1IO_SUBMIT(2) Linux Programmer's Manual IO_SUBMIT(2)
2
3
4
6 io_submit - submit asynchronous I/O blocks for processing
7
9 #include <linux/aio_abi.h> /* Defines needed types */
10
11 int io_submit(aio_context_t ctx_id, long nr, struct iocb **iocbpp);
12
13 Note: There is no glibc wrapper for this system call; see NOTES.
14
16 The io_submit() system call queues nr I/O request blocks for processing
17 in the AIO context ctx_id. The iocbpp argument should be an array of
18 nr AIO control blocks, which will be submitted to context ctx_id.
19
20 The iocb (I/O control block) structure defined in linux/aio_abi.h
21 defines the parameters that control the I/O operation.
22
23 #include <linux/aio_abi.h>
24
25 struct iocb {
26 __u64 aio_data;
27 __u32 PADDED(aio_key, aio_rw_flags);
28 __u16 aio_lio_opcode;
29 __s16 aio_reqprio;
30 __u32 aio_fildes;
31 __u64 aio_buf;
32 __u64 aio_nbytes;
33 __s64 aio_offset;
34 __u64 aio_reserved2;
35 __u32 aio_flags;
36 __u32 aio_resfd;
37 };
38
39 The fields of this structure are as follows:
40
41 aio_data
42 This is an internal field used by the kernel. Do not modify
43 this field after an io_submit(2) call.
44
45 aio_key
46 This is an internal field used by the kernel. Do not modify
47 this field after an io_submit(2) call.
48
49 aio_rw_flags
50 This defines the R/W flags passed with structure. The valid
51 values are:
52
53 RWF_APPEND (since Linux 4.16)
54 Append data to the end of the file. See the description
55 of the flag of the same name in pwritev2(2) as well as
56 the description of O_APPEND in open(2). The aio_offset
57 field is ignored. The file offset is not changed.
58
59 RWF_DSYNC (since Linux 4.7)
60 Write operation complete according to requirement of syn‐
61 chronized I/O data integrity. See the description of the
62 flag of the same name in pwritev2(2) as well the descrip‐
63 tion of O_DSYNC in open(2).
64
65 RWF_HIPRI (since Linux 4.6)
66 High priority request, poll if possible
67
68 RWF_NOWAIT (since Linux 4.14)
69 Don't wait if the I/O will block for operations such as
70 file block allocations, dirty page flush, mutex locks, or
71 a congested block device inside the kernel. If any of
72 these conditions are met, the control block is returned
73 immediately with a return value of -EAGAIN in the res
74 field of the io_event structure (see io_getevents(2)).
75
76 RWF_SYNC (since Linux 4.7)
77 Write operation complete according to requirement of syn‐
78 chronized I/O file integrity. See the description of the
79 flag of the same name in pwritev2(2) as well the descrip‐
80 tion of O_SYNC in open(2).
81
82 aio_lio_opcode
83 This defines the type of I/O to be performed by the iocb struc‐
84 ture. The valid values are defined by the enum defined in
85 linux/aio_abi.h:
86
87 enum {
88 IOCB_CMD_PREAD = 0,
89 IOCB_CMD_PWRITE = 1,
90 IOCB_CMD_FSYNC = 2,
91 IOCB_CMD_FDSYNC = 3,
92 IOCB_CMD_NOOP = 6,
93 IOCB_CMD_PREADV = 7,
94 IOCB_CMD_PWRITEV = 8,
95 };
96
97 aio_reqprio
98 This defines the requests priority.
99
100 aio_filedes
101 The file descriptor on which the I/O operation is to be per‐
102 formed.
103
104 aio_buf
105 This is the buffer used to transfer data for a read or write
106 operation.
107
108 aio_nbytes
109 This is the size of the buffer pointed to by aio_buf.
110
111 aio_offset
112 This is the file offset at which the I/O operation is to be per‐
113 formed.
114
115 aio_flags
116 This is the flag to be passed iocb structure. The only valid
117 value is IOCB_FLAG_RESFD, which indicates that the asynchronous
118 I/O control must signal the file descriptor mentioned in
119 aio_resfd upon completion.
120
121 aio_resfd
122 The file descriptor to signal in the event of asynchronous I/O
123 completion.
124
126 On success, io_submit() returns the number of iocbs submitted (which
127 may be less than nr, or 0 if nr is zero). For the failure return, see
128 NOTES.
129
131 EAGAIN Insufficient resources are available to queue any iocbs.
132
133 EBADF The file descriptor specified in the first iocb is invalid.
134
135 EFAULT One of the data structures points to invalid data.
136
137 EINVAL The AIO context specified by ctx_id is invalid. nr is less than
138 0. The iocb at *iocbpp[0] is not properly initialized, or the
139 operation specified is invalid for the file descriptor in the
140 iocb.
141
142 ENOSYS io_submit() is not implemented on this architecture.
143
145 The asynchronous I/O system calls first appeared in Linux 2.5.
146
148 io_submit() is Linux-specific and should not be used in programs that
149 are intended to be portable.
150
152 Glibc does not provide a wrapper function for this system call. You
153 could invoke it using syscall(2). But instead, you probably want to
154 use the io_submit() wrapper function provided by libaio.
155
156 Note that the libaio wrapper function uses a different type (io_con‐
157 text_t) for the ctx_id argument. Note also that the libaio wrapper
158 does not follow the usual C library conventions for indicating errors:
159 on error it returns a negated error number (the negative of one of the
160 values listed in ERRORS). If the system call is invoked via
161 syscall(2), then the return value follows the usual conventions for
162 indicating an error: -1, with errno set to a (positive) value that
163 indicates the error.
164
166 io_cancel(2), io_destroy(2), io_getevents(2), io_setup(2), aio(7)
167
169 This page is part of release 4.16 of the Linux man-pages project. A
170 description of the project, information about reporting bugs, and the
171 latest version of this page, can be found at
172 https://www.kernel.org/doc/man-pages/.
173
174
175
176Linux 2018-04-30 IO_SUBMIT(2)