1CREDENTIALS(7) Linux Programmer's Manual CREDENTIALS(7)
2
6 credentials - process identifiers
7
9 Process ID (PID)
10 Each process has a unique nonnegative integer identifier that is
11 assigned when the process is created using fork(2). A process can
12 obtain its PID using getpid(2). A PID is represented using the type
13 pid_t (defined in <sys/types.h>).
14 PIDs are used in a range of system calls to identify the process
16 affected by the call, for example: kill(2), ptrace(2), setpriority(2)
17 setpgid(2), setsid(2), sigqueue(2), and waitpid(2).
18 A process's PID is preserved across an execve(2).
20 Parent Process ID (PPID)
22 A process's parent process ID identifies the process that created this
23 process using fork(2). A process can obtain its PPID using getppid(2).
24 A PPID is represented using the type pid_t.
25 A process's PPID is preserved across an execve(2).
27 Process Group ID and Session ID
29 Each process has a session ID and a process group ID, both represented
30 using the type pid_t. A process can obtain its session ID using get‐
31 sid(2), and its process group ID using getpgrp(2).
32 A child created by fork(2) inherits its parent's session ID and process
34 group ID. A process's session ID and process group ID are preserved
35 across an execve(2).
36 Sessions and process groups are abstractions devised to support shell
38 job control. A process group (sometimes called a "job") is a collec‐
39 tion of processes that share the same process group ID; the shell cre‐
40 ates a new process group for the process(es) used to execute single
41 command or pipeline (e.g., the two processes created to execute the
42 command "ls | wc" are placed in the same process group). A process's
43 group membership can be set using setpgid(2). The process whose
44 process ID is the same as its process group ID is the process group
45 leader for that group.
46 A session is a collection of processes that share the same session ID.
48 All of the members of a process group also have the same session ID
49 (i.e., all of the members of a process group always belong to the same
50 session, so that sessions and process groups form a strict two-level
51 hierarchy of processes.) A new session is created when a process calls
52 setsid(2), which creates a new session whose session ID is the same as
53 the PID of the process that called setsid(2). The creator of the ses‐
54 sion is called the session leader.
55 User and Group Identifiers
57 Each process has various associated user and groups IDs. These IDs are
58 integers, respectively represented using the types uid_t and gid_t
59 (defined in <sys/types.h>).
60 On Linux, each process has the following user and group identifiers:
62 * Real user ID and real group ID. These IDs determine who owns the
64 process. A process can obtain its real user (group) ID using
65 getuid(2) (getgid(2)).
66 * Effective user ID and effective group ID. These IDs are used by the
68 kernel to determine the permissions that the process will have when
69 accessing shared resources such as message queues, shared memory,
70 and semaphores. On most Unix systems, these IDs also determine the
71 permissions when accessing files. However, Linux uses the file sys‐
72 tem IDs described below for this task. A process can obtain its
73 effective user (group) ID using geteuid(2) (getegid(2)).
74 * Saved set-user-ID and saved set-group-ID. These IDs are used in
76 set-user-ID and set-group-ID programs to save a copy of the corre‐
77 sponding effective IDs that were set when the program was executed
78 (see execve(2)). A set-user-ID program can assume and drop privi‐
79 leges by switching its effective user ID back and forth between the
80 values in its real user ID and saved set-user-ID. This switching is
81 done via calls to seteuid(2), setreuid(2), or setresuid(2). A set-
82 group-ID program performs the analogous tasks using setegid(2),
83 setregid(2), or setresgid(2). A process can obtain its saved set-
84 user-ID (set-group-ID) using getresuid(2) (getresgid(2)).
85 * File system user ID and file system group ID (Linux-specific).
87 These IDs, in conjunction with the supplementary group IDs described
88 below, are used to determine permissions for accessing files; see
89 path_resolution(7) for details. Whenever a process's effective user
90 (group) ID is changed, the kernel also automatically changes the
91 file system user (group) ID to the same value. Consequently, the
92 file system IDs normally have the same values as the corresponding
93 effective ID, and the semantics for file-permission checks are thus
94 the same on Linux as on other Unix systems. The file system IDs can
95 be made to differ from the effective IDs by calling setfsuid(2) and
96 setfsgid(2).
97 * Supplementary group IDs. This is a set of additional group IDs that
99 are used for permission checks when accessing files and other shared
100 resources. On Linux kernels before 2.6.4, a process can be a member
101 of up to 32 supplementary groups; since kernel 2.6.4, a process can
102 be a member of up to 65536 supplementary groups. The call
103 sysconf(_SC_NGROUPS_MAX) can be used to determine the number of sup‐
104 plementary groups of which a process may be a member. A process can
105 obtain its set of supplementary group IDs using getgroups(2), and
106 can modify the set using setgroups(2).
107 A child process created by fork(2) inherits copies of its parent's user
109 and groups IDs. During an execve(2), a process's real user and group
110 ID and supplementary group IDs are preserved; the effective and saved
111 set IDs may be changed, as described in execve(2).
112 Aside from the purposes noted above, a process's user IDs are also
114 employed in a number of other contexts:
115 * when determining the permissions for sending signals — see kill(2);
117 * when determining the permissions for setting process-scheduling
119 parameters (nice value, real time scheduling policy and priority,
120 CPU affinity, I/O priority) using setpriority(2), sched_setaffin‐
121 ity(2), sched_setscheduler(2), sched_setparam(2), and ioprio_set(2);
122 * when checking resource limits; see getrlimit(2);
124 * when checking the limit on the number of inotify instances that the
126 process may create; see inotify(7).
127
129 Process IDs, parent process IDs, process group IDs, and session IDs are
130 specified in POSIX.1-2001. The real, effective, and saved set user and
131 groups IDs, and the supplementary group IDs, are specified in
132 POSIX.1-2001. The file system user and group IDs are a Linux exten‐
133 sion.
134
136 The POSIX threads specification requires that credentials are shared by
137 all of the threads in a process. However, at the kernel level, Linux
138 maintains separate user and group credentials for each thread. The
139 NPTL threading implementation does some work to ensure that any change
140 to user or group credentials (e.g., calls to setuid(2), setresuid(2),
141 etc.) is carried through to all of the POSIX threads in a process.
142
144 bash(1), csh(1), ps(1), access(2), execve(2), faccessat(2), fork(2),
145 getpgrp(2), getpid(2), getppid(2), getsid(2), kill(2), killpg(2), sete‐
146 gid(2), seteuid(2), setfsgid(2), setfsuid(2), setgid(2), setgroups(2),
147 setresgid(2), setresuid(2), setuid(2), waitpid(2), euidaccess(3), init‐
148 groups(3), tcgetpgrp(3), tcsetpgrp(3), capabilities(7), path_resolu‐
149 tion(7), unix(7)
150
152 This page is part of release 3.25 of the Linux man-pages project. A
153 description of the project, and information about reporting bugs, can
154 be found at http://www.kernel.org/doc/man-pages/.
155Linux 2008-06-03 CREDENTIALS(7)