6. Input and Output Primitives Manager¶

6.1. Introduction¶

The input and output primitives manager is …

The directives provided by the input and output primitives manager are:

pipe - Create an Inter-Process Channel
dup - Duplicates an open file descriptor
dup2 - Duplicates an open file descriptor
close - Closes a file
read - Reads from a file
write - Writes to a file
fcntl - Manipulates an open file descriptor
lseek - Reposition read/write file offset
fsync - Synchronize file complete in-core state with that on disk
fdatasync - Synchronize file in-core data with that on disk
sync - Schedule file system updates
mount - Mount a file system
unmount - Unmount file systems
readv - Vectored read from a file
writev - Vectored write to a file
aio_read - Asynchronous Read
aio_write - Asynchronous Write
lio_listio - List Directed I/O
aio_error - Retrieve Error Status of Asynchronous I/O Operation
aio_return_ - Retrieve Return Status Asynchronous I/O Operation
aio_cancel - Cancel Asynchronous I/O Request
aio_suspend - Wait for Asynchronous I/O Request
aio_fsync - Asynchronous File Synchronization

6.2. Background¶

There is currently no text in this section.

6.3. Operations¶

There is currently no text in this section.

6.4. Directives¶

This section details the input and output primitives manager’s directives. A subsection is dedicated to each of this manager’s directives and describes the calling sequence, related constants, usage, and status codes.

6.4.1. pipe - Create an Inter-Process Channel¶

CALLING SEQUENCE:

#include <unistd.h>
int pipe(
    int fildes[2]
);

STATUS CODES:

E

The

DESCRIPTION:

NOTES:

This routine is not currently supported by RTEMS but could be in a future version.

6.4.2. dup - Duplicates an open file descriptor¶

CALLING SEQUENCE:

#include <unistd.h>
int dup(
    int fildes
);

STATUS CODES:

`EBADF`	Invalid file descriptor.
`EINTR`	Function was interrupted by a signal.
`EMFILE`	The process already has the maximum number of file descriptors open and tried to open a new one.

DESCRIPTION:

The dup function returns the lowest numbered available file descriptor. This new desciptor refers to the same open file as the original descriptor and shares any locks.

NOTES:

NONE

6.4.3. dup2 - Duplicates an open file descriptor¶

CALLING SEQUENCE:

#include <unistd.h>
int dup2(
    int fildes,
    int fildes2
);

STATUS CODES:

`EBADF`	Invalid file descriptor.
`EINTR`	Function was interrupted by a signal.
`EMFILE`	The process already has the maximum number of file descriptors open and tried to open a new one.

DESCRIPTION:

dup2 creates a copy of the file descriptor oldfd.

The old and new descriptors may be used interchangeably. They share locks, file position pointers and flags; for example, if the file position is modified by using lseek on one of the descriptors, the position is also changed for the other.

NOTES:

NONE

6.4.4. close - Closes a file¶

CALLING SEQUENCE:

#include <unistd.h>
int close(
    int fildes
);

STATUS CODES:

`EBADF`	Invalid file descriptor
`EINTR`	Function was interrupted by a signal.

DESCRIPTION:

The close() function deallocates the file descriptor named by fildes and makes it available for reuse. All outstanding record locks owned by this process for the file are unlocked.

NOTES:

A signal can interrupt the close() function. In that case, close() returns -1 with errno set to EINTR. The file may or may not be closed.

6.4.5. read - Reads from a file¶

CALLING SEQUENCE:

#include <unistd.h>
ssize_t read(
    int fildes,
    void *buf,
    size_t nbyte
);

STATUS CODES:

On error, this routine returns -1 and sets errno to one of the following:

`EAGAIN`	The O_NONBLOCK flag is set for a file descriptor and the process would be delayed in the I/O operation.
`EBADF`	Invalid file descriptor
`EINTR`	Function was interrupted by a signal.
`EIO`	Input or output error
`EINVAL`	Bad buffer pointer

DESCRIPTION:

The read() function reads nbyte bytes from the file associated with fildes into the buffer pointed to by buf.

The read() function returns the number of bytes actually read and placed in the buffer. This will be less than nbyte if:

The number of bytes left in the file is less than nbyte.
The read() request was interrupted by a signal.
The file is a pipe or FIFO or special file with less than nbytes immediately available for reading.

When attempting to read from any empty pipe or FIFO:

If no process has the pipe open for writing, zero is returned to indicate end-of-file.
If some process has the pipe open for writing and O_NONBLOCK is set, -1 is returned and errno is set to EAGAIN.
If some process has the pipe open for writing and O_NONBLOCK is clear, read() waits for some data to be written or the pipe to be closed.

When attempting to read from a file other than a pipe or FIFO and no data is available.

If O_NONBLOCK is set, -1 is returned and errno is set to EAGAIN.
If O_NONBLOCK is clear, read() waits for some data to become available.
The O_NONBLOCK flag is ignored if data is available.

NOTES:

NONE

6.4.6. write - Writes to a file¶

CALLING SEQUENCE:

#include <unistd.h>
ssize_t write(
    int fildes,
    const void *buf,
    size_t nbyte
);

STATUS CODES:

`EAGAIN`	The O_NONBLOCK flag is set for a file descriptor and the process would be delayed in the I/O operation.
`EBADF`	Invalid file descriptor
`EFBIG`	An attempt was made to write to a file that exceeds the maximum file size
`EINTR`	The function was interrupted by a signal.
`EIO`	Input or output error.
`ENOSPC`	No space left on disk.
`EPIPE`	Attempt to write to a pope or FIFO with no reader.
`EINVAL`	Bad buffer pointer

DESCRIPTION:

The write() function writes nbyte from the array pointed to by buf into the file associated with fildes.

If nybte is zero and the file is a regular file, the write() function returns zero and has no other effect. If nbyte is zero and the file is a special file, te results are not portable.

The write() function returns the number of bytes written. This number will be less than nbytes if there is an error. It will never be greater than nbytes.

NOTES:

NONE

6.4.7. fcntl - Manipulates an open file descriptor¶

CALLING SEQUENCE:

#include <fcntl.h>
int fcntl(
    int fildes,
    int cmd,
    ...
);

STATUS CODES:

`EACCESS`	Search permission is denied for a direcotry in a file’s path prefix.
`EAGAIN`	The O_NONBLOCK flag is set for a file descriptor and the process would be delayed in the I/O operation.
`EBADF`	Invalid file descriptor
`EDEADLK`	An `fcntl` with function `F_SETLKW` would cause a deadlock.
`EINTR`	The functioin was interrupted by a signal.
`EINVAL`	Invalid argument
`EMFILE`	Too many file descriptor or in use by the process.
`ENOLCK`	No locks available

DESCRIPTION:

fcntl() performs one of various miscellaneous operations on``fd``. The operation in question is determined by cmd:

`F_DUPFD`	Makes `arg` be a copy of `fd`, closing `fd` first if necessary. The same functionality can be more easily achieved by using `dup2()`. The old and new descriptors may be used interchangeably. They share locks, file position pointers and flags; for example, if the file position is modified by using `lseek()` on one of the descriptors, the position is also changed for the other. The two descriptors do not share the close-on-exec flag, however. The close-on-exec flag of the copy is off, meaning that it will be closed on exec. On success, the new descriptor is returned.
`F_GETFD`	Read the close-on-exec flag. If the low-order bit is 0, the file will remain open across exec, otherwise it will be closed.
`F_SETFD`	Set the close-on-exec flag to the value specified by `arg` (only the least significant bit is used).
`F_GETFL`	Read the descriptor’s flags (all flags (as set by open()) are returned).
`F_SETFL`	Set the descriptor’s flags to the value specified by `arg`. Only``O_APPEND`` and `O_NONBLOCK` may be set. The flags are shared between copies (made with `dup()` etc.) of the same file descriptor. The flags and their semantics are described in `open()`.
`F_GETLK`, `F_SETLK` and `F_SETLKW`	Manage discretionary file locks. The third argument `arg` is a pointer to a struct flock (that may be overwritten by this call).
`F_GETLK`	Return the flock structure that prevents us from obtaining the lock, or set the``l_type`` field of the lock to `F_UNLCK` if there is no obstruction.
`F_SETLK`	The lock is set (when `l_type` is `F_RDLCK` or `F_WRLCK`) or cleared (when it is `F_UNLCK`. If lock is held by someone else, this call returns -1 and sets `errno` to EACCES or EAGAIN.
`F_SETLKW`	Like `F_SETLK`, but instead of returning an error we wait for the lock to be released.
`F_GETOWN`	Get the process ID (or process group) of the owner of a socket. Process groups are returned as negative values.
`F_SETOWN`	Set the process or process group that owns a socket. For these commands, ownership means receiving `SIGIO` or `SIGURG` signals. Process groups are specified using negative values.

NOTES:

The errors returned by dup2 are different from those returned by F_DUPFD.

6.4.8. lseek - Reposition read/write file offset¶

CALLING SEQUENCE:

#include <unistd.h>
off_t lseek(
    int fildes,
    off_t offset,
    int whence
);

STATUS CODES:

`EBADF`	`fildes` is not an open file descriptor.
`ESPIPE`	`fildes` is associated with a pipe, socket or FIFO.
`EINVAL`	`whence` is not a proper value.

DESCRIPTION:

The lseek function repositions the offset of the file descriptor fildes to the argument offset according to the directive whence. The argument fildes must be an open file descriptor. Lseek repositions the file pointer fildes as follows:

If whence is SEEK_SET, the offset is set to offset bytes.
If whence is SEEK_CUR, the offset is set to its current location plus offset bytes.
If whence is SEEK_END, the offset is set to the size of the file plus offset bytes.

The lseek function allows the file offset to be set beyond the end of the existing end-of-file of the file. If data is later written at this point, subsequent reads of the data in the gap return bytes of zeros (until data is actually written into the gap).

Some devices are incapable of seeking. The value of the pointer associated with such a device is undefined.

NOTES:

NONE

6.4.9. fsync - Synchronize file complete in-core state with that on disk¶

CALLING SEQUENCE:

#include <unistd.h>
int fsync(
    int fildes
);

STATUS CODES:

On success, zero is returned. On error, -1 is returned, and errno is set appropriately.

`EBADF`	`fd` is not a valid descriptor open for writing
`EINVAL`	`fd` is bound to a special file which does not support support synchronization
`EROFS`	`fd` is bound to a special file which does not support support synchronization
`EIO`	An error occurred during synchronization

DESCRIPTION:

fsync copies all in-core parts of a file to disk.

NOTES:

NONE

6.4.10. fdatasync - Synchronize file in-core data with that on disk¶

CALLING SEQUENCE:

#include <unistd.h>
int fdatasync(
    int fildes
);

STATUS CODES:

On success, zero is returned. On error, -1 is returned, and errno is set appropriately.

`EBADF`	`fd` is not a valid file descriptor open for writing.
`EINVAL`	`fd` is bound to a special file which does not support synchronization.
`EIO`	An error occurred during synchronization.
`EROFS`	`fd` is bound to a special file which dows not support synchronization.

DESCRIPTION:

fdatasync flushes all data buffers of a file to disk (before the system call returns). It resembles fsync but is not required to update the metadata such as access time.

Applications that access databases or log files often write a tiny data fragment (e.g., one line in a log file) and then call fsync immediately in order to ensure that the written data is physically stored on the harddisk. Unfortunately, fsync will always initiate two write operations: one for the newly written data and another one in order to update the modification time stored in the inode. If the modification time is not a part of the transaction concept fdatasync can be used to avoid unnecessary inode disk write operations.

NOTES:

NONE

6.4.11. sync - Schedule file system updates¶

CALLING SEQUENCE:

#include <unistd.h>
void sync(
    void
);

STATUS CODES:

NONE

DESCRIPTION:

The sync service causes all information in memory that updates file systems to be scheduled for writing out to all file systems.

NOTES:

The writing of data to the file systems is only guaranteed to be scheduled upon return. It is not necessarily complete upon return from sync.

6.4.12. mount - Mount a file system¶

CALLING SEQUENCE:

#include <libio.h>
int mount(
    rtems_filesystem_mount_table_entry_t **mt_entry,
    rtems_filesystem_operations_table *fs_ops,
    rtems_filesystem_options_t fsoptions,
    char *device,
    char *mount_point
);

STATUS CODES:

ENOMEM

Unable to allocate memory needed.

EINVAL

The filesystem does not support being mounted.

EINVAL

Attempt to mount a read-only filesystem as writeable.

DESCRIPTION:

The mount routines mounts the filesystem class which uses the filesystem operations specified by fs_ops and fsoptions. The filesystem is mounted at the directory mount_point and the mode of the mounted filesystem is specified by fsoptions. If this filesystem class requires a device, then the name of the device must be specified by device.

If this operation succeeds, the mount table entry for the mounted filesystem is returned in mt_entry.

NOTES:

This method is not defined in the POSIX standard.

6.4.13. unmount - Unmount file systems¶

CALLING SEQUENCE:

#include <libio.h>
int unmount(
    const char *mount_path
);

STATUS CODES:

EBUSY

Filesystem is in use or the root filesystem.

EACCESS

Unable to allocate memory needed.

DESCRIPTION:

The unmount routine removes the attachment of the filesystem specified by mount_path.

NOTES:

This method is not defined in the POSIX standard.

6.4.14. readv - Vectored read from a file¶

CALLING SEQUENCE:

#include <sys/uio.h>
ssize_t readv(
    int fildes,
    const struct iovec *iov,
    int iovcnt
);

STATUS CODES:

In addition to the errors detected by Input and Output Primitives Manager read - Reads from a file, read(), this routine may return -1 and sets errno based upon the following errors:

`EINVAL`	The sum of the `iov_len` values in the iov array overflowed an `ssize_t`.
`EINVAL`	The `iovcnt` argument was less than or equal to 0, or greater than `IOV_MAX`.

DESCRIPTION:

The readv() function is equivalent to read() except as described here. The readv() function shall place the input data into the iovcnt buffers specified by the members of the iov array: iov[0], iov[1], ..., iov[iovcnt-1].

Each iovec entry specifies the base address and length of an area in memory where data should be placed. The readv() function always fills an area completely before proceeding to the next.

NOTES:

NONE

6.4.15. writev - Vectored write to a file¶

CALLING SEQUENCE:

#include <sys/uio.h>
ssize_t writev(
    int fildes,
    const struct iovec *iov,
    int iovcnt
);

STATUS CODES:

In addition to the errors detected by Input and Output Primitives Manager write - Write to a file, write(), this routine may return -1 and sets errno based upon the following errors:

`EINVAL`	The sum of the `iov_len` values in the iov array overflowed an `ssize_t`.
`EINVAL`	The `iovcnt` argument was less than or equal to 0, or greater than `IOV_MAX`.

DESCRIPTION:

The writev() function is equivalent to write(), except as noted here. The writev() function gathers output data from the iovcnt buffers specified by the members of the iov array: iov[0], iov[1], ..., iov[iovcnt-1]. The iovcnt argument is valid if greater than 0 and less than or equal to IOV_MAX.

Each iovec entry specifies the base address and length of an area in memory from which data should be written. The writev() function always writes a complete area before proceeding to the next.

If fd refers to a regular file and all of the iov_len members in the array pointed to by iov are 0, writev() returns 0 and has no other effect. For other file types, the behavior is unspecified by POSIX.

NOTES:

NONE

6.4.16. aio_read - Asynchronous Read¶

CALLING SEQUENCE:

#include <aio.h>
int aio_read(
    struct aiocb *aiocbp
);

STATUS CODES:

If the request is successfully enqueued, zero is returned. On error, this routine returns -1 and sets errno to one of the following:

`EBADF`	The file descriptor is not open for reading.
`EINVAL`	Invalid aio_reqprio, aio_offset, or aio_nbytes.
`EAGAIN`	Not enough memory to queue the request.
`EAGAIN`	the addition of a new request to the queue would violate the `RTEMS_AIO_MAX` limit.
`EINVAL`	The starting position is past the maximum offset for the file.
`EINVAL`	`aiocbp->aio_sigevent` does not point to a valid sigevent struct.
`EINVAL`	`aiocbp` is a NULL pointer.

DESCRIPTION:

The aio_read() function is the asynchronous equivalent of read(). This function returns immediately, the request is serviced by thread(s) running in the background.

The parameters for the read are specified in the aiocbp structure.

NOTES:

NONE

6.4.17. aio_write - Asynchronous Write¶

CALLING SEQUENCE:

#include <aio.h>
int aio_write(
    struct aiocb *aiocbp
);

STATUS CODES:

If the request is successfully enqueued, zero is returned. On error, this routine returns -1 and sets errno to one of the following:

`EBADF`	The file descriptor is not open for writing.
`EINVAL`	Invalid aio_reqprio, aio_offset, or aio_nbytes.
`EAGAIN`	Not enough memory to queue the request.
`EAGAIN`	the addition of a new request to the queue would violate the `RTEMS_AIO_MAX` limit.
`EINVAL`	`aiocbp->aio_sigevent` does not point to a valid sigevent struct.
`EINVAL`	`aiocbp` is a NULL pointer.

DESCRIPTION:

The aio_write() function is the asynchronous equivalent of write(). This function returns immediately, the request is serviced by thread(s) running in the background.

The parameters for the write are specified in the aiocbp structure.

NOTES:

NONE

6.4.18. lio_listio - List Directed I/O¶

CALLING SEQUENCE:

#include <aio.h>
int lio_listio(
    int mode,
    struct aiocb *restrict const list[restrict],
    int nent,
    struct sigevent *restrict sig
);

STATUS CODES:

If the call to lio_listio is successful, zero is returned. On error, this routine returns -1 and sets errno to one of the following:

`ENOSYS`	The project has been build with RTEMS_POSIX_API not defined.
`EAGAIN`	The call failed due to resources limitations.
`EAGAIN`	The number of entries indicated by nent value would cause the `RTEMS_AIO_MAX` limit to be excedeed.
`EINVAL`	list is a `NULL` pointer.
`EINVAL`	`mode` is not a valid value.
`EINVAL`	the value of `nent` is not valid or higher than `AIO_LISTIO_MAX`.
`EINVAL`	the `sigevent` struct pointed by `sig` is not valid.
`EINTR`	The wait for list completion during a `LIO_WAIT` operation was interrupted by an external event.
`EIO`	One or more of the individual I/O operations failed.

DESCRIPTION:

The lio_listio() function allows for the simultaneous initiation of multiple asynchronous I/O operations.

Each operation is described by an aiocb structure in the array list.

The mode parameter determines when the function will return. If mode is LIO_WAIT the function returns when the I/O operation have completed, if mode is LIO_NOWAIT the function returns after enqueueing the operations.

If mode is LIO_NOWAIT, the sigevent struct pointed by sig is used to notify list completion.

NOTES:

When the mode is LIO_NOWAIT and the sigev_notify field of sig is set to SIGEV_SIGNAL, a signal is sent to the process to notify the completion of the list. Since each RTEMS application is logically a single POSIX process, if the user wants to wait for the signal (using, for example, sigwait()), it is necessary to ensure that the signal is blocked by every thread.

This function is only available when RTEMS_POSIX_API is defined. To do so, it’s necessary to add RTEMS_POSIX_API = True to the config.ini file.

6.4.19. aio_error - Retrieve Error Status of Asynchronous I/O Operation¶

CALLING SEQUENCE:

#include <aio.h>
int aio_error(
    const struct aiocb *aiocbp
);

STATUS CODES:

The function return the error status of the request, if 0 is returned the operation completed without errors.

If the request is still in progress, the function returns EINPROGRESS.

On error, this routine returns -1 and sets errno to one of the following:

`EINVAL`	The return status for the request has already been retrieved.
`EINVAL`	`aiocbp` is a NULL pointer.

DESCRIPTION:

The aio_error() function retrieves the error status of the request.

aiocbp is a pointer to the request control block.

NOTES:

NONE .. _aio_return:

6.4.20. aio_return - Retrieve Return Status of Asynchronous I/O Operation¶

CALLING SEQUENCE:

#include <aio.h>
ssize_t aio_return(
    struct aiocb *aiocbp
);

STATUS CODES:

If the result can be returned, it is returned as defined by the various operations.

On error, this routine returns -1 and sets errno to one of the following:

`EINVAL`	The return status for the request has already been retrieved.
`EINVAL`	`aiocbp` is a NULL pointer.

DESCRIPTION:

The aio_return() function retrieves the return status of the asynchronous I/O operation. aiocbp is a pointer to the request control block.

NOTES:

NONE

6.4.21. aio_cancel - Cancel Asynchronous I/O Request¶

CALLING SEQUENCE:

#include <aio.h>
int aio_cancel(
    int fildes,
    struct aiocb *aiocbp
);

STATUS CODES:

If the function terminated without errors, the return value has one of the following values:

`AIO_CANCELED`	The requested operation(s) were canceled.
`AIO_NOTCANCELED`	Some operations could not be canceled because they are in progress.
`AIO_ALLDONE`	None of the operations could be canceled because they are already complete.

If the file descriptor is invalid, -1 is returned and errno is set to EBADF

DESCRIPTION:

The aio_cancel() function attempts to cancel asynchronous I/O operations.

filedes is the file descriptor associated with the operations to be canceled. aiocbp is a pointer to an asynchronous I/O control block.

If aiocbp is NULL, the function will attempt to eliminate all the operations enqueued for the specified filedes.

If aiocbp points to a control block, then only the referenced operation shall be eliminated. The aio_filedef value of aiocbp must be equal to filedes, otherwise the function will return with an error.

NOTES:

NONE

6.4.22. aio_suspend - Wait for Asynchronous I/O Request¶

CALLING SEQUENCE:

#include <aio.h>
int aio_suspend(
    const struct aiocb *const list[],
    int nent,
    const struct timespec *timeout
);

STATUS CODES:

On success, zero is returned. On error, -1 is returned, and errno is set appropriately.

E

The

DESCRIPTION:

The aio_suspend() function suspends the calling process until one or more operations have completed or until the specified timeout has expired. list contains the requests that must complete.

NOTES:

This routine is not currently supported by RTEMS.

6.4.23. aio_fsync - Asynchronous File Synchronization¶

CALLING SEQUENCE:

#include <aio.h>
int aio_fsync(
    int op,
    struct aiocb *aiocbp
);

STATUS CODES:

If the requests are succesfully enqueued, zero is returned. On error, this routine returns -1 and sets errno to one of the following:

`EAGAIN`	The operation could not be queued due to temporary resource limitations.
`EAGAIN`	the addition of a new request to the queue would violate the `RTEMS_AIO_MAX` limit.
`EBADF`	The aio_fildes member of aiocbp is not a valid file descriptor.
`EINVAL`	A value of op other than O_SYNC or O_DSYNC was specified.
`EINVAL`	`aiocbp->aio_sigevent` does not point to a valid sigevent struct.
`EINVAL`	`aiocbp` is a NULL pointer.

DESCRIPTION:

The aio_fsync() function initiates an asynchronous file sync operation. op specifies what kind of synchronization should be performed. If op is O_SYNC, all currently queued I/O operations shall be synchronized as if by a call to fsync(). If op is O_DSYNC, all currently queued I/O operations shall be synchronized as if by a call to fdatasync().

NOTES:

Currently, O_DSYNC and O_SYNC are mapped to the same value. As a result, every file will be synced as if by a call to fsync().