path: root/fs/umsdos/namei.c

/*
 *  linux/fs/umsdos/namei.c
 *
 *	Written 1993 by Jacques Gelinas 
 *	Inspired from linux/fs/msdos/... by Werner Almesberger
 *
 * Maintain and access the --linux alternate directory file.
 */

#include <linux/errno.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/stat.h>
#include <linux/string.h>
#include <linux/msdos_fs.h>
#include <linux/umsdos_fs.h>
#include <linux/malloc.h>

#define PRINTK(x)
#define Printk(x)	printk x

#if 1
/*
	Wait for creation exclusivity.
	Return 0 if the dir was already available.
	Return 1 if a wait was necessary.
		When 1 is return, it means a wait was done. It does not
		mean the directory is available.
*/
static int umsdos_waitcreate(struct inode *dir)
{
  int ret = 0;
  if (dir->u.umsdos_i.u.dir_info.creating
      && dir->u.umsdos_i.u.dir_info.pid != current->pid){
    sleep_on(&dir->u.umsdos_i.u.dir_info.p);
    ret = 1;
  }
  return ret;
}

/*
	Wait for any lookup process to finish
*/
static void umsdos_waitlookup (struct inode *dir)
{
  while (dir->u.umsdos_i.u.dir_info.looking){
    sleep_on(&dir->u.umsdos_i.u.dir_info.p);
  }
}

/*
	Lock all other process out of this directory.
*/
void umsdos_lockcreate (struct inode *dir)
{
  /* #Specification: file creation / not atomic
     File creation is a two step process. First we create (allocate)
     an entry in the EMD file and then (using the entry offset) we
     build a unique name for MSDOS. We create this name in the msdos
     space.
     
     We have to use semaphore (sleep_on/wake_up) to prevent lookup
     into a directory when we create a file or directory and to
     prevent creation while a lookup is going on. Since many lookup
     may happen at the same time, the semaphore is a counter.

     Only one creation is allowed at the same time. This protection
     may not be necessary. The problem arise mainly when a lookup
     or a readdir is done while a file is partially created. The
     lookup process see that as a "normal" problem and silently
     erase the file from the EMD file. Normal because a file
     may be erased during a MSDOS session, but not removed from
     the EMD file.
     
     The locking is done on a directory per directory basis. Each
     directory inode has its wait_queue.
     
     For some operation like hard link, things even get worse. Many
     creation must occur at once (atomic). To simplify the design
     a process is allowed to recursively lock the directory for
     creation. The pid of the locking process is kept along with
     a counter so a second level of locking is granted or not.
  */
  /*
    Wait for any creation process to finish except
    if we (the process) own the lock
  */
  while (umsdos_waitcreate(dir)!=0);
  dir->u.umsdos_i.u.dir_info.creating++;
  dir->u.umsdos_i.u.dir_info.pid = current->pid;
  umsdos_waitlookup (dir);
}

/*
	Lock all other process out of those two directories.
*/
static void umsdos_lockcreate2 (struct inode *dir1, struct inode *dir2)
{
  /*
    We must check that both directory are available before
    locking anyone of them. This is to avoid some deadlock.
    Thanks to dglaude@is1.vub.ac.be (GLAUDE DAVID) for pointing
    this to me.
  */
  while (1){
    if (umsdos_waitcreate(dir1)==0
	&& umsdos_waitcreate(dir2)==0){
      /* We own both now */
      dir1->u.umsdos_i.u.dir_info.creating++;
      dir1->u.umsdos_i.u.dir_info.pid = current->pid;
      dir2->u.umsdos_i.u.dir_info.creating++;
      dir2->u.umsdos_i.u.dir_info.pid = current->pid;
      break;
    }
  }
  umsdos_waitlookup(dir1);
  umsdos_waitlookup(dir2);
}

/*
	Wait until creation is finish in this directory.
*/
void umsdos_startlookup (struct inode *dir)
{
  while (umsdos_waitcreate (dir) != 0);
  dir->u.umsdos_i.u.dir_info.looking++;
}

/*
	Unlock the directory.
*/
void umsdos_unlockcreate (struct inode *dir)
{
  dir->u.umsdos_i.u.dir_info.creating--;
  if (dir->u.umsdos_i.u.dir_info.creating < 0){
    printk ("UMSDOS: dir->u.umsdos_i.u.dir_info.creating < 0: %d"
	    ,dir->u.umsdos_i.u.dir_info.creating);
  }
  wake_up (&dir->u.umsdos_i.u.dir_info.p);
}

/*
	Tell directory lookup is over.
*/
void umsdos_endlookup (struct inode *dir)
{
  dir->u.umsdos_i.u.dir_info.looking--;
  if (dir->u.umsdos_i.u.dir_info.looking < 0){
    printk ("UMSDOS: dir->u.umsdos_i.u.dir_info.looking < 0: %d"
	    ,dir->u.umsdos_i.u.dir_info.looking);
  }
  wake_up (&dir->u.umsdos_i.u.dir_info.p);
}

#else
static void umsdos_lockcreate (struct inode *dir){}
static void umsdos_lockcreate2 (struct inode *dir1, struct inode *dir2){}
void umsdos_startlookup (struct inode *dir){}
static void umsdos_unlockcreate (struct inode *dir){}
void umsdos_endlookup (struct inode *dir){}
#endif

static int umsdos_nevercreat(
			     struct inode *dir,
			     struct dentry *dentry,
			     int errcod)				/* Length of the name */
{
  const char *name = dentry->d_name.name;
  int len = dentry->d_name.len;
  int ret = 0;
  if (umsdos_is_pseudodos(dir,dentry)){
    /* #Specification: pseudo root / any file creation /DOS
       The pseudo sub-directory /DOS can't be created!
       EEXIST is returned.
       
       The pseudo sub-directory /DOS can't be removed!
       EPERM is returned.
    */
    ret = -EPERM;
    ret = errcod;
  }else if (name[0] == '.'
	    && (len == 1 || (len == 2 && name[1] == '.'))){
    /* #Specification: create / . and ..
       If one try to creates . or .., it always fail and return
       EEXIST.
       
       If one try to delete . or .., it always fail and return
       EPERM.
       
       This should be test at the VFS layer level to avoid
       duplicating this in all file systems. Any comments ?
    */
    ret = errcod;
  }
  return ret;
}
	
/*
  Add a new file (ordinary or special) into the alternate directory.
  The file is added to the real MSDOS directory. If successful, it
  is then added to the EDM file.
  
  Return the status of the operation. 0 mean success.
*/
static int umsdos_create_any (
   struct inode *dir,
   struct dentry *dentry,  /* name/length etc*/
   int mode,		   /* Permission bit + file type ??? */
   int rdev,		   /* major, minor or 0 for ordinary file */
   /* and symlinks */
   char flags
   )	/* Will hold the inode of the newly created */
     /* file */
{
    
  int ret;
  struct dentry *fake;
  
  Printk (("umsdos_create_any /mn/: create %.*s in dir=%lu - nevercreat=/", (int) dentry->d_name.len, dentry->d_name.name, dir->i_ino));
  ret = umsdos_nevercreat(dir,dentry,-EEXIST);
  Printk (("%d/\n", ret));
  if (ret == 0){
    struct umsdos_info info;
    ret = umsdos_parse(dentry->d_name.name,dentry->d_name.len,&info);
  
    if (ret == 0){
      info.entry.mode = mode;
      info.entry.rdev = rdev;
      info.entry.flags = flags;
      info.entry.uid = current->fsuid;
      info.entry.gid = (dir->i_mode & S_ISGID)
	? dir->i_gid : current->fsgid;
      info.entry.ctime = info.entry.atime = info.entry.mtime
	= CURRENT_TIME;
      info.entry.nlink = 1;
      umsdos_lockcreate(dir);
      ret = umsdos_newentry (dir,&info);
      if (ret == 0){
	dir->i_count++;
        fake = creat_dentry (info.fake.fname, info.fake.len, NULL, dentry->d_parent);	/* create short name dentry */
	ret = msdos_create (dir, fake, S_IFREG|0777);
	if (ret == 0){
	  struct inode *inode = fake->d_inode;
	  umsdos_lookup_patch (dir,inode,&info.entry,info.f_pos);
	  Printk (("inode %p[%lu], count=%d ",inode, inode->i_ino, inode->i_count));
	  Printk (("Creation OK: [dir %lu] %.*s pid=%d pos %ld\n", dir->i_ino,
		   info.fake.len, info.fake.fname, current->pid, info.f_pos));
	  
	  d_instantiate(dentry, inode);	/* long name also gets inode info */
	}else{
	  /* #Specification: create / file exist in DOS
	     Here is a situation. Trying to create a file with
	     UMSDOS. The file is unknown to UMSDOS but already
	     exist in the DOS directory.
	     
	     Here is what we are NOT doing:
	     
	     We could silently assume that everything is fine
	     and allows the creation to succeed.
	     
	     It is possible not all files in the partition
	     are mean to be visible from linux. By trying to create
	     those file in some directory, one user may get access
	     to those file without proper permissions. Looks like
	     a security hole to me. Off course sharing a file system
	     with DOS is some kind of security hole :-)
	     
	     So ?
	     
	     We return EEXIST in this case.
	     The same is true for directory creation.
	  */
	  if (ret == -EEXIST){
	    printk ("UMSDOS: out of sync, Creation error [%ld], "
		    "deleting %.*s %d %d pos %ld\n",dir->i_ino
		    ,info.fake.len,info.fake.fname,-ret,current->pid,info.f_pos);
	  }
	  umsdos_delentry (dir,&info,0);
	}
	Printk (("umsdos_create %.*s ret = %d pos %ld\n",
		 info.fake.len, info.fake.fname, ret, info.f_pos));
      }
      umsdos_unlockcreate(dir);
    }
  }
  /* d_add(dentry,dir); /mn/ FIXME: msdos_create already did this for short name ! */
  return ret;
}

/*
	Initialise the new_entry from the old for a rename operation.
	(Only useful for umsdos_rename_f() below).
*/
static void umsdos_ren_init(
      struct umsdos_info *new_info,
      struct umsdos_info *old_info,
      int flags)		/* 0 == copy flags from old_name */
     /* != 0, this is the value of flags */
{
  new_info->entry.mode = old_info->entry.mode;
  new_info->entry.rdev = old_info->entry.rdev;
  new_info->entry.uid = old_info->entry.uid;
  new_info->entry.gid = old_info->entry.gid;
  new_info->entry.ctime = old_info->entry.ctime;
  new_info->entry.atime = old_info->entry.atime;
  new_info->entry.mtime = old_info->entry.mtime;
  new_info->entry.flags = flags ? flags : old_info->entry.flags;
  new_info->entry.nlink = old_info->entry.nlink;
}

#define chkstk() \
if (STACK_MAGIC != *(unsigned long *)current->kernel_stack_page){\
    printk(KERN_ALERT "UMSDOS: %s magic %x != %lx ligne %d\n" \
	   , current->comm,STACK_MAGIC \
	   ,*(unsigned long *)current->kernel_stack_page \
	   ,__LINE__); \
}

#undef chkstk
#define chkstk() do { } while (0)
	
/*
	Rename a file (move) in the file system.
*/
static int umsdos_rename_f(
			   struct inode * old_dir,
			   struct dentry *old_dentry,
			   struct inode * new_dir,
			   struct dentry *new_dentry,
			   int flags)		/* 0 == copy flags from old_name */
     /* != 0, this is the value of flags */
{
  int ret = -EPERM;
  struct umsdos_info old_info;
  int old_ret = umsdos_parse (old_dentry->d_name.name,
			      old_dentry->d_name.len,&old_info);
  struct umsdos_info new_info;
  int new_ret = umsdos_parse (new_dentry->d_name.name,
			      new_dentry->d_name.len,&new_info);
  chkstk();
  Printk (("umsdos_rename %d %d ",old_ret,new_ret));
  if (old_ret == 0 && new_ret == 0){
    umsdos_lockcreate2(old_dir,new_dir);
    chkstk();
    Printk (("old findentry "));
    ret = umsdos_findentry(old_dir,&old_info,0);
    chkstk();
    Printk (("ret %d ",ret));
    if (ret == 0){
      /* check sticky bit on old_dir */
      if ( !(old_dir->i_mode & S_ISVTX) || capable(CAP_FOWNER) ||
	   current->fsuid == old_info.entry.uid ||
	   current->fsuid == old_dir->i_uid ) {
	/* Does new_name already exist? */
	PRINTK(("new findentry "));
	ret = umsdos_findentry(new_dir,&new_info,0);
	if (ret != 0 || /* if destination file exists, are we allowed to replace it ? */
	    !(new_dir->i_mode & S_ISVTX) || capable(CAP_FOWNER) ||
	    current->fsuid == new_info.entry.uid ||
	    current->fsuid == new_dir->i_uid ) {
	  PRINTK (("new newentry "));
	  umsdos_ren_init(&new_info,&old_info,flags);
	  ret = umsdos_newentry (new_dir,&new_info);
	  chkstk();
	  PRINTK (("ret %d %d ",ret,new_info.fake.len));
	  if (ret == 0){
	    struct dentry *old, *new;
	    old = creat_dentry (old_info.fake.fname, old_info.fake.len, NULL, NULL);
	    new = creat_dentry (new_info.fake.fname, new_info.fake.len, NULL, NULL);

	    PRINTK (("msdos_rename "));
	    old_dir->i_count++;
	    new_dir->i_count++;	/* Both inode are needed later */
	    ret = msdos_rename (old_dir,
				old,
				new_dir,
				new);
	    chkstk();
	    PRINTK (("after m_rename ret %d ",ret));
	    if (ret != 0){
	      umsdos_delentry (new_dir,&new_info
			       ,S_ISDIR(new_info.entry.mode));
	      chkstk();
	    }else{
	      ret = umsdos_delentry (old_dir,&old_info
				     ,S_ISDIR(old_info.entry.mode));
	      chkstk();
	      if (ret == 0){
		/*
		  This umsdos_lookup_x does not look very useful.
		  It makes sure that the inode of the file will
		  be correctly setup (umsdos_patch_inode()) in
		  case it is already in use.
		  
		  Not very efficient ...
		*/
		struct inode *inode;
		new_dir->i_count++;
		PRINTK ((KERN_DEBUG "rename lookup len %d %d -- ",new_len,new_info.entry.flags));
		ret = umsdos_lookup_x (new_dir, new_dentry, 0);
		inode = new_dentry->d_inode;
		chkstk();
		if (ret != 0){
		  printk ("UMSDOS: partial rename for file %.*s\n"
			  ,new_info.entry.name_len,new_info.entry.name);
		}else{
		  /*
		    Update f_pos so notify_change will succeed
		    if the file was already in use.
		  */
		  umsdos_set_dirinfo (inode,new_dir,new_info.f_pos);
		  chkstk();
		  /* iput (inode); FIXME */
		}
	      }
	    }
	  }
	}else{
	  /* sticky bit set on new_dir */
	  Printk(("sticky set on new "));
	  ret = -EPERM;
	}
      }else{
	/* sticky bit set on old_dir */
	Printk(("sticky set on old "));
	ret = -EPERM;
      }
    }
    umsdos_unlockcreate(old_dir);
    umsdos_unlockcreate(new_dir);
  }
  d_move(old_dentry,new_dentry);
  Printk (("\n"));
  return ret;
}

/*
	Setup un Symbolic link or a (pseudo) hard link
	Return a negative error code or 0 if ok.
*/
static int umsdos_symlink_x(
			    struct inode * dir,
			    struct dentry *dentry,
			    const char * symname,	/* name will point to this path */
			    int mode,
			    char flags)
{
  /* #Specification: symbolic links / strategy
     A symbolic link is simply a file which hold a path. It is
     implemented as a normal MSDOS file (not very space efficient :-()
     
     I see 2 different way to do it. One is to place the link data
     in unused entry of the EMD file. The other is to have a separate
     file dedicated to hold all symbolic links data.
     
     Let's go for simplicity...
  */
  
  
  int ret;
  dir->i_count++;/* We keep the inode in case we need it */
  /* later */
  ret = umsdos_create_any (dir,dentry,mode,0,flags);
  Printk (("umsdos_symlink ret %d ",ret));
  if (ret == 0){
    int len = strlen(symname);
    struct file filp;
    loff_t myofs=0;
    fill_new_filp (&filp, dentry);

    /* Make the inode acceptable to MSDOS FIXME */
    Printk ((KERN_WARNING "umsdos_symlink_x: /mn/ Is this ok?\n"));
    Printk ((KERN_WARNING "   symname=%s ; dentry name=%.*s (ino=%lu)\n", symname, (int) dentry->d_name.len, dentry->d_name.name, dentry->d_inode->i_ino));
    ret = umsdos_file_write_kmem_real (&filp, symname, len, &myofs);
    /* dput(dentry); ?? where did this come from FIXME */
    if (ret >= 0){
      if (ret != len){
	ret = -EIO;
	printk ("UMSDOS: "
		"Can't write symbolic link data\n");
      }else{
	ret = 0;
      }
    }
    if (ret != 0){
      UMSDOS_unlink (dir,dentry);
      dir = NULL;
    }
  }
  /* d_instantiate(dentry,dir);   //already done in umsdos_create_any */
  Printk (("\n"));
  return ret;
}

/*
  Setup un Symbolic link.
  Return a negative error code or 0 if ok.
*/
int UMSDOS_symlink(
   struct inode * dir,
   struct dentry *dentry,
   const char * symname
   )
{
  return umsdos_symlink_x (dir,dentry,symname,S_IFLNK|0777,0);
}

/*
  Add a link to an inode in a directory
*/
int UMSDOS_link (
		 struct dentry * olddentry,
		 struct inode * dir,
		 struct dentry *dentry)
{
    struct inode *oldinode = olddentry->d_inode;
  /* #Specification: hard link / strategy
     Well ... hard link are difficult to implement on top of an
     MsDOS fat file system. Unlike UNIX file systems, there are no
     inode. A directory entry hold the functionality of the inode
     and the entry.
     
     We will used the same strategy as a normal Unix file system
     (with inode) except we will do it symbolically (using paths).
     
     Because anything can happen during a DOS session (defragment,
     directory sorting, etc...), we can't rely on MsDOS pseudo
     inode number to record the link. For this reason, the link
     will be done using hidden symbolic links. The following
     scenario illustrate how it work.
     
     Given a file /foo/file
     
     #
     ln /foo/file /tmp/file2
     
     become internally
     
     mv /foo/file /foo/-LINK1
     ln -s /foo/-LINK1 /foo/file
     ln -s /foo/-LINK1 /tmp/file2
     #
     
     Using this strategy, we can operate on /foo/file or /foo/file2.
     We can remove one and keep the other, like a normal Unix hard link.
     We can rename /foo/file or /tmp/file2 independently.
     
     The entry -LINK1 will be hidden. It will hold a link count.
     When all link are erased, the hidden file is erased too.
  */
  /* #Specification: weakness / hard link
     The strategy for hard link introduces a side effect that
     may or may not be acceptable. Here is the sequence
     
     #
     mkdir subdir1
     touch subdir1/file
     mkdir subdir2
     ln    subdir1/file subdir2/file
     rm    subdir1/file
     rmdir subdir1
     rmdir: subdir1: Directory not empty
     #
     
     This happen because there is an invisible file (--link) in
     subdir1 which is referenced by subdir2/file.
     
     Any idea ?
  */
  /* #Specification: weakness / hard link / rename directory
     Another weakness of hard link come from the fact that
     it is based on hidden symbolic links. Here is an example.
     
     #
     mkdir /subdir1
     touch /subdir1/file
     mkdir /subdir2
     ln    /subdir1/file subdir2/file
     mv    /subdir1 subdir3
     ls -l /subdir2/file
     #
     
     Since /subdir2/file is a hidden symbolic link
     to /subdir1/..hlinkNNN, accessing it will fail since
     /subdir1 does not exist anymore (has been renamed).
  */
  int ret = 0;
  if (S_ISDIR(oldinode->i_mode)){
    /* #Specification: hard link / directory
       A hard link can't be made on a directory. EPERM is returned
       in this case.
    */
    ret = -EPERM;
  }else if ((ret = umsdos_nevercreat(dir,dentry,-EPERM))==0){
    struct inode *olddir;
    ret = umsdos_get_dirowner(oldinode,&olddir);
    Printk (("umsdos_link dir_owner = %lu -> %p [%d] ",
	     oldinode->u.umsdos_i.i_dir_owner, olddir, olddir->i_count));
    if (ret == 0){
      struct umsdos_dirent entry;
      umsdos_lockcreate2(dir,olddir);
      ret = umsdos_inode2entry (olddir,oldinode,&entry);
      if (ret == 0){
	Printk (("umsdos_link :%.*s: ino %lu flags %d "
		 ,entry.name_len, entry.name
		 ,oldinode->i_ino, entry.flags));
	if (!(entry.flags & UMSDOS_HIDDEN)){
	  /* #Specification: hard link / first hard link
	     The first time a hard link is done on a file, this
	     file must be renamed and hidden. Then an internal
	     symbolic link must be done on the hidden file.
	     
	     The second link is done after on this hidden file.
	     
	     It is expected that the Linux MSDOS file system
	     keeps the same pseudo inode when a rename operation
	     is done on a file in the same directory.
	  */
	  struct umsdos_info info;
	  ret = umsdos_newhidden (olddir,&info);
	  if (ret == 0){
	    Printk (("olddir[%d] ",olddir->i_count));
	    ret = umsdos_rename_f( 
				  olddentry->d_inode,
				  olddentry,
				  dir,
				  dentry,
				   UMSDOS_HIDDEN);
	    if (ret == 0){
	      char *path = (char*)kmalloc(PATH_MAX,GFP_KERNEL);
	      if (path == NULL){
		ret = -ENOMEM;
	      }else{
		struct dentry *temp;
		temp = creat_dentry (entry.name, entry.name_len, NULL, NULL);
		Printk (("olddir[%d] ",olddir->i_count));
		ret = umsdos_locate_path (oldinode,path);
		Printk (("olddir[%d] ",olddir->i_count));
		if (ret == 0){
		  olddir->i_count++;
		  ret = umsdos_symlink_x (olddir
					  ,temp
					  ,path
					  ,S_IFREG|0777,UMSDOS_HLINK);
		  if (ret == 0){
		    dir->i_count++;
		    ret = umsdos_symlink_x (dir,dentry,
					    path,
					    S_IFREG|0777,UMSDOS_HLINK);
		  }
		}
		kfree (path);
	      }
	    }
	  }
	}else{
	  char *path = (char*)kmalloc(PATH_MAX,GFP_KERNEL);
	  if (path == NULL){
	    ret = -ENOMEM;
	  }else{
	    ret = umsdos_locate_path (oldinode,path);
	    if (ret == 0){
	      dir->i_count++;
	      ret = umsdos_symlink_x (dir,dentry,path
				      ,S_IFREG|0777,UMSDOS_HLINK);
	    }
	    kfree (path);
	  }
	}
      }
      umsdos_unlockcreate(olddir);
      umsdos_unlockcreate(dir);
    }
    /* iput (olddir); FIXME */
  }
  if (ret == 0){
    struct iattr newattrs;
    oldinode->i_nlink++;
    newattrs.ia_valid = 0;
    ret = UMSDOS_notify_change(olddentry, &newattrs);
  }

/*  dput (olddentry);
  dput (dentry); FIXME.... */
  
  Printk (("umsdos_link %d\n",ret));
  return ret;
}



/*
  Add a new file into the alternate directory.
  The file is added to the real MSDOS directory. If successful, it
  is then added to the EDM file.
  
  Return the status of the operation. 0 mean success.
*/
int UMSDOS_create (
		   struct inode *dir,
		   struct dentry *dentry,
		   int mode		  /* Permission bit + file type ??? */
		   )	/* Will hold the inode of the newly created */
							/* file */
{
  return umsdos_create_any (dir,dentry,mode,0,0);
}



/*
	Add a sub-directory in a directory
*/
int UMSDOS_mkdir(
		 struct inode * dir,
		 struct dentry *dentry,
		 int mode)
{
  int ret = umsdos_nevercreat(dir,dentry,-EEXIST);
  if (ret == 0){
    struct umsdos_info info;
    ret = umsdos_parse (dentry->d_name.name,dentry->d_name.len,&info);
    Printk (("umsdos_mkdir %d\n",ret));
    if (ret == 0){
      info.entry.mode = mode | S_IFDIR;
      info.entry.rdev = 0;
      info.entry.uid = current->fsuid;
      info.entry.gid = (dir->i_mode & S_ISGID)
	? dir->i_gid : current->fsgid;
      info.entry.ctime = info.entry.atime = info.entry.mtime
	= CURRENT_TIME;
      info.entry.flags = 0;
      umsdos_lockcreate(dir);
      info.entry.nlink = 1;
      ret = umsdos_newentry (dir,&info);
      Printk (("newentry %d ",ret));
      if (ret == 0){
	struct dentry *temp, *tdir;
	tdir = creat_dentry ("mkd-dir", 7, dir, NULL);
	temp = creat_dentry (info.fake.fname, info.fake.len, NULL, tdir);
	dir->i_count++;
	ret = msdos_mkdir (dir, temp, mode);
	
	if (ret != 0){
	  umsdos_delentry (dir,&info,1);
	  /* #Specification: mkdir / Directory already exist in DOS
	     We do the same thing as for file creation.
	     For all user it is an error.
	  */
	}else{
	  /* #Specification: mkdir / umsdos directory / create EMD
	     When we created a new sub-directory in a UMSDOS
	     directory (one with full UMSDOS semantic), we
	     create immediately an EMD file in the new
	     sub-directory so it inherit UMSDOS semantic.
	  */
	  struct inode *subdir;
	  ret = compat_umsdos_real_lookup (dir,info.fake.fname,
					   info.fake.len,&subdir);
	  if (ret == 0){
	    struct dentry *tdentry, *tdsub;
	    tdsub = creat_dentry ("mkd-emd", 7, subdir, NULL);
	    tdentry = creat_dentry (UMSDOS_EMD_FILE, UMSDOS_EMD_NAMELEN, NULL, tdsub);
	    ret = msdos_create (subdir, tdentry, S_IFREG|0777);
	    kill_dentry (tdentry);	/* we don't want empty EMD file to be visible ! too bad kill_dentry does nothing at the moment :-)  FIXME */
	    kill_dentry (tdsub);
	    umsdos_setup_dir_inode (subdir);	/* this should setup dir so it is promoted to EMD, and EMD file is not visible inside it */
	    subdir = NULL;
	    d_instantiate(dentry, temp->d_inode);
	    /* iput (result); FIXME */
	  }
	  if (ret < 0){
	    printk ("UMSDOS: Can't create empty --linux-.---\n");
	  }
	  
	
	  /* iput (subdir); FIXME */
	}
      }
      umsdos_unlockcreate(dir);
    }
  }
  Printk (("umsdos_mkdir %d\n",ret));
  /* dput (dentry); / * FIXME /mn/ */
  return ret;
}

/*
	Add a new device special file into a directory.
*/
int UMSDOS_mknod(
	struct inode * dir,
	struct dentry *dentry,         
	int mode,
	int rdev)
{
  /* #Specification: Special files / strategy
     Device special file, pipes, etc ... are created like normal
     file in the msdos file system. Of course they remain empty.
     
     One strategy was to create those files only in the EMD file
     since they were not important for MSDOS. The problem with
     that, is that there were not getting inode number allocated.
     The MSDOS filesystems is playing a nice game to fake inode
     number, so why not use it.
     
     The absence of inode number compatible with those allocated
     for ordinary files was causing major trouble with hard link
     in particular and other parts of the kernel I guess.
  */

  int ret = umsdos_create_any (dir,dentry,mode,rdev,0);
  /* dput(dentry); / * /mn/ FIXME! */
  return ret;
}

/*
  Remove a sub-directory.
*/
int UMSDOS_rmdir(
	struct inode * dir,
	struct dentry *dentry)
{
  /* #Specification: style / iput strategy
     In the UMSDOS project, I am trying to apply a single
     programming style regarding inode management. Many
     entry point are receiving an inode to act on, and must
     do an iput() as soon as they are finished with
     the inode.
     
     For simple case, there is no problem. When you introduce
     error checking, you end up with many iput placed around the
     code.
     
     The coding style I use all around is one where I am trying
     to provide independent flow logic (I don't know how to
     name this). With this style, code is easier to understand
     but you rapidly get iput() all around. Here is an exemple
     of what I am trying to avoid.
     
     #
     if (a){
     ...
     if(b){
     ...
     }
     ...
     if (c){
     // Complex state. Was b true ? 
     ...
     }
     ...
     }
     // Weird state
     if (d){
     // ...
     }
     // Was iput finally done ?
     return status;
     #
     
     Here is the style I am using. Still sometime I do the
     first when things are very simple (or very complicated :-( )
     
     #
     if (a){
     if (b){
     ...
     }else if (c){
     // A single state gets here
     }
     }else if (d){
     ...
     }
     return status;
     #
     
     Again, while this help clarifying the code, I often get a lot
     of iput(), unlike the first style, where I can place few 
     "strategic" iput(). "strategic" also mean, more difficult
     to place.
     
     So here is the style I will be using from now on in this project.
     There is always an iput() at the end of a function (which has
     to do an iput()). One iput by inode. There is also one iput()
     at the places where a successful operation is achieved. This
     iput() is often done by a sub-function (often from the msdos
     file system). So I get one too many iput() ? At the place
     where an iput() is done, the inode is simply nulled, disabling
     the last one.
     
     #
     if (a){
     if (b){
     ...
     }else if (c){
     msdos_rmdir(dir,...);
     dir = NULL;
     }
     }else if (d){
     ...
     }
     iput (dir);
     return status;
     #
     
     Note that the umsdos_lockcreate() and umsdos_unlockcreate() function
     pair goes against this practice of "forgetting" the inode as soon
     as possible.
  */		

  int ret;
  
  ret = umsdos_nevercreat(dir,dentry,-EPERM);
  if (ret == 0){
    volatile struct inode *sdir;
    dir->i_count++;
    ret = umsdos_lookup_x (dir, dentry, 0);
    sdir = dentry->d_inode;
    Printk (("rmdir lookup %d ",ret));
    if (ret == 0){
      int empty;
      umsdos_lockcreate(dir);
      if (sdir->i_count > 1){
        Printk ((" /mn/ rmdir: FIXME EBUSY: hmmm, i_count is %d > 1\n", sdir->i_count));
	ret = -EBUSY;
      }else if ((empty = umsdos_isempty (sdir)) != 0){
	struct dentry *tdentry, *tedir;
	tedir = creat_dentry ("emd-rmd", 7, dir, NULL);
	tdentry = creat_dentry (UMSDOS_EMD_FILE, UMSDOS_EMD_NAMELEN, NULL, tedir);
        umsdos_real_lookup (dir, tdentry);	/* fill inode part */
	Printk (("isempty %d i_count %d ",empty,sdir->i_count));
				/* check sticky bit */
	if ( !(dir->i_mode & S_ISVTX) || capable(CAP_FOWNER) ||
	     current->fsuid == sdir->i_uid ||
	     current->fsuid == dir->i_uid ) {
	  if (empty == 1){
	    /* We have to remove the EMD file */
	    ret = msdos_unlink (sdir, tdentry);
	    Printk (("UMSDOS_rmdir: unlinking empty EMD ret=%d", ret));
	    sdir = NULL;
	  }
	  /* sdir must be free before msdos_rmdir() */
	  /* iput (sdir); FIXME */
	  sdir = NULL;
	  Printk (("isempty ret %d nlink %d ",ret,dir->i_nlink));
	  if (ret == 0){
	    struct umsdos_info info;
	    struct dentry *temp, *tdir;
	    dir->i_count++;
	    umsdos_parse (dentry->d_name.name,dentry->d_name.len,&info);
	    /* The findentry is there only to complete */
	    /* the mangling */
	    umsdos_findentry (dir,&info,2);

            tdir = creat_dentry ("dir-rmd", 7, dir, NULL);
            temp = creat_dentry (info.fake.fname, info.fake.len, NULL, tdir);
            umsdos_real_lookup (dir, temp);	/* fill inode part */
            
	      Printk ((KERN_ERR "  rmdir start dir=%lu, dir->sb=%p\n", dir->i_ino, dir->i_sb)); /* FIXME: /mn/ debug only */
	      Printk ((KERN_ERR "    dentry=%.*s d_count=%d ino=%lu\n", (int) temp->d_name.len, temp->d_name.name, temp->d_count, temp->d_inode->i_ino));
	      Printk ((KERN_ERR "    d_parent=%.*s d_count=%d ino=%lu\n", (int) temp->d_parent->d_name.len, temp->d_parent->d_name.name, temp->d_parent->d_count, temp->d_parent->d_inode->i_ino));

	    ret = msdos_rmdir (dir, temp);

	      Printk ((KERN_ERR "  rmdir passed %d\n", ret)); /* FIXME: /mn/ debug only */
	      Printk ((KERN_ERR "  rmdir end dir=%lu, dir->sb=%p\n", dir->i_ino, dir->i_sb));
	      Printk ((KERN_ERR "    dentry=%.*s d_count=%d ino=%p\n", (int) temp->d_name.len, temp->d_name.name, temp->d_count, temp->d_inode));
	      Printk ((KERN_ERR "    d_parent=%.*s d_count=%d ino=%lu\n", (int) temp->d_parent->d_name.len, temp->d_parent->d_name.name, temp->d_parent->d_count, temp->d_parent->d_inode->i_ino));

            kill_dentry (tdir);
            kill_dentry (temp);
            
	    if (ret == 0){
	      ret = umsdos_delentry (dir,&info,1);
              d_delete (dentry);
	    }
	  }
	}else{
	  /* sticky bit set and we don't have permission */
	  Printk(("sticky set "));
	  ret = -EPERM;
	}
      }else{	
	/*
	  The subdirectory is not empty, so leave it there
	*/
	ret = -ENOTEMPTY;
      }
      /* iput(sdir); FIXME */
      umsdos_unlockcreate(dir);
    }	
  }
 /*  dput(dentry); FIXME /mn/ */
  Printk (("umsdos_rmdir %d\n",ret));
  return ret;
}



/*
  Remove a file from the directory.
*/
int UMSDOS_unlink (
	struct inode * dir,
	struct dentry *dentry)
{
  int ret;
  Printk ((" *** UMSDOS_unlink entering /mn/ *** \n"));

  ret = umsdos_nevercreat(dir,dentry,-EPERM);
  
  Printk (("UMSDOS_unlink /mn/: nevercreat=%d\n", ret));
  
  if (ret == 0){
    struct umsdos_info info;
    ret = umsdos_parse (dentry->d_name.name,dentry->d_name.len,&info);
    if (ret == 0){
      umsdos_lockcreate(dir);
      ret = umsdos_findentry(dir,&info,1);
      Printk (("UMSDOS_unlink: findentry returned %d\n", ret));
      if (ret == 0){
	Printk (("UMSDOS_unlink %.*s ",info.fake.len,info.fake.fname));
				/* check sticky bit */
	if ( !(dir->i_mode & S_ISVTX) || capable(CAP_FOWNER) ||
	     current->fsuid == info.entry.uid ||
	     current->fsuid == dir->i_uid ) {
	  if (info.entry.flags & UMSDOS_HLINK){
	    /* #Specification: hard link / deleting a link
	       When we deletes a file, and this file is a link
	       we must subtract 1 to the nlink field of the
	       hidden link.
	       
	       If the count goes to 0, we delete this hidden
	       link too.
	    */
	    /*
	      First, get the inode of the hidden link
	      using the standard lookup function.
	    */
	    struct inode *inode;
	    dir->i_count++;
	    ret = umsdos_lookup_x (dir, dentry, 0);
	    inode = dentry->d_inode;
	    if (ret == 0){
	      Printk (("unlink nlink = %d ",inode->i_nlink));
	      inode->i_nlink--;
	      if (inode->i_nlink == 0){
		struct inode *hdir = iget(inode->i_sb
					  ,inode->u.umsdos_i.i_dir_owner);
		struct umsdos_dirent entry;
		ret = umsdos_inode2entry (hdir,inode,&entry);
		if (ret == 0){
		  ret = UMSDOS_unlink (hdir,dentry);
		}else{
		    /* iput (hdir); FIXME */
		}
	      }else{
		struct iattr newattrs;
		newattrs.ia_valid = 0;
		ret = UMSDOS_notify_change (dentry, &newattrs);
	      }
	      /* iput (inode); FIXME */
	    }
	  }
	  if (ret == 0){
	    ret = umsdos_delentry (dir,&info,0);
	    if (ret == 0){
	      struct dentry *temp, *tdir;
	      Printk (("Avant msdos_unlink %.*s ",info.fake.len,info.fake.fname));
	      dir->i_count++;  /* FIXME /mn/ is this needed anymore now that msdos_unlink locks dir using d_parent ? */
	      tdir = creat_dentry ("dir-del", 7, dir, NULL);	/* FIXME /mn/: do we need iget(dir->i_ino) or would dir itself suffice ? */
	      temp = creat_dentry (info.fake.fname, info.fake.len, NULL, tdir);
	      umsdos_real_lookup (dir, temp);	/* fill inode part */
	      
	      ret = msdos_unlink_umsdos (dir, temp);
	      Printk (("msdos_unlink %.*s %o ret %d ",info.fake.len,info.fake.fname
		       ,info.entry.mode,ret));

	      d_delete (dentry);
	      
	      kill_dentry (tdir);
	      kill_dentry (temp);
	    }
	  }
	}else{
	  /* sticky bit set and we've not got permission */
	  Printk(("sticky set "));
	  ret = -EPERM;
	}
      }
      umsdos_unlockcreate(dir);
    }
  }	
  /* dput(dentry); FIXME: shouldn't this be done in msdos_unlink ? */
  Printk (("umsdos_unlink %d\n",ret));
  return ret;
}



/*
	Rename a file (move) in the file system.
*/
int UMSDOS_rename(
		  struct inode * old_dir,
		  struct dentry * old_dentry,
		  struct inode * new_dir,
		  struct dentry * new_dentry)
{
  /* #Specification: weakness / rename
     There is a case where UMSDOS rename has a different behavior
     than normal UNIX file system. Renaming an open file across
     directory boundary does not work. Renaming an open file within
     a directory does work however.
     
     The problem (not sure) is in the linux VFS msdos driver.
     I believe this is not a bug but a design feature, because
     an inode number represent some sort of directory address
     in the MSDOS directory structure. So moving the file into
     another directory does not preserve the inode number.
  */
  int ret = umsdos_nevercreat(new_dir,new_dentry,-EEXIST);
  if (ret == 0){
    /* umsdos_rename_f eat the inode and we may need those later */
    old_dir->i_count++;
    new_dir->i_count++;
    ret = umsdos_rename_f (old_dir,old_dentry,new_dir,new_dentry,0);
    if (ret == -EEXIST){
      /* #Specification: rename / new name exist
	 If the destination name already exist, it will
	 silently be removed. EXT2 does it this way
	 and this is the spec of SUNOS. So does UMSDOS.
	 
	 If the destination is an empty directory it will
	 also be removed.
      */
      /* #Specification: rename / new name exist / possible flaw
	 The code to handle the deletion of the target (file
	 and directory) use to be in umsdos_rename_f, surrounded
	 by proper directory locking. This was insuring that only
	 one process could achieve a rename (modification) operation
	 in the source and destination directory. This was also
	 insuring the operation was "atomic".
	 
	 This has been changed because this was creating a kernel
	 stack overflow (stack is only 4k in the kernel). To avoid
	 the code doing the deletion of the target (if exist) has
	 been moved to a upper layer. umsdos_rename_f is tried
	 once and if it fails with EEXIST, the target is removed
	 and umsdos_rename_f is done again.
	 
	 This makes the code cleaner and (not sure) solve a
	 deadlock problem one tester was experiencing.
	 
	 The point is to mention that possibly, the semantic of
	 "rename" may be wrong. Anyone dare to check that :-)
	 Be aware that IF it is wrong, to produce the problem you
	 will need two process trying to rename a file to the
	 same target at the same time. Again, I am not sure it
	 is a problem at all.
      */
      /* This is not super efficient but should work */
      new_dir->i_count++;
      ret = UMSDOS_unlink (new_dir,new_dentry);
      chkstk();
      Printk (("rename unlink ret %d -- ",ret));
      if (ret == -EISDIR){
	new_dir->i_count++;
	ret = UMSDOS_rmdir (new_dir,new_dentry);
	chkstk();
	Printk (("rename rmdir ret %d -- ",ret));
      }
      if (ret == 0){
	ret = umsdos_rename_f(old_dir,old_dentry,
			      new_dir,new_dentry,0);
	new_dir = old_dir = NULL;
      }
    }
  }
  /*
  dput (new_dentry);
  dput (old_dentry); FIXME /mn/ */
  return ret;
}