1SYNC_FILE_RANGE(2) Linux Programmer's Manual SYNC_FILE_RANGE(2)
2
6 sync_file_range - sync a file segment with disk
7
9 #define _GNU_SOURCE /* See feature_test_macros(7) */
10 #include <fcntl.h>
11 int sync_file_range(int fd, off64_t offset, off64_t nbytes,
13 unsigned int flags);
14
16 sync_file_range() permits fine control when synchronizing the open file
17 referred to by the file descriptor fd with disk.
18 offset is the starting byte of the file range to be synchronized.
20 nbytes specifies the length of the range to be synchronized, in bytes;
21 if nbytes is zero, then all bytes from offset through to the end of
22 file are synchronized. Synchronization is in units of the system page
23 size: offset is rounded down to a page boundary; (offset+nbytes-1) is
24 rounded up to a page boundary.
25 The flags bit-mask argument can include any of the following values:
27 SYNC_FILE_RANGE_WAIT_BEFORE
29 Wait upon write-out of all pages in the specified range that
30 have already been submitted to the device driver for write-out
31 before performing any write.
32 SYNC_FILE_RANGE_WRITE
34 Initiate write-out of all dirty pages in the specified range
35 which are not presently submitted write-out. Note that even
36 this may block if you attempt to write more than request queue
37 size.
38 SYNC_FILE_RANGE_WAIT_AFTER
40 Wait upon write-out of all pages in the range after performing
41 any write.
42 Specifying flags as 0 is permitted, as a no-op.
44 Warning
46 This system call is extremely dangerous and should not be used in por‐
47 table programs. None of these operations writes out the file's meta‐
48 data. Therefore, unless the application is strictly performing over‐
49 writes of already-instantiated disk blocks, there are no guarantees
50 that the data will be available after a crash. There is no user inter‐
51 face to know if a write is purely an overwrite. On filesystems using
52 copy-on-write semantics (e.g., btrfs) an overwrite of existing allo‐
53 cated blocks is impossible. When writing into preallocated space, many
54 filesystems also require calls into the block allocator, which this
55 system call does not sync out to disk. This system call does not flush
56 disk write caches and thus does not provide any data integrity on sys‐
57 tems with volatile disk write caches.
58 Some details
60 SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect
61 any I/O errors or ENOSPC conditions and will return these to the
62 caller.
63 Useful combinations of the flags bits are:
65 SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE
67 Ensures that all pages in the specified range which were dirty
68 when sync_file_range() was called are placed under write-out.
69 This is a start-write-for-data-integrity operation.
70 SYNC_FILE_RANGE_WRITE
72 Start write-out of all dirty pages in the specified range which
73 are not presently under write-out. This is an asynchronous
74 flush-to-disk operation. This is not suitable for data integ‐
75 rity operations.
76 SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER)
78 Wait for completion of write-out of all pages in the specified
79 range. This can be used after an earlier
80 SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE operation to
81 wait for completion of that operation, and obtain its result.
82 SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE |
84 SYNC_FILE_RANGE_WAIT_AFTER
85 This is a write-for-data-integrity operation that will ensure
86 that all pages in the specified range which were dirty when
87 sync_file_range() was called are committed to disk.
88
90 On success, sync_file_range() returns 0; on failure -1 is returned and
91 errno is set to indicate the error.
92
94 EBADF fd is not a valid file descriptor.
95 EINVAL flags specifies an invalid bit; or offset or nbytes is invalid.
97 EIO I/O error.
99 ENOMEM Out of memory.
101 ENOSPC Out of disk space.
103 ESPIPE fd refers to something other than a regular file, a block de‐
105 vice, or a directory.
106
108 sync_file_range() appeared on Linux in kernel 2.6.17.
109
111 This system call is Linux-specific, and should be avoided in portable
112 programs.
113
115 sync_file_range2()
116 Some architectures (e.g., PowerPC, ARM) need 64-bit arguments to be
117 aligned in a suitable pair of registers. On such architectures, the
118 call signature of sync_file_range() shown in the SYNOPSIS would force a
119 register to be wasted as padding between the fd and offset arguments.
120 (See syscall(2) for details.) Therefore, these architectures define a
121 different system call that orders the arguments suitably:
122 int sync_file_range2(int fd, unsigned int flags,
124 off64_t offset, off64_t nbytes);
125 The behavior of this system call is otherwise exactly the same as
127 sync_file_range().
128 A system call with this signature first appeared on the ARM architec‐
130 ture in Linux 2.6.20, with the name arm_sync_file_range(). It was re‐
131 named in Linux 2.6.22, when the analogous system call was added for
132 PowerPC. On architectures where glibc support is provided, glibc
133 transparently wraps sync_file_range2() under the name
134 sync_file_range().
135
137 fdatasync(2), fsync(2), msync(2), sync(2)
138
140 This page is part of release 5.13 of the Linux man-pages project. A
141 description of the project, information about reporting bugs, and the
142 latest version of this page, can be found at
143 https://www.kernel.org/doc/man-pages/.
144Linux 2021-03-22 SYNC_FILE_RANGE(2)