Table of Contents
This article describes the underlying principles and mechanisms of the Pluggable Authentication Modules (PAM) library, and explains how to configure PAM, how to integrate PAM into applications, and how to write PAM modules.
See Section D.3.2, “Networks Associates Technology's license on the PAM article” for the license of this chapter.
The Pluggable Authentication Modules (PAM) library is a generalized API for authentication-related services which allows a system administrator to add new authentication methods simply by installing new PAM modules, and to modify authentication policies by editing configuration files.
PAM was defined and developed in 1995 by Vipin Samar and Charlie Lai of Sun Microsystems, and has not changed much since. In 1997, the Open Group published the X/Open Single Sign-on (XSSO) preliminary specification, which standardized the PAM API and added extensions for single (or rather integrated) sign-on. At the time of this writing, this specification has not yet been adopted as a standard.
Although this article focuses primarily on FreeBSD 5.x and NetBSD 3.x, which both use OpenPAM, it should be equally applicable to FreeBSD 4.x, which uses Linux-PAM, and other operating systems such as Linux and Solaris™.
The terminology surrounding PAM is rather confused. Neither Samar and Lai's original paper nor the XSSO specification made any attempt at formally defining terms for the various actors and entities involved in PAM, and the terms that they do use (but do not define) are sometimes misleading and ambiguous. The first attempt at establishing a consistent and unambiguous terminology was a whitepaper written by Andrew G. Morgan (author of Linux-PAM) in 1999. While Morgan's choice of terminology was a huge leap forward, it is in this author's opinion by no means perfect. What follows is an attempt, heavily inspired by Morgan, to define precise and unambiguous terms for all actors and entities involved in PAM.
The set of credentials the applicant is requesting from the arbitrator.
The user or entity requesting authentication.
The user or entity who has the privileges necessary to verify the applicant's credentials and the authority to grant or deny the request.
A sequence of modules that will be invoked in response to a PAM request. The chain includes information about the order in which to invoke the modules, what arguments to pass to them, and how to interpret the results.
The application responsible for initiating an authentication request on behalf of the applicant and for obtaining the necessary authentication information from him.
One of the four basic groups of functionality provided by PAM: authentication, account management, session management and authentication token update.
A collection of one or more related functions implementing a particular authentication facility, gathered into a single (normally dynamically loadable) binary file and identified by a single name.
The complete set of configuration statements describing how to handle PAM requests for a particular service. A policy normally consists of four chains, one for each facility, though some services do not use all four facilities.
The application acting on behalf of the arbitrator to converse with the client, retrieve authentication information, verify the applicant's credentials and grant or deny requests.
A class of servers providing similar or related functionality and requiring similar authentication. PAM policies are defined on a per-service basis, so all servers that claim the same service name will be subject to the same policy.
The context within which service is rendered to the applicant by the server. One of PAM's four facilities, session management, is concerned exclusively with setting up and tearing down this context.
A chunk of information associated with the account, such as a password or passphrase, which the applicant must provide to prove his identity.
A sequence of requests from the same applicant to the same instance of the same server, beginning with authentication and session set-up and ending with session tear-down.
This section aims to illustrate the meanings of some of the terms defined above by way of a handful of simple examples.
This simple example shows alice
	  su(1)'ing to root.
$whoamialice$ls -l `which su`-r-sr-xr-x 1 root wheel 10744 Dec 6 19:06 /usr/bin/su$su -Password:xi3kiune#whoami root
The example below shows eve try to
	  initiate an ssh(1) connection to
	  login.example.com, ask to log in as
	  bob, and succeed.  Bob should have chosen
	  a better password!
$whoamieve$ssh bob@login.example.combob@login.example.com's password:godLast login: Thu Oct 11 09:52:57 2001 from 192.168.0.1 NetBSD 3.0 (LOGIN) #1: Thu Mar 10 18:22:36 WET 2005 Welcome to NetBSD!$
The following is FreeBSD's default policy for
	  sshd:
sshd auth required pam_nologin.so no_warn sshd auth required pam_unix.so no_warn try_first_pass sshd account required pam_login_access.so sshd account required pam_unix.so sshd session required pam_lastlog.so no_fail sshd password required pam_permit.so
This policy applies to the sshd
	      service (which is not necessarily restricted to the
	      sshd(8) server.)
auth, account,
	      session and
	      password are facilities.
pam_nologin.so,
	      pam_unix.so,
	      pam_login_access.so,
	      pam_lastlog.so and
	      pam_permit.so are modules.  It is
	      clear from this example that
	      pam_unix.so provides at least two
	      facilities (authentication and account
	      management.)
There are some differences between FreeBSD and NetBSD PAM policies:
By default, every configuration is done
	    under /etc/pam.d.
If configuration is non-existent, you will not have access to the system, in contrast with FreeBSD that has a default policy of allowing authentication.
For authentication, NetBSD forces at least one
	    required, requisite or
	    binding module to be present.
The PAM API offers six different authentication primitives grouped in four facilities, which are described below.
authAuthentication. This facility concerns itself with authenticating the applicant and establishing the account credentials. It provides two primitives:
pam_authenticate(3) authenticates the applicant, usually by requesting an authentication token and comparing it with a value stored in a database or obtained from an authentication server.
pam_setcred(3) establishes account credentials such as user ID, group membership and resource limits.
accountAccount management. This facility handles non-authentication-related issues of account availability, such as access restrictions based on the time of day or the server's work load. It provides a single primitive:
pam_acct_mgmt(3) verifies that the requested account is available.
sessionSession management. This facility handles tasks associated with session set-up and tear-down, such as login accounting. It provides two primitives:
pam_open_session(3) performs tasks
		  associated with session set-up: add an entry in the
		  utmp and
		  wtmp databases, start an SSH
		  agent, etc.
pam_close_session(3) performs tasks
		  associated with session tear-down: add an entry in
		  the utmp and
		  wtmp databases, stop the SSH
		  agent, etc.
passwordPassword management. This facility is used to change the authentication token associated with an account, either because it has expired or because the user wishes to change it. It provides a single primitive:
pam_chauthtok(3) changes the authentication token, optionally verifying that it is sufficiently hard to guess, has not been used previously, etc.
Modules are a very central concept in PAM; after all, they are the “M” in “PAM”. A PAM module is a self-contained piece of program code that implements the primitives in one or more facilities for one particular mechanism; possible mechanisms for the authentication facility, for instance, include the UNIX® password database, NIS, LDAP and Radius.
FreeBSD and NetBSD implement each mechanism in a single module,
	  named
	  pam_
	  (for instance, mechanism.sopam_unix.so for the UNIX®
	  mechanism.)  Other implementations sometimes have separate
	  modules for separate facilities, and include the facility
	  name as well as the mechanism name in the module name.  To
	  name one example, Solaris™ has a
	  pam_dial_auth.so.1 module which is
	  commonly used to authenticate dialup users.
	  Also, almost every module has a man page with the same name,
	  i.e.: pam_unix(8) explains how the
	  pam_unix.so module works.
FreeBSD's original PAM implementation, based on Linux-PAM, did not use version numbers for PAM modules. This would commonly cause problems with legacy applications, which might be linked against older versions of the system libraries, as there was no way to load a matching version of the required modules.
OpenPAM, on the other hand, looks for modules that have the same version number as the PAM library (currently 2 in FreeBSD and 0 in NetBSD), and only falls back to an unversioned module if no versioned module could be loaded. Thus legacy modules can be provided for legacy applications, while allowing new (or newly built) applications to take advantage of the most recent modules.
Although Solaris™ PAM modules commonly have a version number, they're not truly versioned, because the number is a part of the module name and must be included in the configuration.
When a server initiates a PAM transaction, the PAM library tries to load a policy for the service specified in the pam_start(3) call. The policy specifies how authentication requests should be processed, and is defined in a configuration file. This is the other central concept in PAM: the possibility for the admin to tune the system security policy (in the wider sense of the word) simply by editing a text file.
A policy consists of four chains, one for each of the four PAM facilities. Each chain is a sequence of configuration statements, each specifying a module to invoke, some (optional) parameters to pass to the module, and a control flag that describes how to interpret the return code from the module.
Understanding the control flags is essential to understanding PAM configuration files. There are a number of different control flags:
bindingIf the module succeeds and no earlier module in the chain has failed, the chain is immediately terminated and the request is granted. If the module fails, the rest of the chain is executed, but the request is ultimately denied.
This control flag was introduced by Sun in Solaris™ 9 (SunOS™ 5.9), and is also supported by OpenPAM.
requiredIf the module succeeds, the rest of the chain is executed, and the request is granted unless some other module fails. If the module fails, the rest of the chain is also executed, but the request is ultimately denied.
requisiteIf the module succeeds, the rest of the chain is executed, and the request is granted unless some other module fails. If the module fails, the chain is immediately terminated and the request is denied.
sufficientIf the module succeeds and no earlier module in the chain has failed, the chain is immediately terminated and the request is granted. If the module fails, the module is ignored and the rest of the chain is executed.
As the semantics of this flag may be somewhat
	      confusing, especially when it is used for the last
	      module in a chain, it is recommended that the
	      binding control flag be used instead
	      if the implementation supports it.
optionalThe module is executed, but its result is ignored.
	      If all modules in a chain are marked
	      optional, all requests will always be
	      granted.
When a server invokes one of the six PAM primitives, PAM
	retrieves the chain for the facility the primitive belongs to,
	and invokes each of the modules listed in the chain, in the
	order they are listed, until it reaches the end, or determines
	that no further processing is necessary (either because a
	binding or
	sufficient module succeeded, or because a
	requisite module failed.)  The request is
	granted if and only if at least one module was invoked, and
	all non-optional modules succeeded.
Note that it is possible, though not very common, to have the same module listed several times in the same chain. For instance, a module that looks up user names and passwords in a directory server could be invoked multiple times with different parameters specifying different directory servers to contact. PAM treat different occurrences of the same module in the same chain as different, unrelated modules.
The lifecycle of a typical PAM transaction is described below. Note that if any of these steps fails, the server should report a suitable error message to the client and abort the transaction.
If necessary, the server obtains arbitrator
	    credentials through a mechanism independent of
	    PAM—most commonly by virtue of having been started
	    by root, or of being setuid
	    root.
The server calls pam_start(3) to initialize the PAM library and specify its service name and the target account, and register a suitable conversation function.
The server obtains various information relating to the transaction (such as the applicant's user name and the name of the host the client runs on) and submits it to PAM using pam_set_item(3).
The server calls pam_authenticate(3) to authenticate the applicant.
The server calls pam_acct_mgmt(3) to verify that the
	    requested account is available and valid.  If the password
	    is correct but has expired, pam_acct_mgmt(3) will
	    return PAM_NEW_AUTHTOK_REQD instead of
	    PAM_SUCCESS.
If the previous step returned
	    PAM_NEW_AUTHTOK_REQD, the server now
	    calls pam_chauthtok(3) to force the client to change
	    the authentication token for the requested account.
Now that the applicant has been properly authenticated, the server calls pam_setcred(3) to establish the credentials of the requested account. It is able to do this because it acts on behalf of the arbitrator, and holds the arbitrator's credentials.
Once the correct credentials have been established, the server calls pam_open_session(3) to set up the session.
The server now performs whatever service the client requested—for instance, provide the applicant with a shell.
Once the server is done serving the client, it calls pam_close_session(3) to tear down the session.
Finally, the server calls pam_end(3) to notify the PAM library that it is done and that it can release whatever resources it has allocated in the course of the transaction.
The traditional PAM policy file is
	  /etc/pam.conf.  This file contains all
	  the PAM policies for your system.  Each line of the file
	  describes one step in a chain, as shown below:
login auth required pam_nologin.so no_warn
The fields are, in order: service name, facility name, control flag, module name, and module arguments. Any additional fields are interpreted as additional module arguments.
A separate chain is constructed for each service /
	  facility pair, so while the order in which lines for the
	  same service and facility appear is significant, the order
	  in which the individual services and facilities are listed
	  is not.  The examples in the original PAM paper grouped
	  configuration lines by facility, and the Solaris™ stock
	  pam.conf still does that, but FreeBSD's
	  stock configuration groups configuration lines by service.
	  Either way is fine; either way makes equal sense.
OpenPAM and Linux-PAM support an alternate configuration
	  mechanism, which is the preferred mechanism in FreeBSD and
	  NetBSD.
	  In this scheme, each policy is contained in a separate file
	  bearing the name of the service it applies to.  These files
	  are stored in /etc/pam.d/.
These per-service policy files have only four fields
	  instead of pam.conf's five: the service
	  name field is omitted.  Thus, instead of the sample
	  pam.conf line from the previous
	  section, one would have the following line in
	  /etc/pam.d/login:
auth required pam_nologin.so no_warn
As a consequence of this simplified syntax, it is
	  possible to use the same policy for multiple services by
	  linking each service name to a same policy file.  For
	  instance, to use the same policy for the
	  su and sudo services,
	  one could do as follows:
#cd /etc/pam.d#ln -s su sudo
This works because the service name is determined from the file name rather than specified in the policy file, so the same file can be used for multiple differently-named services.
Since each service's policy is stored in a separate
	  file, the pam.d mechanism also makes it
	  very easy to install additional policies for third-party
	  software packages.
As we have seen above, PAM policies can be found in a
	  number of places. If no configuration file is found for a
	  particular service, the /etc/pam.d/other
	  is used instead. If that file does not exist,
	  /etc/pam.conf is searched for entries
	  matching he specified service or, failing that, the "other"
	  service.
It is essential to understand that PAM's configuration system is centered on chains.
As explained in the PAM policy files section, each line in
	/etc/pam.conf consists of four or more
	fields: the service name, the facility name, the control flag,
	the module name, and zero or more module arguments.
The service name is generally (though not always) the name of the application the statement applies to. If you are unsure, refer to the individual application's documentation to determine what service name it uses.
Note that if you use /etc/pam.d/
	instead of /etc/pam.conf, the service
	name is specified by the name of the policy file, and omitted
	from the actual configuration lines, which then start with the
	facility name.
The facility is one of the four facility keywords described in the Facilities and primitives section.
Likewise, the control flag is one of the four keywords described in the Chains and policies section, describing how to interpret the return code from the module. Linux-PAM supports an alternate syntax that lets you specify the action to associate with each possible return code, but this should be avoided as it is non-standard and closely tied in with the way Linux-PAM dispatches service calls (which differs greatly from the way Solaris™ and OpenPAM do it.) Unsurprisingly, OpenPAM does not support this syntax.
To configure PAM correctly, it is essential to understand how policies are interpreted.
When an application calls pam_start(3), the PAM
	library loads the policy for the specified service and
	constructs four module chains (one for each facility.)  If one
	or more of these chains are empty, the corresponding chains
	from the policy for the other service are
	substituted.
When the application later calls one of the six PAM primitives, the PAM library retrieves the chain for the corresponding facility and calls the appropriate service function in each module listed in the chain, in the order in which they were listed in the configuration. After each call to a service function, the module type and the error code returned by the service function are used to determine what happens next. With a few exceptions, which we discuss below, the following table applies:
Table 18.1. PAM chain execution summary
| PAM_SUCCESS | PAM_IGNORE | other | |
|---|---|---|---|
| binding | if (!fail) break; | - | fail = true; | 
| required | - | - | fail = true; | 
| requisite | - | - | fail = true; break; | 
| sufficient | if (!fail) break; | - | - | 
| optional | - | - | - | 
If fail is true at the end of a chain,
	or when a “break” is reached, the dispatcher
	returns the error code returned by the first module that
	failed.  Otherwise, it returns
	PAM_SUCCESS.
The first exception of note is that the error code
	PAM_NEW_AUTHTOK_REQD is treated like a
	success, except that if no module failed, and at least one
	module returned PAM_NEW_AUTHTOK_REQD, the
	dispatcher will return
	PAM_NEW_AUTHTOK_REQD.
The second exception is that pam_setcred(3) treats
	binding and
	sufficient modules as if they were
	required.
The third and final exception is that
	pam_chauthtok(3) runs the entire chain twice (once for
	preliminary checks and once to actually set the password), and
	in the preliminary phase it treats
	binding and
	sufficient modules as if they were
	required.
The pam_deny(8) module is one of the simplest modules
  	  available; it responds to any request with
	  PAM_AUTH_ERR.  It is useful for quickly
	  disabling a service (add it to the top of every chain), or for
	  terminating chains of sufficient
	  modules.
The pam_echo(8) module simply passes its arguments to
	  the conversation function as a
	  PAM_TEXT_INFO message.  It is mostly useful
	  for debugging, but can also serve to display messages such as
	  “Unauthorized access will be prosecuted” before
	  starting the authentication procedure.
The pam_exec(8) module takes its first argument to be the name of a program to execute, and the remaining arguments are passed to that program as command-line arguments. One possible application is to use it to run a program at login time which mounts the user's home directory.
The pam_ftpusers(8) module successes if and only if
	the user is listed in /etc/ftpusers.
	Currently, in NetBSD, this module doesn't understand the
	extended syntax of ftpd(8), but this will be fixed in
	the future.
	
The pam_group(8) module accepts or rejects applicants
  	  on the basis of their membership in a particular file group
	  (normally wheel for su(1)).  It is
	  primarily intended for maintaining the traditional behaviour
	  of BSD su(1), but has many other uses, such as excluding
	  certain groups of users from a particular service.
In NetBSD, there is an argument called
	  authenticate in which the user is asked to
	  authenticate using his own password.
The pam_guest(8) module allows guest logins using fixed login names. Various requirements can be placed on the password, but the default behaviour is to allow any password as long as the login name is that of a guest account. The pam_guest(8) module can easily be used to implement anonymous FTP logins.
The pam_krb5(8) module provides functions to verify the identity of a user and to set user specific credentials using Kerberos 5. It prompts the user for a password and obtains a new Kerberos TGT for the principal. The TGT is verified by obtaining a service ticket for the local host. The newly acquired credentials are stored in a credential cache and the environment variable KRB5CCNAME is set appropriately. The credentials cache should be destroyed by the user at logout with kdestroy(1).
The pam_ksu(8) module provides only authentication services for Kerberos 5 to determine whether or not the applicant is authorized to obtain the privileges of the target account.
The pam_lastlog(8) module provides only session management services. It records the session in utmp(5), utmpx(5), wtmp(5), wtmpx(5), lastlog(5) and lastlogx(5) databases.
The pam_login_access(8) module provides an implementation of the account management primitive which enforces the login restrictions specified in the login.access(5) table.
The pam_nologin(8) module refuses non-root logins
	  when /var/run/nologin exists.  This file
	  is normally created by shutdown(8) when less than five
	  minutes remain until the scheduled shutdown time.
The pam_permit(8) module is one of the simplest
	  modules available; it responds to any request with
	  PAM_SUCCESS.  It is useful as a placeholder
	  for services where one or more chains would otherwise be
	  empty.
The pam_radius(8) module provides authentication services based upon the RADIUS (Remote Authentication Dial In User Service) protocol.
The pam_rhosts(8) module provides only
	  authentication services. It reports success if and only if the
	  target user's ID is not 0 and the remote host and user are
	  listed in /etc/hosts.equiv or in the
	  target user's ~/.rhosts.
The pam_rootok(8) module reports success if and only
	  if the real user id of the process calling it (which is
	  assumed to be run by the applicant) is 0.  This is useful for
	  non-networked services such as su(1) or passwd(1),
	  to which the root should have automatic
	  access.
The pam_securetty(8) module provides only account services. It is used when the applicant is attempting to authenticate as superuser, and the process is attached to an insecure TTY.
The pam_self(8) module reports success if and only if the names of the applicant matches that of the target account. It is most useful for non-networked services such as su(1), where the identity of the applicant can be easily verified.
The pam_ssh(8) module provides both authentication
	  and session services.  The authentication service allows users
	  who have passphrase-protected SSH secret keys in their
	  ~/.ssh directory to authenticate
	  themselves by typing their passphrase.  The session service
	  starts ssh-agent(1) and preloads it with the keys that
	  were decrypted in the authentication phase.  This feature is
	  particularly useful for local logins, whether in X (using
	  xdm(1) or another PAM-aware X login manager) or at the
	  console.
This module implements what is fundamentally a password authentication scheme. Care should be taken to only use this module over a secure session (secure TTY, encrypted session, etc.), otherwise the user's SSH passphrase could be compromised.
Additional consideration should be given to the use of pam_ssh(8). Users often assume that file permissions are sufficient to protect their SSH keys, and thus use weak or no passphrases. Since the system administrator has no effective means of enforcing SSH passphrase quality, this has the potential to expose the system to security risks.
The pam_unix(8) module implements traditional UNIX® password authentication, using getpwnam(3) under FreeBSD or getpwnam_r(3) under NetBSD to obtain the target account's password and compare it with the one provided by the applicant. It also provides account management services (enforcing account and password expiration times) and password-changing services. This is probably the single most useful module, as the great majority of admins will want to maintain historical behaviour for at least some services.
The pam_skey(8) module implements S/Key One Time
	  Password (OTP) authentication methods, using the
	  /etc/skeykeys database.
The following is a minimal implementation of su(1)
      using PAM.  Note that it uses the OpenPAM-specific
      openpam_ttyconv(3) conversation function, which is
      prototyped in
      security/openpam.h.
      If you wish
      build this application on a system with a different PAM library,
      you will have to provide your own conversation function.  A
      robust conversation function is surprisingly difficult to
      implement; the one presented in the
      Sample PAM Conversation
      Function sub-chapter is a good
      starting point, but should not be used in real-world
      applications.
#include <sys/param.h>
#include <sys/wait.h>
#include <err.h>
#include <pwd.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <syslog.h>
#include <unistd.h>
#include <security/pam_appl.h>
#include <security/openpam.h>	/* for openpam_ttyconv() */
extern char **environ;
static pam_handle_t *pamh;
static struct pam_conv pamc;
static void
usage(void)
{
	fprintf(stderr, "Usage: su [login [args]]\n");
	exit(1);
}
int
main(int argc, char *argv[])
{
	char hostname[MAXHOSTNAMELEN];
	const char *user, *tty;
	char **args, **pam_envlist, **pam_env;
	struct passwd *pwd;
	int o, pam_err, status;
	pid_t pid;
	while ((o = getopt(argc, argv, "h")) != -1)
		switch (o) {
		case 'h':
		default:
			usage();
		}
	argc -= optind;
	argv += optind;
	if (argc > 0) {
		user = *argv;
		--argc;
		++argv;
	} else {
		user = "root";
	}
	/* initialize PAM */
	pamc.conv = &openpam_ttyconv;
	pam_start("su", user, &pamc, &pamh);
	/* set some items */
	gethostname(hostname, sizeof(hostname));
	if ((pam_err = pam_set_item(pamh, PAM_RHOST, hostname)) != PAM_SUCCESS)
		goto pamerr;
	user = getlogin();
	if ((pam_err = pam_set_item(pamh, PAM_RUSER, user)) != PAM_SUCCESS)
		goto pamerr;
	tty = ttyname(STDERR_FILENO);
	if ((pam_err = pam_set_item(pamh, PAM_TTY, tty)) != PAM_SUCCESS)
		goto pamerr;
	/* authenticate the applicant */
	if ((pam_err = pam_authenticate(pamh, 0)) != PAM_SUCCESS)
		goto pamerr;
	if ((pam_err = pam_acct_mgmt(pamh, 0)) == PAM_NEW_AUTHTOK_REQD)
		pam_err = pam_chauthtok(pamh, PAM_CHANGE_EXPIRED_AUTHTOK);
	if (pam_err != PAM_SUCCESS)
		goto pamerr;
	/* establish the requested credentials */
	if ((pam_err = pam_setcred(pamh, PAM_ESTABLISH_CRED)) != PAM_SUCCESS)
		goto pamerr;
	/* authentication succeeded; open a session */
	if ((pam_err = pam_open_session(pamh, 0)) != PAM_SUCCESS)
		goto pamerr;
	/* get mapped user name; PAM may have changed it */
	pam_err = pam_get_item(pamh, PAM_USER, (const void **)&user);
	if (pam_err != PAM_SUCCESS || (pwd = getpwnam(user)) == NULL)
		goto pamerr;
	/* export PAM environment */
	if ((pam_envlist = pam_getenvlist(pamh)) != NULL) {
		for (pam_env = pam_envlist; *pam_env != NULL; ++pam_env) {
			putenv(*pam_env);
			free(*pam_env);
		}
		free(pam_envlist);
	}
	/* build argument list */
	if ((args = calloc(argc + 2, sizeof *args)) == NULL) {
		warn("calloc()");
		goto err;
	}
	*args = pwd->pw_shell;
	memcpy(args + 1, argv, argc * sizeof *args);
	/* fork and exec */
	switch ((pid = fork())) {
	case -1:
		warn("fork()");
		goto err;
	case 0:
		/* child: give up privs and start a shell */
		/* set uid and groups */
		if (initgroups(pwd->pw_name, pwd->pw_gid) == -1) {
			warn("initgroups()");
			_exit(1);
		}
		if (setgid(pwd->pw_gid) == -1) {
			warn("setgid()");
			_exit(1);
		}
		if (setuid(pwd->pw_uid) == -1) {
			warn("setuid()");
			_exit(1);
		}
		execve(*args, args, environ);
		warn("execve()");
		_exit(1);
	default:
		/* parent: wait for child to exit */
		waitpid(pid, &status, 0);
		/* close the session and release PAM resources */
		pam_err = pam_close_session(pamh, 0);
		pam_end(pamh, pam_err);
		exit(WEXITSTATUS(status));
	}
pamerr:
	fprintf(stderr, "Sorry\n");
err:
	pam_end(pamh, pam_err);
	exit(1);
}
The following is a minimal implementation of pam_unix(8), offering only authentication services. It should build and run with most PAM implementations, but takes advantage of OpenPAM extensions if available: note the use of pam_get_authtok(3), which enormously simplifies prompting the user for a password.
#include <sys/param.h>
#include <pwd.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <security/pam_modules.h>
#include <security/pam_appl.h>
#ifndef _OPENPAM
static char password_prompt[] = "Password:";
#endif
#ifndef PAM_EXTERN
#define PAM_EXTERN
#endif
PAM_EXTERN int
pam_sm_authenticate(pam_handle_t *pamh, int flags,
	int argc, const char *argv[])
{
#ifndef _OPENPAM
	const void *ptr;
	const struct pam_conv *conv;
	struct pam_message msg;
	const struct pam_message *msgp;
	struct pam_response *resp;
#endif
	struct passwd *pwd;
	const char *user;
	char *crypt_password, *password;
	int pam_err, retry;
	/* identify user */
	if ((pam_err = pam_get_user(pamh, &user, NULL)) != PAM_SUCCESS)
		return (pam_err);
	if ((pwd = getpwnam(user)) == NULL)
		return (PAM_USER_UNKNOWN);
	/* get password */
#ifndef _OPENPAM
	pam_err = pam_get_item(pamh, PAM_CONV, &ptr);
	if (pam_err != PAM_SUCCESS)
		return (PAM_SYSTEM_ERR);
	conv = ptr;
	msg.msg_style = PAM_PROMPT_ECHO_OFF;
	msg.msg = password_prompt;
	msgp = &msg;
#endif
	password = NULL;
	for (retry = 0; retry < 3; ++retry) {
#ifdef _OPENPAM
		pam_err = pam_get_authtok(pamh, PAM_AUTHTOK,
		    (const char **)&password, NULL);
#else
		resp = NULL;
		pam_err = (*conv->conv)(1, &msgp, &resp, conv->appdata_ptr);
		if (resp != NULL) {
			if (pam_err == PAM_SUCCESS)
				password = resp->resp;
			else
				free(resp->resp);
			free(resp);
		}
#endif
		if (pam_err == PAM_SUCCESS)
			break;
	}
	if (pam_err == PAM_CONV_ERR)
		return (pam_err);
	if (pam_err != PAM_SUCCESS)
		return (PAM_AUTH_ERR);
	/* compare passwords */
	if ((!pwd->pw_passwd[0] && (flags & PAM_DISALLOW_NULL_AUTHTOK)) ||
	    (crypt_password = crypt(password, pwd->pw_passwd)) == NULL ||
	    strcmp(crypt_password, pwd->pw_passwd) != 0)
		pam_err = PAM_AUTH_ERR;
	else
		pam_err = PAM_SUCCESS;
#ifndef _OPENPAM
	free(password);
#endif
	return (pam_err);
}
PAM_EXTERN int
pam_sm_setcred(pam_handle_t *pamh, int flags,
	int argc, const char *argv[])
{
	return (PAM_SUCCESS);
}
PAM_EXTERN int
pam_sm_acct_mgmt(pam_handle_t *pamh, int flags,
	int argc, const char *argv[])
{
	return (PAM_SUCCESS);
}
PAM_EXTERN int
pam_sm_open_session(pam_handle_t *pamh, int flags,
	int argc, const char *argv[])
{
	return (PAM_SUCCESS);
}
PAM_EXTERN int
pam_sm_close_session(pam_handle_t *pamh, int flags,
	int argc, const char *argv[])
{
	return (PAM_SUCCESS);
}
PAM_EXTERN int
pam_sm_chauthtok(pam_handle_t *pamh, int flags,
	int argc, const char *argv[])
{
	return (PAM_SERVICE_ERR);
}
#ifdef PAM_MODULE_ENTRY
PAM_MODULE_ENTRY("pam_unix");
#endif
The conversation function presented below is a greatly simplified version of OpenPAM's openpam_ttyconv(3). It is fully functional, and should give the reader a good idea of how a conversation function should behave, but it is far too simple for real-world use. Even if you're not using OpenPAM, feel free to download the source code and adapt openpam_ttyconv(3) to your uses; we believe it to be as robust as a tty-oriented conversation function can reasonably get.
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <security/pam_appl.h>
int
converse(int n, const struct pam_message **msg,
	struct pam_response **resp, void *data)
{
	struct pam_response *aresp;
	char buf[PAM_MAX_RESP_SIZE];
	int i;
	data = data;
	if (n <= 0 || n > PAM_MAX_NUM_MSG)
		return (PAM_CONV_ERR);
	if ((aresp = calloc(n, sizeof *aresp)) == NULL)
		return (PAM_BUF_ERR);
	for (i = 0; i < n; ++i) {
		aresp[i].resp_retcode = 0;
		aresp[i].resp = NULL;
		switch (msg[i]->msg_style) {
		case PAM_PROMPT_ECHO_OFF:
			aresp[i].resp = strdup(getpass(msg[i]->msg));
			if (aresp[i].resp == NULL)
				goto fail;
			break;
		case PAM_PROMPT_ECHO_ON:
			fputs(msg[i]->msg, stderr);
			if (fgets(buf, sizeof buf, stdin) == NULL)
				goto fail;
			aresp[i].resp = strdup(buf);
			if (aresp[i].resp == NULL)
				goto fail;
			break;
		case PAM_ERROR_MSG:
			fputs(msg[i]->msg, stderr);
			if (strlen(msg[i]->msg) > 0 &&
			    msg[i]->msg[strlen(msg[i]->msg) - 1] != '\n')
				fputc('\n', stderr);
			break;
		case PAM_TEXT_INFO:
			fputs(msg[i]->msg, stdout);
			if (strlen(msg[i]->msg) > 0 &&
			    msg[i]->msg[strlen(msg[i]->msg) - 1] != '\n')
				fputc('\n', stdout);
			break;
		default:
			goto fail;
		}
	}
	*resp = aresp;
	return (PAM_SUCCESS);
 fail:
        for (i = 0; i < n; ++i) {
                if (aresp[i].resp != NULL) {
                        memset(aresp[i].resp, 0, strlen(aresp[i].resp));
                        free(aresp[i].resp);
                }
        }
        memset(aresp, 0, n * sizeof *aresp);
	*resp = NULL;
	return (PAM_CONV_ERR);
}
[sun-pam] Making Login Services Independent of Authentication Technologies. Sun Microsystems.
[opengroup-singlesignon] X/Open Single Sign-on Preliminary Specification. The Open Group. 1-85912-144-6. June 1997.
[kernelorg-pamdraft] Pluggable Authentication Modules. October 6, 1999.
[sun-pamadmin] PAM Administration. Sun Microsystems.
[openpam-website] OpenPAM homepage. ThinkSec AS.
[linuxpam-website] Linux-PAM homepage.
[solarispam-website] Solaris PAM homepage. Sun Microsystems.