dup(2)

✨ dup(2) ✨

okay so bestie, this is the System Calls Manual page for dup(2) and honestly? it's giving "how to clone a file descriptor" and it ate 💅 no cap.

---

🌸 NAME 🌸

dup, dup2, dup3 — duplicate a file descriptor

🎀 LIBRARY 🎀

Standard C library (libc, -lc) 🫶

✨ SYNOPSIS ✨

#include <unistd.h>

int dup(int oldfd);
int dup2(int oldfd, int newfd);

#define _GNU_SOURCE             /* See feature_test_macros(7) */
#include <fcntl.h>              /* Definition of O_* constants */
#include <unistd.h>

int dup3(int oldfd, int newfd, int flags);

the _GNU_SOURCE thing is literally the vibe check you gotta pass before dup3 will even look at you, iykyk 💀

---

💖 DESCRIPTION 💖

so the dup() system call is the main character here: it allocates a brand-new file descriptor that refers to the same open file description as the descriptor oldfd. (and if you're sitting there like "bestie what even is an open file description", go read open(2), it explains it, trust 🌈.)

the new file descriptor number is guaranteed to be the lowest-numbered file descriptor that was unused in the calling process. like, it literally goes looking for the first free spot. so efficient. so iconic. 👑

after a successful return, the old and new file descriptors may be used interchangeably. and because the two file descriptors refer to the same open file description, they share the file offset and the file status flags. for example: if the file offset is modified by using lseek(2) on one of the file descriptors, the offset is also changed for the other one. they're linked. soulmates, basically 🫶💞

BUT — and this is the plot twist — the two file descriptors do not share file descriptor flags. that's the close-on-exec flag. the close-on-exec flag (FD_CLOEXEC; see fcntl(2)) for the duplicate descriptor is off. so it starts fresh, no baggage 💅

🦋 dup2() 🦋

the dup2() system call does the same task as dup(), but instead of grabbing the lowest-numbered unused file descriptor, it uses the file descriptor number you specify in newfd. in other words: newfd gets adjusted so that it now refers to the same open file description as oldfd. you're not getting a hand-me-down slot, you're claiming a specific one ✨

if the file descriptor newfd was previously open, it gets closed before being reused — and that close is performed silently, which means any errors during the close are not reported by dup2(). it just... doesn't tell you. 🙈

the steps of closing and reusing the file descriptor newfd are performed atomically. this is important, bestie, because if you tried to build this yourself with close(2) and dup() you'd be living in race condition city, where newfd might get reused between those two steps. that reuse could happen because the main program got interrupted by a signal handler that allocates a file descriptor, or because a parallel thread allocates a file descriptor. chaos. not the cute kind either 💀

note the following points:

• if oldfd is not a valid file descriptor, then the call fails, and newfd is not closed. (protective king energy, honestly.)
• if oldfd is a valid file descriptor, and newfd has the same value as oldfd, then dup2() does nothing, and returns newfd. it simply refuses to do the thing. periodt.

🌈 dup3() 🌈

dup3() is the same as dup2(), except that:

• the caller can force the close-on-exec flag to be set for the new file descriptor by specifying O_CLOEXEC in flags. (see the description of the same flag in open(2) for reasons why this may be useful 👀)
• if oldfd equals newfd, then dup3() fails with the error EINVAL. no silent no-op for you, sorry babe.

---

👑 RETURN VALUE 👑

on success, these system calls return the new file descriptor. on error, -1 is returned, and errno is set to indicate the error. classic. 🤷‍♀️

💀 ERRORS 💀

• EBADF — oldfd isn't an open file descriptor.
• EBADF — newfd is out of the allowed range for file descriptors (see the discussion of RLIMIT_NOFILE in getrlimit(2)).
• EBUSY — (Linux only) this may be returned by dup2() or dup3() during a race condition with open(2) and dup().
• EINTR — the dup2() or dup3() call was interrupted by a signal; see signal(7) 😭
• EINVAL — (dup3()) flags contain an invalid value.
• EINVAL — (dup3()) oldfd was equal to newfd.
• EMFILE — the per-process limit on the number of open file descriptors has been reached (see the discussion of RLIMIT_NOFILE in getrlimit(2)). you've officially collected too many. hoarder behavior. respect. 🍓
• ENOMEM — insufficient kernel memory was available.

---

⭐ STANDARDS ⭐

POSIX.1-2024. she standards-certified, your honor ✨

🧚 HISTORY 🧚

dup() , dup2() — 4.3BSD, SVr4, POSIX.1-1988.

dup3() — POSIX.1-2024. Linux 2.6.27, glibc 2.9.

look at the lore. the ancestry. we love a documented glow-up 🌸

---

💜 NOTES 💜

the error returned by dup2() is different from the error returned by fcntl(..., F_DUPFD, ...) when newfd is out of range. on some systems, dup2() also sometimes returns EINVAL like F_DUPFD. it's messy, it's inconsistent, we simply accept it 🫠

if newfd was open, any errors that would have been reported at close(2) time are lost. and if this is of concern to you — then, unless the program is single-threaded and does not allocate file descriptors in signal handlers — the correct approach is not to close newfd before calling dup2(), because of the race condition described above. instead, code something like the following could be used:

/* Obtain a duplicate of 'newfd' that can subsequently
   be used to check for close() errors; an EBADF error
   means that 'newfd' was not open.  */

tmpfd = dup(newfd);
if (tmpfd == -1 && errno != EBADF) {
    /* Handle unexpected dup() error.  */
}

/* Atomically duplicate 'oldfd' on 'newfd'.  */

if (dup2(oldfd, newfd) == -1) {
    /* Handle dup2() error.  */
}

/* Now check for close() errors on the file originally
   referred to by 'newfd'.  */

if (tmpfd != -1) {
    if (close(tmpfd) == -1) {
        /* Handle errors from close.  */
    }
}

bestie this is the correct pattern, do not fight it, do not "optimize" it, just copy it 💅

---

🔥 SEE ALSO 🔥

close(2), fcntl(2), open(2), pidfd_getfd(2)

---

and that's the whole page, no crumbs left 🎀 this certified banger of a man page comes to you from Linux man-pages 6.20, dated 2026-02-08 — go forth and duplicate responsibly, my love 💖✨