1INITRD(4) Linux Programmer's Manual INITRD(4)
2
3
4
6 initrd - boot loader initialized RAM disk
7
9 The /dev/initrd is a read-only block device assigned major number 1 and
10 minor number 250. Typically /dev/initrd is owned by root.disk with
11 mode 0400 (read access by root only). If the Linux system does not
12 have /dev/initrd already created, it can be created with the following
13 commands:
14 mknod -m 400 /dev/initrd b 1 250
15 chown root:disk /dev/initrd
16
17 Also, support for both "RAM disk" and "Initial RAM disk" (e.g., CON‐
18 FIG_BLK_DEV_RAM=y and CONFIG_BLK_DEV_INITRD=y) must be compiled
19 directly into the Linux kernel to use /dev/initrd. When using
20 /dev/initrd, the RAM disk driver cannot be loaded as a module.
21
23 The special file /dev/initrd is a read-only block device. This device
24 is a RAM disk that is initialized (e.g., loaded) by the boot loader
25 before the kernel is started. The kernel then can use /dev/initrd's
26 contents for a two-phase system boot-up.
27
28 In the first boot-up phase, the kernel starts up and mounts an initial
29 root file-system from the contents of /dev/initrd (e.g., RAM disk ini‐
30 tialized by the boot loader). In the second phase, additional drivers
31 or other modules are loaded from the initial root device's contents.
32 After loading the additional modules, a new root file system (i.e., the
33 normal root file system) is mounted from a different device.
34
35 Boot-up operation
36 When booting up with initrd, the system boots as follows:
37
38 1. The boot loader loads the kernel program and /dev/initrd's contents
39 into memory.
40
41 2. On kernel startup, the kernel uncompresses and copies the contents
42 of the device /dev/initrd onto device /dev/ram0 and then frees the
43 memory used by /dev/initrd.
44
45 3. The kernel then read-write mounts the device /dev/ram0 as the ini‐
46 tial root file system.
47
48 4. If the indicated normal root file system is also the initial root
49 file-system (e.g., /dev/ram0) then the kernel skips to the last step
50 for the usual boot sequence.
51
52 5. If the executable file /linuxrc is present in the initial root file-
53 system, /linuxrc is executed with UID 0. (The file /linuxrc must
54 have executable permission. The file /linuxrc can be any valid exe‐
55 cutable, including a shell script.)
56
57 6. If /linuxrc is not executed or when /linuxrc terminates, the normal
58 root file system is mounted. (If /linuxrc exits with any file-sys‐
59 tems mounted on the initial root file-system, then the behavior of
60 the kernel is UNSPECIFIED. See the NOTES section for the current
61 kernel behavior.)
62
63 7. If the normal root file system has a directory /initrd, the device
64 /dev/ram0 is moved from / to /initrd. Otherwise if the directory
65 /initrd does not exist, the device /dev/ram0 is unmounted. (When
66 moved from / to /initrd, /dev/ram0 is not unmounted and therefore
67 processes can remain running from /dev/ram0. If directory /initrd
68 does not exist on the normal root file system and any processes
69 remain running from /dev/ram0 when /linuxrc exits, the behavior of
70 the kernel is UNSPECIFIED. See the NOTES section for the current
71 kernel behavior.)
72
73 8. The usual boot sequence (e.g., invocation of /sbin/init) is per‐
74 formed on the normal root file system.
75
76 Options
77 The following boot loader options, when used with initrd, affect the
78 kernel's boot-up operation:
79
80 initrd=filename
81 Specifies the file to load as the contents of /dev/initrd. For
82 LOADLIN this is a command-line option. For LILO you have to use
83 this command in the LILO configuration file /etc/lilo.config.
84 The filename specified with this option will typically be a
85 gzipped file-system image.
86
87 noinitrd
88 This boot option disables the two-phase boot-up operation. The
89 kernel performs the usual boot sequence as if /dev/initrd was
90 not initialized. With this option, any contents of /dev/initrd
91 loaded into memory by the boot loader contents are preserved.
92 This option permits the contents of /dev/initrd to be any data
93 and need not be limited to a file system image. However, device
94 /dev/initrd is read-only and can be read only one time after
95 system startup.
96
97 root=device-name
98 Specifies the device to be used as the normal root file system.
99 For LOADLIN this is a command-line option. For LILO this is a
100 boot time option or can be used as an option line in the LILO
101 configuration file /etc/lilo.config. The device specified by
102 the this option must be a mountable device having a suitable
103 root file-system.
104
105 Changing the normal root file system
106 By default, the kernel's settings (e.g., set in the kernel file with
107 rdev(8) or compiled into the kernel file), or the boot loader option
108 setting is used for the normal root file systems. For an NFS-mounted
109 normal root file system, one has to use the nfs_root_name and
110 nfs_root_addrs boot options to give the NFS settings. For more infor‐
111 mation on NFS-mounted root see the kernel documentation file Documenta‐
112 tion/filesystems/nfsroot.txt. For more information on setting the root
113 file system see also the LILO and LOADLIN documentation.
114
115 It is also possible for the /linuxrc executable to change the normal
116 root device. For /linuxrc to change the normal root device, /proc must
117 be mounted. After mounting /proc, /linuxrc changes the normal root
118 device by writing into the proc files /proc/sys/kernel/real-root-dev,
119 /proc/sys/kernel/nfs-root-name, and /proc/sys/kernel/nfs-root-addrs.
120 For a physical root device, the root device is changed by having /lin‐
121 uxrc write the new root file system device number into /proc/sys/ker‐
122 nel/real-root-dev. For an NFS root file system, the root device is
123 changed by having /linuxrc write the NFS setting into files
124 /proc/sys/kernel/nfs-root-name and /proc/sys/kernel/nfs-root-addrs and
125 then writing 0xff (e.g., the pseudo-NFS-device number) into file
126 /proc/sys/kernel/real-root-dev. For example, the following shell com‐
127 mand line would change the normal root device to /dev/hdb1:
128
129 echo 0x365 >/proc/sys/kernel/real-root-dev
130
131 For an NFS example, the following shell command lines would change the
132 normal root device to the NFS directory /var/nfsroot on a local net‐
133 worked NFS server with IP number 193.8.232.7 for a system with IP num‐
134 ber 193.8.232.2 and named "idefix":
135
136 echo /var/nfsroot >/proc/sys/kernel/nfs-root-name
137 echo 193.8.232.2:193.8.232.7::255.255.255.0:idefix \
138 >/proc/sys/kernel/nfs-root-addrs
139 echo 255 >/proc/sys/kernel/real-root-dev
140
141 Note: The use of /proc/sys/kernel/real-root-dev to change the root file
142 system is obsolete. See the Linux kernel source file Documenta‐
143 tion/initrd.txt as well as pivot_root(2) and pivot_root(8) for informa‐
144 tion on the modern method of changing the root file system.
145
146 Usage
147 The main motivation for implementing initrd was to allow for modular
148 kernel configuration at system installation.
149
150 A possible system installation scenario is as follows:
151
152 1. The loader program boots from floppy or other media with a minimal
153 kernel (e.g., support for /dev/ram, /dev/initrd, and the ext2 file-
154 system) and loads /dev/initrd with a gzipped version of the initial
155 file-system.
156
157 2. The executable /linuxrc determines what is needed to (1) mount the
158 normal root file-system (i.e., device type, device drivers, file
159 system) and (2) the distribution media (e.g., CD-ROM, network, tape,
160 ...). This can be done by asking the user, by auto-probing, or by
161 using a hybrid approach.
162
163 3. The executable /linuxrc loads the necessary modules from the initial
164 root file-system.
165
166 4. The executable /linuxrc creates and populates the root file system.
167 (At this stage the normal root file system does not have to be a
168 completed system yet.)
169
170 5. The executable /linuxrc sets /proc/sys/kernel/real-root-dev, unmount
171 /proc, the normal root file system and any other file systems it has
172 mounted, and then terminates.
173
174 6. The kernel then mounts the normal root file system.
175
176 7. Now that the file system is accessible and intact, the boot loader
177 can be installed.
178
179 8. The boot loader is configured to load into /dev/initrd a file system
180 with the set of modules that was used to bring up the system.
181 (e.g., Device /dev/ram0 can be modified, then unmounted, and
182 finally, the image is written from /dev/ram0 to a file.)
183
184 9. The system is now bootable and additional installation tasks can be
185 performed.
186
187 The key role of /dev/initrd in the above is to reuse the configuration
188 data during normal system operation without requiring initial kernel
189 selection, a large generic kernel or, recompiling the kernel.
190
191 A second scenario is for installations where Linux runs on systems with
192 different hardware configurations in a single administrative network.
193 In such cases, it may be desirable to use only a small set of kernels
194 (ideally only one) and to keep the system-specific part of configura‐
195 tion information as small as possible. In this case, create a common
196 file with all needed modules. Then, only the /linuxrc file or a file
197 executed by /linuxrc would be different.
198
199 A third scenario is more convenient recovery disks. Because informa‐
200 tion like the location of the root file-system partition is not needed
201 at boot time, the system loaded from /dev/initrd can use a dialog
202 and/or auto-detection followed by a possible sanity check.
203
204 Last but not least, Linux distributions on CD-ROM may use initrd for
205 easy installation from the CD-ROM. The distribution can use LOADLIN to
206 directly load /dev/initrd from CD-ROM without the need of any floppies.
207 The distribution could also use a LILO boot floppy and then bootstrap a
208 bigger RAM disk via /dev/initrd from the CD-ROM.
209
211 /dev/initrd
212 /dev/ram0
213 /linuxrc
214 /initrd
215
217 1. With the current kernel, any file systems that remain mounted when
218 /dev/ram0 is moved from / to /initrd continue to be accessible.
219 However, the /proc/mounts entries are not updated.
220
221 2. With the current kernel, if directory /initrd does not exist, then
222 /dev/ram0 will not be fully unmounted if /dev/ram0 is used by any
223 process or has any file-system mounted on it. If /dev/ram0 is not
224 fully unmounted, then /dev/ram0 will remain in memory.
225
226 3. Users of /dev/initrd should not depend on the behavior give in the
227 above notes. The behavior may change in future versions of the
228 Linux kernel.
229
231 chown(1), mknod(1), ram(4), freeramdisk(8), rdev(8)
232
233 Documentation/initrd.txt in the Linux kernel source tree, the LILO doc‐
234 umentation, the LOADLIN documentation, the SYSLINUX documentation
235
237 This page is part of release 3.53 of the Linux man-pages project. A
238 description of the project, and information about reporting bugs, can
239 be found at http://www.kernel.org/doc/man-pages/.
240
241
242
243Linux 2010-09-04 INITRD(4)