Filesystem: Reference counting for locations

 o A new data structure rtems_filesystem_global_location_t was
   introduced to be used for
    o the mount point location in the mount table entry,
    o the file system root location in the mount table entry,
    o the root directory location in the user environment, and
    o the current directory location in the user environment.
   During the path evaluation global start locations are obtained to
   ensure that the current file system instance will be not unmounted in
   the meantime.
 o The user environment uses now reference counting and is protected
   from concurrent access.
 o The path evaluation process was completely rewritten and simplified.
   The IMFS, RFS, NFS, and DOSFS use now a generic path evaluation
   method.  Recursive calls in the path evaluation have been replaced
   with iteration to avoid stack overflows.  Only the evaluation of
   symbolic links is recursive.  No dynamic memory allocations and
   intermediate buffers are used in the high level path evaluation.  No
   global locks are held during the file system instance specific path
   evaluation process.
 o Recursive symbolic link evaluation is now limited by
   RTEMS_FILESYSTEM_SYMLOOP_MAX.  Applications can retrieve this value
   via sysconf().
 o The device file system (devFS) uses now no global variables and
   allocation from the workspace.  Node names are allocated from the
   heap.
 o The upper layer lseek() performs now some parameter checks.
 o The upper layer ftruncate() performs now some parameter checks.
 o unmask() is now restricted to the RWX flags and protected from
   concurrent access.
 o The fchmod_h and rmnod_h file system node handlers are now a file
   system operation.
 o The unlink_h operation has been removed.  All nodes are now destroyed
   with the rmnod_h operation.
 o New lock_h, unlock_h, clonenod_h, and are_nodes_equal_h file system
   operations.
 o The path evaluation and file system operations are now protected by
   per file system instance lock and unlock operations.
 o Fix and test file descriptor duplicate in fcntl().
 o New test fstests/fsnofs01.

1 files changed, 107 insertions, 154 deletions

diff --git a/cpukit/libcsupport/src/open.c b/cpukit/libcsupport/src/open.c
index a2e1b507e3..6fa7783322 100644
--- a/cpukit/libcsupport/src/open.c
+++ b/cpukit/libcsupport/src/open.c
@@ -4,6 +4,9 @@
  *  COPYRIGHT (c) 1989-2010.
  *  On-Line Applications Research Corporation (OAR).
  *
+ *  Modifications to support reference counting in the file system are
+ *  Copyright (c) 2012 embedded brains GmbH.
+ *
  *  The license and distribution terms for this file may be
  *  found in the file LICENSE in this distribution or at
  *  http://www.rtems.com/license/LICENSE.
@@ -12,185 +15,135 @@
  */
 
 #if HAVE_CONFIG_H
-#include "config.h"
+  #include "config.h"
 #endif
 
-#include <stdarg.h>
+#include <sys/stat.h>
 #include <fcntl.h>
-#include <rtems/libio_.h>
-#include <rtems/seterr.h>
-
+#include <stdarg.h>
 #include <unistd.h>
 
-/*
- *  Returns file descriptor on success or -1 and errno set to one of the
- *  following:
- *
- *    EACCESS  - Seach permission is denied on a component of the path prefix,
- *               or the file exists and the permissions specified by the
- *               flags are denied, or the file does not exist and write
- *               permission is denied for the parent directory of the file
- *               to be created, or O_TRUNC is specified and write permission
- *               is denied.
- *    EEXIST   - O_CREAT and O_EXCL are set and the named file exists.
- *    EINTR    - The open( operation was interrupted by a signal.
- *    EINVAL   - This implementation does not support synchronized IO for this
- *               file.
- *    EISDIR   - The named file is a directory and the flags argument
- *               specified write or read/write access.
- *    EMFILE   - Too many file descriptors are in used by this process.
- *    ENAMETOOLONG -
- *               The length of the path exceeds PATH_MAX or a pathname
- *               component is longer than NAME_MAX while POSIX_NO_TRUNC
- *               is in effect.
- *    ENFILE   - Too many files are open in the system.
- *    ENOENT   - O_CREAT is not set and and the anmed file does not exist,
- *               or O_CREAT is set and either the path prefix does not exist
- *               or the path argument points to an empty string.
- *    ENOSPC   - The directory or file system that would contain the new file
- *               cannot be extended.
- *    ENOTDIR  - A component of the path prefix is not a directory.
- *    ENXIO    - O_NONBLOCK is set, the named file is a FIFO, O_WRONLY is
- *               set, and no process has the file open for reading.
- *    EROFS    - The named file resides on a read-only file system and either
- *               O_WRONLY, O_RDWR, O_CREAT (if the file does not exist), or
- *               O_TRUNC is set in the flags argument.
- */
+#include <rtems/libio_.h>
 
-int open(
-  const char   *pathname,
-  int           flags,
-  ...
+static void create_regular_file(
+  rtems_filesystem_eval_path_context_t *ctx,
+  mode_t mode
 )
 {
-  va_list                             ap;
-  mode_t                              mode;
-  int                                 rc;
-  rtems_libio_t                      *iop = 0;
-  int                                 status;
-  rtems_filesystem_location_info_t    loc;
-  rtems_filesystem_location_info_t   *loc_to_free = NULL;
-  int                                 eval_flags;
-
-  /*
-   * Set the Evaluation flags
-   */
-  eval_flags = 0;
-  status = flags + 1;
-  if ( ( status & _FREAD ) == _FREAD )
-    eval_flags |= RTEMS_LIBIO_PERMS_READ;
-  if ( ( status & _FWRITE ) == _FWRITE )
-    eval_flags |= RTEMS_LIBIO_PERMS_WRITE;
-
-  va_start(ap, flags);
-
-  mode = va_arg( ap, mode_t );
+  int rv = 0;
+  const rtems_filesystem_location_info_t *currentloc = 
+    rtems_filesystem_eval_path_get_currentloc( ctx );
+  const char *token = rtems_filesystem_eval_path_get_token( ctx );
+  size_t tokenlen = rtems_filesystem_eval_path_get_tokenlen( ctx );
+
+  rv = rtems_filesystem_mknod(
+    currentloc,
+    token,
+    tokenlen,
+    S_IFREG | mode,
+    0
+  );
+
+  if ( rv == 0 ) {
+    /* The mode only applies to future accesses of the newly created file */
+    rtems_filesystem_eval_path_set_flags( ctx, 0 );
+
+    rtems_filesystem_eval_path_set_path( ctx, token, tokenlen );
+    rtems_filesystem_eval_path_continue( ctx );
+  } else {
+    rtems_filesystem_eval_path_error( ctx, 0 );
+  }
+}
 
-  /*
-   * NOTE: This comment is OBSOLETE.  The proper way to do this now
-   *       would be to support a magic mounted file system.
-   *
-   *             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().
-   */
-
-  /* allocate a file control block */
-  iop = rtems_libio_allocate();
-  if ( iop == 0 ) {
-    rc = ENFILE;
-    goto done;
+static int do_open(
+  rtems_libio_t *iop,
+  const char *path,
+  int oflag,
+  mode_t mode
+)
+{
+  int rv = 0;
+  int fd = iop - rtems_libio_iops;
+  int rwflag = oflag + 1;
+  bool read_access = (rwflag & _FREAD) == _FREAD;
+  bool write_access = (rwflag & _FWRITE) == _FWRITE;
+  bool make = (oflag & O_CREAT) == O_CREAT;
+  bool exclusive = (oflag & (O_CREAT | O_EXCL)) == (O_CREAT | O_EXCL);
+  bool truncate = (oflag & O_TRUNC) == O_TRUNC;
+  int eval_flags = RTEMS_LIBIO_FOLLOW_LINK
+    | (read_access ? RTEMS_LIBIO_PERMS_READ : 0)
+    | (write_access ? RTEMS_LIBIO_PERMS_WRITE : 0)
+    | (make ? RTEMS_LIBIO_MAKE : 0)
+    | (exclusive ?  RTEMS_LIBIO_EXCLUSIVE : 0);
+  rtems_filesystem_eval_path_context_t ctx;
+
+  rtems_filesystem_eval_path_start( &ctx, path, eval_flags );
+
+  if ( rtems_filesystem_eval_path_has_token( &ctx ) ) {
+    create_regular_file( &ctx, mode );
   }
 
-  /*
-   *  See if the file exists.
-   */
-  status = rtems_filesystem_evaluate_path(
-    pathname, strlen( pathname ), eval_flags, &loc, true );
+  if ( write_access ) {
+    const rtems_filesystem_location_info_t *currentloc =
+      rtems_filesystem_eval_path_get_currentloc( &ctx );
+    rtems_filesystem_node_types_t type =
+      (*currentloc->ops->node_type_h)( currentloc );
 
-  if ( status == -1 ) {
-    if ( errno != ENOENT ) {
-      rc = errno;
-      goto done;
+    if ( type == RTEMS_FILESYSTEM_DIRECTORY ) {
+      rtems_filesystem_eval_path_error( &ctx, EISDIR );
     }
+  }
 
-    /* If the file does not exist and we are not trying to create it--> error */
-    if ( !(flags & O_CREAT) ) {
-      rc = ENOENT;
-      goto done;
-    }
+  iop->flags |= rtems_libio_fcntl_flags( oflag );
+  rtems_filesystem_eval_path_extract_currentloc( &ctx, &iop->pathinfo );
+  rtems_filesystem_eval_path_cleanup( &ctx );
 
-    /* Create the node for the new regular file */
-    rc = mknod( pathname, S_IFREG | mode, 0LL );
-    if ( rc ) {
-      rc = errno;
-      goto done;
+  rv = (*iop->pathinfo.handlers->open_h)( iop, path, oflag, mode );
+
+  if ( rv == 0 ) {
+    if ( truncate ) {
+      rv = ftruncate( fd, 0 );
+      if ( rv != 0 ) {
+        (*iop->pathinfo.handlers->close_h)( iop );
+      }
     }
 
-    /*
-     * After we do the mknod(), we have to evaluate the path to get the
-     * "loc" structure needed to actually have the file itself open.
-     * So we created it, and then we need to have "look it up."
-     */
-    status = rtems_filesystem_evaluate_path(
-      pathname, strlen( pathname ), 0x0, &loc, true );
-    if ( status != 0 ) {   /* The file did not exist */
-      rc = EACCES;
-      goto done;
+    if ( rv == 0 ) {
+      rv = fd;
+    } else {
+      rv = -1;
     }
+  }
 
-  } else if ((flags & (O_EXCL|O_CREAT)) == (O_EXCL|O_CREAT)) {
-    /* We were trying to create a file that already exists */
-    rc = EEXIST;
-    loc_to_free = &loc;
-    goto done;
+  if ( rv < 0 ) {
+    rtems_libio_free( iop );
   }
 
-  loc_to_free = &loc;
+  return rv;
+}
+
+int open( const char *path, int oflag, ... )
+{
+  int rv = 0;
+  va_list ap;
+  mode_t mode = 0;
+  rtems_libio_t *iop = NULL;
 
-  /*
-   *  Fill in the file control block based on the loc structure
-   *  returned by successful path evaluation.
-   */
-  iop->flags     |= rtems_libio_fcntl_flags( flags );
-  iop->pathinfo   = loc;
+  va_start( ap, oflag );
 
-  rc = (*iop->pathinfo.handlers->open_h)( iop, pathname, flags, mode );
-  if ( rc ) {
-    rc = errno;
-    goto done;
-  }
+  mode = va_arg( ap, mode_t );
 
-  /*
-   *  Optionally truncate the file.
-   */
-  if ( (flags & O_TRUNC) == O_TRUNC ) {
-    rc = ftruncate( iop - rtems_libio_iops, 0 );
-    if ( rc ) {
-      if(errno) rc = errno;
-      close( iop - rtems_libio_iops );
-      /* those are released by close(): */
-      iop = 0;
-      loc_to_free = NULL;
-    }
+  iop = rtems_libio_allocate();
+  if ( iop != NULL ) {
+    rv = do_open( iop, path, oflag, mode );
+  } else {
+    errno = ENFILE;
+    rv = -1;
   }
 
-  /*
-   *  Single exit and clean up path.
-   */
-done:
-  va_end(ap);
-
-  if ( rc ) {
-    if ( iop )
-      rtems_libio_free( iop );
-    if ( loc_to_free )
-      rtems_filesystem_freenode( loc_to_free );
-    rtems_set_errno_and_return_minus_one( rc );
-  }
+  va_end( ap );
 
-  return iop - rtems_libio_iops;
+  return rv;
 }
 
 /*
@@ -206,10 +159,10 @@ done:
 int _open_r(
   struct _reent *ptr __attribute__((unused)),
   const char    *buf,
-  int            flags,
+  int            oflag,
   int            mode
 )
 {
-  return open( buf, flags, mode );
+  return open( buf, oflag, mode );
 }
 #endif