1 Mounting the root filesystem via NFS (nfsroot)
2 ===============================================
4 Written 1996 by Gero Kuhlmann <gero@gkminix.han.de>
5 Updated 1997 by Martin Mares <mj@atrey.karlin.mff.cuni.cz>
6 Updated 2006 by Nico Schottelius <nico-kernel-nfsroot@schottelius.org>
7 Updated 2006 by Horms <horms@verge.net.au>
11 In order to use a diskless system, such as an X-terminal or printer server
12 for example, it is necessary for the root filesystem to be present on a
13 non-disk device. This may be an initramfs (see Documentation/filesystems/
14 ramfs-rootfs-initramfs.txt), a ramdisk (see Documentation/initrd.txt) or a
15 filesystem mounted via NFS. The following text describes on how to use NFS
16 for the root filesystem. For the rest of this text 'client' means the
17 diskless system, and 'server' means the NFS server.
22 1.) Enabling nfsroot capabilities
23 -----------------------------
25 In order to use nfsroot, NFS client support needs to be selected as
26 built-in during configuration. Once this has been selected, the nfsroot
27 option will become available, which should also be selected.
29 In the networking options, kernel level autoconfiguration can be selected,
30 along with the types of autoconfiguration to support. Selecting all of
31 DHCP, BOOTP and RARP is safe.
36 2.) Kernel command line
39 When the kernel has been loaded by a boot loader (see below) it needs to be
40 told what root fs device to use. And in the case of nfsroot, where to find
41 both the server and the name of the directory on the server to mount as root.
42 This can be established using the following kernel command line parameters:
47 This is necessary to enable the pseudo-NFS-device. Note that it's not a
48 real device but just a synonym to tell the kernel to use NFS instead of
52 nfsroot=[<server-ip>:]<root-dir>[,<nfs-options>]
54 If the `nfsroot' parameter is NOT given on the command line,
55 the default "/tftpboot/%s" will be used.
57 <server-ip> Specifies the IP address of the NFS server.
58 The default address is determined by the `ip' parameter
59 (see below). This parameter allows the use of different
60 servers for IP autoconfiguration and NFS.
62 <root-dir> Name of the directory on the server to mount as root.
63 If there is a "%s" token in the string, it will be
64 replaced by the ASCII-representation of the client's
67 <nfs-options> Standard NFS options. All options are separated by commas.
68 The following defaults are used:
69 port = as given by server portmap daemon
78 flags = hard, nointr, noposix, cto, ac
81 ip=<client-ip>:<server-ip>:<gw-ip>:<netmask>:<hostname>:<device>:<autoconf>:
84 This parameter tells the kernel how to configure IP addresses of devices
85 and also how to set up the IP routing table. It was originally called
86 `nfsaddrs', but now the boot-time IP configuration works independently of
87 NFS, so it was renamed to `ip' and the old name remained as an alias for
88 compatibility reasons.
90 If this parameter is missing from the kernel command line, all fields are
91 assumed to be empty, and the defaults mentioned below apply. In general
92 this means that the kernel tries to configure everything using
95 The <autoconf> parameter can appear alone as the value to the `ip'
96 parameter (without all the ':' characters before). If the value is
97 "ip=off" or "ip=none", no autoconfiguration will take place, otherwise
98 autoconfiguration will take place. The most common way to use this
101 <client-ip> IP address of the client.
103 Default: Determined using autoconfiguration.
105 <server-ip> IP address of the NFS server. If RARP is used to determine
106 the client address and this parameter is NOT empty only
107 replies from the specified server are accepted.
109 Only required for NFS root. That is autoconfiguration
110 will not be triggered if it is missing and NFS root is not
113 Default: Determined using autoconfiguration.
114 The address of the autoconfiguration server is used.
116 <gw-ip> IP address of a gateway if the server is on a different subnet.
118 Default: Determined using autoconfiguration.
120 <netmask> Netmask for local network interface. If unspecified
121 the netmask is derived from the client IP address assuming
124 Default: Determined using autoconfiguration.
126 <hostname> Name of the client. May be supplied by autoconfiguration,
127 but its absence will not trigger autoconfiguration.
128 If specified and DHCP is used, the user provided hostname will
129 be carried in the DHCP request to hopefully update DNS record.
131 Default: Client IP address is used in ASCII notation.
133 <device> Name of network device to use.
135 Default: If the host only has one device, it is used.
136 Otherwise the device is determined using
137 autoconfiguration. This is done by sending
138 autoconfiguration requests out of all devices,
139 and using the device that received the first reply.
141 <autoconf> Method to use for autoconfiguration. In the case of options
142 which specify multiple autoconfiguration protocols,
143 requests are sent using all protocols, and the first one
146 Only autoconfiguration protocols that have been compiled
147 into the kernel will be used, regardless of the value of
150 off or none: don't use autoconfiguration
151 (do static IP assignment instead)
152 on or any: use any protocol available in the kernel
157 both: use both BOOTP and RARP but not DHCP
158 (old option kept for backwards compatibility)
160 if dhcp is used, the client identifier can be used by following
161 format "ip=dhcp,client-id-type,client-id-value"
165 <dns0-ip> IP address of first nameserver.
166 Value gets exported by /proc/net/pnp which is often linked
167 on embedded systems by /etc/resolv.conf.
169 <dns1-ip> IP address of secound nameserver.
175 This parameter enables debugging messages to appear in the kernel
176 log at boot time so that administrators can verify that the correct
177 NFS mount options, server address, and root path are passed to the
181 rdinit=<executable file>
183 To specify which file contains the program that starts system
184 initialization, administrators can use this command line parameter.
185 The default value of this parameter is "/init". If the specified
186 file exists and the kernel can execute it, root filesystem related
187 kernel command line parameters, including `nfsroot=', are ignored.
189 A description of the process of mounting the root file system can be
192 Documentation/early-userspace/README
200 To get the kernel into memory different approaches can be used.
201 They depend on various facilities being available:
204 3.1) Booting from a floppy using syslinux
206 When building kernels, an easy way to create a boot floppy that uses
207 syslinux is to use the zdisk or bzdisk make targets which use zimage
208 and bzimage images respectively. Both targets accept the
209 FDARGS parameter which can be used to set the kernel command line.
212 make bzdisk FDARGS="root=/dev/nfs"
214 Note that the user running this command will need to have
215 access to the floppy drive device, /dev/fd0
217 For more information on syslinux, including how to create bootdisks
218 for prebuilt kernels, see http://syslinux.zytor.com/
220 N.B: Previously it was possible to write a kernel directly to
221 a floppy using dd, configure the boot device using rdev, and
222 boot using the resulting floppy. Linux no longer supports this
225 3.2) Booting from a cdrom using isolinux
227 When building kernels, an easy way to create a bootable cdrom that
228 uses isolinux is to use the isoimage target which uses a bzimage
229 image. Like zdisk and bzdisk, this target accepts the FDARGS
230 parameter which can be used to set the kernel command line.
233 make isoimage FDARGS="root=/dev/nfs"
235 The resulting iso image will be arch/<ARCH>/boot/image.iso
236 This can be written to a cdrom using a variety of tools including
240 cdrecord dev=ATAPI:1,0,0 arch/x86/boot/image.iso
242 For more information on isolinux, including how to create bootdisks
243 for prebuilt kernels, see http://syslinux.zytor.com/
246 When using LILO all the necessary command line parameters may be
247 specified using the 'append=' directive in the LILO configuration
250 However, to use the 'root=' directive you also need to create
251 a dummy root device, which may be removed after LILO is run.
253 mknod /dev/boot255 c 0 255
255 For information on configuring LILO, please refer to its documentation.
258 When using GRUB, kernel parameter are simply appended after the kernel
259 specification: kernel <kernel> <parameters>
262 loadlin may be used to boot Linux from a DOS command prompt without
263 requiring a local hard disk to mount as root. This has not been
264 thoroughly tested by the authors of this document, but in general
265 it should be possible configure the kernel command line similarly
266 to the configuration of LILO.
268 Please refer to the loadlin documentation for further information.
270 3.5) Using a boot ROM
271 This is probably the most elegant way of booting a diskless client.
272 With a boot ROM the kernel is loaded using the TFTP protocol. The
273 authors of this document are not aware of any no commercial boot
274 ROMs that support booting Linux over the network. However, there
275 are two free implementations of a boot ROM, netboot-nfs and
276 etherboot, both of which are available on sunsite.unc.edu, and both
277 of which contain everything you need to boot a diskless Linux client.
280 Pxelinux may be used to boot linux using the PXE boot loader
281 which is present on many modern network cards.
283 When using pxelinux, the kernel image is specified using
284 "kernel <relative-path-below /tftpboot>". The nfsroot parameters
285 are passed to the kernel by adding them to the "append" line.
286 It is common to use serial console in conjunction with pxeliunx,
287 see Documentation/serial-console.txt for more information.
289 For more information on isolinux, including how to create bootdisks
290 for prebuilt kernels, see http://syslinux.zytor.com/
298 The nfsroot code in the kernel and the RARP support have been written
299 by Gero Kuhlmann <gero@gkminix.han.de>.
301 The rest of the IP layer autoconfiguration code has been written
302 by Martin Mares <mj@atrey.karlin.mff.cuni.cz>.
304 In order to write the initial version of nfsroot I would like to thank
305 Jens-Uwe Mager <jum@anubis.han.de> for his help.