Merge git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm

* git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm:
  dm snapshot: extend exception store functions
  dm snapshot: split out exception store implementations
  dm snapshot: rename struct exception_store
  dm snapshot: separate out exception store interface
  dm mpath: move trigger_event to system workqueue
  dm: add name and uuid to sysfs
  dm table: rework reference counting
  dm: support barriers on simple devices
  dm request: extend target interface
  dm request: add caches
  dm ioctl: allow dm_copy_name_and_uuid to return only one field
  dm log: ensure log bitmap fits on log device
  dm log: move region_size validation
  dm log: avoid reinitialising io_req on every operation
  dm: consolidate target deregistration error handling
  dm raid1: fix error count
  dm log: fix dm_io_client leak on error paths
  dm snapshot: change yield to msleep
  dm table: drop reference at unbind

21 files changed, 1293 insertions, 981 deletions

diff --git a/drivers/md/Makefile b/drivers/md/Makefile
index 1c615804ea7..72880b7e28d 100644
--- a/drivers/md/Makefile
+++ b/drivers/md/Makefile
@@ -3,9 +3,10 @@
 #
 
 dm-mod-objs	:= dm.o dm-table.o dm-target.o dm-linear.o dm-stripe.o \
-		   dm-ioctl.o dm-io.o dm-kcopyd.o
+		   dm-ioctl.o dm-io.o dm-kcopyd.o dm-sysfs.o
 dm-multipath-objs := dm-path-selector.o dm-mpath.o
-dm-snapshot-objs := dm-snap.o dm-exception-store.o
+dm-snapshot-objs := dm-snap.o dm-exception-store.o dm-snap-transient.o \
+		    dm-snap-persistent.o
 dm-mirror-objs	:= dm-raid1.o
 md-mod-objs     := md.o bitmap.o
 raid456-objs	:= raid5.o raid6algos.o raid6recov.o raid6tables.o \
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 3326750ec02..35bda49796f 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -1322,11 +1322,7 @@ static int __init dm_crypt_init(void)
 
 static void __exit dm_crypt_exit(void)
 {
-	int r = dm_unregister_target(&crypt_target);
-
-	if (r < 0)
-		DMERR("unregister failed %d", r);
-
+	dm_unregister_target(&crypt_target);
 	kmem_cache_destroy(_crypt_io_pool);
 }
 
diff --git a/drivers/md/dm-delay.c b/drivers/md/dm-delay.c
index 848b381f117..59ee1b015d2 100644
--- a/drivers/md/dm-delay.c
+++ b/drivers/md/dm-delay.c
@@ -364,11 +364,7 @@ bad_queue:
 
 static void __exit dm_delay_exit(void)
 {
-	int r = dm_unregister_target(&delay_target);
-
-	if (r < 0)
-		DMERR("unregister failed %d", r);
-
+	dm_unregister_target(&delay_target);
 	kmem_cache_destroy(delayed_cache);
 	destroy_workqueue(kdelayd_wq);
 }
diff --git a/drivers/md/dm-exception-store.c b/drivers/md/dm-exception-store.c
index 01590f3e000..dccbfb0e010 100644
--- a/drivers/md/dm-exception-store.c
+++ b/drivers/md/dm-exception-store.c
@@ -1,756 +1,45 @@
 /*
- * dm-exception-store.c
- *
  * Copyright (C) 2001-2002 Sistina Software (UK) Limited.
- * Copyright (C) 2006 Red Hat GmbH
+ * Copyright (C) 2006-2008 Red Hat GmbH
  *
  * This file is released under the GPL.
  */
 
-#include "dm-snap.h"
+#include "dm-exception-store.h"
 
 #include <linux/mm.h>
 #include <linux/pagemap.h>
 #include <linux/vmalloc.h>
 #include <linux/slab.h>
-#include <linux/dm-io.h>
-#include <linux/dm-kcopyd.h>
-
-#define DM_MSG_PREFIX "snapshots"
-#define DM_CHUNK_SIZE_DEFAULT_SECTORS 32	/* 16KB */
-
-/*-----------------------------------------------------------------
- * Persistent snapshots, by persistent we mean that the snapshot
- * will survive a reboot.
- *---------------------------------------------------------------*/
-
-/*
- * We need to store a record of which parts of the origin have
- * been copied to the snapshot device.  The snapshot code
- * requires that we copy exception chunks to chunk aligned areas
- * of the COW store.  It makes sense therefore, to store the
- * metadata in chunk size blocks.
- *
- * There is no backward or forward compatibility implemented,
- * snapshots with different disk versions than the kernel will
- * not be usable.  It is expected that "lvcreate" will blank out
- * the start of a fresh COW device before calling the snapshot
- * constructor.
- *
- * The first chunk of the COW device just contains the header.
- * After this there is a chunk filled with exception metadata,
- * followed by as many exception chunks as can fit in the
- * metadata areas.
- *
- * All on disk structures are in little-endian format.  The end
- * of the exceptions info is indicated by an exception with a
- * new_chunk of 0, which is invalid since it would point to the
- * header chunk.
- */
-
-/*
- * Magic for persistent snapshots: "SnAp" - Feeble isn't it.
- */
-#define SNAP_MAGIC 0x70416e53
-
-/*
- * The on-disk version of the metadata.
- */
-#define SNAPSHOT_DISK_VERSION 1
-
-struct disk_header {
-	uint32_t magic;
-
-	/*
-	 * Is this snapshot valid.  There is no way of recovering
-	 * an invalid snapshot.
-	 */
-	uint32_t valid;
-
-	/*
-	 * Simple, incrementing version. no backward
-	 * compatibility.
-	 */
-	uint32_t version;
-
-	/* In sectors */
-	uint32_t chunk_size;
-};
-
-struct disk_exception {
-	uint64_t old_chunk;
-	uint64_t new_chunk;
-};
-
-struct commit_callback {
-	void (*callback)(void *, int success);
-	void *context;
-};
-
-/*
- * The top level structure for a persistent exception store.
- */
-struct pstore {
-	struct dm_snapshot *snap;	/* up pointer to my snapshot */
-	int version;
-	int valid;
-	uint32_t exceptions_per_area;
-
-	/*
-	 * Now that we have an asynchronous kcopyd there is no
-	 * need for large chunk sizes, so it wont hurt to have a
-	 * whole chunks worth of metadata in memory at once.
-	 */
-	void *area;
-
-	/*
-	 * An area of zeros used to clear the next area.
-	 */
-	void *zero_area;
-
-	/*
-	 * Used to keep track of which metadata area the data in
-	 * 'chunk' refers to.
-	 */
-	chunk_t current_area;
-
-	/*
-	 * The next free chunk for an exception.
-	 */
-	chunk_t next_free;
-
-	/*
-	 * The index of next free exception in the current
-	 * metadata area.
-	 */
-	uint32_t current_committed;
-
-	atomic_t pending_count;
-	uint32_t callback_count;
-	struct commit_callback *callbacks;
-	struct dm_io_client *io_client;
-
-	struct workqueue_struct *metadata_wq;
-};
-
-static unsigned sectors_to_pages(unsigned sectors)
-{
-	return DIV_ROUND_UP(sectors, PAGE_SIZE >> 9);
-}
-
-static int alloc_area(struct pstore *ps)
-{
-	int r = -ENOMEM;
-	size_t len;
-
-	len = ps->snap->chunk_size << SECTOR_SHIFT;
-
-	/*
-	 * Allocate the chunk_size block of memory that will hold
-	 * a single metadata area.
-	 */
-	ps->area = vmalloc(len);
-	if (!ps->area)
-		return r;
-
-	ps->zero_area = vmalloc(len);
-	if (!ps->zero_area) {
-		vfree(ps->area);
-		return r;
-	}
-	memset(ps->zero_area, 0, len);
-
-	return 0;
-}
-
-static void free_area(struct pstore *ps)
-{
-	vfree(ps->area);
-	ps->area = NULL;
-	vfree(ps->zero_area);
-	ps->zero_area = NULL;
-}
-
-struct mdata_req {
-	struct dm_io_region *where;
-	struct dm_io_request *io_req;
-	struct work_struct work;
-	int result;
-};
-
-static void do_metadata(struct work_struct *work)
-{
-	struct mdata_req *req = container_of(work, struct mdata_req, work);
-
-	req->result = dm_io(req->io_req, 1, req->where, NULL);
-}
-
-/*
- * Read or write a chunk aligned and sized block of data from a device.
- */
-static int chunk_io(struct pstore *ps, chunk_t chunk, int rw, int metadata)
-{
-	struct dm_io_region where = {
-		.bdev = ps->snap->cow->bdev,
-		.sector = ps->snap->chunk_size * chunk,
-		.count = ps->snap->chunk_size,
-	};
-	struct dm_io_request io_req = {
-		.bi_rw = rw,
-		.mem.type = DM_IO_VMA,
-		.mem.ptr.vma = ps->area,
-		.client = ps->io_client,
-		.notify.fn = NULL,
-	};
-	struct mdata_req req;
-
-	if (!metadata)
-		return dm_io(&io_req, 1, &where, NULL);
-
-	req.where = &where;
-	req.io_req = &io_req;
-
-	/*
-	 * Issue the synchronous I/O from a different thread
-	 * to avoid generic_make_request recursion.
-	 */
-	INIT_WORK(&req.work, do_metadata);
-	queue_work(ps->metadata_wq, &req.work);
-	flush_workqueue(ps->metadata_wq);
-
-	return req.result;
-}
-
-/*
- * Convert a metadata area index to a chunk index.
- */
-static chunk_t area_location(struct pstore *ps, chunk_t area)
-{
-	return 1 + ((ps->exceptions_per_area + 1) * area);
-}
-
-/*
- * Read or write a metadata area.  Remembering to skip the first
- * chunk which holds the header.
- */
-static int area_io(struct pstore *ps, int rw)
-{
-	int r;
-	chunk_t chunk;
-
-	chunk = area_location(ps, ps->current_area);
-
-	r = chunk_io(ps, chunk, rw, 0);
-	if (r)
-		return r;
-
-	return 0;
-}
-
-static void zero_memory_area(struct pstore *ps)
-{
-	memset(ps->area, 0, ps->snap->chunk_size << SECTOR_SHIFT);
-}
-
-static int zero_disk_area(struct pstore *ps, chunk_t area)
-{
-	struct dm_io_region where = {
-		.bdev = ps->snap->cow->bdev,
-		.sector = ps->snap->chunk_size * area_location(ps, area),
-		.count = ps->snap->chunk_size,
-	};
-	struct dm_io_request io_req = {
-		.bi_rw = WRITE,
-		.mem.type = DM_IO_VMA,
-		.mem.ptr.vma = ps->zero_area,
-		.client = ps->io_client,
-		.notify.fn = NULL,
-	};
-
-	return dm_io(&io_req, 1, &where, NULL);
-}
-
-static int read_header(struct pstore *ps, int *new_snapshot)
-{
-	int r;
-	struct disk_header *dh;
-	chunk_t chunk_size;
-	int chunk_size_supplied = 1;
-
-	/*
-	 * Use default chunk size (or hardsect_size, if larger) if none supplied
-	 */
-	if (!ps->snap->chunk_size) {
-        	ps->snap->chunk_size = max(DM_CHUNK_SIZE_DEFAULT_SECTORS,
-		    bdev_hardsect_size(ps->snap->cow->bdev) >> 9);
-		ps->snap->chunk_mask = ps->snap->chunk_size - 1;
-		ps->snap->chunk_shift = ffs(ps->snap->chunk_size) - 1;
-		chunk_size_supplied = 0;
-	}
-
-	ps->io_client = dm_io_client_create(sectors_to_pages(ps->snap->
-							     chunk_size));
-	if (IS_ERR(ps->io_client))
-		return PTR_ERR(ps->io_client);
-
-	r = alloc_area(ps);
-	if (r)
-		return r;
-
-	r = chunk_io(ps, 0, READ, 1);
-	if (r)
-		goto bad;
-
-	dh = (struct disk_header *) ps->area;
-
-	if (le32_to_cpu(dh->magic) == 0) {
-		*new_snapshot = 1;
-		return 0;
-	}
-
-	if (le32_to_cpu(dh->magic) != SNAP_MAGIC) {
-		DMWARN("Invalid or corrupt snapshot");
-		r = -ENXIO;
-		goto bad;
-	}
-
-	*new_snapshot = 0;
-	ps->valid = le32_to_cpu(dh->valid);
-	ps->version = le32_to_cpu(dh->version);
-	chunk_size = le32_to_cpu(dh->chunk_size);
-
-	if (!chunk_size_supplied || ps->snap->chunk_size == chunk_size)
-		return 0;
-
-	DMWARN("chunk size %llu in device metadata overrides "
-	       "table chunk size of %llu.",
-	       (unsigned long long)chunk_size,
-	       (unsigned long long)ps->snap->chunk_size);
-
-	/* We had a bogus chunk_size. Fix stuff up. */
-	free_area(ps);
-
-	ps->snap->chunk_size = chunk_size;
-	ps->snap->chunk_mask = chunk_size - 1;
-	ps->snap->chunk_shift = ffs(chunk_size) - 1;
-
-	r = dm_io_client_resize(sectors_to_pages(ps->snap->chunk_size),
-				ps->io_client);
-	if (r)
-		return r;
-
-	r = alloc_area(ps);
-	return r;
-
-bad:
-	free_area(ps);
-	return r;
-}
-
-static int write_header(struct pstore *ps)
-{
-	struct disk_header *dh;
-
-	memset(ps->area, 0, ps->snap->chunk_size << SECTOR_SHIFT);
-
-	dh = (struct disk_header *) ps->area;
-	dh->magic = cpu_to_le32(SNAP_MAGIC);
-	dh->valid = cpu_to_le32(ps->valid);
-	dh->version = cpu_to_le32(ps->version);
-	dh->chunk_size = cpu_to_le32(ps->snap->chunk_size);
-
-	return chunk_io(ps, 0, WRITE, 1);
-}
-
-/*
- * Access functions for the disk exceptions, these do the endian conversions.
- */
-static struct disk_exception *get_exception(struct pstore *ps, uint32_t index)
-{
-	BUG_ON(index >= ps->exceptions_per_area);
-
-	return ((struct disk_exception *) ps->area) + index;
-}
 
-static void read_exception(struct pstore *ps,
-			   uint32_t index, struct disk_exception *result)
-{
-	struct disk_exception *e = get_exception(ps, index);
-
-	/* copy it */
-	result->old_chunk = le64_to_cpu(e->old_chunk);
-	result->new_chunk = le64_to_cpu(e->new_chunk);
-}
-
-static void write_exception(struct pstore *ps,
-			    uint32_t index, struct disk_exception *de)
-{
-	struct disk_exception *e = get_exception(ps, index);
-
-	/* copy it */
-	e->old_chunk = cpu_to_le64(de->old_chunk);
-	e->new_chunk = cpu_to_le64(de->new_chunk);
-}
+#define DM_MSG_PREFIX "snapshot exception stores"
 
-/*
- * Registers the exceptions that are present in the current area.
- * 'full' is filled in to indicate if the area has been
- * filled.
- */
-static int insert_exceptions(struct pstore *ps, int *full)
+int dm_exception_store_init(void)
 {
 	int r;
-	unsigned int i;
-	struct disk_exception de;
-
-	/* presume the area is full */
-	*full = 1;
-
-	for (i = 0; i < ps->exceptions_per_area; i++) {
-		read_exception(ps, i, &de);
-
-		/*
-		 * If the new_chunk is pointing at the start of
-		 * the COW device, where the first metadata area
-		 * is we know that we've hit the end of the
-		 * exceptions.  Therefore the area is not full.
-		 */
-		if (de.new_chunk == 0LL) {
-			ps->current_committed = i;
-			*full = 0;
-			break;
-		}
-
-		/*
-		 * Keep track of the start of the free chunks.
-		 */
-		if (ps->next_free <= de.new_chunk)
-			ps->next_free = de.new_chunk + 1;
-
-		/*
-		 * Otherwise we add the exception to the snapshot.
-		 */
-		r = dm_add_exception(ps->snap, de.old_chunk, de.new_chunk);
-		if (r)
-			return r;
-	}
-
-	return 0;
-}
-
-static int read_exceptions(struct pstore *ps)
-{
-	int r, full = 1;
-
-	/*
-	 * Keeping reading chunks and inserting exceptions until
-	 * we find a partially full area.
-	 */
-	for (ps->current_area = 0; full; ps->current_area++) {
-		r = area_io(ps, READ);
-		if (r)
-			return r;
 
-		r = insert_exceptions(ps, &full);
-		if (r)
-			return r;
+	r = dm_transient_snapshot_init();
+	if (r) {
+		DMERR("Unable to register transient exception store type.");
+		goto transient_fail;
 	}
 
-	ps->current_area--;
-
-	return 0;
-}
-
-static struct pstore *get_info(struct exception_store *store)
-{
-	return (struct pstore *) store->context;
-}
-
-static void persistent_fraction_full(struct exception_store *store,
-				     sector_t *numerator, sector_t *denominator)
-{
-	*numerator = get_info(store)->next_free * store->snap->chunk_size;
-	*denominator = get_dev_size(store->snap->cow->bdev);
-}
-
-static void persistent_destroy(struct exception_store *store)
-{
-	struct pstore *ps = get_info(store);
-
-	destroy_workqueue(ps->metadata_wq);
-	dm_io_client_destroy(ps->io_client);
-	vfree(ps->callbacks);
-	free_area(ps);
-	kfree(ps);
-}
-
-static int persistent_read_metadata(struct exception_store *store)
-{
-	int r, uninitialized_var(new_snapshot);
-	struct pstore *ps = get_info(store);
-
-	/*
-	 * Read the snapshot header.
-	 */
-	r = read_header(ps, &new_snapshot);
-	if (r)
-		return r;
-
-	/*
-	 * Now we know correct chunk_size, complete the initialisation.
-	 */
-	ps->exceptions_per_area = (ps->snap->chunk_size << SECTOR_SHIFT) /
-				  sizeof(struct disk_exception);
-	ps->callbacks = dm_vcalloc(ps->exceptions_per_area,
-			sizeof(*ps->callbacks));
-	if (!ps->callbacks)
-		return -ENOMEM;
-
-	/*
-	 * Do we need to setup a new snapshot ?
-	 */
-	if (new_snapshot) {
-		r = write_header(ps);
-		if (r) {
-			DMWARN("write_header failed");
-			return r;
-		}
-
-		ps->current_area = 0;
-		zero_memory_area(ps);
-		r = zero_disk_area(ps, 0);
-		if (r) {
-			DMWARN("zero_disk_area(0) failed");
-			return r;
-		}
-	} else {
-		/*
-		 * Sanity checks.
-		 */
-		if (ps->version != SNAPSHOT_DISK_VERSION) {
-			DMWARN("unable to handle snapshot disk version %d",
-			       ps->version);
-			return -EINVAL;
-		}
-
-		/*
-		 * Metadata are valid, but snapshot is invalidated
-		 */
-		if (!ps->valid)
-			return 1;
-
-		/*
-		 * Read the metadata.
-		 */
-		r = read_exceptions(ps);
-		if (r)
-			return r;
+	r = dm_persistent_snapshot_init();
+	if (r) {
+		DMERR("Unable to register persistent exception store type");
+		goto persistent_fail;
 	}
 
 	return 0;
-}
-
-static int persistent_prepare(struct exception_store *store,
-			      struct dm_snap_exception *e)
-{
-	struct pstore *ps = get_info(store);
-	uint32_t stride;
-	chunk_t next_free;
-	sector_t size = get_dev_size(store->snap->cow->bdev);
-
-	/* Is there enough room ? */
-	if (size < ((ps->next_free + 1) * store->snap->chunk_size))
-		return -ENOSPC;
 
-	e->new_chunk = ps->next_free;
-
-	/*
-	 * Move onto the next free pending, making sure to take
-	 * into account the location of the metadata chunks.
-	 */
-	stride = (ps->exceptions_per_area + 1);
-	next_free = ++ps->next_free;
-	if (sector_div(next_free, stride) == 1)
-		ps->next_free++;
-
-	atomic_inc(&ps->pending_count);
-	return 0;
-}
-
-static void persistent_commit(struct exception_store *store,
-			      struct dm_snap_exception *e,
-			      void (*callback) (void *, int success),
-			      void *callback_context)
-{
-	unsigned int i;
-	struct pstore *ps = get_info(store);
-	struct disk_exception de;
-	struct commit_callback *cb;
-
-	de.old_chunk = e->old_chunk;
-	de.new_chunk = e->new_chunk;
-	write_exception(ps, ps->current_committed++, &de);
-
-	/*
-	 * Add the callback to the back of the array.  This code
-	 * is the only place where the callback array is
-	 * manipulated, and we know that it will never be called
-	 * multiple times concurrently.
-	 */
-	cb = ps->callbacks + ps->callback_count++;
-	cb->callback = callback;
-	cb->context = callback_context;
-
-	/*
-	 * If there are exceptions in flight and we have not yet
-	 * filled this metadata area there's nothing more to do.
-	 */
-	if (!atomic_dec_and_test(&ps->pending_count) &&
-	    (ps->current_committed != ps->exceptions_per_area))
-		return;
-
-	/*
-	 * If we completely filled the current area, then wipe the next one.
-	 */
-	if ((ps->current_committed == ps->exceptions_per_area) &&
-	     zero_disk_area(ps, ps->current_area + 1))
-		ps->valid = 0;
-
-	/*
-	 * Commit exceptions to disk.
-	 */
-	if (ps->valid && area_io(ps, WRITE))
-		ps->valid = 0;
-
-	/*
-	 * Advance to the next area if this one is full.
-	 */
-	if (ps->current_committed == ps->exceptions_per_area) {
-		ps->current_committed = 0;
-		ps->current_area++;
-		zero_memory_area(ps);
-	}
-
-	for (i = 0; i < ps->callback_count; i++) {
-		cb = ps->callbacks + i;
-		cb->callback(cb->context, ps->valid);
-	}
-
-	ps->callback_count = 0;
-}
-
-static void persistent_drop(struct exception_store *store)
-{
-	struct pstore *ps = get_info(store);
-
-	ps->valid = 0;
-	if (write_header(ps))
-		DMWARN("write header failed");
-}
-
-int dm_create_persistent(struct exception_store *store)
-{
-	struct pstore *ps;
-
-	/* allocate the pstore */
-	ps = kmalloc(sizeof(*ps), GFP_KERNEL);
-	if (!ps)
-		return -ENOMEM;
-
-	ps->snap = store->snap;
-	ps->valid = 1;
-	ps->version = SNAPSHOT_DISK_VERSION;
-	ps->area = NULL;
-	ps->next_free = 2;	/* skipping the header and first area */
-	ps->current_committed = 0;
-
-	ps->callback_count = 0;
-	atomic_set(&ps->pending_count, 0);
-	ps->callbacks = NULL;
-
-	ps->metadata_wq = create_singlethread_workqueue("ksnaphd");
-	if (!ps->metadata_wq) {
-		kfree(ps);
-		DMERR("couldn't start header metadata update thread");
-		return -ENOMEM;
-	}
-
-	store->destroy = persistent_destroy;
-	store->read_metadata = persistent_read_metadata;
-	store->prepare_exception = persistent_prepare;
-	store->commit_exception = persistent_commit;
-	store->drop_snapshot = persistent_drop;
-	store->fraction_full = persistent_fraction_full;
-	store->context = ps;
-
-	return 0;
-}
-
-/*-----------------------------------------------------------------
- * Implementation of the store for non-persistent snapshots.
- *---------------------------------------------------------------*/
-struct transient_c {
-	sector_t next_free;
-};
-
-static void transient_destroy(struct exception_store *store)
-{
-	kfree(store->context);
-}
-
-static int transient_read_metadata(struct exception_store *store)
-{
-	return 0;
-}
-
-static int transient_prepare(struct exception_store *store,
-			     struct dm_snap_exception *e)
-{
-	struct transient_c *tc = (struct transient_c *) store->context;
-	sector_t size = get_dev_size(store->snap->cow->bdev);
-
-	if (size < (tc->next_free + store->snap->chunk_size))
-		return -1;
-
-	e->new_chunk = sector_to_chunk(store->snap, tc->next_free);
-	tc->next_free += store->snap->chunk_size;
-
-	return 0;
-}
-
-static void transient_commit(struct exception_store *store,
-			     struct dm_snap_exception *e,
-			     void (*callback) (void *, int success),
-			     void *callback_context)
-{
-	/* Just succeed */
-	callback(callback_context, 1);
-}
-
-static void transient_fraction_full(struct exception_store *store,
-				    sector_t *numerator, sector_t *denominator)
-{
-	*numerator = ((struct transient_c *) store->context)->next_free;
-	*denominator = get_dev_size(store->snap->cow->bdev);
+persistent_fail:
+	dm_persistent_snapshot_exit();
+transient_fail:
+	return r;
 }
 
-int dm_create_transient(struct exception_store *store)
+void dm_exception_store_exit(void)
 {
-	struct transient_c *tc;
-
-	store->destroy = transient_destroy;
-	store->read_metadata = transient_read_metadata;
-	store->prepare_exception = transient_prepare;
-	store->commit_exception = transient_commit;
-	store->drop_snapshot = NULL;
-	store->fraction_full = transient_fraction_full;
-
-	tc = kmalloc(sizeof(struct transient_c), GFP_KERNEL);
-	if (!tc)
-		return -ENOMEM;
-
-	tc->next_free = 0;
-	store->context = tc;
-
-	return 0;
+	dm_persistent_snapshot_exit();
+	dm_transient_snapshot_exit();
 }
diff --git a/drivers/md/dm-exception-store.h b/drivers/md/dm-exception-store.h
new file mode 100644
index 00000000000..bb9f33d5daa
--- /dev/null
+++ b/drivers/md/dm-exception-store.h
@@ -0,0 +1,148 @@
+/*
+ * Copyright (C) 2001-2002 Sistina Software (UK) Limited.
+ * Copyright (C) 2008 Red Hat, Inc. All rights reserved.
+ *
+ * Device-mapper snapshot exception store.
+ *
+ * This file is released under the GPL.
+ */
+
+#ifndef _LINUX_DM_EXCEPTION_STORE
+#define _LINUX_DM_EXCEPTION_STORE
+
+#include <linux/blkdev.h>
+#include <linux/device-mapper.h>
+
+/*
+ * The snapshot code deals with largish chunks of the disk at a
+ * time. Typically 32k - 512k.
+ */
+typedef sector_t chunk_t;
+
+/*
+ * An exception is used where an old chunk of data has been
+ * replaced by a new one.
+ * If chunk_t is 64 bits in size, the top 8 bits of new_chunk hold the number
+ * of chunks that follow contiguously.  Remaining bits hold the number of the
+ * chunk within the device.
+ */
+struct dm_snap_exception {
+	struct list_head hash_list;
+
+	chunk_t old_chunk;
+	chunk_t new_chunk;
+};
+
+/*
+ * Abstraction to handle the meta/layout of exception stores (the
+ * COW device).
+ */
+struct dm_exception_store {
+	/*
+	 * Destroys this object when you've finished with it.
+	 */
+	void (*destroy) (struct dm_exception_store *store);
+
+	/*
+	 * The target shouldn't read the COW device until this is
+	 * called.  As exceptions are read from the COW, they are
+	 * reported back via the callback.
+	 */
+	int (*read_metadata) (struct dm_exception_store *store,
+			      int (*callback)(void *callback_context,
+					      chunk_t old, chunk_t new),
+			      void *callback_context);
+
+	/*
+	 * Find somewhere to store the next exception.
+	 */
+	int (*prepare_exception) (struct dm_exception_store *store,
+				  struct dm_snap_exception *e);
+
+	/*
+	 * Update the metadata with this exception.
+	 */
+	void (*commit_exception) (struct dm_exception_store *store,
+				  struct dm_snap_exception *e,
+				  void (*callback) (void *, int success),
+				  void *callback_context);
+
+	/*
+	 * The snapshot is invalid, note this in the metadata.
+	 */
+	void (*drop_snapshot) (struct dm_exception_store *store);
+
+	int (*status) (struct dm_exception_store *store, status_type_t status,
+		       char *result, unsigned int maxlen);
+
+	/*
+	 * Return how full the snapshot is.
+	 */
+	void (*fraction_full) (struct dm_exception_store *store,
+			       sector_t *numerator,
+			       sector_t *denominator);
+
+	struct dm_snapshot *snap;
+	void *context;
+};
+
+/*
+ * Funtions to manipulate consecutive chunks
+ */
+#  if defined(CONFIG_LBD) || (BITS_PER_LONG == 64)
+#    define DM_CHUNK_CONSECUTIVE_BITS 8
+#    define DM_CHUNK_NUMBER_BITS 56
+
+static inline chunk_t dm_chunk_number(chunk_t chunk)
+{
+	return chunk & (chunk_t)((1ULL << DM_CHUNK_NUMBER_BITS) - 1ULL);
+}
+
+static inline unsigned dm_consecutive_chunk_count(struct dm_snap_exception *e)
+{
+	return e->new_chunk >> DM_CHUNK_NUMBER_BITS;
+}
+
+static inline void dm_consecutive_chunk_count_inc(struct dm_snap_exception *e)
+{
+	e->new_chunk += (1ULL << DM_CHUNK_NUMBER_BITS);
+
+	BUG_ON(!dm_consecutive_chunk_count(e));
+}
+
+#  else
+#    define DM_CHUNK_CONSECUTIVE_BITS 0
+
+static inline chunk_t dm_chunk_number(chunk_t chunk)
+{
+	return chunk;
+}
+
+static inline unsigned dm_consecutive_chunk_count(struct dm_snap_exception *e)
+{
+	return 0;
+}
+
+static inline void dm_consecutive_chunk_count_inc(struct dm_snap_exception *e)
+{
+}
+
+#  endif
+
+int dm_exception_store_init(void);
+void dm_exception_store_exit(void);
+
+/*
+ * Two exception store implementations.
+ */
+int dm_persistent_snapshot_init(void);
+void dm_persistent_snapshot_exit(void);
+
+int dm_transient_snapshot_init(void);
+void dm_transient_snapshot_exit(void);
+
+int dm_create_persistent(struct dm_exception_store *store);
+
+int dm_create_transient(struct dm_exception_store *store);
+
+#endif /* _LINUX_DM_EXCEPTION_STORE */
diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c
index 777c948180f..54d0588fc1f 100644
--- a/drivers/md/dm-ioctl.c
+++ b/drivers/md/dm-ioctl.c
@@ -233,7 +233,7 @@ static void __hash_remove(struct hash_cell *hc)
 	}
 
 	if (hc->new_map)
-		dm_table_put(hc->new_map);
+		dm_table_destroy(hc->new_map);
 	dm_put(hc->md);
 	free_cell(hc);
 }
@@ -827,8 +827,8 @@ static int do_resume(struct dm_ioctl *param)
 
 		r = dm_swap_table(md, new_map);
 		if (r) {
+			dm_table_destroy(new_map);
 			dm_put(md);
-			dm_table_put(new_map);
 			return r;
 		}
 
@@ -836,8 +836,6 @@ static int do_resume(struct dm_ioctl *param)
 			set_disk_ro(dm_disk(md), 0);
 		else
 			set_disk_ro(dm_disk(md), 1);
-
-		dm_table_put(new_map);
 	}
 
 	if (dm_suspended(md))
@@ -1080,7 +1078,7 @@ static int table_load(struct dm_ioctl *param, size_t param_size)
 	}
 
 	if (hc->new_map)
-		dm_table_put(hc->new_map);
+		dm_table_destroy(hc->new_map);
 	hc->new_map =