io_uring_enter2

NAME

io_uring_enter - initiate and/or complete asynchronous I/O

SYNOPSIS

#include <liburing.h>

int io_uring_enter(unsigned int fd, unsigned int to_submit,
                   unsigned int min_complete, unsigned int flags,
                   sigset_t *sig);

int io_uring_enter2(unsigned int fd, unsigned int to_submit,
                    unsigned int min_complete, unsigned int flags,
                    sigset_t *sig, size_t sz);

io_uring_enter(2) is used to initiate and complete I/O using the shared submission and completion queues setup by a call to io_uring_setup(2). A single call can both submit new I/O and wait for completions of I/O initiated by this call or previous calls to io_uring_enter(2).

fd is the file descriptor returned by io_uring_setup(2). to_submit specifies the number of I/Os to submit from the submission queue. flags is a bitmask of the following values:

IORING_ENTER_GETEVENTS: If this flag is set, then the system call will wait for the specified number of events in min_complete before returning. This flag can be set along with to_submit to both submit and complete events in a single system call.

IORING_ENTER_SQ_WAKEUP: If the ring has been created with IORING_SETUP_SQPOLL, then this flag asks the kernel to wakeup the SQ kernel thread to submit IO.

IORING_ENTER_SQ_WAIT: If the ring has been created with IORING_SETUP_SQPOLL, then the application has no real insight into when the SQ kernel thread has consumed entries from the SQ ring. This can lead to a situation where the application can no longer get a free SQE entry to submit, without knowing when it one becomes available as the SQ kernel thread consumes them. If the system call is used with this flag set, then it will wait until at least one entry is free in the SQ ring.

IORING_ENTER_EXT_ARG

Since kernel 5.11, the system calls arguments have been modified to look like the following:

int io_uring_enter2(unsigned int fd, unsigned int to_submit,
                    unsigned int min_complete, unsigned int flags,
                    const void *arg, size_t argsz);

which behaves just like the original definition by default. However, if IORING_ENTER_EXT_ARG is set, then instead of a sigset_t being passed in, a pointer to a struct io_uring_getevents_arg is used instead and argsz must be set to the size of this structure. The definition is as follows:

struct io_uring_getevents_args {
        __u64   sigmask;
        __u32   sigmask_sz;
        __u32   pad;
        __u64   ts;
};

which allows passing in both a signal mask as well as pointer to a struct __kernel_timespec timeout value. If ts is set to a valid pointer, then this time value indicates the timeout for waiting on events. If an application is waiting on events and wishes to stop waiting after a specified amount of time, then this can be accomplished directly in version 5.11 and newer by using this feature.

IORING_ENTER_REGISTERED_RING: If the ring file descriptor has been registered through use of IORING_REGISTER_RING_FDS, then setting this flag will tell the kernel that the ring_fd passed in is the registered ring offset rather than a normal file descriptor.

If the io_uring instance was configured for polling, by specifying IORING_SETUP_IOPOLL in the call to io_uring_setup(2), then min_complete has a slightly different meaning. Passing a value of 0 instructs the kernel to return any events which are already complete, without blocking. If min_complete is a non-zero value, the kernel will still return immediately if any completion events are available. If no event completions are available, then the call will poll either until one or more completions become available, or until the process has exceeded its scheduler time slice.

Note that, for interrupt driven I/O (where IORING_SETUP_IOPOLL was not specified in the call to io_uring_setup(2)), an application may check the completion queue for event completions without entering the kernel at all.

When the system call returns that a certain amount of SQEs have been consumed and submitted, it's safe to reuse SQE entries in the ring. This is true even if the actual IO submission had to be punted to async context, which means that the SQE may in fact not have been submitted yet. If the kernel requires later use of a particular SQE entry, it will have made a private copy of it.

sig is a pointer to a signal mask (see sigprocmask(2)); if sig is not NULL, io_uring_enter(2) first replaces the current signal mask by the one pointed to by sig, then waits for events to become available in the completion queue, and then restores the original signal mask. The following io_uring_enter(2) call:

ret = io_uring_enter(fd, 0, 1, IORING_ENTER_GETEVENTS, &sig);

is equivalent to atomically executing the following calls:

pthread_sigmask(SIG_SETMASK, &sig, &orig);
ret = io_uring_enter(fd, 0, 1, IORING_ENTER_GETEVENTS, NULL);
pthread_sigmask(SIG_SETMASK, &orig, NULL);

See the description of pselect(2) for an explanation of why the sig parameter is necessary.

Submission queue entries are represented using the following data structure:

/*
 * IO submission data structure (Submission Queue Entry)
 */
struct io_uring_sqe {
    __u8    opcode;         /* type of operation for this sqe */
    __u8    flags;          /* IOSQE_ flags */
    __u16   ioprio;         /* ioprio for the request */
    __s32   fd;             /* file descriptor to do IO on */
    union {
        __u64   off;            /* offset into file */
        __u64   addr2;
    };
    union {
        __u64   addr;       /* pointer to buffer or iovecs */
        __u64   splice_off_in;
    }
    __u32   len;            /* buffer size or number of iovecs */
    union {
        __kernel_rwf_t  rw_flags;
        __u32    fsync_flags;
        __u16    poll_events;   /* compatibility */
        __u32    poll32_events; /* word-reversed for BE */
        __u32    sync_range_flags;
        __u32    msg_flags;
        __u32    timeout_flags;
        __u32    accept_flags;
        __u32    cancel_flags;
        __u32    open_flags;
        __u32    statx_flags;
        __u32    fadvise_advice;
        __u32    splice_flags;
        __u32    rename_flags;
        __u32    unlink_flags;
        __u32    hardlink_flags;
    };
    __u64    user_data;     /* data to be passed back at completion time */
    union {
    struct {
        /* index into fixed buffers, if used */
            union {
                /* index into fixed buffers, if used */
                __u16    buf_index;
                /* for grouped buffer selection */
                __u16    buf_group;
            }
        /* personality to use, if used */
        __u16    personality;
        union {
            __s32    splice_fd_in;
            __u32    file_index;
	};
    };
    __u64    __pad2[3];
    };
};

The opcode describes the operation to be performed. It can be one of:

IORING_OP_NOP: Do not perform any I/O. This is useful for testing the performance of the io_uring implementation itself.

IORING_OP_READV

IORING_OP_WRITEV: Vectored read and write operations, similar to preadv2(2) and pwritev2(2). If the file is not seekable, off must be set to zero or -1.

IORING_OP_READ_FIXED

IORING_OP_WRITE_FIXED: Read from or write to pre-mapped buffers. See io_uring_register(2) for details on how to setup a context for fixed reads and writes.

IORING_OP_FSYNC: File sync. See also fsync(2). Note that, while I/O is initiated in the order in which it appears in the submission queue, completions are unordered. For example, an application which places a write I/O followed by an fsync in the submission queue cannot expect the fsync to apply to the write. The two operations execute in parallel, so the fsync may complete before the write is issued to the storage. The same is also true for previously issued writes that have not completed prior to the fsync.

IORING_OP_POLL_ADD: Poll the fd specified in the submission queue entry for the events specified in the poll_events field. Unlike poll or epoll without EPOLLONESHOT, by default this interface always works in one shot mode. That is, once the poll operation is completed, it will have to be resubmitted. If IORING_POLL_ADD_MULTI is set in the SQE len field, then the poll will work in multi shot mode instead. That means it'll repatedly trigger when the requested event becomes true, and hence multiple CQEs can be generated from this single SQE. The CQE flags field will have IORING_CQE_F_MORE set on completion if the application should expect further CQE entries from the original request. If this flag isn't set on completion, then the poll request has been terminated and no further events will be generated. This mode is available since 5.13. If IORING_POLL_UPDATE_EVENTS is set in the SQE len field, then the request will update an existing poll request with the mask of events passed in with this request. The lookup is based on the user_data field of the original SQE submitted, and this values is passed in the addr field of the SQE. This mode is available since 5.13. If IORING_POLL_UPDATE_USER_DATA is set in the SQE len field, then the request will update the user_data of an existing poll request based on the value passed in the off field. This mode is available since 5.13. This command works like an async poll(2) and the completion event result is the returned mask of events. For the variants that update user_data or events , the completion result will be similar to IORING_OP_POLL_REMOVE.

IORING_OP_POLL_REMOVE: Remove an existing poll request. If found, the res field of the struct io_uring_cqe will contain 0. If not found, res will contain -ENOENT, or -EALREADY if the poll request was in the process of completing already.

IORING_OP_EPOLL_CTL: Add, remove or modify entries in the interest list of epoll(7). See epoll_ctl(2) for details of the system call. fd holds the file descriptor that represents the epoll instance, addr holds the file descriptor to add, remove or modify, len holds the operation (EPOLL_CTL_ADD, EPOLL_CTL_DEL, EPOLL_CTL_MOD) to perform and, off holds a pointer to the epoll_events structure. Available since 5.6.

IORING_OP_SYNC_FILE_RANGE: Issue the equivalent of a sync_file_range (2) on the file descriptor. The fd field is the file descriptor to sync, the off field holds the offset in bytes, the len field holds the length in bytes, and the sync_range_flags field holds the flags for the command. See also sync_file_range(2) for the general description of the related system call. Available since 5.2.

IORING_OP_SENDMSG: Issue the equivalent of a sendmsg(2) system call. fd must be set to the socket file descriptor, addr must contain a pointer to the msghdr structure, and msg_flags holds the flags associated with the system call. See also sendmsg(2) for the general description of the related system call. Available since 5.3. This command also supports the following modifiers in ioprio:

IORING_RECVSEND_POLL_FIRST If set, io_uring will assume the socket is currently full and attempting to send data will be unsuccessful. For this case, io_uring will arm internal poll and trigger a send of the data when there is enough space available. This initial send attempt can be wasteful for the case where the socket is expected to be full, setting this flag will bypass the initial send attempt and go straight to arming poll. If poll does indicate that data can be sent, the operation will proceed.

IORING_OP_RECVMSG: Works just like IORING_OP_SENDMSG, except for recvmsg(2) instead. See the description of IORING_OP_SENDMSG. Available since 5.3. This command also supports the following modifiers in ioprio:

IORING_RECVSEND_POLL_FIRST If set, io_uring will assume the socket is currently empty and attempting to receive data will be unsuccessful. For this case, io_uring will arm internal poll and trigger a receive of the data when the socket has data to be read. This initial receive attempt can be wasteful for the case where the socket is expected to be empty, setting this flag will bypass the initial receive attempt and go straight to arming poll. If poll does indicate that data is ready to be received, the operation will proceed.

IORING_OP_SEND: Issue the equivalent of a send(2) system call. fd must be set to the socket file descriptor, addr must contain a pointer to the buffer, len denotes the length of the buffer to send, and msg_flags holds the flags associated with the system call. See also send(2) for the general description of the related system call. Available since 5.6. This command also supports the following modifiers in ioprio:

IORING_OP_RECV: Works just like IORING_OP_SEND, except for recv(2) instead. See the description of IORING_OP_SEND. Available since 5.6. This command also supports the following modifiers in ioprio:

IORING_OP_TIMEOUT: This command will register a timeout operation. The addr field must contain a pointer to a struct timespec64 structure, len must contain 1 to signify one timespec64 structure, timeout_flags may contain IORING_TIMEOUT_ABS for an absolute timeout value, or 0 for a relative timeout. off may contain a completion event count. A timeout will trigger a wakeup event on the completion ring for anyone waiting for events. A timeout condition is met when either the specified timeout expires, or the specified number of events have completed. Either condition will trigger the event. If set to 0, completed events are not counted, which effectively acts like a timer. io_uring timeouts use the CLOCK_MONOTONIC clock source. The request will complete with -ETIME if the timeout got completed through expiration of the timer, or 0 if the timeout got completed through requests completing on their own. If the timeout was canceled before it expired, the request will complete with -ECANCELED. Available since 5.4. Since 5.15, this command also supports the following modifiers in timeout_flags:

IORING_TIMEOUT_BOOTTIME If set, then the clocksource used is CLOCK_BOOTTIME instead of CLOCK_MONOTONIC. This clocksource differs in that it includes time elapsed if the system was suspend while having a timeout request in-flight.

IORING_TIMEOUT_REALTIME If set, then the clocksource used is CLOCK_REALTIME instead of CLOCK_MONOTONIC.

IORING_OP_TIMEOUT_REMOVE: If timeout_flags are zero, then it attempts to remove an existing timeout operation. addr must contain the user_data field of the previously issued timeout operation. If the specified timeout request is found and canceled successfully, this request will terminate with a result value of 0 If the timeout request was found but expiration was already in progress, this request will terminate with a result value of -EBUSY If the timeout request wasn't found, the request will terminate with a result value of -ENOENT Available since 5.5. If timeout_flags contain IORING_TIMEOUT_UPDATE, instead of removing an existing operation, it updates it. addr and return values are same as before. addr2 field must contain a pointer to a struct timespec64 structure. timeout_flags may also contain IORING_TIMEOUT_ABS, in which case the value given is an absolute one, not a relative one. Available since 5.11.

IORING_OP_ACCEPT: Issue the equivalent of an accept4(2) system call. fd must be set to the socket file descriptor, addr must contain the pointer to the sockaddr structure, and addr2 must contain a pointer to the socklen_t addrlen field. Flags can be passed using the accept_flags field. See also accept4(2) for the general description of the related system call. Available since 5.5. If the file_index field is set to a positive number, the file won't be installed into the normal file table as usual but will be placed into the fixed file table at index file_index - 1. In this case, instead of returning a file descriptor, the result will contain either 0 on success or an error. If the index points to a valid empty slot, the installation is guaranteed to not fail. If there is already a file in the slot, it will be replaced, similar to IORING_OP_FILES_UPDATE. Please note that only io_uring has access to such files and no other syscall can use them. See IOSQE_FIXED_FILE and IORING_REGISTER_FILES. Available since 5.5.

IORING_OP_ASYNC_CANCEL: Attempt to cancel an already issued request. addr must contain the user_data field of the request that should be canceled. The cancelation request will complete with one of the following results codes. If found, the res field of the cqe will contain 0. If not found, res will contain -ENOENT. If found and attempted canceled, the res field will contain -EALREADY. In this case, the request may or may not terminate. In general, requests that are interruptible (like socket IO) will get canceled, while disk IO requests cannot be canceled if already started. Available since 5.5.

IORING_OP_LINK_TIMEOUT: This request must be linked with another request through IOSQE_IO_LINK which is described below. Unlike IORING_OP_TIMEOUT, IORING_OP_LINK_TIMEOUT acts on the linked request, not the completion queue. The format of the command is otherwise like IORING_OP_TIMEOUT, except there's no completion event count as it's tied to a specific request. If used, the timeout specified in the command will cancel the linked command, unless the linked command completes before the timeout. The timeout will complete with -ETIME if the timer expired and the linked request was attempted canceled, or -ECANCELED if the timer got canceled because of completion of the linked request. Like IORING_OP_TIMEOUT the clock source used is CLOCK_MONOTONIC Available since 5.5.

IORING_OP_CONNECT: Issue the equivalent of a connect(2) system call. fd must be set to the socket file descriptor, addr must contain the const pointer to the sockaddr structure, and off must contain the socklen_t addrlen field. See also connect(2) for the general description of the related system call. Available since 5.5.

IORING_OP_FALLOCATE: Issue the equivalent of a fallocate(2) system call. fd must be set to the file descriptor, len must contain the mode associated with the operation, off must contain the offset on which to operate, and addr must contain the length. See also fallocate(2) for the general description of the related system call. Available since 5.6.

IORING_OP_FADVISE: Issue the equivalent of a posix_fadvise(2) system call. fd must be set to the file descriptor, off must contain the offset on which to operate, len must contain the length, and fadvise_advice must contain the advice associated with the operation. See also posix_fadvise(2) for the general description of the related system call. Available since 5.6.

IORING_OP_MADVISE: Issue the equivalent of a madvise(2) system call. addr must contain the address to operate on, len must contain the length on which to operate, and fadvise_advice must contain the advice associated with the operation. See also madvise(2) for the general description of the related system call. Available since 5.6.

IORING_OP_OPENAT: Issue the equivalent of a openat(2) system call. fd is the dirfd argument, addr must contain a pointer to the *pathname argument, open_flags should contain any flags passed in, and len is access mode of the file. See also openat(2) for the general description of the related system call. Available since 5.6. If the file_index field is set to a positive number, the file won't be installed into the normal file table as usual but will be placed into the fixed file table at index file_index - 1. In this case, instead of returning a file descriptor, the result will contain either 0 on success or an error. If the index points to a valid empty slot, the installation is guaranteed to not fail. If there is already a file in the slot, it will be replaced, similar to IORING_OP_FILES_UPDATE. Please note that only io_uring has access to such files and no other syscall can use them. See IOSQE_FIXED_FILE and IORING_REGISTER_FILES. Available since 5.15.

IORING_OP_OPENAT2: Issue the equivalent of a openat2(2) system call. fd is the dirfd argument, addr must contain a pointer to the *pathname argument, len should contain the size of the open_how structure, and off should be set to the address of the open_how structure. See also openat2(2) for the general description of the related system call. Available since 5.6. If the file_index field is set to a positive number, the file won't be installed into the normal file table as usual but will be placed into the fixed file table at index file_index - 1. In this case, instead of returning a file descriptor, the result will contain either 0 on success or an error. If the index points to a valid empty slot, the installation is guaranteed to not fail. If there is already a file in the slot, it will be replaced, similar to IORING_OP_FILES_UPDATE. Please note that only io_uring has access to such files and no other syscall can use them. See IOSQE_FIXED_FILE and IORING_REGISTER_FILES. Available since 5.15.

IORING_OP_CLOSE: Issue the equivalent of a close(2) system call. fd is the file descriptor to be closed. See also close(2) for the general description of the related system call. Available since 5.6. If the file_index field is set to a positive number, this command can be used to close files that were direct opened through IORING_OP_OPENAT , IORING_OP_OPENAT2 , or IORING_OP_ACCEPT using the io_uring specific direct descriptors. Note that only one of the descriptor fields may be set. The direct close feature is available since the 5.15 kernel, where direct descriptors were introduced.

IORING_OP_STATX: Issue the equivalent of a statx(2) system call. fd is the dirfd argument, addr must contain a pointer to the *pathname string, statx_flags is the flags argument, len should be the mask argument, and off must contain a pointer to the statxbuf to be filled in. See also statx(2) for the general description of the related system call. Available since 5.6.

IORING_OP_READ

IORING_OP_WRITE: Issue the equivalent of a pread(2) or pwrite(2) system call. fd is the file descriptor to be operated on, addr contains the buffer in question, len contains the length of the IO operation, and offs contains the read or write offset. If fd does not refer to a seekable file, off must be set to zero or -1. If offs is set to -1 , the offset will use (and advance) the file position, like the read(2) and write(2) system calls. These are non-vectored versions of the IORING_OP_READV and IORING_OP_WRITEV opcodes. See also read(2) and write(2) for the general description of the related system call. Available since 5.6.

IORING_OP_SPLICE: Issue the equivalent of a splice(2) system call. splice_fd_in is the file descriptor to read from, splice_off_in is an offset to read from, fd is the file descriptor to write to, off is an offset from which to start writing to. A sentinel value of -1 is used to pass the equivalent of a NULL for the offsets to splice(2). len contains the number of bytes to copy. splice_flags contains a bit mask for the flag field associated with the system call. Please note that one of the file descriptors must refer to a pipe. See also splice(2) for the general description of the related system call. Available since 5.7.

IORING_OP_TEE: Issue the equivalent of a tee(2) system call. splice_fd_in is the file descriptor to read from, fd is the file descriptor to write to, len contains the number of bytes to copy, and splice_flags contains a bit mask for the flag field associated with the system call. Please note that both of the file descriptors must refer to a pipe. See also tee(2) for the general description of the related system call. Available since 5.8.

IORING_OP_FILES_UPDATE: This command is an alternative to using IORING_REGISTER_FILES_UPDATE which then works in an async fashion, like the rest of the io_uring commands. The arguments passed in are the same. addr must contain a pointer to the array of file descriptors, len must contain the length of the array, and off must contain the offset at which to operate. Note that the array of file descriptors pointed to in addr must remain valid until this operation has completed. Available since 5.6.

IORING_OP_PROVIDE_BUFFERS: This command allows an application to register a group of buffers to be used by commands that read/receive data. Using buffers in this manner can eliminate the need to separate the poll + read, which provides a convenient point in time to allocate a buffer for a given request. It's often infeasible to have as many buffers available as pending reads or receive. With this feature, the application can have its pool of buffers ready in the kernel, and when the file or socket is ready to read/receive data, a buffer can be selected for the operation. fd must contain the number of buffers to provide, addr must contain the starting address to add buffers from, len must contain the length of each buffer to add from the range, buf_group must contain the group ID of this range of buffers, and off must contain the starting buffer ID of this range of buffers. With that set, the kernel adds buffers starting with the memory address in addr, each with a length of len. Hence the application should provide len * fd worth of memory in addr. Buffers are grouped by the group ID, and each buffer within this group will be identical in size according to the above arguments. This allows the application to provide different groups of buffers, and this is often used to have differently sized buffers available depending on what the expectations are of the individual request. When submitting a request that should use a provided buffer, the IOSQE_BUFFER_SELECT flag must be set, and buf_group must be set to the desired buffer group ID where the buffer should be selected from. Available since 5.7.

IORING_OP_REMOVE_BUFFERS: Remove buffers previously registered with IORING_OP_PROVIDE_BUFFERS. fd must contain the number of buffers to remove, and buf_group must contain the buffer group ID from which to remove the buffers. Available since 5.7.

IORING_OP_SHUTDOWN: Issue the equivalent of a shutdown(2) system call. fd is the file descriptor to the socket being shutdown, and len must be set to the how argument. No no other fields should be set. Available since 5.11.

IORING_OP_RENAMEAT: Issue the equivalent of a renameat2(2) system call. fd should be set to the olddirfd, addr should be set to the oldpath, len should be set to the newdirfd, addr should be set to the oldpath, addr2 should be set to the newpath, and finally rename_flags should be set to the flags passed in to renameat2(2). Available since 5.11.

IORING_OP_UNLINKAT: Issue the equivalent of a unlinkat2(2) system call. fd should be set to the dirfd, addr should be set to the pathname, and unlink_flags should be set to the flags being passed in to unlinkat(2). Available since 5.11.

IORING_OP_MKDIRAT: Issue the equivalent of a mkdirat2(2) system call. fd should be set to the dirfd, addr should be set to the pathname, and len should be set to the mode being passed in to mkdirat(2). Available since 5.15.

IORING_OP_SYMLINKAT: Issue the equivalent of a symlinkat2(2) system call. fd should be set to the newdirfd, addr should be set to the target and addr2 should be set to the linkpath being passed in to symlinkat(2). Available since 5.15.

IORING_OP_LINKAT: Issue the equivalent of a linkat2(2) system call. fd should be set to the olddirfd, addr should be set to the oldpath, len should be set to the newdirfd, addr2 should be set to the newpath, and hardlink_flags should be set to the flags being passed in to linkat(2). Available since 5.15.

IORING_OP_MSG_RING: Send a message to an io_uring. fd must be set to a file descriptor of a ring that the application has access to, len can be set to any 32-bit value that the application wishes to pass on, and off should be set any 64-bit value that the application wishes to send. On the target ring, a CQE will be posted with the res field matching the len set, and a user_data field matching the off value being passed in. This request type can be used to either just wake or interrupt anyone waiting for completions on the target ring, or it can be used to pass messages via the two fields. Available since 5.18.

IORING_OP_SOCKET: Issue the equivalent of a socket(2) system call. fd must contain the communication domain, off must contain the communication type, len must contain the protocol, and rw_flags is currently unused and must be set to zero. See also socket(2) for the general description of the related system call. Available since 5.19. If the file_index field is set to a positive number, the file won't be installed into the normal file table as usual but will be placed into the fixed file table at index file_index - 1. In this case, instead of returning a file descriptor, the result will contain either 0 on success or an error. If the index points to a valid empty slot, the installation is guaranteed to not fail. If there is already a file in the slot, it will be replaced, similar to IORING_OP_FILES_UPDATE. Please note that only io_uring has access to such files and no other syscall can use them. See IOSQE_FIXED_FILE and IORING_REGISTER_FILES. Available since 5.19.

IORING_OP_SEND_ZC: Issue the zerocopy equivalent of a send(2) system call. Similar to IORING_OP_SEND, but tries to avoid making intermediate copies of data. Zerocopy execution is not guaranteed and may fall back to copying. The request may also fail with -EOPNOTSUPP , when a protocol doesn't support zerocopy, in which case users are recommended to use copying sends instead. The flags field of the first struct io_uring_cqe may likely contain IORING_CQE_F_MORE , which means that there will be a second completion event / notification for the request, with the user_data field set to the same value. The user must not modify the data buffer until the notification is posted. The first cqe follows the usual rules and so its res field will contain the number of bytes sent or a negative error code. The notification's res field will be set to zero and the flags field will contain IORING_CQE_F_NOTIF . The two step model is needed because the kernel may hold on to buffers for a long time, e.g. waiting for a TCP ACK, and having a separate cqe for request completions allows userspace to push more data without extra delays. Note, notifications are only responsible for controlling the lifetime of the buffers, and as such don't mean anything about whether the data has atually been sent out or received by the other end. Even errored requests may generate a notification, and the user must check for IORING_CQE_F_MORE rather than relying on the result. fd must be set to the socket file descriptor, addr must contain a pointer to the buffer, len denotes the length of the buffer to send, and msg_flags holds the flags associated with the system call. When addr2 is non-zero it points to the address of the target with addr_len specifying its size, turning the request into a sendto(2) system call equivalent. Available since 6.0. This command also supports the following modifiers in ioprio:

IORING_RECVSEND_FIXED_BUF If set, instructs io_uring to use a pre-mapped buffer. The buf_index field should contain an index into an array of fixed buffers. See io_uring_register(2) for details on how to setup a context for fixed buffer I/O.

The flags field is a bit mask. The supported flags are:

IOSQE_FIXED_FILE: When this flag is specified, fd is an index into the files array registered with the io_uring instance (see the IORING_REGISTER_FILES section of the io_uring_register(2) man page). Note that this isn't always available for all commands. If used on a command that doesn't support fixed files, the SQE will error with -EBADF. Available since 5.1.

IOSQE_IO_DRAIN: When this flag is specified, the SQE will not be started before previously submitted SQEs have completed, and new SQEs will not be started before this one completes. Available since 5.2.

IOSQE_IO_LINK: When this flag is specified, the SQE forms a link with the next SQE in the submission ring. That next SQE will not be started before the previous request completes. This, in effect, forms a chain of SQEs, which can be arbitrarily long. The tail of the chain is denoted by the first SQE that does not have this flag set. Chains are not supported across submission boundaries. Even if the last SQE in a submission has this flag set, it will still terminate the current chain. This flag has no effect on previous SQE submissions, nor does it impact SQEs that are outside of the chain tail. This means that multiple chains can be executing in parallel, or chains and individual SQEs. Only members inside the chain are serialized. A chain of SQEs will be broken, if any request in that chain ends in error. io_uring considers any unexpected result an error. This means that, eg, a short read will also terminate the remainder of the chain. If a chain of SQE links is broken, the remaining unstarted part of the chain will be terminated and completed with -ECANCELED as the error code. Available since 5.3.

IOSQE_IO_HARDLINK: Like IOSQE_IO_LINK, but it doesn't sever regardless of the completion result. Note that the link will still sever if we fail submitting the parent request, hard links are only resilient in the presence of completion results for requests that did submit correctly. IOSQE_IO_HARDLINK implies IOSQE_IO_LINK. Available since 5.5.

IOSQE_ASYNC: Normal operation for io_uring is to try and issue an sqe as non-blocking first, and if that fails, execute it in an async manner. To support more efficient overlapped operation of requests that the application knows/assumes will always (or most of the time) block, the application can ask for an sqe to be issued async from the start. Available since 5.6.

IOSQE_BUFFER_SELECT: Used in conjunction with the IORING_OP_PROVIDE_BUFFERS command, which registers a pool of buffers to be used by commands that read or receive data. When buffers are registered for this use case, and this flag is set in the command, io_uring will grab a buffer from this pool when the request is ready to receive or read data. If successful, the resulting CQE will have IORING_CQE_F_BUFFER set in the flags part of the struct, and the upper IORING_CQE_BUFFER_SHIFT bits will contain the ID of the selected buffers. This allows the application to know exactly which buffer was selected for the operation. If no buffers are available and this flag is set, then the request will fail with -ENOBUFS as the error code. Once a buffer has been used, it is no longer available in the kernel pool. The application must re-register the given buffer again when it is ready to recycle it (eg has completed using it). Available since 5.7.

IOSQE_CQE_SKIP_SUCCESS: Don't generate a CQE if the request completes successfully. If the request fails, an appropriate CQE will be posted as usual and if there is no IOSQE_IO_HARDLINK, CQEs for all linked requests will be omitted. The notion of failure/success is opcode specific and is the same as with breaking chains of IOSQE_IO_LINK. One special case is when the request has a linked timeout, then the CQE generation for the linked timeout is decided solely by whether it has IOSQE_CQE_SKIP_SUCCESS set, regardless whether it timed out or was canceled. In other words, if a linked timeout has the flag set, it's guaranteed to not post a CQE. The semantics are chosen to accommodate several use cases. First, when all but the last request of a normal link without linked timeouts are marked with the flag, only one CQE per lin is posted. Additionally, it enables suppression of CQEs in cases where the side effects of a successfully executed operation is enough for userspace to know the state of the system. One such example would be writing to a synchronisation file. This flag is incompatible with IOSQE_IO_DRAIN. Using both of them in a single ring is undefined behavior, even when they are not used together in a single request. Currently, after the first request with IOSQE_CQE_SKIP_SUCCESS, all subsequent requests marked with drain will be failed at submission time. Note that the error reporting is best effort only, and restrictions may change in the future. Available since 5.17.

ioprio specifies the I/O priority. See ioprio_get(2) for a description of Linux I/O priorities.

fd specifies the file descriptor against which the operation will be performed, with the exception noted above.

If the operation is one of IORING_OP_READ_FIXED or IORING_OP_WRITE_FIXED, addr and len must fall within the buffer located at buf_index in the fixed buffer array. If the operation is either IORING_OP_READV or IORING_OP_WRITEV, then addr points to an iovec array of len entries.

rw_flags, specified for read and write operations, contains a bitwise OR of per-I/O flags, as described in the preadv2(2) man page.

The fsync_flags bit mask may contain either 0, for a normal file integrity sync, or IORING_FSYNC_DATASYNC to provide data sync only semantics. See the descriptions of O_SYNC and O_DSYNC in the open(2) manual page for more information.

The bits that may be set in poll_events are defined in <poll.h>, and documented in poll(2).

user_data is an application-supplied value that will be copied into the completion queue entry (see below). buf_index is an index into an array of fixed buffers, and is only valid if fixed buffers were registered. personality is the credentials id to use for this operation. See io_uring_register(2) for how to register personalities with io_uring. If set to 0, the current personality of the submitting task is used.

Once the submission queue entry is initialized, I/O is submitted by placing the index of the submission queue entry into the tail of the submission queue. After one or more indexes are added to the queue, and the queue tail is advanced, the io_uring_enter(2) system call can be invoked to initiate the I/O.

Completions use the following data structure:

/*
 * IO completion data structure (Completion Queue Entry)
 */
struct io_uring_cqe {
    __u64    user_data; /* sqe->data submission passed back */
    __s32    res;       /* result code for this event */
    __u32    flags;
};

user_data is copied from the field of the same name in the submission queue entry. The primary use case is to store data that the application will need to access upon completion of this particular I/O. The flags is used for certain commands, like IORING_OP_POLL_ADD or in conjunction with IOSQE_BUFFER_SELECT or IORING_OP_MSG_RING, , see those entries for details. res is the operation-specific result, but io_uring-specific errors (e.g. flags or opcode invalid) are returned through this field. They are described in section CQE ERRORS.

For read and write opcodes, the return values match errno values documented in the preadv2(2) and pwritev2(2) man pages, with res holding the equivalent of -errno for error cases, or the transferred number of bytes in case the operation is successful. Hence both error and success return can be found in that field in the CQE. For other request types, the return values are documented in the matching man page for that type, or in the opcodes section above for io_uring-specific opcodes.

RETURN VALUE

io_uring_enter(2) returns the number of I/Os successfully consumed. This can be zero if to_submit was zero or if the submission queue was empty. Note that if the ring was created with IORING_SETUP_SQPOLL specified, then the return value will generally be the same as to_submit as submission happens outside the context of the system call.

The errors related to a submission queue entry will be returned through a completion queue entry (see section CQE ERRORS), rather than through the system call itself.

Errors that occur not on behalf of a submission queue entry are returned via the system call directly. On such an error, a negative error code is returned. The caller should not rely on errno variable.

ERRORS

These are the errors returned by io_uring_enter(2) system call.

EAGAIN: The kernel was unable to allocate memory for the request, or otherwise ran out of resources to handle it. The application should wait for some completions and try again.

EBADF: fd is not a valid file descriptor.

EBADFD: fd is a valid file descriptor, but the io_uring ring is not in the right state (enabled). See io_uring_register(2) for details on how to enable the ring.

EBADR: At least one CQE was dropped even with the IORING_FEAT_NODROP feature, and there are no otherwise available CQEs. This clears the error state and so with no other changes the next call to io_uring_setup(2) will not have this error. This error should be extremely rare and indicates the machine is running critically low on memory and. It may be reasonable for the application to terminate running unless it is able to safely handle any CQE being lost.

EBUSY: If the IORING_FEAT_NODROP feature flag is set, then EBUSY will be returned if there were overflow entries, IORING_ENTER_GETEVENTS flag is set and not all of the overflow entries were able to be flushed to the CQ ring. Without IORING_FEAT_NODROP the application is attempting to overcommit the number of requests it can have pending. The application should wait for some completions and try again. May occur if the application tries to queue more requests than we have room for in the CQ ring, or if the application attempts to wait for more events without having reaped the ones already present in the CQ ring.

EINVAL: Some bits in the flags argument are invalid.

EFAULT: An invalid user space address was specified for the sig argument.

ENXIO: The io_uring instance is in the process of being torn down.

EOPNOTSUPP: fd does not refer to an io_uring instance.

EINTR: The operation was interrupted by a delivery of a signal before it could complete; see signal(7). Can happen while waiting for events with IORING_ENTER_GETEVENTS.

CQE ERRORS

These io_uring-specific errors are returned as a negative value in the res field of the completion queue entry.

EACCES: The flags field or opcode in a submission queue entry is not allowed due to registered restrictions. See io_uring_register(2) for details on how restrictions work.

EBADF: The fd field in the submission queue entry is invalid, or the IOSQE_FIXED_FILE flag was set in the submission queue entry, but no files were registered with the io_uring instance.

EFAULT: buffer is outside of the process' accessible address space

EFAULT: IORING_OP_READ_FIXED or IORING_OP_WRITE_FIXED was specified in the opcode field of the submission queue entry, but either buffers were not registered for this io_uring instance, or the address range described by addr and len does not fit within the buffer registered at buf_index.

EINVAL: The flags field or opcode in a submission queue entry is invalid.

EINVAL: The buf_index member of the submission queue entry is invalid.

EINVAL: The personality field in a submission queue entry is invalid.

EINVAL: IORING_OP_NOP was specified in the submission queue entry, but the io_uring context was setup for polling (IORING_SETUP_IOPOLL was specified in the call to io_uring_setup).

EINVAL: IORING_OP_READV or IORING_OP_WRITEV was specified in the submission queue entry, but the io_uring instance has fixed buffers registered.

EINVAL: IORING_OP_READ_FIXED or IORING_OP_WRITE_FIXED was specified in the submission queue entry, and the buf_index is invalid.

EINVAL: IORING_OP_READV, IORING_OP_WRITEV, IORING_OP_READ_FIXED, IORING_OP_WRITE_FIXED or IORING_OP_FSYNC was specified in the submission queue entry, but the io_uring instance was configured for IOPOLLing, or any of addr, ioprio, off, len, or buf_index was set in the submission queue entry.

EINVAL: IORING_OP_POLL_ADD or IORING_OP_POLL_REMOVE was specified in the opcode field of the submission queue entry, but the io_uring instance was configured for busy-wait polling (IORING_SETUP_IOPOLL), or any of ioprio, off, len, or buf_index was non-zero in the submission queue entry.

EINVAL: IORING_OP_POLL_ADD was specified in the opcode field of the submission queue entry, and the addr field was non-zero.

EOPNOTSUPP: opcode is valid, but not supported by this kernel.

EOPNOTSUPP: IOSQE_BUFFER_SELECT was set in the flags field of the submission queue entry, but the opcode doesn't support buffer selection.

2019-01-22

Linux

io_uring_enter2

NAME

SYNOPSIS

DESCRIPTION

RETURN VALUE

ERRORS

CQE ERRORS

Recommended readings

Questions & Answers