This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


RE: [Xen-devel] [PATCH] MAC Matching vif-route & tip

To: xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: RE: [Xen-devel] [PATCH] MAC Matching vif-route & tip
From: Andy Lee <andylee@xxxxxxxx>
Date: Wed, 07 Sep 2005 11:52:22 -0700
Delivery-date: Wed, 07 Sep 2005 18:50:45 +0000
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
At 10:44 AM 9/7/2005, Ian wrote:

> This patch will setup iptable rules for vif's that does MAC
> matching, as PHYSDEV matching only works for bridged devices.

Is that definitely the case? I don't see why physdev shouldn't work with
routed setups.

Yes. I first noticed the the byte counters of Physdev iptable rules being stuck at 0. They don't catch routed packets. Then I read:

CONFIG_IP_NF_MATCH_PHYSDEV: "Physdev packet matching matches against the physical bridge ports the IP packet arrived on or will leave by."

Using menuconfig, as I disable "802.1d Ethernet Bridging" from Networking, Physdev disappears from the netfilter menu too.

> On a separate but still networking issue, I noticed that my
> XenU's outbound traffic is sporadic (between 800KB/sec and
> 2MB/sec).  The problem is the default qdisc being inadequate.
>  So I fix it with:
> XenU# tc qdisc add dev eth0 root tbf rate 50mbit latency 20ms
> burst 50k

How much more buffering does this add?

As I understand it, the buffer must complement the network speed. Default i386 kernel uses 100 HZ, or 10ms time slices. Thus 50mbits/sec fills 50KB in 10ms.

Xen-devel mailing list

<Prev in Thread] Current Thread [Next in Thread>