dup2(2) - f39

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

NAME

6       dup, dup2, dup3 - duplicate a file descriptor
7

LIBRARY

9       Standard C library (libc, -lc)
10

SYNOPSIS

12       #include <unistd.h>
13
14       int dup(int oldfd);
15       int dup2(int oldfd, int newfd);
16
17       #define _GNU_SOURCE             /* See feature_test_macros(7) */
18       #include <fcntl.h>              /* Definition of O_* constants */
19       #include <unistd.h>
20
21       int dup3(int oldfd, int newfd, int flags);
22

DESCRIPTION

24       The  dup()  system  call allocates a new file descriptor that refers to
25       the same open file description as the descriptor oldfd.  (For an expla‐
26       nation  of open file descriptions, see open(2).)  The new file descrip‐
27       tor number is guaranteed to be the lowest-numbered file descriptor that
28       was unused in the calling process.
29
30       After a successful return, the old and new file descriptors may be used
31       interchangeably.  Since the two file descriptors refer to the same open
32       file description, they share file offset and file status flags; for ex‐
33       ample, if the file offset is modified by using lseek(2) on one  of  the
34       file  descriptors,  the  offset  is also changed for the other file de‐
35       scriptor.
36
37       The two file descriptors do not share file descriptor flags (the close-
38       on-exec  flag).   The close-on-exec flag (FD_CLOEXEC; see fcntl(2)) for
39       the duplicate descriptor is off.
40
41   dup2()
42       The dup2() system call performs the same task as dup(), but instead  of
43       using  the lowest-numbered unused file descriptor, it uses the file de‐
44       scriptor number specified in newfd.  In other words, the file  descrip‐
45       tor  newfd  is adjusted so that it now refers to the same open file de‐
46       scription as oldfd.
47
48       If the file descriptor newfd was previously open, it is  closed  before
49       being  reused; the close is performed silently (i.e., any errors during
50       the close are not reported by dup2()).
51
52       The steps of closing and reusing the file  descriptor  newfd  are  per‐
53       formed  atomically.   This  is  important,  because trying to implement
54       equivalent functionality using close(2) and dup() would be  subject  to
55       race  conditions,  whereby newfd might be reused between the two steps.
56       Such reuse could happen because the main program is  interrupted  by  a
57       signal  handler that allocates a file descriptor, or because a parallel
58       thread allocates a file descriptor.
59
60       Note the following points:
61
62       •  If oldfd is not a valid file descriptor, then the  call  fails,  and
63          newfd is not closed.
64
65       •  If oldfd is a valid file descriptor, and newfd has the same value as
66          oldfd, then dup2() does nothing, and returns newfd.
67
68   dup3()
69       dup3() is the same as dup2(), except that:
70
71       •  The caller can force the close-on-exec flag to be set  for  the  new
72          file  descriptor by specifying O_CLOEXEC in flags.  See the descrip‐
73          tion of the same flag in open(2) for reasons why this may be useful.
74
75       •  If oldfd equals newfd, then dup3() fails with the error EINVAL.
76

RETURN VALUE

78       On success, these system calls return the new file descriptor.  On  er‐
79       ror, -1 is returned, and errno is set to indicate the error.
80

ERRORS

82       EBADF  oldfd isn't an open file descriptor.
83
84       EBADF  newfd  is out of the allowed range for file descriptors (see the
85              discussion of RLIMIT_NOFILE in getrlimit(2)).
86
87       EBUSY  (Linux only) This may be returned by dup2() or dup3()  during  a
88              race condition with open(2) and dup().
89
90       EINTR  The  dup2() or dup3() call was interrupted by a signal; see sig‐
91              nal(7).
92
93       EINVAL (dup3()) flags contain an invalid value.
94
95       EINVAL (dup3()) oldfd was equal to newfd.
96
97       EMFILE The per-process limit on the number of open file descriptors has
98              been  reached  (see  the  discussion  of  RLIMIT_NOFILE in getr‐
99              limit(2)).
100

STANDARDS

102       dup()
103       dup2() POSIX.1-2008.
104
105       dup3() Linux.
106

HISTORY

108       dup()
109       dup2() POSIX.1-2001, SVr4, 4.3BSD.
110
111       dup3() Linux 2.6.27, glibc 2.9.
112

NOTES

114       The error returned by dup2() is different from  that  returned  by  fc‐
115       ntl(...,  F_DUPFD,  ...)  when newfd is out of range.  On some systems,
116       dup2() also sometimes returns EINVAL like F_DUPFD.
117
118       If newfd was open, any errors that would have been reported at close(2)
119       time  are lost.  If this is of concern, then—unless the program is sin‐
120       gle-threaded and does not allocate file descriptors in signal handlers—
121       the  correct  approach is not to close newfd before calling dup2(), be‐
122       cause of the race condition described above.  Instead,  code  something
123       like the following could be used:
124
125           /* Obtain a duplicate of 'newfd' that can subsequently
126              be used to check for close() errors; an EBADF error
127              means that 'newfd' was not open. */
128
129           tmpfd = dup(newfd);
130           if (tmpfd == -1 && errno != EBADF) {
131               /* Handle unexpected dup() error. */
132           }
133
134           /* Atomically duplicate 'oldfd' on 'newfd'. */
135
136           if (dup2(oldfd, newfd) == -1) {
137               /* Handle dup2() error. */
138           }
139
140           /* Now check for close() errors on the file originally
141              referred to by 'newfd'. */
142
143           if (tmpfd != -1) {
144               if (close(tmpfd) == -1) {
145                   /* Handle errors from close. */
146               }
147           }
148