.com>
     [not found] ` <20040729210401.A32456@electric-eye.fr.zoreil.com>
2004-07-29 22:41   ` Bart Alewijnse [this message]
2004-07-30 15:15     ` gigabit trouble Bart Alewijnse
2004-07-30 18:54       ` Francois Romieu
2004-07-30 21:03         ` Bart Alewijnse
2004-07-30 21:41           ` Francois Romieu
2004-07-31 19:51             ` Bart Alewijnse
2004-07-31 21:18               ` Francois Romieu
2004-08-01 19:03                 ` Bart Alewijnse
2004-08-03  2:47                   ` Bart Alewijnse
2004-08-03  7:48                     ` Francois Romieu
2004-08-03 12:32                       ` Bart Alewijnse

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=b71082d80407291541f9d6f93@mail.gmail.com \
    --to=scarfboy@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@oss.sgi.com \
    /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;
as well as URLs for NNTP newsgroup(s).