2 (___()'`; Rusty's Remarkably Unreliable Guide to Lguest
3 /, /` - or, A Young Coder's Illustrated Hypervisor
4 \\"--\\ http://lguest.ozlabs.org
6 Lguest is designed to be a minimal 32-bit x86 hypervisor for the Linux kernel,
7 for Linux developers and users to experiment with virtualization with the
8 minimum of complexity. Nonetheless, it should have sufficient features to
9 make it useful for specific tasks, and, of course, you are encouraged to fork
10 and enhance it (see drivers/lguest/README).
14 - Kernel module which runs in a normal kernel.
15 - Simple I/O model for communication.
16 - Simple program to create new guests.
17 - Logo contains cute puppies: http://lguest.ozlabs.org
22 - No ABI: being tied to a specific kernel anyway, you can change anything.
23 - Many opportunities for improvement or feature implementation.
27 - The easiest way to run lguest is to use same kernel as guest and host.
28 You can configure them differently, but usually it's easiest not to.
30 You will need to configure your kernel with the following options:
33 "Prompt for development and/or incomplete code/drivers" = Y
34 (CONFIG_EXPERIMENTAL=y)
36 "Processor type and features":
37 "Paravirtualized guest support" = Y
38 "Lguest guest support" = Y
39 "High Memory Support" = off/4GB
40 "Alignment value to which kernel should be aligned" = 0x100000
41 (CONFIG_PARAVIRT=y, CONFIG_LGUEST_GUEST=y, CONFIG_HIGHMEM64G=n and
42 CONFIG_PHYSICAL_ALIGN=0x100000)
46 "Virtio block driver (EXPERIMENTAL)" = M/Y
47 "Network device support"
48 "Universal TUN/TAP device driver support" = M/Y
49 "Virtio network driver (EXPERIMENTAL)" = M/Y
50 (CONFIG_VIRTIO_BLK=m, CONFIG_VIRTIO_NET=m and CONFIG_TUN=m)
53 "Linux hypervisor example code" = M/Y
56 - A tool called "lguest" is available in this directory: type "make"
57 to build it. If you didn't build your kernel in-tree, use "make
60 - Create or find a root disk image. There are several useful ones
61 around, such as the xm-test tiny root image at
62 http://xm-test.xensource.com/ramdisks/initrd-1.1-i386.img
64 For more serious work, I usually use a distribution ISO image and
65 install it under qemu, then make multiple copies:
67 dd if=/dev/zero of=rootfile bs=1M count=2048
68 qemu -cdrom image.iso -hda rootfile -net user -net nic -boot d
70 Make sure that you install a getty on /dev/hvc0 if you want to log in on the
73 - "modprobe lg" if you built it as a module.
75 - Run an lguest as root:
77 Documentation/lguest/lguest 64 vmlinux --tunnet=192.168.19.1 --block=rootfile root=/dev/vda
80 64: the amount of memory to use, in MB.
82 vmlinux: the kernel image found in the top of your build directory. You
83 can also use a standard bzImage.
85 --tunnet=192.168.19.1: configures a "tap" device for networking with this
88 --block=rootfile: a file or block device which becomes /dev/vda
91 root=/dev/vda: this (and anything else on the command line) are
92 kernel boot parameters.
94 - Configuring networking. I usually have the host masquerade, using
95 "iptables -t nat -A POSTROUTING -o eth0 -j MASQUERADE" and "echo 1 >
96 /proc/sys/net/ipv4/ip_forward". In this example, I would configure
97 eth0 inside the guest at 192.168.19.2.
99 Another method is to bridge the tap device to an external interface
100 using --tunnet=bridge:<bridgename>, and perhaps run dhcp on the guest
101 to obtain an IP address. The bridge needs to be configured first:
102 this option simply adds the tap interface to it.
104 A simple example on my system:
106 ifconfig eth0 0.0.0.0
112 Then use --tunnet=bridge:lg0 when launching the guest.
116 http://www.linuxfoundation.org/collaborate/workgroups/networking/bridge
118 for general information on how to get bridging to work.
120 - Random number generation. Using the --rng option will provide a
121 /dev/hwrng in the guest that will read from the host's /dev/random.
122 Use this option in conjunction with rng-tools (see ../hw_random.txt)
123 to provide entropy to the guest kernel's /dev/random.
125 There is a helpful mailing list at http://ozlabs.org/mailman/listinfo/lguest
128 Rusty Russell rusty@rustcorp.com.au.