Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 28 Mar 2006 07:40:18 -0300
From:      JoaoBR <joao@matik.com.br>
To:        freebsd-stable@freebsd.org
Cc:        Pyun YongHyeon <pyunyh@gmail.com>, Pieter de Goeje <pieter@degoeje.nl>
Subject:   Re: new sk driver [was: nve timeout (and down) regression?]
Message-ID:  <200603280740.18951.joao@matik.com.br>
In-Reply-To: <200603280221.28996.pieter@degoeje.nl>
References:  <20060324223317.2069564f@it.buh.tecnik93.com> <200603271848.19010.joao@matik.com.br> <200603280221.28996.pieter@degoeje.nl>

next in thread | previous in thread | raw e-mail | index | archive | help
On Monday 27 March 2006 21:21, Pieter de Goeje wrote:

> > IMO it would make sense to be replaced by Pyuns driver because the
> > original is not functional
> >
> > kind of strange getting a new release with a known not functional driver
> >
> > Jo=E3o
>
> This is not quite true; I am running 2 systems with the sk driver. One of
> them is a dual core athlon64, and it's network connection is indeed flawed
> unless I change debug.mpsafenet to 0. The other machine is a regular
> athlon64 and has no problems at all. Both systems are running
> 6.1-PRERELEASE amd64.
>
> With debug.mpsafenet=3D1, cvsup failes with "protocol error"s on the SMP =
box.
> Also, I run several Half-Life: Counter-Strike servers on it and clients
> occasionaly receive corrupted UDP packets which cause them to drop the
> connection. However as soon as I changed debug.mpsafenet to 0, all proble=
ms
> were gone.
>
> So, maybe you should try changing debug.mpsafenet.
> FYI: the SMP box has an Asus A8V (agp) mobo, and the other box an Asus K8V
> SE.


probably you do not have the traffic to make the box crash or less then 1/2=
GB=20
of RAM in use

in fact the problem does not happen on UP machines, only some times a devic=
e=20
timeout which only ocasionally cause rx/tx to stop

The problem is appearing on SMP machines

when you have less then 2Gb of RAM the problem ocurres once a day or so and=
=20
seems to depend on memory use and amount of traffic

soon the traffic reaches more than 1Mbit/s the crash is predictable and you=
=20
can wait to see

on 4GB of Ram machines and more traffic the crash is imediatly and worse wh=
en=20
the box crashed under load (4-6Mbit/s) and comes back then the high demand=
=20
strokes it and it crashes in minutes or imediatly soon the network is up

so probably mpsafenet may help by not processing concurrent packets but thi=
s=20
is a workaround not a solution (for me)

last time I checked mpsafenet=3D0 almost cut 1Mbit/s of traffic and the ove=
rall=20
performance/response was bad, higher HZ did not resolved anything and=20
disabling polling made it still worse (I have other NICs installed), the=20
machines are working as GW

until january the machines didn't crashed, only timeouts and rx/tx stops
I used Pyun's driver and the timeouts went away, thank's again!

so then I got confused by some if_sk talks on stable and thought the driver=
=20
was comitted and the boxes started crashing until I got it last week and=20
reused Pyun's driver again and my sk problems are gone again, the machines=
=20
are stable for 4/5 days now

Jo=E3o









A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200603280740.18951.joao>