On Thu, 11 Aug 2005, Oliver Gerlich wrote:
Couldn't we avoid these incompatibilities if we would route packets only on
the Ethernet level? If the Qemu networking setup on the host involves IP
addresses or such things, we're already on the wrong OSI layer I think...
From what I can tell there is no way userspace can send Ethernet frames to
the host as if coming from the eth0 interface, short of writing a kernel
module to make this magics happen.
I'm not very experienced in networking, but IMHO the network should be set up
like this:
(eth0 on Host OS)
/
( LAN ) <----> (real eth0) <-> VDE
\
(NIC on Guest OS)
This is what you do with bridgeing.
(br0 on Host OS)
/
( LAN ) <---> eth0 <-> bridge
\
(tap0 to VDE/qemu)
Hmm.. actually the bridge setup can be simplified somewhat on Linux hosts
by using an early boot script run before the eth0 is initialized, renaming
eth0 to something else and creating a bridge named eth0, including the
renamed eth0 interface..
modprobe eth0
ip link set eth0 name eth0real
brctl addbr eth0
brctl addif eth0 eth0real
ip link set eth0real up
So this is a summary of what I think we should try to achieve (sorted by
priority):
-provide Ethernet-layer access from the Guest to the LAN (so that it's
transparent for IP traffic)
Done.
-allow using a LAN-wide DHCP server for the Guest
Done by the above.
-the host should be able to set up his LAN interface over the LAN-wide DHCP
server.
Done.
-the host networking still works as before (at least on the layers above
Ethernet)
Done.
-the whole thing is easy to set up
Done.
-the whole thing can be started by any user at any time (so it's not started
at boot time, but whenever a user starts Qemu).
Requires root rights.
The restrictions to the above is that the guest can not reach the host in
such setup.
Regards
Henrik
_______________________________________________
Qemu-devel mailing list
Qemu-devel@nongnu.org
http://lists.nongnu.org/mailman/listinfo/qemu-devel
