GETLOGIN(2) BSD System Calls Manual GETLOGIN(2)
getlogin, getlogin_r, setlogin -- get/set login name
getlogin_r(char *name, size_t namelen);
setlogin(const char *name);
The getlogin() routine returns the login name of the user associated with
the current session, as previously set by setlogin(). The name is nor-
mally associated with a login shell at the time a session is created, and
is inherited by all processes descended from the login shell. (This is
true even if some of those processes assume another user ID, for example
when su(1) is used.)
The getlogin_r() routine is a reentrant version of getlogin(). It is
functionally identical to getlogin() except that the caller must provide
a buffer, name, in which to store the user's login name and a correspond-
ing length parameter, namelen, that specifies the size of the buffer.
The buffer should be large enough to store the login name and a trailing
NUL (typically LOGIN_NAME_MAX bytes).
setlogin() sets the login name of the user associated with the current
session to name. This call is restricted to the superuser, and is nor-
mally used only when a new session is being created on behalf of the
named user (for example, at login time, or when a remote shell is
NOTE: There is only one login name per session.
It is CRITICALLY important to ensure that setlogin() is only ever called
after the process has taken adequate steps to ensure that it is detached
from its parent's session. The ONLY way to do this is via the setsid()
function. The daemon() function calls setsid() which is an ideal way of
detaching from a controlling terminal and forking into the background.
In particular, neither ioctl(ttyfd, TIOCNOTTY, ...) nor setpgrp(...) is
sufficient to create a new session.
Once a parent process has called setsid(), it is acceptable for some
child of that process to then call setlogin(), even though it is not the
session leader. Beware, however, that ALL processes in the session will
change their login name at the same time, even the parent.
This is different from traditional UNIX privilege inheritance and as such
can be counter-intuitive.
Since the setlogin() routine is restricted to the super-user, it is
assumed that (like all other privileged programs) the programmer has
taken adequate precautions to prevent security violations.
If a call to getlogin() succeeds, it returns a pointer to a NUL-termi-
nated string in a static buffer. If the name has not been set, it
returns NULL. If a call to getlogin_r() succeeds, a value of 0 is
returned, else the error number is returned. If a call to setlogin()
succeeds, a value of 0 is returned. If setlogin() fails, a value of -1
is returned and an error code is placed in the global location errno.
getlogin_r() and setlogin() will succeed unless:
[EFAULT] The name parameter points to an invalid address.
In addition, getlogin_r() may return the following error:
[ERANGE] The value of namelen is not large enough to store the
user's login name and a trailing NUL.
setlogin() may return the following errors:
[EINVAL] The name parameter pointed to a string that was too
long. Login names are limited to LOGIN_NAME_MAX-1
characters, currently 31.
[EPERM] The caller tried to set the login name and was not the
The getlogin() and getlogin_r() functions conform to IEEE Std 1003.1-2008
The getlogin() function first appeared in 4.2BSD.
In earlier versions of the system, getlogin() failed unless the process
was associated with a login terminal. The current implementation (using
setlogin()) allows getlogin to succeed even when the process has no con-
trolling terminal. In earlier versions of the system, the value returned
by getlogin() could not be trusted without checking the user ID. Porta-
ble programs should probably still make this check.
BSD March 30, 2013 BSD