chroot(2)

1chroot(2)                     System Calls Manual                    chroot(2)
2
3
4

NAME

6       chroot - change root directory
7

LIBRARY

9       Standard C library (libc, -lc)
10

SYNOPSIS

12       #include <unistd.h>
13
14       int chroot(const char *path);
15
16   Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
17
18       chroot():
19           Since glibc 2.2.2:
20               _XOPEN_SOURCE && ! (_POSIX_C_SOURCE >= 200112L)
21                   || /* Since glibc 2.20: */ _DEFAULT_SOURCE
22                   || /* glibc <= 2.19: */ _BSD_SOURCE
23           Before glibc 2.2.2:
24               none
25

DESCRIPTION

27       chroot()  changes  the  root  directory  of the calling process to that
28       specified in path.  This directory will be used for pathnames beginning
29       with /.  The root directory is inherited by all children of the calling
30       process.
31
32       Only a privileged process (Linux: one with the CAP_SYS_CHROOT  capabil‐
33       ity in its user namespace) may call chroot().
34
35       This  call changes an ingredient in the pathname resolution process and
36       does nothing else.  In particular, it is not intended to  be  used  for
37       any kind of security purpose, neither to fully sandbox a process nor to
38       restrict filesystem system calls.  In the past, chroot() has been  used
39       by  daemons  to  restrict themselves prior to passing paths supplied by
40       untrusted users to system calls such as open(2).  However, if a  folder
41       is  moved  out of the chroot directory, an attacker can exploit that to
42       get out of the chroot directory as well.  The easiest way to do that is
43       to  chdir(2) to the to-be-moved directory, wait for it to be moved out,
44       then open a path like ../../../etc/passwd.
45
46       A slightly trickier variation also works under  some  circumstances  if
47       chdir(2)  is not permitted.  If a daemon allows a "chroot directory" to
48       be specified, that usually means that if you  want  to  prevent  remote
49       users  from  accessing files outside the chroot directory, you must en‐
50       sure that folders are never moved out of it.
51
52       This call does not change the current working directory, so that  after
53       the call '.' can be outside the tree rooted at '/'.  In particular, the
54       superuser can escape from a "chroot jail" by doing:
55
56           mkdir foo; chroot foo; cd ..
57
58       This call does not close open file descriptors, and such file  descrip‐
59       tors may allow access to files outside the chroot tree.
60

RETURN VALUE

62       On  success,  zero is returned.  On error, -1 is returned, and errno is
63       set to indicate the error.
64

ERRORS

66       Depending on the filesystem, other errors can be  returned.   The  more
67       general errors are listed below:
68
69       EACCES Search  permission  is denied on a component of the path prefix.
70              (See also path_resolution(7).)
71
72       EFAULT path points outside your accessible address space.
73
74       EIO    An I/O error occurred.
75
76       ELOOP  Too many symbolic links were encountered in resolving path.
77
78       ENAMETOOLONG
79              path is too long.
80
81       ENOENT The file does not exist.
82
83       ENOMEM Insufficient kernel memory was available.
84
85       ENOTDIR
86              A component of path is not a directory.
87
88       EPERM  The caller has insufficient privilege.
89

STANDARDS

91       None.
92

HISTORY

94       SVr4, 4.4BSD, SUSv2 (marked LEGACY).  This  function  is  not  part  of
95       POSIX.1-2001.
96

NOTES

98       A  child  process created via fork(2) inherits its parent's root direc‐
99       tory.  The root directory is left unchanged by execve(2).
100
101       The magic symbolic link, /proc/pid/root, can  be  used  to  discover  a
102       process's root directory; see proc(5) for details.
103
104       FreeBSD has a stronger jail() system call.
105