[LogFS] add new flash file system

This is a new flash file system. See
Documentation/filesystems/logfs.txt

Signed-off-by: Joern Engel <joern@logfs.org>

25 files changed, 10446 insertions, 0 deletions

diff --git a/Documentation/filesystems/00-INDEX b/Documentation/filesystems/00-INDEX
index f15621ee5599..d362aa543b27 100644
--- a/Documentation/filesystems/00-INDEX
+++ b/Documentation/filesystems/00-INDEX
@@ -62,6 +62,8 @@ jfs.txt
 	- info and mount options for the JFS filesystem.
 locks.txt
 	- info on file locking implementations, flock() vs. fcntl(), etc.
+logfs.txt
+	- info on the LogFS flash filesystem.
 mandatory-locking.txt
 	- info on the Linux implementation of Sys V mandatory file locking.
 ncpfs.txt
diff --git a/Documentation/filesystems/logfs.txt b/Documentation/filesystems/logfs.txt
new file mode 100644
index 000000000000..e64c94ba401a
--- /dev/null
+++ b/Documentation/filesystems/logfs.txt
@@ -0,0 +1,241 @@
+
+The LogFS Flash Filesystem
+==========================
+
+Specification
+=============
+
+Superblocks
+-----------
+
+Two superblocks exist at the beginning and end of the filesystem.
+Each superblock is 256 Bytes large, with another 3840 Bytes reserved
+for future purposes, making a total of 4096 Bytes.
+
+Superblock locations may differ for MTD and block devices.  On MTD the
+first non-bad block contains a superblock in the first 4096 Bytes and
+the last non-bad block contains a superblock in the last 4096 Bytes.
+On block devices, the first 4096 Bytes of the device contain the first
+superblock and the last aligned 4096 Byte-block contains the second
+superblock.
+
+For the most part, the superblocks can be considered read-only.  They
+are written only to correct errors detected within the superblocks,
+move the journal and change the filesystem parameters through tunefs.
+As a result, the superblock does not contain any fields that require
+constant updates, like the amount of free space, etc.
+
+Segments
+--------
+
+The space in the device is split up into equal-sized segments.
+Segments are the primary write unit of LogFS.  Within each segments,
+writes happen from front (low addresses) to back (high addresses.  If
+only a partial segment has been written, the segment number, the
+current position within and optionally a write buffer are stored in
+the journal.
+
+Segments are erased as a whole.  Therefore Garbage Collection may be
+required to completely free a segment before doing so.
+
+Journal
+--------
+
+The journal contains all global information about the filesystem that
+is subject to frequent change.  At mount time, it has to be scanned
+for the most recent commit entry, which contains a list of pointers to
+all currently valid entries.
+
+Object Store
+------------
+
+All space except for the superblocks and journal is part of the object
+store.  Each segment contains a segment header and a number of
+objects, each consisting of the object header and the payload.
+Objects are either inodes, directory entries (dentries), file data
+blocks or indirect blocks.
+
+Levels
+------
+
+Garbage collection (GC) may fail if all data is written
+indiscriminately.  One requirement of GC is that data is seperated
+roughly according to the distance between the tree root and the data.
+Effectively that means all file data is on level 0, indirect blocks
+are on levels 1, 2, 3 4 or 5 for 1x, 2x, 3x, 4x or 5x indirect blocks,
+respectively.  Inode file data is on level 6 for the inodes and 7-11
+for indirect blocks.
+
+Each segment contains objects of a single level only.  As a result,
+each level requires its own seperate segment to be open for writing.
+
+Inode File
+----------
+
+All inodes are stored in a special file, the inode file.  Single
+exception is the inode file's inode (master inode) which for obvious
+reasons is stored in the journal instead.  Instead of data blocks, the
+leaf nodes of the inode files are inodes.
+
+Aliases
+-------
+
+Writes in LogFS are done by means of a wandering tree.  A naïve
+implementation would require that for each write or a block, all
+parent blocks are written as well, since the block pointers have
+changed.  Such an implementation would not be very efficient.
+
+In LogFS, the block pointer changes are cached in the journal by means
+of alias entries.  Each alias consists of its logical address - inode
+number, block index, level and child number (index into block) - and
+the changed data.  Any 8-byte word can be changes in this manner.
+
+Currently aliases are used for block pointers, file size, file used
+bytes and the height of an inodes indirect tree.
+
+Segment Aliases
+---------------
+
+Related to regular aliases, these are used to handle bad blocks.
+Initially, bad blocks are handled by moving the affected segment
+content to a spare segment and noting this move in the journal with a
+segment alias, a simple (to, from) tupel.  GC will later empty this
+segment and the alias can be removed again.  This is used on MTD only.
+
+Vim
+---
+
+By cleverly predicting the life time of data, it is possible to
+seperate long-living data from short-living data and thereby reduce
+the GC overhead later.  Each type of distinc life expectency (vim) can
+have a seperate segment open for writing.  Each (level, vim) tupel can
+be open just once.  If an open segment with unknown vim is encountered
+at mount time, it is closed and ignored henceforth.
+
+Indirect Tree
+-------------
+
+Inodes in LogFS are similar to FFS-style filesystems with direct and
+indirect block pointers.  One difference is that LogFS uses a single
+indirect pointer that can be either a 1x, 2x, etc. indirect pointer.
+A height field in the inode defines the height of the indirect tree
+and thereby the indirection of the pointer.
+
+Another difference is the addressing of indirect blocks.  In LogFS,
+the first 16 pointers in the first indirect block are left empty,
+corresponding to the 16 direct pointers in the inode.  In ext2 (maybe
+others as well) the first pointer in the first indirect block
+corresponds to logical block 12, skipping the 12 direct pointers.
+So where ext2 is using arithmetic to better utilize space, LogFS keeps
+arithmetic simple and uses compression to save space.
+
+Compression
+-----------
+
+Both file data and metadata can be compressed.  Compression for file
+data can be enabled with chattr +c and disabled with chattr -c.  Doing
+so has no effect on existing data, but new data will be stored
+accordingly.  New inodes will inherit the compression flag of the
+parent directory.
+
+Metadata is always compressed.  However, the space accounting ignores
+this and charges for the uncompressed size.  Failing to do so could
+result in GC failures when, after moving some data, indirect blocks
+compress worse than previously.  Even on a 100% full medium, GC may
+not consume any extra space, so the compression gains are lost space
+to the user.
+
+However, they are not lost space to the filesystem internals.  By
+cheating the user for those bytes, the filesystem gained some slack
+space and GC will run less often and faster.
+
+Garbage Collection and Wear Leveling
+------------------------------------
+
+Garbage collection is invoked whenever the number of free segments
+falls below a threshold.  The best (known) candidate is picked based
+on the least amount of valid data contained in the segment.  All
+remaining valid data is copied elsewhere, thereby invalidating it.
+
+The GC code also checks for aliases and writes then back if their
+number gets too large.
+
+Wear leveling is done by occasionally picking a suboptimal segment for
+garbage collection.  If a stale segments erase count is significantly
+lower than the active segments' erase counts, it will be picked.  Wear
+leveling is rate limited, so it will never monopolize the device for
+more than one segment worth at a time.
+
+Values for "occasionally", "significantly lower" are compile time
+constants.
+
+Hashed directories
+------------------
+
+To satisfy efficient lookup(), directory entries are hashed and
+located based on the hash.  In order to both support large directories
+and not be overly inefficient for small directories, several hash
+tables of increasing size are used.  For each table, the hash value
+modulo the table size gives the table index.
+
+Tables sizes are chosen to limit the number of indirect blocks with a
+fully populated table to 0, 1, 2 or 3 respectively.  So the first
+table contains 16 entries, the second 512-16, etc.
+
+The last table is special in several ways.  First its size depends on
+the effective 32bit limit on telldir/seekdir cookies.  Since logfs
+uses the upper half of the address space for indirect blocks, the size
+is limited to 2^31.  Secondly the table contains hash buckets with 16
+entries each.
+
+Using single-entry buckets would result in birthday "attacks".  At
+just 2^16 used entries, hash collisions would be likely (P >= 0.5).
+My math skills are insufficient to do the combinatorics for the 17x
+collisions necessary to overflow a bucket, but testing showed that in
+10,000 runs the lowest directory fill before a bucket overflow was
+188,057,130 entries with an average of 315,149,915 entries.  So for
+directory sizes of up to a million, bucket overflows should be
+virtually impossible under normal circumstances.
+
+With carefully chosen filenames, it is obviously possible to cause an
+overflow with just 21 entries (4 higher tables + 16 entries + 1).  So
+there may be a security concern if a malicious user has write access
+to a directory.
+
+Open For Discussion
+===================
+
+Device Address Space
+--------------------
+
+A device address space is used for caching.  Both block devices and
+MTD provide functions to either read a single page or write a segment.
+Partial segments may be written for data integrity, but where possible
+complete segments are written for performance on simple block device
+flash media.
+
+Meta Inodes
+-----------
+
+Inodes are stored in the inode file, which is just a regular file for
+most purposes.  At umount time, however, the inode file needs to
+remain open until all dirty inodes are written.  So
+generic_shutdown_super() may not close this inode, but shouldn't
+complain about remaining inodes due to the inode file either.  Same
+goes for mapping inode of the device address space.
+
+Currently logfs uses a hack that essentially copies part of fs/inode.c
+code over.  A general solution would be preferred.
+
+Indirect block mapping
+----------------------
+
+With compression, the block device (or mapping inode) cannot be used
+to cache indirect blocks.  Some other place is required.  Currently
+logfs uses the top half of each inode's address space.  The low 8TB
+(on 32bit) are filled with file data, the high 8TB are used for
+indirect blocks.
+
+One problem is that 16TB files created on 64bit systems actually have
+data in the top 8TB.  But files >16TB would cause problems anyway, so
+only the limit has changed.
diff --git a/fs/Kconfig b/fs/Kconfig
index 64d44efad7a5..7405f071be67 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -177,6 +177,7 @@ source "fs/efs/Kconfig"
 source "fs/jffs2/Kconfig"
 # UBIFS File system configuration
 source "fs/ubifs/Kconfig"
+source "fs/logfs/Kconfig"
 source "fs/cramfs/Kconfig"
 source "fs/squashfs/Kconfig"
 source "fs/freevxfs/Kconfig"
diff --git a/fs/Makefile b/fs/Makefile
index af6d04700d9c..c3633aa46911 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -99,6 +99,7 @@ obj-$(CONFIG_NTFS_FS)		+= ntfs/
 obj-$(CONFIG_UFS_FS)		+= ufs/
 obj-$(CONFIG_EFS_FS)		+= efs/
 obj-$(CONFIG_JFFS2_FS)		+= jffs2/
+obj-$(CONFIG_LOGFS)		+= logfs/
 obj-$(CONFIG_UBIFS_FS)		+= ubifs/
 obj-$(CONFIG_AFFS_FS)		+= affs/
 obj-$(CONFIG_ROMFS_FS)		+= romfs/
diff --git a/fs/logfs/Kconfig b/fs/logfs/Kconfig
new file mode 100644
index 000000000000..daf9a9b32dd3
--- /dev/null
+++ b/fs/logfs/Kconfig
@@ -0,0 +1,17 @@
+config LOGFS
+	tristate "LogFS file system (EXPERIMENTAL)"
+	depends on (MTD || BLOCK) && EXPERIMENTAL
+	select ZLIB_INFLATE
+	select ZLIB_DEFLATE
+	select CRC32
+	select BTREE
+	help
+	  Flash filesystem aimed to scale efficiently to large devices.
+	  In comparison to JFFS2 it offers significantly faster mount
+	  times and potentially less RAM usage, although the latter has
+	  not been measured yet.
+
+	  In its current state it is still very experimental and should
+	  not be used for other than testing purposes.
+
+	  If unsure, say N.
diff --git a/fs/logfs/Makefile b/fs/logfs/Makefile
new file mode 100644
index 000000000000..4820027787ee
--- /dev/null
+++ b/fs/logfs/Makefile
@@ -0,0 +1,13 @@
+obj-$(CONFIG_LOGFS)	+= logfs.o
+
+logfs-y	+= compr.o
+logfs-y	+= dir.o
+logfs-y	+= file.o
+logfs-y	+= gc.o
+logfs-y	+= inode.o
+logfs-y	+= journal.o
+logfs-y	+= readwrite.o
+logfs-y	+= segment.o
+logfs-y	+= super.o
+logfs-$(CONFIG_BLOCK)	+= dev_bdev.o
+logfs-$(CONFIG_MTD)	+= dev_mtd.o
diff --git a/fs/logfs/compr.c b/fs/logfs/compr.c
new file mode 100644
index 000000000000..44bbfd249abc
--- /dev/null
+++ b/fs/logfs/compr.c
@@ -0,0 +1,95 @@
+/*
+ * fs/logfs/compr.c	- compression routines
+ *
+ * As should be obvious for Linux kernel code, license is GPLv2
+ *
+ * Copyright (c) 2005-2008 Joern Engel <joern@logfs.org>
+ */
+#include "logfs.h"
+#include <linux/vmalloc.h>
+#include <linux/zlib.h>
+
+#define COMPR_LEVEL 3
+
+static DEFINE_MUTEX(compr_mutex);
+static struct z_stream_s stream;
+
+int logfs_compress(void *in, void *out, size_t inlen, size_t outlen)
+{
+	int err, ret;
+
+	ret = -EIO;
+	mutex_lock(&compr_mutex);
+	err = zlib_deflateInit(&stream, COMPR_LEVEL);
+	if (err != Z_OK)
+		goto error;
+
+	stream.next_in = in;
+	stream.avail_in = inlen;
+	stream.total_in = 0;
+	stream.next_out = out;
+	stream.avail_out = outlen;
+	stream.total_out = 0;
+
+	err = zlib_deflate(&stream, Z_FINISH);
+	if (err != Z_STREAM_END)
+		goto error;
+
+	err = zlib_deflateEnd(&stream);
+	if (err != Z_OK)
+		goto error;
+
+	if (stream.total_out >= stream.total_in)
+		goto error;
+
+	ret = stream.total_out;
+error:
+	mutex_unlock(&compr_mutex);
+	return ret;
+}
+
+int logfs_uncompress(void *in, void *out, size_t inlen, size_t outlen)
+{
+	int err, ret;
+
+	ret = -EIO;
+	mutex_lock(&compr_mutex);
+	err = zlib_inflateInit(&stream);
+	if (err != Z_OK)
+		goto error;
+
+	stream.next_in = in;
+	stream.avail_in = inlen;
+	stream.total_in = 0;
+	stream.next_out = out;
+	stream.avail_out = outlen;
+	stream.total_out = 0;
+
+	err = zlib_inflate(&stream, Z_FINISH);
+	if (err != Z_STREAM_END)
+		goto error;
+
+	err = zlib_inflateEnd(&stream);
+	if (err != Z_OK)
+		goto error;
+
+	ret = 0;
+error:
+	mutex_unlock(&compr_mutex);
+	return ret;
+}
+
+int __init logfs_compr_init(void)
+{
+	size_t size = max(zlib_deflate_workspacesize(),
+			zlib_inflate_workspacesize());
+	stream.workspace = vmalloc(size);
+	if (!stream.workspace)
+		return -ENOMEM;
+	return 0;
+}
+
+void logfs_compr_exit(void)
+{
+	vfree(stream.workspace);
+}
diff --git a/fs/logfs/dev_bdev.c b/fs/logfs/dev_bdev.c
new file mode 100644
index 000000000000..58a057b6e1af
--- /dev/null
+++ b/fs/logfs/dev_bdev.c
@@ -0,0 +1,263 @@
+/*
+ * fs/logfs/dev_bdev.c	- Device access methods for block devices
+ *
+ * As should be obvious for Linux kernel code, license is GPLv2
+ *
+ * Copyright (c) 2005-2008 Joern Engel <joern@logfs.org>
+ */
+#include "logfs.h"
+#include <linux/bio.h>
+#include <linux/blkdev.h>
+#include <linux/buffer_head.h>
+
+#define PAGE_OFS(ofs) ((ofs) & (PAGE_SIZE-1))
+
+static void request_complete(struct bio *bio, int err)
+{
+	complete((struct completion *)bio->bi_private);
+}
+
+static int sync_request(struct page *page, struct block_device *bdev, int rw)
+{
+	struct bio bio;
+	struct bio_vec bio_vec;
+	struct completion complete;
+
+	bio_init(&bio);
+	bio.bi_io_vec = &bio_vec;
+	bio_vec.bv_page = page;
+	bio_vec.bv_len = PAGE_SIZE;
+	bio_vec.bv_offset = 0;
+	bio.bi_vcnt = 1;
+	bio.bi_idx = 0;
+	bio.bi_size = PAGE_SIZE;
+	bio.bi_bdev = bdev;
+	bio.bi_sector = page->index * (PAGE_SIZE >> 9);
+	init_completion(&complete);
+	bio.bi_private = &complete;
+	bio.bi_end_io = request_complete;
+
+	submit_bio(rw, &bio);
+	generic_unplug_device(bdev_get_queue(bdev));
+	wait_for_completion(&complete);
+	return test_bit(BIO_UPTODATE, &bio.bi_flags) ? 0 : -EIO;
+}
+
+static int bdev_readpage(void *_sb, struct page *page)
+{
+	struct super_block *sb = _sb;
+	struct block_device *bdev = logfs_super(sb)->s_bdev;
+	int err;
+
+	err = sync_request(page, bdev, READ);
+	if (err) {
+		ClearPageUptodate(page);
+		SetPageError(page);
+	} else {
+		SetPageUptodate(page);
+		ClearPageError(page);
+	}
+	unlock_page(page);
+	return err;
+}
+
+static DECLARE_WAIT_QUEUE_HEAD(wq);
+
+static void writeseg_end_io(struct bio *bio, int err)
+{
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct super_block *sb = bio->bi_private;
+	struct logfs_super *super = logfs_super(sb);
+	struct page *page;
+
+	BUG_ON(!uptodate); /* FIXME: Retry io or write elsewhere */
+	BUG_ON(err);
+	BUG_ON(bio->bi_vcnt == 0);
+	do {
+		page = bvec->bv_page;
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+
+		end_page_writeback(page);
+	} while (bvec >= bio->bi_io_vec);
+	bio_put(bio);
+	if (atomic_dec_and_test(&super->s_pending_writes))
+		wake_up(&wq);
+}
+
+static int __bdev_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
+		size_t nr_pages)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	struct bio *bio;
+	struct page *page;
+	struct request_queue *q = bdev_get_queue(sb->s_bdev);
+	unsigned int max_pages = queue_max_hw_sectors(q) >> (PAGE_SHIFT - 9);
+	int i;
+
+	bio = bio_alloc(GFP_NOFS, max_pages);
+	BUG_ON(!bio); /* FIXME: handle this */
+
+	for (i = 0; i < nr_pages; i++) {
+		if (i >= max_pages) {
+			/* Block layer cannot split bios :( */
+			bio->bi_vcnt = i;
+			bio->bi_idx = 0;
+			bio->bi_size = i * PAGE_SIZE;
+			bio->bi_bdev = super->s_bdev;
+			bio->bi_sector = ofs >> 9;
+			bio->bi_private = sb;
+			bio->bi_end_io = writeseg_end_io;
+			atomic_inc(&super->s_pending_writes);
+			submit_bio(WRITE, bio);
+
+			ofs += i * PAGE_SIZE;
+			index += i;
+			nr_pages -= i;
+			i = 0;
+
+			bio = bio_alloc(GFP_NOFS, max_pages);
+			BUG_ON(!bio);
+		}
+		page = find_lock_page(mapping, index + i);
+		BUG_ON(!page);
+		bio->bi_io_vec[i].bv_page = page;
+		bio->bi_io_vec[i].bv_len = PAGE_SIZE;
+		bio->bi_io_vec[i].bv_offset = 0;
+
+		BUG_ON(PageWriteback(page));
+		set_page_writeback(page);
+		unlock_page(page);
+	}
+	bio->bi_vcnt = nr_pages;
+	bio->bi_idx = 0;
+	bio->bi_size = nr_pages * PAGE_SIZE;
+	bio->bi_bdev = super->s_bdev;
+	bio->bi_sector = ofs >> 9;
+	bio->bi_private = sb;
+	bio->bi_end_io = writeseg_end_io;
+	atomic_inc(&super->s_pending_writes);
+	submit_bio(WRITE, bio);
+	return 0;
+}
+
+static void bdev_writeseg(struct super_block *sb, u64 ofs, size_t len)
+{
+	struct logfs_super *super = logfs_super(sb);
+	int head;
+
+	BUG_ON(super->s_flags & LOGFS_SB_FLAG_RO);
+
+	if (len == 0) {
+		/* This can happen when the object fit perfectly into a
+		 * segment, the segment gets written per sync and subsequently
+		 * closed.
+		 */
+		return;
+	}
+	head = ofs & (PAGE_SIZE - 1);
+	if (head) {
+		ofs -= head;
+		len += head;
+	}
+	len = PAGE_ALIGN(len);
+	__bdev_writeseg(sb, ofs, ofs >> PAGE_SHIFT, len >> PAGE_SHIFT);
+	generic_unplug_device(bdev_get_queue(logfs_super(sb)->s_bdev));
+}
+
+static int bdev_erase(struct super_block *sb, loff_t to, size_t len)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	struct page *page;
+	pgoff_t index = to >> PAGE_SHIFT;
+	int i, nr_pages = len >> PAGE_SHIFT;
+
+	BUG_ON(to & (PAGE_SIZE - 1));
+	BUG_ON(len & (PAGE_SIZE - 1));
+
+	if (logfs_super(sb)->s_flags & LOGFS_SB_FLAG_RO)
+		return -EROFS;
+
+	for (i = 0; i < nr_pages; i++) {
+		page = find_get_page(mapping, index + i);
+		if (page) {
+			memset(page_address(page), 0xFF, PAGE_SIZE);
+			page_cache_release(page);
+		}
+	}
+	return 0;
+}
+
+static void bdev_sync(struct super_block *sb)
+{
+	struct logfs_super *super = logfs_super(sb);
+
+	wait_event(wq, atomic_read(&super->s_pending_writes) == 0);
+}
+
+static struct page *bdev_find_first_sb(struct super_block *sb, u64 *ofs)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	filler_t *filler = bdev_readpage;
+
+	*ofs = 0;
+	return read_cache_page(mapping, 0, filler, sb);
+}
+
+static struct page *bdev_find_last_sb(struct super_block *sb, u64 *ofs)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	filler_t *filler = bdev_readpage;
+	u64 pos = (super->s_bdev->bd_inode->i_size & ~0xfffULL) - 0x1000;
+	pgoff_t index = pos >> PAGE_SHIFT;
+
+	*ofs = pos;
+	return read_cache_page(mapping, index, filler, sb);
+}
+
+static int bdev_write_sb(struct super_block *sb, struct page *page)
+{
+	struct block_device *bdev = logfs_super(sb)->s_bdev;
+
+	/* Nothing special to do for block devices. */
+	return sync_request(page, bdev, WRITE);
+}
+
+static void bdev_put_device(struct super_block *sb)
+{
+	close_bdev_exclusive(logfs_super(sb)->s_bdev, FMODE_READ|FMODE_WRITE);
+}
+
+static const struct logfs_device_ops bd_devops = {
+	.find_first_sb	= bdev_find_first_sb,
+	.find_last_sb	= bdev_find_last_sb,
+	.write_sb	= bdev_write_sb,
+	.readpage	= bdev_readpage,
+	.writeseg	= bdev_writeseg,
+	.erase		= bdev_erase,
+	.sync		= bdev_sync,
+	.put_device	= bdev_put_device,
+};
+
+int logfs_get_sb_bdev(struct file_system_type *type, int flags,
+		const char *devname, struct vfsmount *mnt)
+{
+	struct block_device *bdev;
+
+	bdev = open_bdev_exclusive(devname, FMODE_READ|FMODE_WRITE, type);
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
+
+	if (MAJOR(bdev->bd_dev) == MTD_BLOCK_MAJOR) {
+		int mtdnr = MINOR(bdev->bd_dev);
+		close_bdev_exclusive(bdev, FMODE_READ|FMODE_WRITE);
+		return logfs_get_sb_mtd(type, flags, mtdnr, mnt);
+	}
+
+	return logfs_get_sb_device(type, flags, NULL, bdev, &bd_devops, mnt);
+}
diff --git a/fs/logfs/dev_mtd.c b/fs/logfs/dev_mtd.c
new file mode 100644
index 000000000000..68e99d046c23
--- /dev/null
+++ b/fs/logfs/dev_mtd.c
@@ -0,0 +1,253 @@
+/*
+ * fs/logfs/dev_mtd.c	- Device access methods for MTD
+ *
+ * As should be obvious for Linux kernel code, license is GPLv2
+ *
+ * Copyright (c) 2005-2008 Joern Engel <joern@logfs.org>
+ */
+#include "logfs.h"
+#include <linux/completion.h>
+#include <linux/mount.h>
+#include <linux/sched.h>
+
+#define PAGE_OFS(ofs) ((ofs) & (PAGE_SIZE-1))
+
+static int mtd_read(struct super_block *sb, loff_t ofs, size_t len, void *buf)
+{
+	struct mtd_info *mtd = logfs_super(sb)->s_mtd;
+	size_t retlen;
+	int ret;
+
+	ret = mtd->read(mtd, ofs, len, &retlen, buf);
+	BUG_ON(ret == -EINVAL);
+	if (ret)
+		return ret;
+
+	/* Not sure if we should loop instead. */
+	if (retlen != len)
+		return -EIO;
+
+	return 0;
+}
+
+static int mtd_write(struct super_block *sb, loff_t ofs, size_t len, void *buf)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct mtd_info *mtd = super->s_mtd;
+	size_t retlen;
+	loff_t page_start, page_end;
+	int ret;
+
+	if (super->s_flags & LOGFS_SB_FLAG_RO)
+		return -EROFS;
+
+	BUG_ON((ofs >= mtd->size) || (len > mtd->size - ofs));
+	BUG_ON(ofs != (ofs >> super->s_writeshift) << super->s_writeshift);
+	BUG_ON(len > PAGE_CACHE_SIZE);
+	page_start = ofs & PAGE_CACHE_MASK;
+	page_end = PAGE_CACHE_ALIGN(ofs + len) - 1;
+	ret = mtd->write(mtd, ofs, len, &retlen, buf);
+	if (ret || (retlen != len))
+		return -EIO;
+
+	return 0;
+}
+
+/*
+ * For as long as I can remember (since about 2001) mtd->erase has been an
+ * asynchronous interface lacking the first driver to actually use the
+ * asynchronous properties.  So just to prevent the first implementor of such
+ * a thing from breaking logfs in 2350, we do the usual pointless dance to
+ * declare a completion variable and wait for completion before returning
+ * from mtd_erase().  What an excercise in futility!
+ */
+static void logfs_erase_callback(struct erase_info *ei)
+{
+	complete((struct completion *)ei->priv);
+}
+
+static int mtd_erase_mapping(struct super_block *sb, loff_t ofs, size_t len)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	struct page *page;
+	pgoff_t index = ofs >> PAGE_SHIFT;
+
+	for (index = ofs >> PAGE_SHIFT; index < (ofs + len) >> PAGE_SHIFT; index++) {
+		page = find_get_page(mapping, index);
+		if (!page)
+			continue;
+		memset(page_address(page), 0xFF, PAGE_SIZE);
+		page_cache_release(page);
+	}
+	return 0;
+}
+
+static int mtd_erase(struct super_block *sb, loff_t ofs, size_t len)
+{
+	struct mtd_info *mtd = logfs_super(sb)->s_mtd;
+	struct erase_info ei;
+	DECLARE_COMPLETION_ONSTACK(complete);
+	int ret;
+
+	BUG_ON(len % mtd->erasesize);
+	if (logfs_super(sb)->s_flags & LOGFS_SB_FLAG_RO)
+		return -EROFS;
+
+	memset(&ei, 0, sizeof(ei));
+	ei.mtd = mtd;
+	ei.addr = ofs;
+	ei.len = len;
+	ei.callback = logfs_erase_callback;
+	ei.priv = (long)&complete;
+	ret = mtd->erase(mtd, &ei);
+	if (ret)
+		return -EIO;
+
+	wait_for_completion(&complete);
+	if (ei.state != MTD_ERASE_DONE)
+		return -EIO;
+	return mtd_erase_mapping(sb, ofs, len);
+}
+
+static void mtd_sync(struct super_block *sb)
+{
+	struct mtd_info *mtd = logfs_super(sb)->s_mtd;
+
+	if (mtd->sync)
+		mtd->sync(mtd);
+}
+
+static int mtd_readpage(void *_sb, struct page *page)
+{
+	struct super_block *sb = _sb;
+	int err;
+
+	err = mtd_read(sb, page->index << PAGE_SHIFT, PAGE_SIZE,
+			page_address(page));
+	if (err == -EUCLEAN) {
+		err = 0;
+		/* FIXME: force GC this segment */
+	}
+	if (err) {
+		ClearPageUptodate(page);
+		SetPageError(page);
+	} else {
+		SetPageUptodate(page);
+		ClearPageError(page);
+	}
+	unlock_page(page);
+	return err;
+}
+
+static struct page *mtd_find_first_sb(struct super_block *sb, u64 *ofs)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	filler_t *filler = mtd_readpage;
+	struct mtd_info *mtd = super->s_mtd;
+
+	if (!mtd->block_isbad)
+		return NULL;
+
+	*ofs = 0;
+	while (mtd->block_isbad(mtd, *ofs)) {
+		*ofs += mtd->erasesize;
+		if (*ofs >= mtd->size)
+			return NULL;
+	}
+	BUG_ON(*ofs & ~PAGE_MASK);
+	return read_cache_page(mapping, *ofs >> PAGE_SHIFT, filler, sb);
+}
+
+static struct page *mtd_find_last_sb(struct super_block *sb, u64 *ofs)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	filler_t *filler = mtd_readpage;
+	struct mtd_info *mtd = super->s_mtd;
+
+	if (!mtd->block_isbad)
+		return NULL;
+
+	*ofs = mtd->size - mtd->erasesize;
+	while (mtd->block_isbad(mtd, *ofs)) {
+		*ofs -= mtd->erasesize;
+		if (*ofs <= 0)
+			return NULL;
+	}
+	*ofs = *ofs + mtd->erasesize - 0x1000;
+	BUG_ON(*ofs & ~PAGE_MASK);
+	return read_cache_page(mapping, *ofs >> PAGE_SHIFT, filler, sb);
+}
+
+static int __mtd_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
+		size_t nr_pages)
+{
+	struct logfs_super *super = logfs_super(sb);
+	struct address_space *mapping = super->s_mapping_inode->i_mapping;
+	struct page *page;
+	int i, err;
+
+	for (i = 0; i < nr_pages; i++) {
+		page = find_lock_page(mapping, index + i);
+		BUG_ON(!page);
+
+		err = mtd_write(sb, page->index << PAGE_SHIFT, PAGE_SIZE,
+				page_address(page));
+		unlock_page(page);
+		page_cache_release(page);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+static void mtd_writeseg(struct super_block *sb, u64 ofs, size_t len)
+{
+	struct logfs_super *super = logfs_super(sb);
+	int head;
+
+	if (super->s_flags & LOGFS_SB_FLAG_RO)
+		return;
+
+	if (len == 0) {
+		/* This can happen when the object fit perfectly into a
+		 * segment, the segment gets written per sync and subsequently
+		 * closed.
+		 */
+		return;
+	}
+	head = ofs & (PAGE_SIZE - 1);
+	if (head) {
+		ofs -= head;
+		len += head;
+	}
+	len = PAGE_ALIGN(len);
+	__mtd_writeseg(sb, ofs, ofs >> PAGE_SHIFT, len >> PAGE_SHIFT);
+}
+
+static void mtd_put_device(struct super_block *sb)
+{
+	put_mtd_device(logfs_super(sb)->s_mtd);
+}
+
+static const struct logfs_device_ops mtd_devops = {
+	.find_first_sb	= mtd_find_first_sb,
+	.find_last_sb	= mtd_find_last_sb,
+	.readpage	= mtd_readpage,
+	.writese