qemu with tap networking on FreeBSD 6.1

mal content artifact.one at googlemail.com
Fri Jul 28 16:42:05 UTC 2006


Hi.

I've been having serious trouble with this.

I posted a message in the qemu forum, but with only 30-something
threads in the BSD section, I don't think it's really a good place to
get a decent answer about BSD.

A copy of the post follows:

Followed instructions from various places and ended up with
the following procedure:

# kldload bridge.ko
# sysctl net.link.ether.bridge_cfg=fxp0,tap0
# sysctl net.link.ether.bridge.enable=1

I created 'if-up' for qemu:

#!/bin/sh
ifconfig ${1} 0.0.0.0

I have a working OpenBSD image, 3.9. I started it up,
set an IP address and default route, etc. Everything appears
to be fine there.

I reboot the qemu image and just before the login prompt,
qemu goes insane. For some reason, it blasts UDP packets:

(on the host)
# netstat -an
<snip>
udp4       0      0  *.62756                  *.*
udp4       0      0  *.62324                  *.*
udp4       0      0  *.62127                  *.*
udp4       0      0  *.62741                  *.*
udp4       0      0  *.59182                  *.*
udp4       0      0  *.63792                  *.*
</snip>

(this goes on for several screens).

During this time, my ethernet switch is understandably going
mad and the qemu process has slowed to an unusable crawl.
The host doesn't seem particularly affected.

My prime suspicion is that those UDP packets are DNS
queries. It also sends ICMP packets, that I captured with
ethereal (most of them are along the lines of 'host not
reachable).

Did I do something to cause offense to some electronic deity
somewhere?

Anybody here seen anything like this? Qemu without networking
is next to useless.

MC


More information about the freebsd-questions mailing list