.\" This man page is Copyright (C) 1998 Pawel Krawczyk. .\" .\" %%%LICENSE_START(VERBATIM_ONE_PARA) .\" Permission is granted to distribute possibly modified copies .\" of this page provided the header is included verbatim, .\" and in case of nontrivial modification author and date .\" of the modification is added to the header. .\" %%%LICENSE_END .\" .\" $Id: sendfile.2,v 1.5 1999/05/18 11:54:11 freitag Exp $ .\" 2000-11-19 bert hubert : in_fd cannot be socket .\" .\" 2004-12-17, mtk .\" updated description of in_fd and out_fd for 2.6 .\" Various wording and formatting changes .\" .\" 2005-03-31 Martin Pool mmap() improvements .\" .TH SENDFILE 2 2017-09-15 "Linux" "Linux Programmer's Manual" .SH NAME sendfile \- transfer data between file descriptors .SH SYNOPSIS .B #include .PP .BI "ssize_t sendfile(int" " out_fd" ", int" " in_fd" ", off_t *" \ offset ", size_t" " count" ); .\" The below is too ugly. Comments about glibc versions belong .\" in the notes, not in the header. .\" .\" .B #include .\" .br .\" .B #if (__GLIBC__==2 && __GLIBC_MINOR__>=1) || __GLIBC__>2 .\" .br .\" .B #include .\" .br .\" #else .\" .br .\" .B #include .\" .br .\" .B /* No system prototype before glibc 2.1. */ .\" .br .\" .BI "ssize_t sendfile(int" " out_fd" ", int" " in_fd" ", off_t *" \ .\" offset ", size_t" " count" ) .\" .br .\" .B #endif .\" .SH DESCRIPTION .BR sendfile () copies data between one file descriptor and another. Because this copying is done within the kernel, .BR sendfile () is more efficient than the combination of .BR read (2) and .BR write (2), which would require transferring data to and from user space. .PP .I in_fd should be a file descriptor opened for reading and .I out_fd should be a descriptor opened for writing. .PP If .I offset is not NULL, then it points to a variable holding the file offset from which .BR sendfile () will start reading data from .IR in_fd . When .BR sendfile () returns, this variable will be set to the offset of the byte following the last byte that was read. If .I offset is not NULL, then .BR sendfile () does not modify the file offset of .IR in_fd ; otherwise the file offset is adjusted to reflect the number of bytes read from .IR in_fd . .PP If .I offset is NULL, then data will be read from .IR in_fd starting at the file offset, and the file offset will be updated by the call. .PP .I count is the number of bytes to copy between the file descriptors. .PP The .IR in_fd argument must correspond to a file which supports .BR mmap (2)-like operations (i.e., it cannot be a socket). .PP In Linux kernels before 2.6.33, .I out_fd must refer to a socket. Since Linux 2.6.33 it can be any file. If it is a regular file, then .BR sendfile () changes the file offset appropriately. .SH RETURN VALUE If the transfer was successful, the number of bytes written to .I out_fd is returned. Note that a successful call to .BR sendfile () may write fewer bytes than requested; the caller should be prepared to retry the call if there were unsent bytes. See also NOTES. .PP On error, \-1 is returned, and .I errno is set appropriately. .SH ERRORS .TP .B EAGAIN Nonblocking I/O has been selected using .B O_NONBLOCK and the write would block. .TP .B EBADF The input file was not opened for reading or the output file was not opened for writing. .TP .B EFAULT Bad address. .TP .B EINVAL Descriptor is not valid or locked, or an .BR mmap (2)-like operation is not available for .IR in_fd , or .I count is negative. .TP .B EINVAL .I out_fd has the .B O_APPEND flag set. This is not currently supported by .BR sendfile (). .TP .B EIO Unspecified error while reading from .IR in_fd . .TP .B ENOMEM Insufficient memory to read from .IR in_fd . .TP .B EOVERFLOW .I count is too large, the operation would result in exceeding the maximum size of either the input file or the output file. .TP .B ESPIPE .I offset is not NULL but the input file is not .BR seek (2)-able. .SH VERSIONS .BR sendfile () first appeared in Linux 2.2. The include file .I is present since glibc 2.1. .SH CONFORMING TO Not specified in POSIX.1-2001, nor in other standards. .PP Other UNIX systems implement .BR sendfile () with different semantics and prototypes. It should not be used in portable programs. .SH NOTES .BR sendfile () will transfer at most 0x7ffff000 (2,147,479,552) bytes, returning the number of bytes actually transferred. .\" commit e28cc71572da38a5a12c1cfe4d7032017adccf69 (This is true on both 32-bit and 64-bit systems.) .PP If you plan to use .BR sendfile () for sending files to a TCP socket, but need to send some header data in front of the file contents, you will find it useful to employ the .B TCP_CORK option, described in .BR tcp (7), to minimize the number of packets and to tune performance. .PP In Linux 2.4 and earlier, .I out_fd could also refer to a regular file; this possibility went away in the Linux 2.6.x kernel series, but was restored in Linux 2.6.33. .PP The original Linux .BR sendfile () system call was not designed to handle large file offsets. Consequently, Linux 2.4 added .BR sendfile64 (), with a wider type for the .I offset argument. The glibc .BR sendfile () wrapper function transparently deals with the kernel differences. .PP Applications may wish to fall back to .BR read (2)/ write (2) in the case where .BR sendfile () fails with .B EINVAL or .BR ENOSYS . .PP If .I out_fd refers to a socket or pipe with zero-copy support, callers must ensure the transferred portions of the file referred to by .I in_fd remain unmodified until the reader on the other end of .I out_fd has consumed the transferred data. .PP The Linux-specific .BR splice (2) call supports transferring data between arbitrary file descriptors provided one (or both) of them is a pipe. .SH SEE ALSO .BR copy_file_range (2), .BR mmap (2), .BR open (2), .BR socket (2), .BR splice (2) .SH COLOPHON This page is part of release 5.04 of the Linux .I man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at \%https://www.kernel.org/doc/man\-pages/.