1io_uring_sqe_set_flags(3) liburing Manual io_uring_sqe_set_flags(3)
2
6 io_uring_sqe_set_flags - set flags for submission queue entry
7
9 #include <liburing.h>
10 void io_uring_sqe_set_flags(struct io_uring_sqe *sqe,
12 unsigned flags);
13
15 The io_uring_sqe_set_flags(3) function allows the caller to change the
16 behavior of the submission queue entry by specifying flags. It enables
17 the flags belonging to the sqe submission queue entry param.
18 flags is a bit mask of 0 or more of the following values ORed together:
20 IOSQE_FIXED_FILE
22 The file descriptor in the SQE refers to the index of a previ‐
23 ously registered file or direct file descriptor, not a normal
24 file descriptor.
25 IOSQE_ASYNC
27 Normal operation for io_uring is to try and issue an sqe as non-
28 blocking first, and if that fails, execute it in an async man‐
29 ner. To support more efficient overlapped operation of requests
30 that the application knows/assumes will always (or most of the
31 time) block, the application can ask for an sqe to be issued
32 async from the start. Note that this flag immediately causes the
33 SQE to be offloaded to an async helper thread with no initial
34 non-blocking attempt. This may be less efficient and should not
35 be used liberally or without understanding the performance and
36 efficiency tradeoffs.
37 IOSQE_IO_LINK
39 When this flag is specified, the SQE forms a link with the next
40 SQE in the submission ring. That next SQE will not be started
41 before the previous request completes. This, in effect, forms a
42 chain of SQEs, which can be arbitrarily long. The tail of the
43 chain is denoted by the first SQE that does not have this flag
44 set. Chains are not supported across submission boundaries. Even
45 if the last SQE in a submission has this flag set, it will still
46 terminate the current chain. This flag has no effect on previous
47 SQE submissions, nor does it impact SQEs that are outside of the
48 chain tail. This means that multiple chains can be executing in
49 parallel, or chains and individual SQEs. Only members inside the
50 chain are serialized. A chain of SQEs will be broken if any re‐
51 quest in that chain ends in error.
52 IOSQE_IO_HARDLINK
54 Like IOSQE_IO_LINK , except the links aren't severed if an error
55 or unexpected result occurs.
56 IOSQE_IO_DRAIN
58 When this flag is specified, the SQE will not be started before
59 previously submitted SQEs have completed, and new SQEs will not
60 be started before this one completes.
61 IOSQE_CQE_SKIP_SUCCESS
63 Request that no CQE be generated for this request, if it com‐
64 pletes successfully. This can be useful in cases where the ap‐
65 plication doesn't need to know when a specific request com‐
66 pleted, if it completed successfully.
67 IOSQE_BUFFER_SELECT
69 If set, and if the request types supports it, select an IO buf‐
70 fer from the indicated buffer group. This can be used with re‐
71 quests that read or receive data from a file or socket, where
72 buffer selection is deferred until the kernel is ready to trans‐
73 fer data, instead of when the IO is originally submitted. The
74 application must also set the buf_group field in the SQE, indi‐
75 cating which previously registered buffer group to select a buf‐
76 fer from.
77
80 None
81
83 io_uring_submit(3), io_uring_register(3) io_uring_register_buffers(3)
84 io_uring_register_buf_ring(3)
85liburing-2.1 January 25, 2022 io_uring_sqe_set_flags(3)