1io_uring_prep_openat(3) liburing Manual io_uring_prep_openat(3)
2
6 io_uring_prep_openat - prepare an openat request
7
9 #include <sys/types.h>
10 #include <sys/stat.h>
11 #include <fcntl.h>
12 #include <liburing.h>
13 void io_uring_prep_openat(struct io_uring_sqe *sqe,
15 int dfd,
16 const char *path,
17 int flags,
18 mode_t mode);
19 void io_uring_prep_openat_direct(struct io_uring_sqe *sqe,
21 int dfd,
22 const char *path,
23 int flags,
24 mode_t mode,
25 unsigned file_index);
26
28 The io_uring_prep_openat(3) function prepares an openat request. The
29 submission queue entry sqe is setup to use the directory file descrip‐
30 tor dfd to start opening a file described by path and using the open
31 flags in flags and using the file mode bits specified in mode.
32 For a direct descriptor open request, the offset is specified by the
34 file_index argument. Direct descriptors are io_uring private file de‐
35 scriptors. They avoid some of the overhead associated with thread
36 shared file tables, and can be used in any io_uring request that takes
37 a file descriptor. To do so, IOSQE_FIXED_FILE must be set in the SQE
38 flags member, and the SQE fd field should use the direct descriptor
39 value rather than the regular file descriptor. Direct descriptors are
40 managed like registered files.
41 If the direct variant is used, the application must first have regis‐
43 tered a file table using io_uring_register_files(3) of the appropriate
44 size. Once registered, a direct accept request may use any entry in
45 that table, as long as it is within the size of the registered table.
46 If a specified entry already contains a file, the file will first be
47 removed from the table and closed. It's consistent with the behavior of
48 updating an existing file with io_uring_register_files_update(3). Note
49 that old kernels don't check the SQE file_index field, which is not a
50 problem for liburing helpers, but users of the raw io_uring interface
51 need to zero SQEs to avoid unexpected behavior.
52 If IORING_FILE_INDEX_ALLOC is used as the file_index for a direct open,
54 then io_uring will allocate a free direct descriptor in the existing
55 table. The allocated descriptor is returned in the CQE res field just
56 like it would be for a non-direct open request. If no more entries are
57 available in the direct descriptor table, -ENFILE is returned instead.
58 These functions prepare an async openat(2) request. See that man page
60 for details.
61
64 None
65
67 The CQE res field will contain the result of the operation. See the re‐
68 lated man page for details on possible values. Note that where synchro‐
69 nous system calls will return -1 on failure and set errno to the actual
70 error value, io_uring never uses errno. Instead it returns the negated
71 errno directly in the CQE res field.
72
74 As with any request that passes in data in a struct, that data must re‐
75 main valid until the request has been successfully submitted. It need
76 not remain valid until completion. Once a request has been submitted,
77 the in-kernel state is stable. Very early kernels (5.4 and earlier) re‐
78 quired state to be stable until the completion occurred. Applications
79 can test for this behavior by inspecting the IORING_FEAT_SUBMIT_STABLE
80 flag passed back from io_uring_queue_init_params(3).
81
83 io_uring_get_sqe(3), io_uring_submit(3), io_uring_register(2), ope‐
84 nat(2)
85liburing-2.2 March 13, 2022 io_uring_prep_openat(3)