Wrapfs: a null-layer (aka wrapper) stackable file system

Maintainer: Erez Zadok <ezk AT cs DOT stonybrook DOT edu>
Web Site: <http://wrapfs.filesystems.org/>

------------------------------------------------------------------------------
MOTIVATION:

Wrapfs is a small null-layer stackable file system, similar to BSD's Nullfs.
Wrapfs is small, under 1800 lines of code.  Compare that to, say, eCryptfs
and Unionfs, each of which are over 10,000 LoC.  As such, Wrapfs is simple
and easy to read and understand.  Wrapfs is useful for several reasons:

1. Many people like to experiment with in-kernel file system ideas as a
   prototype; Wrapfs is an ideal small template from which one could modify
   the code to create new file system functionality incrementally.

2. As a platform to test and debug generic stacking problems in other Linux
   stackable file systems (e.g., ecryptfs).

3. As a way to test VFS enhancements to better support stacking in Linux.

4. Wrapfs is a very useful instructional tool, often used as a starting
   point for course assignments, for people who want a small example of who
   the Linux VFS works, or for those who want to learn to write new Linux
   file systems.

Various versions of Wrapfs appeared as part of the "fistgen" package since
1994, and have been used by numerous users world-wide.  For a more detailed
history of Wrapfs, and list of most of its known users, see the section
marked "HISTORY" below.

------------------------------------------------------------------------------
OPERATION:

This is a brief description of how Wrapfs operates.  For more information,
see the full paper published in Linux Expo 1999, titled "A Stackable File
System Interface For Linux":

	<http://www.fsl.cs.sunysb.edu/docs/linux-stacking/linux.pdf>

The basic function of a stackable file system is to pass an operation and
its arguments to the lower-level file system.  For every VFS object (inode,
dentry, file, superblock, etc.), Wrapfs keeps a one-to-one mapping of a
Wrapfs-level object to the lower one.  We call the Wrapfs object the "upper"
one, and the one below we call the "lower" one.  Wrapfs stores these
mappings as simple pointers inside the private field of the existing VFS
objects (e.g., dentry->d_fsdata, sb->s_fs_info, and a container for inodes).

There are two kinds of stackable operations: those that create new VFS
objects and those that don't.

The following distilled code snippet shows a method which doesn't create a
new object.  The method just has to pass it to the lower layer and propagate
any errors back up the VFS:

int wrapfs_unlink(struct inode *dir, struct dentry *dentry)
{
	int err;
	struct inode *lower_dir;
	struct dentry *lower_dentry;
	lower_dir = get_lower_inode(dir);
	lower_dentry = get_lower_dentry(dentry);
	err = lower_dir->i_op->unlink(lower_dir, lower_dentry);
	return err;
}

The following code snippet shows a method which creates a new object.  After
a lower object gets created, Wrapfs has to also create its own object, and
make the pointer connections between the upper and lower objects (the latter
is done via a helper routine called "interpose"):

int wrapfs_create(struct inode *dir, struct dentry *dentry, int mode)
{
	int err;
	struct dentry *lower_dentry;
	struct inode *lower_dir;
	lower_dir = wrapfs_lower_inode(dir);
	lower_dentry = wrapfs_lower_dentry(dentry);
	err = vfs_create(lower_dir, lower_dentry, mode);
	if (!err)
		err = wrapfs_interpose(dentry, dir->i_sb);
	return err;
}

The wrapfs_unlink code snippet above can be easily modified to change the
behavior of unlink(2).  For example, if an ->unlink operation is changed to
->rename, this could become the basis for an "undo" file system; or if the
lower_dentry's name gets encrypted before calling the lower ->unlink, this
could be part of an encryption file system.

------------------------------------------------------------------------------
USAGE:

First, you have to have some pre-existing directory already mounted from any
other file system, say /some/lower/path.  Then, to mount wrapfs in
/mnt/wrapfs, on that lower directory, issue this command:

# mount -t wrapfs /some/lower/path /mnt/wrapfs

To access the files via Wrapfs, use the mount point /mnt/wrapfs.

------------------------------------------------------------------------------
CAVEATS:

Stacking on NFS.  Wrapfs has been tested with LTP, racer, fsx, parallel
compile, and more.  It's been tested on top of ext2, ext3, xfs, reiserfs,
and tmpfs -- and passed all tests.  However, on top of nfs3, wrapfs has to
treat silly-deleted files as if they don't exist: in ->unlink, if we try to
vfs_unlink an NFS silly-deleted file, NFS returns EBUSY; so we simply ignore
it and return 0 (success) to the VFS.  NFS will delete this file later on
anyway.  As the VFS also has special handling for silly-deleted files, this
isn't unusual.  A cleaner way to handle this in the future is if the VFS
were to handle silly-deleted (aka "delayed-delete") files entirely at the
VFS.

------------------------------------------------------------------------------
HISTORY:

Wrapfs was developed initially in 1994 for Linux 2.1, as part of Erez
Zadok's graduate work at Columbia University.  It was designed to be a
flexible null-layer, pass-through, stackable file system, from which other
file systems would be developed and even instantiated automatically using a
high-level language.  One of the first file systems developed from Wrapfs
was a simple encryption file system called Cryptfs (eCryptfs is based on
Cryptfs).  Other examples include Gzipfs, a stackable compression file
system, and Unionfs, a stackable unification file system.  Wrapfs was
integrated into a larger package called fistgen (see www.filesystems.org),
and ported to FreeBSD and Solaris.  Wrapfs and fistgen continued to be
maintained for newer versions of kernels, but remained largely standalone
until recently: this release of Wrapfs for Linux represents a clean version
written from scratch.

Over the past 15+ years, versions of Wrapfs had been used by many users and
companies.  At one point or another, the following groups have used stacking
code based on Wrapfs.

1. PROJECTS: eCryptfs, Unionfs, mini_fo, Aufs, FindFS, StoreCompress,
   TestFS, ToPAS, and MFS.

2. COMPANIES AND RESEARCH LABS: Bell Labs's Plan 9 group, EMC,
   Hewlett-Packard, IBM Research Almaden, IBM Research Austin, Red Hat,
   SuSE, Sun Microsystems, Veritas, Booyaka, CalSoft (India), Computer Farm,
   Deutsche Bank (Germany), DreamWorks LLC, Eli Lilly and Company, FAME
   Information Services, GMX AG (Germany), IBM global services (India), IDA
   Center for Communications Research, Indra Networks, Inc., Kavi
   Corporation, Mendepie, Mitsubishi Electric (Japan), Mobile-Mind, Monster
   Labs, Morning Network (Russia), NeST Technologies, Packet General
   Networks, Inc., Outstep Technologies, Reflective Systems Group, River
   Styx Internet, SARAI Net, Saint-Petersburg Official Web Site (Russia),
   Shadow Island Games, TISCover (Germany), Trymedia Systems, Uber Admin,
   Videsh Sanchar Nigam Limited (India), Wanadoo (France), and iNsu
   Innovations.

3. UNIVERSITIES: Georgia Institute of Technology, Stanford University, UC
   Berkeley, UCLA, University of Maryland, College Park, University of
   Michigan, Ben Gurion University (Israel), Clarkson University, Clemson
   University, Deutsches Elektronen Synchrotron (Germany), Electronics and
   Telecommunications Research Institute (South Korea), Indian Institute of
   Technology (India), National Taiwan University, Pune University (India),
   The College of William \& Mary, Trinity College (Ireland), Universitaet
   Frankfurt am Main (Germany), University Hospital Olomouc (Czech
   Republic), and University of Salermo (Italy).

------------------------------------------------------------------------------
