RTOS/rtems
1
0
Fork 0
forked from Imagelibrary/rtems
Code Pull Requests Activity

Added base version of file system infrastructure. This includes a major

Browse Source 
overhaul of the RTEMS system call interface.  This base file system is
the "In-Memory File System" aka IMFS.

The design and implementation was done by the following people:

  + Joel Sherrill (joel@OARcorp.com)
  + Jennifer Averett (jennifer@OARcorp.com)
  + Steve "Mr Mount" Salitasc (salitasc@OARcorp.com)
  + Kerwin Wade (wade@OARcorp.com)

PROBLEMS
========
  + It is VERY likely that merging this will break the UNIX port.  This
    can/will be fixed.

  + There is likely some reentrancy/mutual exclusion needed.

  + Eventually, there should be a "mini-IMFS" description table to
    eliminate links, symlinks, etc to save memory.  All you need to
    have "classic RTEMS" functionality is technically directories
    and device IO.  All the rest could be left out to save memory.
This commit is contained in:
Joel Sherrill
1998-11-23 19:07:58 +00:00
parent cd3fb80a26
commit 07a3253de2
330 changed files with 31931 additions and 2593 deletions
Download Patch File Download Diff File Expand all files Collapse all files

715
c/src/lib/libc/libio.c
View File

@@ -1,524 +1,295 @@
/*
* Provide UNIX/POSIX-like io system calls for RTEMS using the
* RTEMS IO manager
* This file contains the support infrastructure used to manage the
* table of integer style file descriptors used by the low level
* POSIX system calls like open(), read, fstat(), etc.
*
* This provides the foundation for POSIX compliant IO system calls
* for RTEMS.
*
* COPYRIGHT (c) 1989-1998.
* On-Line Applications Research Corporation (OAR).
* Copyright assigned to U.S. Government, 1994.
*
* The license and distribution terms for this file may be
* found in the file LICENSE in this distribution or at
* http://www.OARcorp.com/rtems/license.html.
*
* $Id$
*/
#include <rtems.h>
#include <rtems/assoc.h> /* assoc.h not included by rtems.h */
#include <stdio.h> /* O_RDONLY, et.al. */
#include <fcntl.h> /* O_RDONLY, et.al. */
#include <assert.h>
#if ! defined(O_NDELAY)
# if defined(solaris2)
# define O_NDELAY O_NONBLOCK
# elif defined(RTEMS_NEWLIB)
# define O_NDELAY _FNBIO
# endif
#endif
#include <errno.h>
#include <string.h> /* strcmp */
#include <unistd.h>
#include <stdlib.h> /* calloc() */
#include "libio.h" /* libio.h not pulled in by rtems */
#include "libio_.h" /* libio_.h pulls in rtems */
/*
* Semaphore to protect the io table
* Global variables used to manage the File Descriptor Table.
* IOP = IO Pointer.
*/
Objects_Id rtems_libio_semaphore;
#define RTEMS_LIBIO_SEM rtems_build_name('L', 'B', 'I', 'O')
#define RTEMS_LIBIO_IOP_SEM(n) rtems_build_name('L', 'B', 'I', n)
extern unsigned32 rtems_libio_number_iops;
rtems_libio_t *rtems_libio_iops;
rtems_libio_t *rtems_libio_last_iop;
#define rtems_libio_iop(fd) ((((unsigned32)(fd)) < rtems_libio_number_iops) ? \
&rtems_libio_iops[fd] : 0)
#define rtems_libio_check_fd(fd) \
do { \
if ((unsigned32) (fd) >= rtems_libio_number_iops) \
{ \
errno = EBADF; \
return -1; \
} \
} while (0)
#define rtems_libio_check_buffer(buffer) \
do { \
if ((buffer) == 0) \
{ \
errno = EINVAL; \
return -1; \
} \
} while (0)
#define rtems_libio_check_count(count) \
do { \
if ((count) == 0) \
{ \
return 0; \
} \
} while (0)
#define rtems_libio_check_permissions(iop, flag) \
do { \
if (((iop)->flags & (flag)) == 0) \
{ \
errno = EINVAL; \
return -1; \
} \
} while (0)
rtems_id rtems_libio_semaphore;
rtems_libio_t *rtems_libio_iops;
rtems_libio_t *rtems_libio_last_iop;
rtems_libio_handler_t rtems_libio_handlers[15];
/*
* External I/O handlers
*
* Space for all possible handlers is preallocated
* to speed up dispatch to external handlers.
* rtems_register_libio_handler
*
* This function registers an external IO handler set. This lets
* other subsystems have their own versions of many of the system
* calls. For example, the networking code registers handlers which
* map the system calls for read() and write() to socket calls.
*
*/
static rtems_libio_handler_t handlers[15];
void
rtems_register_libio_handler(
int handler_flag,
const rtems_libio_handler_t *handler
void rtems_register_libio_handler(
int handler_flag,
const rtems_libio_handler_t *handler
)
{
int handler_index = rtems_file_descriptor_type_index(handler_flag);
int handler_index = rtems_file_descriptor_type_index( handler_flag );
if ((handler_index < 0) || (handler_index >= 15))
rtems_fatal_error_occurred( RTEMS_INVALID_NUMBER );
handlers[handler_index] = *handler;
rtems_libio_handlers[handler_index] = *handler;
}
/*
* Called by bsp startup code to init the libio area.
* rtems_libio_init
*
* Called by BSP startup code to initialize the libio subsystem.
*/
void
rtems_libio_init(void)
void rtems_libio_init( void )
{
rtems_status_code rc;
rtems_status_code rc;
if (rtems_libio_number_iops > 0)
{
rtems_libio_iops = (rtems_libio_t *) calloc(rtems_libio_number_iops,
sizeof(rtems_libio_t));
if (rtems_libio_iops == NULL)
rtems_fatal_error_occurred(RTEMS_NO_MEMORY);
/*
* Allocate memory for the IOP Table
*/
rtems_libio_last_iop = rtems_libio_iops + (rtems_libio_number_iops - 1);
}
if ( rtems_libio_number_iops > 0 ) {
rtems_libio_iops =
(rtems_libio_t *) calloc(rtems_libio_number_iops, sizeof(rtems_libio_t));
rc = rtems_semaphore_create(
RTEMS_LIBIO_SEM,
1,
RTEMS_BINARY_SEMAPHORE | RTEMS_INHERIT_PRIORITY | RTEMS_PRIORITY,
RTEMS_NO_PRIORITY,
&rtems_libio_semaphore
);
if (rc != RTEMS_SUCCESSFUL)
rtems_fatal_error_occurred(rc);
if (rtems_libio_iops == NULL)
rtems_fatal_error_occurred( RTEMS_NO_MEMORY );
rtems_libio_last_iop = rtems_libio_iops + (rtems_libio_number_iops - 1);
}
/*
* Create the binary semaphore used to provide mutual exclusion
* on the IOP Table.
*/
rc = rtems_semaphore_create(
RTEMS_LIBIO_SEM,
1,
RTEMS_BINARY_SEMAPHORE | RTEMS_INHERIT_PRIORITY | RTEMS_PRIORITY,
RTEMS_NO_PRIORITY,
&rtems_libio_semaphore
);
if ( rc != RTEMS_SUCCESSFUL )
rtems_fatal_error_occurred( rc );
/*
* Initialize the base file system infrastructure.
*/
rtems_filesystem_initialize();
}
/*
* Convert RTEMS status to a UNIX errno
*/
rtems_assoc_t errno_assoc[] = {
{ "OK", RTEMS_SUCCESSFUL, 0 },
{ "BUSY", RTEMS_RESOURCE_IN_USE, EBUSY },
{ "INVALID NAME", RTEMS_INVALID_NAME, EINVAL },
{ "NOT IMPLEMENTED", RTEMS_NOT_IMPLEMENTED, ENOSYS },
{ "TIMEOUT", RTEMS_TIMEOUT, ETIMEDOUT },
{ "NO MEMORY", RTEMS_NO_MEMORY, ENOMEM },
{ "NO DEVICE", RTEMS_UNSATISFIED, ENODEV },
{ "INVALID NUMBER", RTEMS_INVALID_NUMBER, EBADF},
{ "NOT RESOURCE OWNER", RTEMS_NOT_OWNER_OF_RESOURCE, EPERM},
{ "IO ERROR", RTEMS_IO_ERROR, EIO},
{ 0, 0, 0 },
};
static unsigned32
rtems_libio_errno(rtems_status_code code)
{
int rc;
if ((rc = rtems_assoc_remote_by_local(errno_assoc, (unsigned32) code)))
{
errno = rc;
return -1;
}
return -1;
}
/*
* Convert UNIX fnctl(2) flags to ones that RTEMS drivers understand
* rtems_libio_fcntl_flags
*
* Convert UNIX fnctl(2) flags to ones that RTEMS drivers understand
*/
rtems_assoc_t access_modes_assoc[] = {
{ "READ", LIBIO_FLAGS_READ, O_RDONLY },
{ "WRITE", LIBIO_FLAGS_WRITE, O_WRONLY },
{ "READ/WRITE", LIBIO_FLAGS_READ_WRITE, O_RDWR },
{ 0, 0, 0 },
{ "READ", LIBIO_FLAGS_READ, O_RDONLY },
{ "WRITE", LIBIO_FLAGS_WRITE, O_WRONLY },
{ "READ/WRITE", LIBIO_FLAGS_READ_WRITE, O_RDWR },
{ 0, 0, 0 },
};
rtems_assoc_t status_flags_assoc[] = {
{ "NO DELAY", LIBIO_FLAGS_NO_DELAY, O_NDELAY },
{ "APPEND", LIBIO_FLAGS_APPEND, O_APPEND },
{ "CREATE", LIBIO_FLAGS_CREATE, O_CREAT },
{ 0, 0, 0 },
{ "NO DELAY", LIBIO_FLAGS_NO_DELAY, O_NDELAY },
{ "APPEND", LIBIO_FLAGS_APPEND, O_APPEND },
{ "CREATE", LIBIO_FLAGS_CREATE, O_CREAT },
{ 0, 0, 0 },
};
static unsigned32
rtems_libio_fcntl_flags(unsigned32 fcntl_flags)
unsigned32 rtems_libio_fcntl_flags(
unsigned32 fcntl_flags
)
{
unsigned32 flags = 0;
unsigned32 access_modes;
unsigned32 flags = 0;
unsigned32 access_modes;
/*
* Access mode is a small integer
*/
access_modes = fcntl_flags & O_ACCMODE;
fcntl_flags &= ~O_ACCMODE;
flags = rtems_assoc_local_by_remote(access_modes_assoc, access_modes);
/*
* Everything else is single bits
*/
flags |= rtems_assoc_local_by_remote_bitfield(status_flags_assoc, fcntl_flags);
return flags;
}
static rtems_libio_t *
rtems_libio_allocate(void)
{
rtems_libio_t *iop;
rtems_status_code rc;
rtems_semaphore_obtain(rtems_libio_semaphore, RTEMS_WAIT, RTEMS_NO_TIMEOUT);
for (iop = rtems_libio_iops; iop <= rtems_libio_last_iop; iop++)
if ((iop->flags & LIBIO_FLAGS_OPEN) == 0)
{
/*
* Got one; create a semaphore for it
*/
rc = rtems_semaphore_create(
RTEMS_LIBIO_IOP_SEM(iop - rtems_libio_iops),
1,
RTEMS_BINARY_SEMAPHORE | RTEMS_INHERIT_PRIORITY | RTEMS_PRIORITY,
RTEMS_NO_PRIORITY,
&iop->sem
);
if (rc != RTEMS_SUCCESSFUL)
goto failed;
iop->flags = LIBIO_FLAGS_OPEN;
goto done;
}
failed:
iop = 0;
done:
rtems_semaphore_release(rtems_libio_semaphore);
return iop;
}
static void
rtems_libio_free(rtems_libio_t *iop)
{
rtems_semaphore_obtain(rtems_libio_semaphore, RTEMS_WAIT, RTEMS_NO_TIMEOUT);
if (iop->sem)
rtems_semaphore_delete(iop->sem);
(void) memset(iop, 0, sizeof(*iop));
rtems_semaphore_release(rtems_libio_semaphore);
}
int
__rtems_open(
const char *pathname,
unsigned32 flag,
unsigned32 mode)
{
rtems_status_code rc;
rtems_libio_t *iop = 0;
rtems_driver_name_t *np;
rtems_libio_open_close_args_t args;
/*
* Additional external I/O handlers would be supported by
* adding code to pick apart the pathname appropriately.
* The networking code does not require changes here since
* network file descriptors are obtained using socket(), not
* open().
*/
if ((rc = rtems_io_lookup_name(pathname, &np)) != RTEMS_SUCCESSFUL)
goto done;
iop = rtems_libio_allocate();
if (iop == 0)
{
rc = RTEMS_TOO_MANY;
goto done;
}
iop->driver = np;
iop->pathname = (char *) pathname;
iop->flags |= rtems_libio_fcntl_flags(flag);
args.iop = iop;
args.flags = iop->flags;
args.mode = mode;
rc = rtems_io_open(np->major, np->minor, (void *) &args);
done:
/*
* Access mode is a small integer
*/
if (rc != RTEMS_SUCCESSFUL)
{
if (iop)
rtems_libio_free(iop);
return rtems_libio_errno(rc);
}
return iop - rtems_libio_iops;
}
int
__rtems_close(
int fd
)
{
rtems_status_code rc;
rtems_driver_name_t *np;
rtems_libio_t *iop;
rtems_libio_open_close_args_t args;
int status;
access_modes = fcntl_flags & O_ACCMODE;
fcntl_flags &= ~O_ACCMODE;
flags = rtems_assoc_local_by_remote( access_modes_assoc, access_modes );
if (rtems_file_descriptor_type(fd)) {
int (*fp)(int fd);
/*
* Everything else is single bits
*/
fp = handlers[rtems_file_descriptor_type_index(fd)].close;
if (fp == NULL) {
errno = EBADF;
return -1;
}
status = (*fp)(fd);
return status;
}
iop = rtems_libio_iop(fd);
rtems_libio_check_fd(fd);
np = iop->driver;
args.iop = iop;
args.flags = 0;
args.mode = 0;
rc = rtems_io_close(np->major, np->minor, (void *) &args);
rtems_libio_free(iop);
if (rc != RTEMS_SUCCESSFUL)
return rtems_libio_errno(rc);
return 0;
}
int
__rtems_read(
int fd,
void * buffer,
unsigned32 count
)
{
rtems_status_code rc;
rtems_driver_name_t *np;
rtems_libio_t *iop;
rtems_libio_rw_args_t args;
if (rtems_file_descriptor_type(fd)) {
int (*fp)(int fd, void *buffer, unsigned32 count);
fp = handlers[rtems_file_descriptor_type_index(fd)].read;
if (fp == NULL) {
errno = EBADF;
return -1;
}
return (*fp)(fd, buffer, count);
}
iop = rtems_libio_iop(fd);
rtems_libio_check_fd(fd);
rtems_libio_check_buffer(buffer);
rtems_libio_check_count(count);
rtems_libio_check_permissions(iop, LIBIO_FLAGS_READ);
np = iop->driver;
args.iop = iop;
args.offset = iop->offset;
args.buffer = buffer;
args.count = count;
args.flags = iop->flags;
args.bytes_moved = 0;
rc = rtems_io_read(np->major, np->minor, (void *) &args);
iop->offset += args.bytes_moved;
if (rc != RTEMS_SUCCESSFUL)
return rtems_libio_errno(rc);
return args.bytes_moved;
flags |=
rtems_assoc_local_by_remote_bitfield(status_flags_assoc, fcntl_flags);
return flags;
}
int
__rtems_write(
int fd,
const void *buffer,
unsigned32 count
)
{
rtems_status_code rc;
rtems_driver_name_t *np;
rtems_libio_t *iop;
rtems_libio_rw_args_t args;
if (rtems_file_descriptor_type(fd)) {
int (*fp)(int fd, const void *buffer, unsigned32 count);
fp = handlers[rtems_file_descriptor_type_index(fd)].write;
if (fp == NULL) {
errno = EBADF;
return -1;
}
return (*fp)(fd, buffer, count);
}
iop = rtems_libio_iop(fd);
rtems_libio_check_fd(fd);
rtems_libio_check_buffer(buffer);
rtems_libio_check_count(count);
rtems_libio_check_permissions(iop, LIBIO_FLAGS_WRITE);
np = iop->driver;
args.iop = iop;
args.offset = iop->offset;
args.buffer = (void *) buffer;
args.count = count;
args.flags = iop->flags;
args.bytes_moved = 0;
rc = rtems_io_write(np->major, np->minor, (void *) &args);
iop->offset += args.bytes_moved;
if (rc != RTEMS_SUCCESSFUL)
return rtems_libio_errno(rc);
return args.bytes_moved;
}
int
__rtems_ioctl(
int fd,
unsigned32 command,
void * buffer)
{
rtems_status_code rc;
rtems_driver_name_t *np;
rtems_libio_t *iop;
rtems_libio_ioctl_args_t args;
if (rtems_file_descriptor_type(fd)) {
int (*fp)(int fd, unsigned32 command, void *buffer);
fp = handlers[rtems_file_descriptor_type_index(fd)].ioctl;
if (fp == NULL) {
errno = EBADF;
return -1;
}
return (*fp)(fd, command, buffer);
}
iop = rtems_libio_iop(fd);
rtems_libio_check_fd(fd);
np = iop->driver;
args.iop = iop;
args.command = command;
args.buffer = buffer;
rc = rtems_io_control(np->major, np->minor, (void *) &args);
if (rc != RTEMS_SUCCESSFUL)
return rtems_libio_errno(rc);
return args.ioctl_return;
}
/*
* internal only??
* rtems_libio_allocate
*
* This routine searches the IOP Table for an unused entry. If it
* finds one, it returns it. Otherwise, it returns NULL.
*/
int
__rtems_lseek(
int fd,
rtems_libio_offset_t offset,
int whence
)
rtems_libio_t *rtems_libio_allocate( void )
{
rtems_libio_t *iop;
rtems_libio_t *iop;
rtems_status_code rc;
rtems_semaphore_obtain( rtems_libio_semaphore, RTEMS_WAIT, RTEMS_NO_TIMEOUT );
if (rtems_file_descriptor_type(fd)) {
int (*fp)(int fd, rtems_libio_offset_t offset, int whence);
for (iop = rtems_libio_iops; iop <= rtems_libio_last_iop; iop++)
if ((iop->flags & LIBIO_FLAGS_OPEN) == 0) {
/*
* Got an IOP -- create a semaphore for it.
*/
fp = handlers[rtems_file_descriptor_type_index(fd)].lseek;
if (fp == NULL) {
errno = EBADF;
return -1;
}
return (*fp)(fd, offset, whence);
}
iop = rtems_libio_iop(fd);
rtems_libio_check_fd(fd);
switch (whence)
{
case SEEK_SET:
iop->offset = offset;
break;
case SEEK_CUR:
iop->offset += offset;
break;
case SEEK_END:
iop->offset = iop->size - offset;
break;
default:
errno = EINVAL;
return -1;
}
return 0;
rc = rtems_semaphore_create(
RTEMS_LIBIO_IOP_SEM(iop - rtems_libio_iops),
1,
RTEMS_BINARY_SEMAPHORE | RTEMS_INHERIT_PRIORITY | RTEMS_PRIORITY,
RTEMS_NO_PRIORITY,
&iop->sem
);
if ( rc != RTEMS_SUCCESSFUL )
goto failed;
iop->flags = LIBIO_FLAGS_OPEN;
goto done;
}
failed:
iop = 0;
done:
rtems_semaphore_release( rtems_libio_semaphore );
return iop;
}
/*
* rtems_libio_free
*
* This routine frees the resources associated with an IOP (file descriptor)
* and clears the slot in the IOP Table.
*/
void rtems_libio_free(
rtems_libio_t *iop
)
{
rtems_semaphore_obtain( rtems_libio_semaphore, RTEMS_WAIT, RTEMS_NO_TIMEOUT );
if (iop->sem)
rtems_semaphore_delete(iop->sem);
(void) memset(iop, 0, sizeof(*iop));
rtems_semaphore_release( rtems_libio_semaphore );
}
/*
* rtems_libio_is_open_files_in_fs
*
* This routine scans the entire file descriptor table to determine if the
* are any active file descriptors that refer to the atleast one node in the
* file system that we are trying to dismount.
*
* If there is at least one node in the file system referenced by the mount
* table entry a 1 is returned, otherwise a 0 is returned.
*/
int rtems_libio_is_open_files_in_fs(
rtems_filesystem_mount_table_entry_t * fs_mt_entry
)
{
rtems_libio_t *iop;
int result = 0;
rtems_semaphore_obtain( rtems_libio_semaphore, RTEMS_WAIT, RTEMS_NO_TIMEOUT );
/*
* Look for any active file descriptor entry.
*/
for ( iop=rtems_libio_iops ; iop <= rtems_libio_last_iop ; iop++ ) {
if ((iop->flags & LIBIO_FLAGS_OPEN) != 0) {
/*
* Check if this node is under the file system that we
* are trying to dismount.
*/
if ( iop->pathinfo.mt_entry == fs_mt_entry ) {
result = 1;
break;
}
}
}
rtems_semaphore_release( rtems_libio_semaphore );
return result;
}
/*
* rtems_libio_is_file_open
*
* This routine scans the entire file descriptor table to determine if the
* given file refers to an active file descriptor.
*
* If the given file is open a 1 is returned, otherwise a 0 is returned.
*/
int rtems_libio_is_file_open(
void *node_access
)
{
rtems_libio_t *iop;
int result=0;
rtems_semaphore_obtain( rtems_libio_semaphore, RTEMS_WAIT, RTEMS_NO_TIMEOUT );
/*
* Look for any active file descriptor entry.
*/
for ( iop=rtems_libio_iops ; iop <= rtems_libio_last_iop ; iop++ ) {
if ((iop->flags & LIBIO_FLAGS_OPEN) != 0) {
/*
* Check if this node is under the file system that we
* are trying to dismount.
*/
if ( iop->pathinfo.node_access == node_access ) {
result = 1;
break;
}
}
}
rtems_semaphore_release( rtems_libio_semaphore );
return result;
}