022ca2fc7f
POSIX AIO is great, but it lacks vectored I/O functions. This commit fixes that shortcoming by adding aio_writev and aio_readv. They aren't part of the standard, but they're an obvious extension. They work just like their synchronous equivalents pwritev and preadv. It isn't yet possible to use vectored aiocbs with lio_listio, but that could be added in the future. Reviewed by: jhb, kib, bcr Relnotes: yes Differential Revision: https://reviews.freebsd.org/D27743
259 lines
6.5 KiB
Groff
259 lines
6.5 KiB
Groff
.\" Copyright (c) 1999 Softweyr LLC.
|
|
.\" All rights reserved.
|
|
.\"
|
|
.\" Redistribution and use in source and binary forms, with or without
|
|
.\" modification, are permitted provided that the following conditions
|
|
.\" are met:
|
|
.\" 1. Redistributions of source code must retain the above copyright
|
|
.\" notice, this list of conditions and the following disclaimer.
|
|
.\" 2. Redistributions in binary form must reproduce the above copyright
|
|
.\" notice, this list of conditions and the following disclaimer in the
|
|
.\" documentation and/or other materials provided with the distribution.
|
|
.\"
|
|
.\" THIS SOFTWARE IS PROVIDED BY Softweyr LLC AND CONTRIBUTORS ``AS IS'' AND
|
|
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
.\" ARE DISCLAIMED. IN NO EVENT SHALL Softweyr LLC OR CONTRIBUTORS BE LIABLE
|
|
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
.\" SUCH DAMAGE.
|
|
.\"
|
|
.\" $FreeBSD$
|
|
.\"
|
|
.Dd January 2, 2021
|
|
.Dt AIO_WRITE 2
|
|
.Os
|
|
.Sh NAME
|
|
.Nm aio_write ,
|
|
.Nm aio_writev
|
|
.Nd asynchronous write to a file (REALTIME)
|
|
.Sh LIBRARY
|
|
.Lb libc
|
|
.Sh SYNOPSIS
|
|
.In aio.h
|
|
.Ft int
|
|
.Fn aio_write "struct aiocb *iocb"
|
|
.In sys/uio.h
|
|
.Ft int
|
|
.Fn aio_writev "struct aiocb *iocb"
|
|
.Sh DESCRIPTION
|
|
The
|
|
.Fn aio_write
|
|
and
|
|
.Fn aio_writev
|
|
system calls allow the calling process to write
|
|
to the descriptor
|
|
.Fa iocb->aio_fildes .
|
|
.Fn aio_write
|
|
will write
|
|
.Fa iocb->aio_nbytes
|
|
from the buffer pointed to by
|
|
.Fa iocb->aio_buf ,
|
|
whereas
|
|
.Fn aio_writev
|
|
gathers the data from the
|
|
.Fa iocb->aio_iovcnt
|
|
buffers specified by the members of the
|
|
.Fa iocb->aio_iov
|
|
array.
|
|
Both syscalls return immediately after the write request has been enqueued
|
|
to the descriptor; the write may or may not have completed at the time
|
|
the call returns.
|
|
If the request could not be enqueued, generally due
|
|
to invalid arguments, the call returns without having enqueued the
|
|
request.
|
|
.Pp
|
|
For
|
|
.Fn aio_writev
|
|
the
|
|
.Fa iovec
|
|
structure is defined in
|
|
.Xr writev 2 .
|
|
.Pp
|
|
If
|
|
.Dv O_APPEND
|
|
is set for
|
|
.Fa iocb->aio_fildes ,
|
|
write operations append to the file in the same order as the calls were
|
|
made.
|
|
If
|
|
.Dv O_APPEND
|
|
is not set for the file descriptor, the write operation will occur at
|
|
the absolute position from the beginning of the file plus
|
|
.Fa iocb->aio_offset .
|
|
.Pp
|
|
If
|
|
.Dv _POSIX_PRIORITIZED_IO
|
|
is defined, and the descriptor supports it, then the enqueued
|
|
operation is submitted at a priority equal to that of the calling
|
|
process minus
|
|
.Fa iocb->aio_reqprio .
|
|
.Pp
|
|
The
|
|
.Fa iocb
|
|
pointer may be subsequently used as an argument to
|
|
.Fn aio_return
|
|
and
|
|
.Fn aio_error
|
|
in order to determine return or error status for the enqueued operation
|
|
while it is in progress.
|
|
.Pp
|
|
If the request is successfully enqueued, the value of
|
|
.Fa iocb->aio_offset
|
|
can be modified during the request as context, so this value must not
|
|
be referenced after the request is enqueued.
|
|
.Pp
|
|
The
|
|
.Fa iocb->aio_sigevent
|
|
structure can be used to request notification of the operation's
|
|
completion as described in
|
|
.Xr aio 4 .
|
|
.Sh RESTRICTIONS
|
|
The Asynchronous I/O Control Block structure pointed to by
|
|
.Fa iocb
|
|
and the buffer that the
|
|
.Fa iocb->aio_buf
|
|
member of that structure references must remain valid until the
|
|
operation has completed.
|
|
.Pp
|
|
The asynchronous I/O control buffer
|
|
.Fa iocb
|
|
should be zeroed before the
|
|
.Fn aio_write
|
|
or
|
|
.Fn aio_writev
|
|
system call to avoid passing bogus context information to the kernel.
|
|
.Pp
|
|
Modifications of the Asynchronous I/O Control Block structure or the
|
|
buffer contents are not allowed while the request is queued.
|
|
.Pp
|
|
If the file offset in
|
|
.Fa iocb->aio_offset
|
|
is past the offset maximum for
|
|
.Fa iocb->aio_fildes ,
|
|
no I/O will occur.
|
|
.Sh RETURN VALUES
|
|
.Rv -std aio_write aio_writev
|
|
.Sh ERRORS
|
|
The
|
|
.Fn aio_write
|
|
and
|
|
.Fn aio_writev
|
|
system calls will fail if:
|
|
.Bl -tag -width Er
|
|
.It Bq Er EAGAIN
|
|
The request was not queued because of system resource limitations.
|
|
.It Bq Er EFAULT
|
|
Part of
|
|
.Fa aio_iov
|
|
points outside the process's allocated address space.
|
|
.It Bq Er EINVAL
|
|
The asynchronous notification method in
|
|
.Fa iocb->aio_sigevent.sigev_notify
|
|
is invalid or not supported.
|
|
.It Bq Er EOPNOTSUPP
|
|
Asynchronous write operations on the file descriptor
|
|
.Fa iocb->aio_fildes
|
|
are unsafe and unsafe asynchronous I/O operations are disabled.
|
|
.El
|
|
.Pp
|
|
The following conditions may be synchronously detected when the
|
|
.Fn aio_write
|
|
or
|
|
.Fn aio_writev
|
|
system call is made, or asynchronously, at any time thereafter.
|
|
If they
|
|
are detected at call time,
|
|
.Fn aio_write
|
|
or
|
|
.Fn aio_writev
|
|
returns -1 and sets
|
|
.Va errno
|
|
appropriately; otherwise the
|
|
.Fn aio_return
|
|
system call must be called, and will return -1, and
|
|
.Fn aio_error
|
|
must be called to determine the actual value that would have been
|
|
returned in
|
|
.Va errno .
|
|
.Bl -tag -width Er
|
|
.It Bq Er EBADF
|
|
The
|
|
.Fa iocb->aio_fildes
|
|
argument
|
|
is invalid, or is not opened for writing.
|
|
.It Bq Er EINVAL
|
|
The offset
|
|
.Fa iocb->aio_offset
|
|
is not valid, the priority specified by
|
|
.Fa iocb->aio_reqprio
|
|
is not a valid priority, or the number of bytes specified by
|
|
.Fa iocb->aio_nbytes
|
|
is not valid.
|
|
.El
|
|
.Pp
|
|
If the request is successfully enqueued, but subsequently canceled
|
|
or an error occurs, the value returned by the
|
|
.Fn aio_return
|
|
system call is per the
|
|
.Xr write 2
|
|
system call, and the value returned by the
|
|
.Fn aio_error
|
|
system call is either one of the error returns from the
|
|
.Xr write 2
|
|
system call, or one of:
|
|
.Bl -tag -width Er
|
|
.It Bq Er EBADF
|
|
The
|
|
.Fa iocb->aio_fildes
|
|
argument
|
|
is invalid for writing.
|
|
.It Bq Er ECANCELED
|
|
The request was explicitly canceled via a call to
|
|
.Fn aio_cancel .
|
|
.It Bq Er EINVAL
|
|
The offset
|
|
.Fa iocb->aio_offset
|
|
would be invalid.
|
|
.El
|
|
.Sh SEE ALSO
|
|
.Xr aio_cancel 2 ,
|
|
.Xr aio_error 2 ,
|
|
.Xr aio_return 2 ,
|
|
.Xr aio_suspend 2 ,
|
|
.Xr aio_waitcomplete 2 ,
|
|
.Xr sigevent 3 ,
|
|
.Xr siginfo 3 ,
|
|
.Xr aio 4
|
|
.Sh STANDARDS
|
|
The
|
|
.Fn aio_write
|
|
system call
|
|
is expected to conform to the
|
|
.St -p1003.1
|
|
standard.
|
|
.Pp
|
|
The
|
|
.Fn aio_writev
|
|
system call is a FreeBSD extension, and should not be used in portable code.
|
|
.Sh HISTORY
|
|
The
|
|
.Fn aio_write
|
|
system call first appeared in
|
|
.Fx 3.0 .
|
|
The
|
|
.Fn aio_writev
|
|
system call first appeared in
|
|
.Fx 13.0 .
|
|
.Sh AUTHORS
|
|
This manual page was written by
|
|
.An Wes Peters Aq Mt wes@softweyr.com .
|
|
.Sh BUGS
|
|
Invalid information in
|
|
.Fa iocb->_aiocb_private
|
|
may confuse the kernel.
|