dup3(2) - f31

1DUP(2)                     Linux Programmer's Manual                    DUP(2)
2
3
4

NAME

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

SYNOPSIS

9       #include <unistd.h>
10
11       int dup(int oldfd);
12       int dup2(int oldfd, int newfd);
13
14       #define _GNU_SOURCE             /* See feature_test_macros(7) */
15       #include <fcntl.h>              /* Obtain O_* constant definitions */
16       #include <unistd.h>
17
18       int dup3(int oldfd, int newfd, int flags);
19

DESCRIPTION

21       The  dup()  system  call  creates  a copy of the file descriptor oldfd,
22       using the lowest-numbered unused file descriptor for the  new  descrip‐
23       tor.
24
25       After a successful return, the old and new file descriptors may be used
26       interchangeably.  They refer to the same  open  file  description  (see
27       open(2)) and thus share file offset and file status flags; for example,
28       if the file offset is modified by using lseek(2) on  one  of  the  file
29       descriptors, the offset is also changed for the other.
30
31       The two file descriptors do not share file descriptor flags (the close-
32       on-exec flag).  The close-on-exec flag (FD_CLOEXEC; see  fcntl(2))  for
33       the duplicate descriptor is off.
34
35   dup2()
36       The  dup2() system call performs the same task as dup(), but instead of
37       using the lowest-numbered unused file  descriptor,  it  uses  the  file
38       descriptor number specified in newfd.  If the file descriptor newfd was
39       previously open, it is silently closed before being reused.
40
41       The steps of closing and reusing the file  descriptor  newfd  are  per‐
42       formed  atomically.   This  is  important,  because trying to implement
43       equivalent functionality using close(2) and dup() would be  subject  to
44       race  conditions,  whereby newfd might be reused between the two steps.
45       Such reuse could happen because the main program is  interrupted  by  a
46       signal  handler that allocates a file descriptor, or because a parallel
47       thread allocates a file descriptor.
48
49       Note the following points:
50
51       *  If oldfd is not a valid file descriptor, then the  call  fails,  and
52          newfd is not closed.
53
54       *  If oldfd is a valid file descriptor, and newfd has the same value as
55          oldfd, then dup2() does nothing, and returns newfd.
56
57   dup3()
58       dup3() is the same as dup2(), except that:
59
60       *  The caller can force the close-on-exec flag to be set  for  the  new
61          file  descriptor by specifying O_CLOEXEC in flags.  See the descrip‐
62          tion of the same flag in open(2) for reasons why this may be useful.
63
64       *  If oldfd equals newfd, then dup3() fails with the error EINVAL.
65

RETURN VALUE

67       On success, these system calls return  the  new  file  descriptor.   On
68       error, -1 is returned, and errno is set appropriately.
69

ERRORS

71       EBADF  oldfd isn't an open file descriptor.
72
73       EBADF  newfd  is out of the allowed range for file descriptors (see the
74              discussion of RLIMIT_NOFILE in getrlimit(2)).
75
76       EBUSY  (Linux only) This may be returned by dup2() or dup3()  during  a
77              race condition with open(2) and dup().
78
79       EINTR  The  dup2() or dup3() call was interrupted by a signal; see sig‐
80              nal(7).
81
82       EINVAL (dup3()) flags contain an invalid value.
83
84       EINVAL (dup3()) oldfd was equal to newfd.
85
86       EMFILE The per-process limit on the number of open file descriptors has
87              been  reached  (see  the  discussion  of  RLIMIT_NOFILE in getr‐
88              limit(2)).
89

VERSIONS

91       dup3() was added to Linux in version 2.6.27; glibc support is available
92       starting with version 2.9.
93

CONFORMING TO

95       dup(), dup2(): POSIX.1-2001, POSIX.1-2008, SVr4, 4.3BSD.
96
97       dup3() is Linux-specific.
98

NOTES

100       The  error  returned  by  dup2()  is  different  from  that returned by
101       fcntl(..., F_DUPFD, ...)  when newfd is out of range.  On some systems,
102       dup2() also sometimes returns EINVAL like F_DUPFD.
103
104       If newfd was open, any errors that would have been reported at close(2)
105       time are lost.  If this is of concern, then—unless the program is  sin‐
106       gle-threaded and does not allocate file descriptors in signal handlers—
107       the correct approach is not  to  close  newfd  before  calling  dup2(),
108       because of the race condition described above.  Instead, code something
109       like the following could be used:
110
111           /* Obtain a duplicate of 'newfd' that can subsequently
112              be used to check for close() errors; an EBADF error
113              means that 'newfd' was not open. */
114
115           tmpfd = dup(newfd);
116           if (tmpfd == -1 && errno != EBADF) {
117               /* Handle unexpected dup() error */
118           }
119
120           /* Atomically duplicate 'oldfd' on 'newfd' */
121
122           if (dup2(oldfd, newfd) == -1) {
123               /* Handle dup2() error */
124           }
125
126           /* Now check for close() errors on the file originally
127              referred to by 'newfd' */
128
129           if (tmpfd != -1) {
130               if (close(tmpfd) == -1) {
131                   /* Handle errors from close */
132               }
133           }
134

COLOPHON

139       This page is part of release 5.02 of the Linux  man-pages  project.   A
140       description  of  the project, information about reporting bugs, and the
141       latest    version    of    this    page,    can     be     found     at
142       https://www.kernel.org/doc/man-pages/.
143
144
145
146Linux                             2017-09-15                            DUP(2)