allows a permitted user to execute a
as the superuser or another user, as specified by the security policy. The invoking user's real
user-ID is used to determine the user name with which to query the security policy.
supports a plugin architecture for security policies and input/output logging. Third parties can develop and distribute their own policy and I/O logging plugins to work seamlessly with the
front end. The default security policy is
which is configured via the file
or via LDAP. See the
section for more information.
The security policy determines what privileges, if any, a user has to run
The policy may require that users authenticate themselves with a password or another authentication mechanism. If authentication is required,
will exit if the user's password is not entered within a configurable time limit. This limit is policy-specific; the default password prompt timeout for the
security policy is
minutes.
Security policies may support credential caching to allow the user to run
again for a period of time without requiring authentication. By default, the
policy caches credentials on a per-terminal basis for
minutes. See the
and
options in
for more information. By running
with the
option, a user can update the cached credentials without running a
On systems where
is the primary method of gaining superuser privileges, it is imperative to avoid syntax errors in the security policy configuration files. For the default security policy,
changes to the configuration files should be made using the
utility which will ensure that no syntax errors are introduced.
When invoked as
the
option (described below), is implied.
Security policies may log successful and failed attempts to use
If an I/O plugin is configured, the running command's input and output may be logged as well.
The options are as follows:
Normally, if
requires a password, it will read it from the user's terminal. If the
option is specified, a (possibly graphical) helper program is executed to read the user's password and output the password to the standard output. If the
environment variable is set, it specifies the path to the helper program. Otherwise, if
contains a line specifying the askpass program, that value will be used. For example:
# Path to askpass helper program Path askpass /usr/X11R6/bin/ssh-askpass
If no askpass program is available,
will exit with an error.
Ring the bell as part of the password promp when a terminal is present. This option has no effect if an askpass program is used.
Run the given command in the background. Note that it is not possible to use shell job control to manipulate background processes started by
Most interactive commands will fail to work properly in background mode.
Close all file descriptors greater than or equal to
before executing a command. Values less than three are not permitted. By default,
will close all open file descriptors other than standard input, standard output and standard error when executing a command. The security policy may restrict the user's ability to use this option. The
policy only permits use of the
option when the administrator has enabled the
option.
Run the command in the specified
instead of the current working directory. The security policy may return an error if the user does not have permission to specify the working directory.
Indicates to the security policy that the user wishes to preserve their existing environment variables. The security policy may return an error if the user does not have permission to preserve the environment.
Indicates to the security policy that the user wishes to add the comma-separated list of environment variables to those preserved from the user's environment. The security policy may return an error if the user does not have permission to preserve the environment. This option may be specified multiple times.
Edit one or more files instead of running a command. In lieu of a path name, the string "sudoedit" is used when consulting the security policy. If the user is authorized by the policy, the following steps are taken:
Temporary copies are made of the files to be edited with the owner set to the invoking user.
The editor specified by the policy is run to edit the temporary files. The
policy uses the
and
environment variables (in that order). If none of
or
are set, the first program listed in the
option is used.
If they have been modified, the temporary files are copied back to their original location and the temporary versions are removed.
To help prevent the editing of unauthorized files, the following restrictions are enforced unless explicitly allowed by the security policy:
Symbolic links may not be edited (version 1.8.15 and higher).
Symbolic links along the path to be edited are not followed when the parent directory is writable by the invoking user unless that user is root (version 1.8.16 and higher).
Files located in a directory that is writable by the invoking user may not be edited unless that user is root (version 1.8.16 and higher).
Users are never allowed to edit device special files.
If the specified file does not exist, it will be created. Note that unlike most commands run by
the editor is run with the invoking user's environment unmodified. If the temporary file becomes empty after editing, the user will be prompted before it is installed. If, for some reason,
is unable to update a file with its edited version, the user will receive a warning and the edited copy will remain in a temporary file.
Run the command with the primary group set to
instead of the primary group specified by the target user's password database entry. The
may be either a group name or a numeric group-ID
prefixed with the
character (e.g.,
for GID 0). When running a command as a GID, many shells require that the
be escaped with a backslash
If no
option is specified, the command will be run as the invoking user. In either case, the primary group will be set to
The
policy permits any of the target user's groups to be specified via the
option as long as the
option is not in use.
Request that the security policy set the
environment variable to the home directory specified by the target user's password database entry. Depending on the policy, this may be the default behavior.
Display a short help message to the standard output and exit.
Run the command on the specified
if the security policy plugin supports remote commands. Note that the
plugin does not currently support running remote commands. This may also be used in conjunction with the
option to list a user's privileges for the remote host.
Run the shell specified by the target user's password database entry as a login shell. This means that login-specific resource files such as
or
will be read by the shell. If a command is specified, it is passed to the shell for execution via the shell's
option. If no command is specified, an interactive shell is executed.
attempts to change to that user's home directory before running the shell. The command is run with an environment similar to the one a user would receive at log in. Note that most shells behave differently when a command is specified as compared to an interactive session; consult the shell's manual for details. The
section in the
manual documents how the
option affects the environment in which a command is run when the
policy is in use.
Similar to the
option, except that it removes the user's cached credentials entirely and may not be used in conjunction with a command or other option. This option does not require a password. Not all security policies support credential caching.
When used without a command, invalidates the user's cached credentials. In other words, the next time
is run a password will be required. This option does not require a password and was added to allow a user to revoke
permissions from a
file.
When used in conjunction with a command or an option that may require a password, this option will cause
to ignore the user's cached credentials. As a result,
will prompt for a password (if one is required by the security policy) and will not update the user's cached credentials.
Not all security policies support credential caching.
If no
is specified, list the allowed (and forbidden) commands for the invoking user (or the user specified by the
option) on the current host. A longer list format is used if this option is specified multiple times and the security policy supports a verbose output format.
If a
is specified and is permitted by the security policy, the fully-qualified path to the command is displayed along with any command line arguments. If a
is specified but not allowed by the policy,
will exit with a status value of 1.
Avoid prompting the user for input of any kind. If a password is required for the command to run,
will display an error message and exit.
Preserve the invoking user's group vector unaltered. By default, the
policy will initialize the group vector to the list of groups the target user is a member of. The real and effective group-IDs, however, are still set to match the target user.
Use a custom password prompt with optional escape sequences. The following percent
escape sequences are supported by the
policy:
expanded to the host name including the domain name (on if the machine's host name is fully qualified or the
option is set in
expanded to the local host name without the domain name
expanded to the name of the user whose password is being requested (respects the
and
flags in
expanded to the login name of the user the command will be run as (defaults to root unless the
option is also specified)
expanded to the invoking user's login name
two consecutive
characters are collapsed into a single
character
The custom prompt will override the default prompt specified by either the security policy or the
environment variable. On systems that use PAM, the custom prompt will also override the prompt specified by a PAM module unless the
flag is disabled in
Change to the specified root
(see
before running the command. The security policy may return an error if the user does not have permission to specify the root directory.
Write the prompt to the standard error and read the password from the standard input instead of using the terminal device.
Run the shell specified by the
environment variable if it is set or the shell specified by the invoking user's password database entry. If a command is specified, it is passed to the shell for execution via the shell's
option. If no command is specified, an interactive shell is executed. Note that most shells behave differently when a command is specified as compared to an interactive session; consult the shell's manual for details.
Used in conjunction with the
option to list the privileges for
instead of for the invoking user. The security policy may restrict listing other users' privileges. The
policy only allows root or a user with the
privilege on the current host to use this option.
Used to set a timeout for the command. If the timeout expires before the command has exited, the command will be terminated. The security policy may restrict the ability to set command timeouts. The
policy requires that user-specified timeouts be explicitly enabled.
Run the command as a user other than the default target user (usually
The
may be either a user name or a numeric user-ID
prefixed with the
character (e.g.,
for UID 0). When running commands as a UID, many shells require that the
be escaped with a backslash
Some security policies may restrict UIDs to those listed in the password database. The
policy allows UIDs that are not in the password database as long as the
option is not set. Other security policies may not support this.
Print the
version string as well as the version string of the security policy plugin and any I/O plugins. If the invoking user is already root the
option will display the arguments passed to configure when
was built and plugins may display more verbose information such as default options.
Update the user's cached credentials, authenticating the user if necessary. For the
plugin, this extends the
timeout for another
minutes by default, but does not run a command. Not all security policies support cached credentials.
The
option indicates that
should stop processing command line arguments.
Options that take a value may only be specified once unless otherwise indicated in the description. This is to help guard against problems caused by poorly written scripts that invoke
with user-controlled input.
Environment variables to be set for the command may also be passed on the command line in the form of
e.g.,
Variables passed on the command line are subject to restrictions imposed by the security policy plugin. The
policy subjects variables passed on the command line to the same restrictions as normal environment variables with one important exception. If the
option is set in
the command to be run has the
tag set or the command matched is
the user may set variables that would otherwise be forbidden. See
for more information.
When
executes a command, the security policy specifies the execution environment for the command. Typically, the real and effective user and group and IDs are set to match those of the target user, as specified in the password database, and the group vector is initialized based on the group database (unless the
option was specified).
The following parameters may be specified by security policy:
real and effective user-ID
real and effective group-ID
supplementary group-IDs
the environment list
current working directory
file creation mode mask (umask)
scheduling priority (aka nice value)
There are two distinct ways
can run a command.
If an I/O logging plugin is configured or if the security policy explicitly requests it, a new pseudo-terminal
is allocated and
is used to create a second
process, referred to as the
The
creates a new terminal session with itself as the leader and the pty as its controlling terminal, calls
sets up the execution environment as described above, and then uses the
system call to run the command in the child process. The
exists to relay job control signals between the user's existing terminal and the pty the command is being run in. This makes it possible to suspend and resume the command. Without the monitor, the command would be in what POSIX terms an
and it would not receive any job control signals from the kernel. When the command exits or is terminated by a signal, the
passes the command's exit status to the main
process and exits. After receiving the command's exit status, the main
passes the command's exit status to the security policy's close function and exits.
If no pty is used,
calls
sets up the execution environment as described above, and uses the
system call to run the command in the child process. The main
process waits until the command has completed, then passes the command's exit status to the security policy's close function and exits. As a special case, if the policy plugin does not define a close function,
will execute the command directly instead of calling
first. The
policy plugin will only define a close function when I/O logging is enabled, a pty is required, or the
or
options are enabled. Note that
and
are enabled by default on systems using PAM.
On systems that use PAM, the security policy's close function is responsible for closing the PAM session. It may also log the command's exit status.
When the command is run as a child of the
process,
will relay signals it receives to the command. The
and
signals are only relayed when the command is being run in a new pty or when the signal was sent by a user process, not the kernel. This prevents the command from receiving
twice each time the user enters control-C. Some signals, such as
and
cannot be caught and thus will not be relayed to the command. As a general rule,
should be used instead of
when you wish to suspend a command being run by
As a special case,
will not relay signals that were sent by the command it is running. This prevents the command from accidentally killing itself. On some systems, the
command sends
to all non-system processes other than itself before rebooting the system. This prevents
from relaying the
signal it received back to
which might then exit before the system was actually rebooted, leaving it in a half-dead state similar to single user mode. Note, however, that this check only applies to the command run by
and not any other processes that the command may create. As a result, running a script that calls
or
via
may cause the system to end up in this undefined state unless the
or
are run using the
family of functions instead of
(which interposes a shell between the command and the calling process).
If no I/O logging plugins are loaded and the policy plugin has not defined a
function, set a command timeout or required that the command be run in a new pty,
may execute the command directly instead of running it as a child process.
Plugins may be specified via
directives in the
file. They may be loaded as dynamic shared objects (on systems that support them), or compiled directly into the
binary. If no
file is present, or if it doesn't contain any
lines,
will use
for the policy, auditing and I/O logging plugins. See the
manual for details of the
file and the
manual for more information about the
plugin architecture.
Upon successful execution of a command, the exit status from
will be the exit status of the program that was executed. If the command terminated due to receipt of a signal,
will send itself the same signal that terminated the command.
If the
option was specified without a command,
will exit with a value of 0 if the user is allowed to run
and they authenticated successfully (as required by the security policy). If a command is specified with the
option, the exit value will only be 0 if the command is permitted by the security policy, otherwise it will be 1.
If there is an authentication failure, a configuration/permission problem or if the given command cannot be executed,
exits with a value of 1. In the latter case, the error string is printed to the standard error. If
cannot
one or more entries in the user's
an error is printed to the standard error. (If the directory does not exist or if it is not really a directory, the entry is ignored and no error is printed.) This should not happen under normal circumstances. The most common reason for
to return
is if you are running an automounter and one of the directories in your
is on a machine that is currently unreachable.
tries to be safe when executing external commands.
To prevent command spoofing,
checks "." and "" (both denoting current directory) last when searching for a command in the user's
(if one or both are in the
Note, however, that the actual
environment variable is
modified and is passed unchanged to the program that
executes.
Users should
be granted
privileges to execute files that are writable by the user or that reside in a directory that is writable by the user. If the user can modify or replace the command there is no way to limit what additional commands they can run.
Please note that
will normally only log the command it explicitly runs. If a user runs a command such as
or
subsequent commands run from that shell are not subject to
security policy. The same is true for commands that offer shell escapes (including most editors). If I/O logging is enabled, subsequent commands will have their input and/or output logged, but there will not be traditional logs for those commands. Because of this, care must be taken when giving users access to commands via
to verify that the command does not inadvertently give the user an effective root shell. For more information, please see the
section in
To prevent the disclosure of potentially sensitive information,
disables core dumps by default while it is executing (they are re-enabled for the command that is run). This historical practice dates from a time when most operating systems allowed set-user-ID processes to dump core by default. To aid in debugging
crashes, you may wish to re-enable core dumps by setting
to false in the
file as follows:
Set disable_coredump false
See the
manual for more information.
utilizes the following environment variables. The security policy has control over the actual content of the command's environment.
Default editor to use in
(sudoedit) mode if neither
nor
is set.
Set to the mail spool of the target user when the
option is specified or when
is enabled in
(unless
is present in the
list).
Set to the home directory of the target user when the
or
options are specified, when the
option is specified and
is set in
when
is enabled in
or when
is enabled in
and
is not present in the
list.
Set to the login name of the target user when the
option is specified, when the
option is enabled in
or when the
option is enabled in
(unless
is present in the
list).
May be overridden by the security policy.
Used to determine shell to run with
option.
Specifies the path to a helper program used to read the password if no terminal is available or if the
option is specified.
Set to the command run by sudo, including command line arguments. The command line arguments are truncated at 4096 characters to prevent a potential execution error.
Default editor to use in
(sudoedit) mode.
Set to the group-ID of the user who invoked sudo.
Used as the default password prompt unless the
option was specified.
If set,
will be set to its value for the program being run.
Set to the user-ID of the user who invoked sudo.
Set to the login name of the user who invoked sudo.
Set to the same value as
described above.
Default editor to use in
(sudoedit) mode if
is not set.
front end configuration
Note: the following examples assume a properly configured security policy.
To get a file listing of an unreadable directory:
$ sudo ls /usr/local/protected
To list the home directory of user yaz on a machine where the file system holding ~yaz is not exported as root:
$ sudo -u yaz ls ~yaz
To edit the
file as user www:
$ sudoedit -u www ~www/htdocs/index.html
To view system logs only accessible to root and users in the adm group:
$ sudo -g adm more /var/log/syslog
To run an editor as jim with a different primary group:
$ sudoedit -u jim -g audio ~jim/sound.txt
To shut down a machine:
$ sudo shutdown -r +15 "quick reboot"
To make a usage listing of the directories in the /home partition. Note that this runs the commands in a sub-shell to make the
and file redirection work.
$ sudo sh -c "cd /home ; du -s * | sort -rn > USAGE"
Error messages produced by
include:
By default,
does not permit editing a file when any of the parent directories are writable by the invoking user. This avoids a race condition that could allow the user to overwrite an arbitrary file. See the
option in
for more information.
By default,
does not follow symbolic links when opening files. See the
option in
for more information.
was not run with root privileges. The
binary must be owned by the root user and have the set-user-ID bit set. Also, it must not be located on a file system mounted with the
option or on an NFS file system that maps uid 0 to an unprivileged uid.
was not run with root privileges. The
binary has the proper owner and permissions but it still did not run with root privileges. The most common reason for this is that the file system the
binary is located on is mounted with the
option or it is an NFS file system that maps uid 0 to an unprivileged uid.
An error occurred while loading or initializing the plugins specified in
One or more environment variable names specified via the
option contained an equal sign
The arguments to the
option should be environment variable names without an associated value.
When
tried to read the password, it did not receive any characters. This may happen if no terminal is available (or the
option is specified) and the standard input has been redirected from
needs to read the password but there is no mechanism available for it to do so. A terminal is not present to read the password from,
has not been configured to read from the standard input, the
option was not used, and no askpass helper has been specified either via the
file or the
environment variable.
was unable to find a usable temporary directory in which to store its intermediate files.
was not run with root privileges. The
binary does not have the correct owner or permissions. It must be owned by the root user and have the set-user-ID bit set.
It is only possible to run
on systems that support setting the effective user-ID.
The user did not enter a password before the password timeout (5 minutes by default) expired.
Your user-ID does not appear in the system passwd database.
It is only possible to specify environment variables when running a command. When editing a file, the editor is run with the user's environment unmodified.
See the HISTORY file in the
distribution (https://www.sudo.ws/history.html) for a brief history of sudo.
Many people have worked on
over the years; this version consists of code written primarily by:
See the CONTRIBUTORS file in the
distribution (https://www.sudo.ws/contributors.html) for an exhaustive list of people who have contributed to
There is no easy way to prevent a user from gaining a root shell if that user is allowed to run arbitrary commands via
Also, many programs (such as editors) allow the user to run commands via shell escapes, thus avoiding
checks. However, on most systems it is possible to prevent shell escapes with the
plugin's
functionality.
It is not meaningful to run the
command directly via sudo, e.g.,
$ sudo cd /usr/local/protected
since when the command exits the parent process (your shell) will still be the same. Please see the
section for more information.
Running shell scripts via
can expose the same kernel bugs that make set-user-ID shell scripts unsafe on some operating systems (if your OS has a /dev/fd/ directory, set-user-ID shell scripts are generally safe).
If you feel you have found a bug in
please submit a bug report at https://bugzilla.sudo.ws/
Limited free support is available via the sudo-users mailing list, see https://www.sudo.ws/mailman/listinfo/sudo-users to subscribe or search the archives.
is provided
and any express or implied warranties, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose are disclaimed. See the LICENSE file distributed with
or https://www.sudo.ws/license.html for complete details.