89 files changed, 6256 insertions, 2574 deletions

diff --git a/fs/bio-integrity.c b/fs/bio-integrity.c
index 9c5e6b2cd11a..c2183f3917cd 100644
--- a/fs/bio-integrity.c
+++ b/fs/bio-integrity.c
@@ -22,6 +22,7 @@
 
 #include <linux/blkdev.h>
 #include <linux/mempool.h>
+#include <linux/export.h>
 #include <linux/bio.h>
 #include <linux/workqueue.h>
 #include <linux/slab.h>
diff --git a/fs/bio.c b/fs/bio.c
index 9bfade8a609b..41c93c722244 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -255,7 +255,6 @@ void bio_init(struct bio *bio)
 {
         memset(bio, 0, sizeof(*bio));
         bio->bi_flags = 1 << BIO_UPTODATE;
-        bio->bi_comp_cpu = -1;
         atomic_set(&bio->bi_cnt, 1);
 }
 EXPORT_SYMBOL(bio_init);
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 95f786ec7f08..b07f1da1de4e 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -971,7 +971,7 @@ static void flush_disk(struct block_device *bdev, bool kill_dirty)
 
         if (!bdev->bd_disk)
                 return;
-        if (disk_partitionable(bdev->bd_disk))
+        if (disk_part_scan_enabled(bdev->bd_disk))
                 bdev->bd_invalidated = 1;
 }
 
@@ -1085,6 +1085,7 @@ static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part);
 static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
 {
         struct gendisk *disk;
+        struct module *owner;
         int ret;
         int partno;
         int perm = 0;
@@ -1110,6 +1111,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
         disk = get_gendisk(bdev->bd_dev, &partno);
         if (!disk)
                 goto out;
+        owner = disk->fops->owner;
 
         disk_block_events(disk);
         mutex_lock_nested(&bdev->bd_mutex, for_part);
@@ -1137,8 +1139,8 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
                                         bdev->bd_disk = NULL;
                                         mutex_unlock(&bdev->bd_mutex);
                                         disk_unblock_events(disk);
-                                        module_put(disk->fops->owner);
                                         put_disk(disk);
+                                        module_put(owner);
                                         goto restart;
                                 }
                         }
@@ -1194,8 +1196,8 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
                                 goto out_unlock_bdev;
                 }
                 /* only one opener holds refs to the module and disk */
-                module_put(disk->fops->owner);
                 put_disk(disk);
+                module_put(owner);
         }
         bdev->bd_openers++;
         if (for_part)
@@ -1215,8 +1217,8 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
  out_unlock_bdev:
         mutex_unlock(&bdev->bd_mutex);
         disk_unblock_events(disk);
-        module_put(disk->fops->owner);
         put_disk(disk);
+        module_put(owner);
  out:
         bdput(bdev);
 
@@ -1442,14 +1444,15 @@ static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part)
         if (!bdev->bd_openers) {
                 struct module *owner = disk->fops->owner;
 
-                put_disk(disk);
-                module_put(owner);
                 disk_put_part(bdev->bd_part);
                 bdev->bd_part = NULL;
                 bdev->bd_disk = NULL;
                 if (bdev != bdev->bd_contains)
                         victim = bdev->bd_contains;
                 bdev->bd_contains = NULL;
+
+                put_disk(disk);
+                module_put(owner);
         }
         mutex_unlock(&bdev->bd_mutex);
         bdput(bdev);
diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 40e6ac08c21f..c0ddfd29c5e5 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -7,6 +7,7 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
            extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
            extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
            export.o tree-log.o free-space-cache.o zlib.o lzo.o \
-           compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o
+           compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \
+           reada.o backref.o
 
 btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o
diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c
index eb159aaa5a11..89b156d85d63 100644
--- a/fs/btrfs/acl.c
+++ b/fs/btrfs/acl.c
@@ -59,22 +59,19 @@ struct posix_acl *btrfs_get_acl(struct inode *inode, int type)
                 if (!value)
                         return ERR_PTR(-ENOMEM);
                 size = __btrfs_getxattr(inode, name, value, size);
-                if (size > 0) {
-                        acl = posix_acl_from_xattr(value, size);
-                        if (IS_ERR(acl)) {
-                                kfree(value);
-                                return acl;
-                        }
-                        set_cached_acl(inode, type, acl);
-                }
-                kfree(value);
+        }
+        if (size > 0) {
+                acl = posix_acl_from_xattr(value, size);
         } else if (size == -ENOENT || size == -ENODATA || size == 0) {
                 /* FIXME, who returns -ENOENT?  I think nobody */
                 acl = NULL;
-                set_cached_acl(inode, type, acl);
         } else {
                 acl = ERR_PTR(-EIO);
         }
+        kfree(value);
+
+        if (!IS_ERR(acl))
+                set_cached_acl(inode, type, acl);
 
         return acl;
 }
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
new file mode 100644
index 000000000000..8855aad3929c
--- /dev/null
+++ b/fs/btrfs/backref.c
@@ -0,0 +1,776 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "backref.h"
+
+struct __data_ref {
+        struct list_head list;
+        u64 inum;
+        u64 root;
+        u64 extent_data_item_offset;
+};
+
+struct __shared_ref {
+        struct list_head list;
+        u64 disk_byte;
+};
+
+static int __inode_info(u64 inum, u64 ioff, u8 key_type,
+                        struct btrfs_root *fs_root, struct btrfs_path *path,
+                        struct btrfs_key *found_key)
+{
+        int ret;
+        struct btrfs_key key;
+        struct extent_buffer *eb;
+
+        key.type = key_type;
+        key.objectid = inum;
+        key.offset = ioff;
+
+        ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0);
+        if (ret < 0)
+                return ret;
+
+        eb = path->nodes[0];
+        if (ret && path->slots[0] >= btrfs_header_nritems(eb)) {
+                ret = btrfs_next_leaf(fs_root, path);
+                if (ret)
+                        return ret;
+                eb = path->nodes[0];
+        }
+
+        btrfs_item_key_to_cpu(eb, found_key, path->slots[0]);
+        if (found_key->type != key.type || found_key->objectid != key.objectid)
+                return 1;
+
+        return 0;
+}
+
+/*
+ * this makes the path point to (inum INODE_ITEM ioff)
+ */
+int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
+                        struct btrfs_path *path)
+{
+        struct btrfs_key key;
+        return __inode_info(inum, ioff, BTRFS_INODE_ITEM_KEY, fs_root, path,
+                                &key);
+}
+
+static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
+                                struct btrfs_path *path,
+                                struct btrfs_key *found_key)
+{
+        return __inode_info(inum, ioff, BTRFS_INODE_REF_KEY, fs_root, path,
+                                found_key);
+}
+
+/*
+ * this iterates to turn a btrfs_inode_ref into a full filesystem path. elements
+ * of the path are separated by '/' and the path is guaranteed to be
+ * 0-terminated. the path is only given within the current file system.
+ * Therefore, it never starts with a '/'. the caller is responsible to provide
+ * "size" bytes in "dest". the dest buffer will be filled backwards. finally,
+ * the start point of the resulting string is returned. this pointer is within
+ * dest, normally.
+ * in case the path buffer would overflow, the pointer is decremented further
+ * as if output was written to the buffer, though no more output is actually
+ * generated. that way, the caller can determine how much space would be
+ * required for the path to fit into the buffer. in that case, the returned
+ * value will be smaller than dest. callers must check this!
+ */
+static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
+                                struct btrfs_inode_ref *iref,
+                                struct extent_buffer *eb_in, u64 parent,
+                                char *dest, u32 size)
+{
+        u32 len;
+        int slot;
+        u64 next_inum;
+        int ret;
+        s64 bytes_left = size - 1;
+        struct extent_buffer *eb = eb_in;
+        struct btrfs_key found_key;
+
+        if (bytes_left >= 0)
+                dest[bytes_left] = '\0';
+
+        while (1) {
+                len = btrfs_inode_ref_name_len(eb, iref);
+                bytes_left -= len;
+                if (bytes_left >= 0)
+                        read_extent_buffer(eb, dest + bytes_left,
+                                                (unsigned long)(iref + 1), len);
+                if (eb != eb_in)
+                        free_extent_buffer(eb);
+                ret = inode_ref_info(parent, 0, fs_root, path, &found_key);
+                if (ret)
+                        break;
+                next_inum = found_key.offset;
+
+                /* regular exit ahead */
+                if (parent == next_inum)
+                        break;
+
+                slot = path->slots[0];
+                eb = path->nodes[0];
+                /* make sure we can use eb after releasing the path */
+                if (eb != eb_in)
+                        atomic_inc(&eb->refs);
+                btrfs_release_path(path);
+
+                iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
+                parent = next_inum;
+                --bytes_left;
+                if (bytes_left >= 0)
+                        dest[bytes_left] = '/';
+        }
+
+        btrfs_release_path(path);
+
+        if (ret)
+                return ERR_PTR(ret);
+
+        return dest + bytes_left;
+}
+
+/*
+ * this makes the path point to (logical EXTENT_ITEM *)
+ * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for
+ * tree blocks and <0 on error.
+ */
+int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
+                        struct btrfs_path *path, struct btrfs_key *found_key)
+{
+        int ret;
+        u64 flags;
+        u32 item_size;
+        struct extent_buffer *eb;
+        struct btrfs_extent_item *ei;
+        struct btrfs_key key;
+
+        key.type = BTRFS_EXTENT_ITEM_KEY;
+        key.objectid = logical;
+        key.offset = (u64)-1;
+
+        ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
+        if (ret < 0)
+                return ret;
+        ret = btrfs_previous_item(fs_info->extent_root, path,
+                                        0, BTRFS_EXTENT_ITEM_KEY);
+        if (ret < 0)
+                return ret;
+
+        btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]);
+        if (found_key->type != BTRFS_EXTENT_ITEM_KEY ||
+            found_key->objectid > logical ||
+            found_key->objectid + found_key->offset <= logical)
+                return -ENOENT;
+
+        eb = path->nodes[0];
+        item_size = btrfs_item_size_nr(eb, path->slots[0]);
+        BUG_ON(item_size < sizeof(*ei));
+
+        ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+        flags = btrfs_extent_flags(eb, ei);
+
+        if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
+                return BTRFS_EXTENT_FLAG_TREE_BLOCK;
+        if (flags & BTRFS_EXTENT_FLAG_DATA)
+                return BTRFS_EXTENT_FLAG_DATA;
+
+        return -EIO;
+}
+
+/*
+ * helper function to iterate extent inline refs. ptr must point to a 0 value
+ * for the first call and may be modified. it is used to track state.
+ * if more refs exist, 0 is returned and the next call to
+ * __get_extent_inline_ref must pass the modified ptr parameter to get the
+ * next ref. after the last ref was processed, 1 is returned.
+ * returns <0 on error
+ */
+static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb,
+                                struct btrfs_extent_item *ei, u32 item_size,
+                                struct btrfs_extent_inline_ref **out_eiref,
+                                int *out_type)
+{
+        unsigned long end;
+        u64 flags;
+        struct btrfs_tree_block_info *info;
+
+        if (!*ptr) {
+                /* first call */
+                flags = btrfs_extent_flags(eb, ei);
+                if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+                        info = (struct btrfs_tree_block_info *)(ei + 1);
+                        *out_eiref =
+                                (struct btrfs_extent_inline_ref *)(info + 1);
+                } else {
+                        *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1);
+                }
+                *ptr = (unsigned long)*out_eiref;
+                if ((void *)*ptr >= (void *)ei + item_size)
+                        return -ENOENT;
+        }
+
+        end = (unsigned long)ei + item_size;
+        *out_eiref = (struct btrfs_extent_inline_ref *)*ptr;
+        *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref);
+
+        *ptr += btrfs_extent_inline_ref_size(*out_type);
+        WARN_ON(*ptr > end);
+        if (*ptr == end)
+                return 1; /* last */
+
+        return 0;
+}
+
+/*
+ * reads the tree block backref for an extent. tree level and root are returned
+ * through out_level and out_root. ptr must point to a 0 value for the first
+ * call and may be modified (see __get_extent_inline_ref comment).
+ * returns 0 if data was provided, 1 if there was no more data to provide or
+ * <0 on error.
+ */
+int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
+                                struct btrfs_extent_item *ei, u32 item_size,
+                                u64 *out_root, u8 *out_level)
+{
+        int ret;
+        int type;
+        struct btrfs_tree_block_info *info;
+        struct btrfs_extent_inline_ref *eiref;
+
+        if (*ptr == (unsigned long)-1)
+                return 1;
+
+        while (1) {
+                ret = __get_extent_inline_ref(ptr, eb, ei, item_size,
+                                                &eiref, &type);
+                if (ret < 0)
+                        return ret;
+
+                if (type == BTRFS_TREE_BLOCK_REF_KEY ||
+                    type == BTRFS_SHARED_BLOCK_REF_KEY)
+                        break;
+
+                if (ret == 1)
+                        return 1;
+        }
+
+        /* we can treat both ref types equally here */
+        info = (struct btrfs_tree_block_info *)(ei + 1);
+        *out_root = btrfs_extent_inline_ref_offset(eb, eiref);
+        *out_level = btrfs_tree_block_level(eb, info);
+
+        if (ret == 1)
+                *ptr = (unsigned long)-1;
+
+        return 0;
+}
+
+static int __data_list_add(struct list_head *head, u64 inum,
+                                u64 extent_data_item_offset, u64 root)
+{
+        struct __data_ref *ref;
+
+        ref = kmalloc(sizeof(*ref), GFP_NOFS);
+        if (!ref)
+                return -ENOMEM;
+
+        ref->inum = inum;
+        ref->extent_data_item_offset = extent_data_item_offset;
+        ref->root = root;
+        list_add_tail(&ref->list, head);
+
+        return 0;
+}
+
+static int __data_list_add_eb(struct list_head *head, struct extent_buffer *eb,
+                                struct btrfs_extent_data_ref *dref)
+{
+        return __data_list_add(head, btrfs_extent_data_ref_objectid(eb, dref),
+                                btrfs_extent_data_ref_offset(eb, dref),
+                                btrfs_extent_data_ref_root(eb, dref));
+}
+
+static int __shared_list_add(struct list_head *head, u64 disk_byte)
+{
+        struct __shared_ref *ref;
+
+        ref = kmalloc(sizeof(*ref), GFP_NOFS);
+        if (!ref)
+                return -ENOMEM;
+
+        ref->disk_byte = disk_byte;
+        list_add_tail(&ref->list, head);
+
+        return 0;
+}
+
+static int __iter_shared_inline_ref_inodes(struct btrfs_fs_info *fs_info,
+                                           u64 logical, u64 inum,
+                                           u64 extent_data_item_offset,
+                                           u64 extent_offset,
+                                           struct btrfs_path *path,
+                                           struct list_head *data_refs,
+                                           iterate_extent_inodes_t *iterate,
+                                           void *ctx)
+{
+        u64 ref_root;
+        u32 item_size;
+        struct btrfs_key key;
+        struct extent_buffer *eb;
+        struct btrfs_extent_item *ei;
+        struct btrfs_extent_inline_ref *eiref;
+        struct __data_ref *ref;
+        int ret;
+        int type;
+        int last;
+        unsigned long ptr = 0;
+
+        WARN_ON(!list_empty(data_refs));
+        ret = extent_from_logical(fs_info, logical, path, &key);
+        if (ret & BTRFS_EXTENT_FLAG_DATA)
+                ret = -EIO;
+        if (ret < 0)
+                goto out;
+
+        eb = path->nodes[0];
+        ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+        item_size = btrfs_item_size_nr(eb, path->slots[0]);
+
+        ret = 0;
+        ref_root = 0;
+        /*
+         * as done in iterate_extent_inodes, we first build a list of refs to
+         * iterate, then free the path and then iterate them to avoid deadlocks.
+         */
+        do {
+                last = __get_extent_inline_ref(&ptr, eb, ei, item_size,
+                                                &eiref, &type);
+                if (last < 0) {
+                        ret = last;
+                        goto out;
+                }
+                if (type == BTRFS_TREE_BLOCK_REF_KEY ||
+                    type == BTRFS_SHARED_BLOCK_REF_KEY) {
+                        ref_root = btrfs_extent_inline_ref_offset(eb, eiref);
+                        ret = __data_list_add(data_refs, inum,
+                                                extent_data_item_offset,
+                                                ref_root);
+                }
+        } while (!ret && !last);
+
+        btrfs_release_path(path);
+
+        if (ref_root == 0) {
+                printk(KERN_ERR "btrfs: failed to find tree block ref "
+                        "for shared data backref %llu\n", logical);
+                WARN_ON(1);
+                ret = -EIO;
+        }
+
+out:
+        while (!list_empty(data_refs)) {
+                ref = list_first_entry(data_refs, struct __data_ref, list);
+                list_del(&ref->list);
+                if (!ret)
+                        ret = iterate(ref->inum, extent_offset +
+                                        ref->extent_data_item_offset,
+                                        ref->root, ctx);
+                kfree(ref);
+        }
+
+        return ret;
+}
+
+static int __iter_shared_inline_ref(struct btrfs_fs_info *fs_info,
+                                    u64 logical, u64 orig_extent_item_objectid,
+                                    u64 extent_offset, struct btrfs_path *path,
+                                    struct list_head *data_refs,
+                                    iterate_extent_inodes_t *iterate,
+                                    void *ctx)
+{
+        u64 disk_byte;
+        struct btrfs_key key;
+        struct btrfs_file_extent_item *fi;
+        struct extent_buffer *eb;
+        int slot;
+        int nritems;
+        int ret;
+        int found = 0;
+
+        eb = read_tree_block(fs_info->tree_root, logical,
+                                fs_info->tree_root->leafsize, 0);
+        if (!eb)
+                return -EIO;
+
+        /*
+         * from the shared data ref, we only have the leaf but we need
+         * the key. thus, we must look into all items and see that we
+         * find one (some) with a reference to our extent item.
+         */
+        nritems = btrfs_header_nritems(eb);
+        for (slot = 0; slot < nritems; ++slot) {
+                btrfs_item_key_to_cpu(eb, &key, slot);
+                if (key.type != BTRFS_EXTENT_DATA_KEY)
+                        continue;
+                fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
+                if (!fi) {
+                        free_extent_buffer(eb);
+                        return -EIO;
+                }
+                disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
+                if (disk_byte != orig_extent_item_objectid) {
+                        if (found)
+                                break;
+                        else
+                                continue;
+                }
+                ++found;
+                ret = __iter_shared_inline_ref_inodes(fs_info, logical,
+                                                        key.objectid,
+                                                        key.offset,
+                                                        extent_offset, path,
+                                                        data_refs,
+                                                        iterate, ctx);
+                if (ret)
+                        break;
+        }
+
+        if (!found) {
+                printk(KERN_ERR "btrfs: failed to follow shared data backref "
+                        "to parent %llu\n", logical);
+                WARN_ON(1);
+                ret = -EIO;
+        }
+
+        free_extent_buffer(eb);
+        return ret;
+}
+
+/*
+ * calls iterate() for every inode that references the extent identified by
+ * the given parameters. will use the path given as a parameter and return it
+ * released.
+ * when the iterator function returns a non-zero value, iteration stops.
+ */
+int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
+                                struct btrfs_path *path,
+                                u64 extent_item_objectid,
+                                u64 extent_offset,
+                                iterate_extent_inodes_t *iterate, void *ctx)
+{
+        unsigned long ptr = 0;
+        int last;
+        int ret;
+        int type;
+        u64 logical;
+        u32 item_size;
+        struct btrfs_extent_inline_ref *eiref;
+        struct btrfs_extent_data_ref *dref;
+        struct extent_buffer *eb;
+        struct btrfs_extent_item *ei;
+        struct btrfs_key key;
+        struct list_head data_refs = LIST_HEAD_INIT(data_refs);
+        struct list_head shared_refs = LIST_HEAD_INIT(shared_refs);
+        struct __data_ref *ref_d;
+        struct __shared_ref *ref_s;
+
+        eb = path->nodes[0];
+        ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+        item_size = btrfs_item_size_nr(eb, path->slots[0]);
+
+        /* first we iterate the inline refs, ... */
+        do {
+                last = __get_extent_inline_ref(&ptr, eb, ei, item_size,
+                                                &eiref, &type);
+                if (last == -ENOENT) {
+                        ret = 0;
+                        break;
+                }
+                if (last < 0) {
+                        ret = last;
+                        break;
+                }
+
+                if (type == BTRFS_EXTENT_DATA_REF_KEY) {
+                        dref = (struct btrfs_extent_data_ref *)(&eiref->offset);
+                        ret = __data_list_add_eb(&data_refs, eb, dref);
+                } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
+                        logical = btrfs_extent_inline_ref_offset(eb, eiref);
+                        ret = __shared_list_add(&shared_refs, logical);
+                }
+        } while (!ret && !last);
+
+        /* ... then we proceed to in-tree references and ... */
+        while (!ret) {
+                ++path->slots[0];
+                if (path->slots[0] > btrfs_header_nritems(eb)) {
+                        ret = btrfs_next_leaf(fs_info->extent_root, path);
+                        if (ret) {
+                                if (ret == 1)
+                                        ret = 0; /* we're done */
+                                break;
+                        }
+                        eb = path->nodes[0];
+                }
+                btrfs_item_key_to_cpu(eb, &key, path->slots[0]);
+                if (key.objectid != extent_item_objectid)
+                        break;
+                if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
+                        dref = btrfs_item_ptr(eb, path->slots[0],
+                                                struct btrfs_extent_data_ref);
+                        ret = __data_list_add_eb(&data_refs, eb, dref);
+                } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
+                        ret = __shared_list_add(&shared_refs, key.offset);
+                }
+        }
+
+        btrfs_release_path(path);
+
+        /*
+         * ... only at the very end we can process the refs we found. this is
+         * because the iterator function we call is allowed to make tree lookups
+         * and we have to avoid deadlocks. additionally, we need more tree
+         * lookups ourselves for shared data refs.
+         */
+        while (!list_empty(&data_refs)) {
+                ref_d = list_first_entry(&data_refs, struct __data_ref, list);
+                list_del(&ref_d->list);
+                if (!ret)
+                        ret = iterate(ref_d->inum, extent_offset +
+                                        ref_d->extent_data_item_offset,
+                                        ref_d->root, ctx);
+                kfree(ref_d);
+        }
+
+        while (!list_empty(&shared_refs)) {
+                ref_s = list_first_entry(&shared_refs, struct __shared_ref,
+                                        list);
+                list_del(&ref_s->list);
+                if (!ret)
+                        ret = __iter_shared_inline_ref(fs_info,
+                                                        ref_s->disk_byte,
+                                                        extent_item_objectid,
+                                                        extent_offset, path,
+                                                        &data_refs,
+                                                        iterate, ctx);
+                kfree(ref_s);
+        }
+
+        return ret;
+}
+
+int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
+                                struct btrfs_path *path,
+                                iterate_extent_inodes_t *iterate, void *ctx)
+{
+        int ret;
+        u64 offset;
+        struct btrfs_key found_key;
+
+        ret = extent_from_logical(fs_info, logical, path,
+                                        &found_key);
+        if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK)
+                ret = -EINVAL;
+        if (ret < 0)
+                return ret;
+
+        offset = logical - found_key.objectid;
+        ret = iterate_extent_inodes(fs_info, path, found_key.objectid,
+                                        offset, iterate, ctx);
+
+        return ret;
+}
+
+static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
+                                struct btrfs_path *path,
+                                iterate_irefs_t *iterate, void *ctx)
+{
+        int ret;
+        int slot;
+        u32 cur;
+        u32 len;
+        u32 name_len;
+        u64 parent = 0;
+        int found = 0;
+        struct extent_buffer *eb;
+        struct btrfs_item *item;
+        struct btrfs_inode_ref *iref;
+        struct btrfs_key found_key;
+
+        while (1) {
+                ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
+                                        &found_key);
+                if (ret < 0)
+                        break;
+                if (ret) {
+                        ret = found ? 0 : -ENOENT;
+                        break;
+                }
+                ++found;
+
+                parent = found_key.offset;
+                slot = path->slots[0];
+                eb = path->nodes[0];
+                /* make sure we can use eb after releasing the path */
+                atomic_inc(&eb->refs);
+                btrfs_release_path(path);
+
+                item = btrfs_item_nr(eb, slot);
+                iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
+
+                for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) {
+                        name_len = btrfs_inode_ref_name_len(eb, iref);
+                        /* path must be released before calling iterate()! */
+                        ret = iterate(parent, iref, eb, ctx);
+                        if (ret) {
+                                free_extent_buffer(eb);
+                                break;
+                        }
+                        len = sizeof(*iref) + name_len;
+                        iref = (struct btrfs_inode_ref *)((char *)iref + len);
+                }
+                free_extent_buffer(eb);
+        }
+
+        btrfs_release_path(path);
+
+        return ret;
+}
+
+/*
+ * returns 0 if the path could be dumped (probably truncated)
+ * returns <0 in case of an error
+ */
+static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref,
+                                struct extent_buffer *eb, void *ctx)
+{
+        struct inode_fs_paths *ipath = ctx;
+        char *fspath;
+        char *fspath_min;
+        int i = ipath->fspath->elem_cnt;
+        const int s_ptr = sizeof(char *);
+        u32 bytes_left;
+
+        bytes_left = ipath->fspath->bytes_left > s_ptr ?
+                                        ipath->fspath->bytes_left - s_ptr : 0;
+
+        fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
+        fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb,
+                                inum, fspath_min, bytes_left);
+        if (IS_ERR(fspath))
+                return PTR_ERR(fspath);
+
+        if (fspath > fspath_min) {
+                ipath->fspath->val[i] = (u64)fspath;
+                ++ipath->fspath->elem_cnt;
+                ipath->fspath->bytes_left = fspath - fspath_min;
+        } else {
+                ++ipath->fspath->elem_missed;
+                ipath->fspath->bytes_missing += fspath_min - fspath;
+                ipath->fspath->bytes_left = 0;
+        }
+
+        return 0;
+}
+
+/*
+ * this dumps all file system paths to the inode into the ipath struct, provided
+ * is has been created large enough. each path is zero-terminated and accessed
+ * from ipath->fspath->val[i].
+ * when it returns, there are ipath->fspath->elem_cnt number of paths available
+ * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the
+ * number of missed paths in recored in ipath->fspath->elem_missed, otherwise,
+ * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would
+ * have been needed to return all paths.
+ */
+int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
+{
+        return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
+                                inode_to_path, ipath);
+}
+
+/*
+ * allocates space to return multiple file system paths for an inode.
+ * total_bytes to allocate are passed, note that space usable for actual path
+ * information will be total_bytes - sizeof(struct inode_fs_paths).
+ * the returned pointer must be freed with free_ipath() in the end.
+ */
+struct btrfs_data_container *init_data_container(u32 total_bytes)
+{
+        struct btrfs_data_container *data;
+        size_t alloc_bytes;
+
+        alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
+        data = kmalloc(alloc_bytes, GFP_NOFS);
+        if (!data)
+                return ERR_PTR(-ENOMEM);
+
+        if (total_bytes >= sizeof(*data)) {
+                data->bytes_left = total_bytes - sizeof(*data);
+                data->bytes_missing = 0;
+        } else {
+                data->bytes_missing = sizeof(*data) - total_bytes;
+                data->bytes_left = 0;
+        }
+
+        data->elem_cnt = 0;
+        data->elem_missed = 0;
+
+        return data;
+}
+
+/*
+ * allocates space to return multiple file system paths for an inode.
+ * total_bytes to allocate are passed, note that space usable for actual path
+ * information will be total_bytes - sizeof(struct inode_fs_paths).
+ * the returned pointer must be freed with free_ipath() in the end.
+ */
+struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
+                                        struct btrfs_path *path)
+{
+        struct inode_fs_paths *ifp;
+        struct btrfs_data_container *fspath;
+
+        fspath = init_data_container(total_bytes);
+        if (IS_ERR(fspath))
+                return (void *)fspath;
+
+        ifp = kmalloc(sizeof(*ifp), GFP_NOFS);
+        if (!ifp) {
+                kfree(fspath);
+                return ERR_PTR(-ENOMEM);
+        }
+
+        ifp->btrfs_path = path;
+        ifp->fspath = fspath;
+        ifp->fs_root = fs_root;
+
+        return ifp;
+}
+
+void free_ipath(struct inode_fs_paths *ipath)
+{
+        kfree(ipath);
+}
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
new file mode 100644
index 000000000000..92618837cb8f
--- /dev/null
+++ b/fs/btrfs/backref.h
@@ -0,0 +1,62 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_BACKREF__
+#define __BTRFS_BACKREF__
+
+#include "ioctl.h"
+
+struct inode_fs_paths {
+        struct btrfs_path               *btrfs_path;
+        struct btrfs_root               *fs_root;
+        struct btrfs_data_container     *fspath;
+};
+
+typedef int (iterate_extent_inodes_t)(u64 inum, u64 offset, u64 root,
+                void *ctx);
+typedef int (iterate_irefs_t)(u64 parent, struct btrfs_inode_ref *iref,
+                                struct extent_buffer *eb, void *ctx);
+
+int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
+                        struct btrfs_path *path);
+
+int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
+                        struct btrfs_path *path, struct btrfs_key *found_key);
+
+int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
+                                struct btrfs_extent_item *ei, u32 item_size,
+                                u64 *out_root, u8 *out_level);
+
+int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
+                                struct btrfs_path *path,
+                                u64 extent_item_objectid,
+                                u64 extent_offset,
+                                iterate_extent_inodes_t *iterate, void *ctx);
+
+int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
+                                struct btrfs_path *path,
+                                iterate_extent_inodes_t *iterate, void *ctx);
+
+int paths_from_inode(u64 inum, struct inode_fs_paths *ipath);
+
+struct btrfs_data_container *init_data_container(u32 total_bytes);
+struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
+                                        struct btrfs_path *path);
+void free_ipath(struct inode_fs_paths *ipath);
+
+#endif
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index d9f99a16edd6..5a5d325a3935 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -103,11 +103,6 @@ struct btrfs_inode {
          */
         u64 delalloc_bytes;
 
-        /* total number of bytes that may be used for this inode for
-         * delalloc
-         */
-        u64 reserved_bytes;
-
         /*
          * the size of the file stored in the metadata on disk.  data=ordered
          * means the in-memory i_size might be larger than the size on disk
@@ -115,9 +110,6 @@ struct btrfs_inode {
          */
         u64 disk_i_size;
 
-        /* flags field from the on disk inode */
-        u32 flags;
-
         /*
          * if this is a directory then index_cnt is the counter for the index
          * number for new files that are created
@@ -132,6 +124,15 @@ struct btrfs_inode {
         u64 last_unlink_trans;
 
         /*
+         * Number of bytes outstanding that are going to need csums.  This is
+         * used in ENOSPC accounting.
+         */
+        u64 csum_bytes;
+
+        /* flags field from the on disk inode */
+        u32 flags;
+
+        /*
          * Counters to keep track of the number of extent item's we may use due
          * to delalloc and such.  outstanding_extents is the number of extent
          * items we think we'll end up using, and reserved_extents is the number
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index 8ec5d86f1734..14f1c5a0b2d2 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -85,7 +85,8 @@ struct compressed_bio {
 static inline int compressed_bio_size(struct btrfs_root *root,
                                       unsigned long disk_size)
 {
-        u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
+
         return sizeof(struct compressed_bio) +
                 ((disk_size + root->sectorsize - 1) / root->sectorsize) *
                 csum_size;
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 011cab3aca8d..0fe615e4ea38 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -902,9 +902,10 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 
         orig_ptr = btrfs_node_blockptr(mid, orig_slot);
 
-        if (level < BTRFS_MAX_LEVEL - 1)
+        if (level < BTRFS_MAX_LEVEL - 1) {
                 parent = path->nodes[level + 1];
-        pslot = path->slots[level + 1];
+                pslot = path->slots[level + 1];
+        }
 
         /*
          * deal with the case where there is only one pointer in the root
@@ -1107,9 +1108,10 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
         mid = path->nodes[level];
         WARN_ON(btrfs_header_generation(mid) != trans->transid);
 
-        if (level < BTRFS_MAX_LEVEL - 1)
+        if (level < BTRFS_MAX_LEVEL - 1) {
                 parent = path->nodes[level + 1];
-        pslot = path->slots[level + 1];
+                pslot = path->slots[level + 1];
+        }
 
         if (!parent)
                 return 1;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 03912c5c6f49..b9ba59ff9292 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -30,6 +30,7 @@
 #include <linux/kobject.h>
 #include <trace/events/btrfs.h>
 #include <asm/kmap_types.h>
+#include <linux/pagemap.h>
 #include "extent_io.h"
 #include "extent_map.h"
 #include "async-thread.h"
@@ -360,6 +361,47 @@ struct btrfs_header {
 #define BTRFS_LABEL_SIZE 256
 
 /*
+ * just in case we somehow lose the roots and are not able to mount,
+ * we store an array of the roots from previous transactions
+ * in the super.
+ */
+#define BTRFS_NUM_BACKUP_ROOTS 4
+struct btrfs_root_backup {
+        __le64 tree_root;
+        __le64 tree_root_gen;
+
+        __le64 chunk_root;
+        __le64 chunk_root_gen;
+
+        __le64 extent_root;
+        __le64 extent_root_gen;
+
+        __le64 fs_root;
+        __le64 fs_root_gen;
+
+        __le64 dev_root;
+        __le64 dev_root_gen;
+
+        __le64 csum_root;
+        __le64 csum_root_gen;
+
+        __le64 total_bytes;
+        __le64 bytes_used;
+        __le64 num_devices;
+        /* future */
+        __le64 unsed_64[4];
+
+        u8 tree_root_level;
+        u8 chunk_root_level;
+        u8 extent_root_level;
+        u8 fs_root_level;
+        u8 dev_root_level;
+        u8 csum_root_level;
+        /* future and to align */
+        u8 unused_8[10];
+} __attribute__ ((__packed__));
+
+/*
  * the super block basically lists the main trees of the FS
  * it currently lacks any block count etc etc
  */
@@ -405,6 +447,7 @@ struct btrfs_super_block {
         /* future expansion */
         __le64 reserved[31];
         u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE];
+        struct btrfs_root_backup super_roots[BTRFS_NUM_BACKUP_ROOTS];
 } __attribute__ ((__packed__));
 
 /*
@@ -772,14 +815,8 @@ struct btrfs_space_info {
 struct btrfs_block_rsv {
         u64 size;
         u64 reserved;
-        u64 freed[2];
         struct btrfs_space_info *space_info;
-        struct list_head list;
         spinlock_t lock;
-        atomic_t usage;
-        unsigned int priority:8;
-        unsigned int durable:1;
-        unsigned int refill_used:1;
         unsigned int full:1;
 };
 
@@ -840,10 +877,10 @@ struct btrfs_block_group_cache {
         spinlock_t lock;
         u64 pinned;
         u64 reserved;
-        u64 reserved_pinned;
         u64 bytes_super;
         u64 flags;
         u64 sectorsize;
+        u64 cache_generation;
         unsigned int ro:1;
         unsigned int dirty:1;
         unsigned int iref:1;
@@ -899,6 +936,10 @@ struct btrfs_fs_info {
         spinlock_t block_group_cache_lock;
         struct rb_root block_group_cache_tree;
 
+        /* keep track of unallocated space */
+        spinlock_t free_chunk_lock;
+        u64 free_chunk_space;
+
         struct extent_io_tree freed_extents[2];
         struct extent_io_tree *pinned_extents;
 
@@ -916,14 +957,11 @@ struct btrfs_fs_info {
         struct btrfs_block_rsv trans_block_rsv;
         /* block reservation for chunk tree */
         struct btrfs_block_rsv chunk_block_rsv;
+        /* block reservation for delayed operations */
+        struct btrfs_block_rsv delayed_block_rsv;
 
         struct btrfs_block_rsv empty_block_rsv;
 
-        /* list of block reservations that cross multiple transactions */
-        struct list_head durable_block_rsv_list;
-
-        struct mutex durable_block_rsv_mutex;
-
         u64 generation;
         u64 last_trans_committed;
 
@@ -942,8 +980,8 @@ struct btrfs_fs_info {
         wait_queue_head_t transaction_blocked_wait;
         wait_queue_head_t async_submit_wait;
 
-        struct btrfs_super_block super_copy;
-        struct btrfs_super_block super_for_commit;
+        struct btrfs_super_block *super_copy;
+        struct btrfs_super_block *super_for_commit;
         struct block_device *__bdev;
         struct super_block *sb;
         struct inode *btree_inode;
@@ -1036,6 +1074,7 @@ struct btrfs_fs_info {
         struct btrfs_workers endio_freespace_worker;
         struct btrfs_workers submit_workers;
         struct btrfs_workers caching_workers;
+        struct btrfs_workers readahead_workers;
 
         /*
          * fixup workers take dirty pages that didn't properly go through
@@ -1119,6 +1158,13 @@ struct btrfs_fs_info {
         u64 fs_state;
 
         struct btrfs_delayed_root *delayed_root;
+
+        /* readahead tree */
+        spinlock_t reada_lock;
+        struct radix_tree_root reada_tree;
+
+        /* next backup root to be overwritten */
+        int backup_root_index;
 };
 
 /*
@@ -1363,6 +1409,7 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_ENOSPC_DEBUG         (1 << 15)
 #define BTRFS_MOUNT_AUTO_DEFRAG         (1 << 16)
 #define BTRFS_MOUNT_INODE_MAP_CACHE     (1 << 17)
+#define BTRFS_MOUNT_RECOVERY            (1 << 18)
 
 #define btrfs_clear_opt(o, opt)         ((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)           ((o) |= BTRFS_MOUNT_##opt)
@@ -1978,6 +2025,55 @@ static inline bool btrfs_root_readonly(struct btrfs_root *root)
         return root->root_item.flags & BTRFS_ROOT_SUBVOL_RDONLY;
 }
 
+/* struct btrfs_root_backup */
+BTRFS_SETGET_STACK_FUNCS(backup_tree_root, struct btrfs_root_backup,
+                   tree_root, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_tree_root_gen, struct btrfs_root_backup,
+                   tree_root_gen, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_tree_root_level, struct btrfs_root_backup,
+                   tree_root_level, 8);
+
+BTRFS_SETGET_STACK_FUNCS(backup_chunk_root, struct btrfs_root_backup,
+                   chunk_root, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_chunk_root_gen, struct btrfs_root_backup,
+                   chunk_root_gen, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_chunk_root_level, struct btrfs_root_backup,
+                   chunk_root_level, 8);
+
+BTRFS_SETGET_STACK_FUNCS(backup_extent_root, struct btrfs_root_backup,
+                   extent_root, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_extent_root_gen, struct btrfs_root_backup,
+                   extent_root_gen, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_extent_root_level, struct btrfs_root_backup,
+                   extent_root_level, 8);
+
+BTRFS_SETGET_STACK_FUNCS(backup_fs_root, struct btrfs_root_backup,
+                   fs_root, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_fs_root_gen, struct btrfs_root_backup,
+                   fs_root_gen, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_fs_root_level, struct btrfs_root_backup,
+                   fs_root_level, 8);
+
+BTRFS_SETGET_STACK_FUNCS(backup_dev_root, struct btrfs_root_backup,
+                   dev_root, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_dev_root_gen, struct btrfs_root_backup,
+                   dev_root_gen, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_dev_root_level, struct btrfs_root_backup,
+                   dev_root_level, 8);
+
+BTRFS_SETGET_STACK_FUNCS(backup_csum_root, struct btrfs_root_backup,
+                   csum_root, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_csum_root_gen, struct btrfs_root_backup,
+                   csum_root_gen, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_csum_root_level, struct btrfs_root_backup,
+                   csum_root_level, 8);
+BTRFS_SETGET_STACK_FUNCS(backup_total_bytes, struct btrfs_root_backup,
+                   total_bytes, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_bytes_used, struct btrfs_root_backup,
+                   bytes_used, 64);
+BTRFS_SETGET_STACK_FUNCS(backup_num_devices, struct btrfs_root_backup,
+                   num_devices, 64);
+
 /* struct btrfs_super_block */
 
 BTRFS_SETGET_STACK_FUNCS(super_bytenr, struct btrfs_super_block, bytenr, 64);
@@ -2129,6 +2225,11 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info)
                 (space_info->flags & BTRFS_BLOCK_GROUP_DATA));
 }
 
+static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping)
+{
+        return mapping_gfp_mask(mapping) & ~__GFP_FS;
+}
+
 /* extent-tree.c */
 static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root,
                                                  unsigned num_items)
@@ -2137,6 +2238,17 @@ static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root,
                 3 * num_items;
 }
 
+/*
+ * Doing a truncate won't result in new nodes or leaves, just what we need for
+ * COW.
+ */
+static inline u64 btrfs_calc_trunc_metadata_size(struct btrfs_root *root,
+                                                 unsigned num_items)
+{
+        return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) *
+                num_items;
+}
+
 void btrfs_put_block_group(struct btrfs_block_group_cache *cache);
 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
                            struct btrfs_root *root, unsigned long count);
@@ -2146,6 +2258,9 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
                              u64 num_bytes, u64 *refs, u64 *flags);
 int btrfs_pin_extent(struct btrfs_root *root,
                      u64 bytenr, u64 num, int reserved);
+int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans,
+                                    struct btrfs_root *root,
+                                    u64 bytenr, u64 num_bytes);
 int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
                           struct btrfs_root *root,
                           u64 objectid, u64 offset, u64 bytenr);
@@ -2196,8 +2311,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans,
                       u64 root_objectid, u64 owner, u64 offset);
 
 int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len);
-int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
-                                u64 num_bytes, int reserve, int sinfo);
+int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root,
+                                       u64 start, u64 len);
 int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
                                 struct btrfs_root *root);
 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
@@ -2240,25 +2355,23 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv);
 struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root);
 void btrfs_free_block_rsv(struct btrfs_root *root,
                           struct btrfs_block_rsv *rsv);
-void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info,
-                                 struct btrfs_block_rsv *rsv);
-int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
-                        struct btrfs_root *root,
+int btrfs_block_rsv_add(struct btrfs_root *root,
                         struct btrfs_block_rsv *block_rsv,
                         u64 num_bytes);
-int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
-                          struct btrfs_root *root,
+int btrfs_block_rsv_add_noflush(struct btrfs_root *root,
+                                struct btrfs_block_rsv *block_rsv,
+                                u64 num_bytes);
+int btrfs_block_rsv_check(struct btrfs_root *root,
+                          struct btrfs_block_rsv *block_rsv, int min_factor);
+int btrfs_block_rsv_refill(struct btrfs_root *root,
                           struct btrfs_block_rsv *block_rsv,
-                          u64 min_reserved, int min_factor);
+                          u64 min_reserved);
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
                             struct btrfs_block_rsv *dst_rsv,
                             u64 num_bytes);
 void btrfs_block_rsv_release(struct btrfs_root *root,
                              struct btrfs_block_rsv *block_rsv,
                              u64 num_bytes);
-int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans,
-                                    struct btrfs_root *root,
-                                    struct btrfs_block_rsv *rsv);
 int btrfs_set_block_group_ro(struct btrfs_root *root,
                              struct btrfs_block_group_cache *cache);
 int btrfs_set_block_group_rw(struct btrfs_root *root,
@@ -2379,6 +2492,18 @@ static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info)
         smp_mb();
         return fs_info->closing;
 }
+static inline void free_fs_info(struct btrfs_fs_info *fs_info)
+{
+        kfree(fs_info->delayed_root);
+        kfree(fs_info->extent_root);
+        kfree(fs_info->tree_root);
+        kfree(fs_info->chunk_root);
+        kfree(fs_info->dev_root);
+        kfree(fs_info->csum_root);
+        kfree(fs_info->super_copy);
+        kfree(fs_info->super_for_commit);
+        kfree(fs_info);
+}
 
 /* root-item.c */
 int btrfs_find_root_ref(struct btrfs_root *tree_root,
@@ -2579,11 +2704,6 @@ int btrfs_update_inode(struct btrfs_trans_handle *trans,
 int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode);
 int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode);
 int btrfs_orphan_cleanup(struct btrfs_root *root);
-void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans,
-                                struct btrfs_pending_snapshot *pending,
-                                u64 *bytes_to_reserve);
-void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans,
-                                struct btrfs_pending_snapshot *pending);
 void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans,
                               struct btrfs_root *root);
 int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size);
@@ -2697,4 +2817,20 @@ int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid);
 int btrfs_scrub_progress(struct btrfs_root *root, u64 devid,
                          struct btrfs_scrub_progress *progress);
 
+/* reada.c */
+struct reada_control {
+        struct btrfs_root       *root;          /* tree to prefetch */
+        struct btrfs_key        key_start;
+        struct btrfs_key        key_end;        /* exclusive */
+        atomic_t                elems;
+        struct kref             refcnt;
+        wait_queue_head_t       wait;
+};
+struct reada_control *btrfs_reada_add(struct btrfs_root *root,
+                              struct btrfs_key *start, struct btrfs_key *end);
+int btrfs_reada_wait(void *handle);
+void btrfs_reada_detach(void *handle);
+int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
+                         u64 start, int err);
+
 #endif
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index ae4d9cd10961..3a1b939c9ae2 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -591,7 +591,7 @@ static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans,
                 return 0;
 
         src_rsv = trans->block_rsv;
-        dst_rsv = &root->fs_info->global_block_rsv;
+        dst_rsv = &root->fs_info->delayed_block_rsv;
 
         num_bytes = btrfs_calc_trans_metadata_size(root, 1);
         ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
@@ -609,7 +609,7 @@ static void btrfs_delayed_item_release_metadata(struct btrfs_root *root,
         if (!item->bytes_reserved)
                 return;
 
-        rsv = &root->fs_info->global_block_rsv;
+        rsv = &root->fs_info->delayed_block_rsv;
         btrfs_block_rsv_release(root, rsv,
                                 item->bytes_reserved);
 }
@@ -624,13 +624,36 @@ static int btrfs_delayed_inode_reserve_metadata(
         u64 num_bytes;
         int ret;
 
-        if (!trans->bytes_reserved)
-                return 0;
-
         src_rsv = trans->block_rsv;
-        dst_rsv = &root->fs_info->global_block_rsv;
+        dst_rsv = &root->fs_info->delayed_block_rsv;
 
         num_bytes = btrfs_calc_trans_metadata_size(root, 1);
+
+        /*
+         * btrfs_dirty_inode will update the inode under btrfs_join_transaction
+         * which doesn't reserve space for speed.  This is a problem since we
+         * still need to reserve space for this update, so try to reserve the
+         * space.
+         *
+         * Now if src_rsv == delalloc_block_rsv we'll let it just steal since
+         * we're accounted for.
+         */
+        if (!trans->bytes_reserved &&
+            src_rsv != &root->fs_info->delalloc_block_rsv) {
+                ret = btrfs_block_rsv_add_noflush(root, dst_rsv, num_bytes);
+                /*
+                 * Since we're under a transaction reserve_metadata_bytes could
+                 * try to commit the transaction which will make it return
+                 * EAGAIN to make us stop the transaction we have, so return
+                 * ENOSPC instead so that btrfs_dirty_inode knows what to do.
+                 */
+                if (ret == -EAGAIN)
+                        ret = -ENOSPC;
+                if (!ret)
+                        node->bytes_reserved = num_bytes;
+                return ret;
+        }
+
         ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
         if (!ret)
                 node->bytes_reserved = num_bytes;
@@ -646,7 +669,7 @@ static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root,
         if (!node->bytes_reserved)
                 return;
 
-        rsv = &root->fs_info->global_block_rsv;
+        rsv = &root->fs_info->delayed_block_rsv;
         btrfs_block_rsv_release(root, rsv,
                                 node->bytes_reserved);
         node->bytes_reserved = 0;
@@ -1026,7 +1049,7 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
         path->leave_spinning = 1;
 
         block_rsv = trans->block_rsv;
-        trans->block_rsv = &root->fs_info->global_block_rsv;
+        trans->block_rsv = &root->fs_info->delayed_block_rsv;
 
         delayed_root = btrfs_get_delayed_root(root);
 
@@ -1069,7 +1092,7 @@ static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
         path->leave_spinning = 1;
 
         block_rsv = trans->block_rsv;
-        trans->block_rsv = &node->root->fs_info->global_block_rsv;
+        trans->block_rsv = &node->root->fs_info->delayed_block_rsv;
 
         ret = btrfs_insert_delayed_items(trans, path, node->root, node);
         if (!ret)
@@ -1149,7 +1172,7 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
                 goto free_path;
 
         block_rsv = trans->block_rsv;
-        trans->block_rsv = &root->fs_info->global_block_rsv;
+        trans->block_rsv = &root->fs_info->delayed_block_rsv;
 
         ret = btrfs_insert_delayed_items(trans, path, root, delayed_node);
         if (!ret)
@@ -1686,11 +1709,8 @@ int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,
         }
 
         ret = btrfs_delayed_inode_reserve_metadata(trans, root, delayed_node);
-        /*
-         * we must reserve enough space when we start a new transaction,
-         * so reserving metadata failure is impossible
-         */
-        BUG_ON(ret);
+        if (ret)
+                goto release_node;
 
         fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
         delayed_node->inode_dirty = 1;
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 07ea91879a91..102c176fc29c 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -256,8 +256,7 @@ void btrfs_csum_final(u32 crc, char *result)
 static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
                            int verify)
 {
-        u16 csum_size =
-                btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
         char *result = NULL;
         unsigned long len;
         unsigned long cur_len;
@@ -367,7 +366,8 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
         clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
         io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
         while (1) {
-                ret = read_extent_buffer_pages(io_tree, eb, start, 1,
+                ret = read_extent_buffer_pages(io_tree, eb, start,
+                                               WAIT_COMPLETE,
                                                btree_get_extent, mirror_num);
                 if (!ret &&
                     !verify_parent_transid(io_tree, eb, parent_transid))
@@ -608,11 +608,48 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
         end = min_t(u64, eb->len, PAGE_CACHE_SIZE);
         end = eb->start + end - 1;
 err:
+        if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) {
+                clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags);
+                btree_readahead_hook(root, eb, eb->start, ret);
+        }
+
         free_extent_buffer(eb);
 out:
         return ret;
 }
 
+static int btree_io_failed_hook(struct bio *failed_bio,
+                         struct page *page, u64 start, u64 end,
+                         u64 mirror_num, struct extent_state *state)
+{
+        struct extent_io_tree *tree;
+        unsigned long len;
+        struct extent_buffer *eb;
+        struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
+
+        tree = &BTRFS_I(page->mapping->host)->io_tree;
+        if (page->private == EXTENT_PAGE_PRIVATE)
+                goto out;
+        if (!page->private)
+                goto out;
+
+        len = page->private >> 2;
+        WARN_ON(len == 0);
+
+        eb = alloc_extent_buffer(tree, start, len, page);
+        if (eb == NULL)
+                goto out;
+
+        if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) {
+                clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags);
+                btree_readahead_hook(root, eb, eb->start, -EIO);
+        }
+        free_extent_buffer(eb);
+
+out:
+        return -EIO;    /* we fixed nothing */
+}
+
 static void end_workqueue_bio(struct bio *bio, int err)
 {
         struct end_io_wq *end_io_wq = bio->bi_private;
@@ -908,7 +945,7 @@ static int btree_readpage(struct file *file, struct page *page)
 {
         struct extent_io_tree *tree;
         tree = &BTRFS_I(page->mapping->host)->io_tree;
-        return extent_read_full_page(tree, page, btree_get_extent);
+        return extent_read_full_page(tree, page, btree_get_extent, 0);
 }
 
 static int btree_releasepage(struct page *page, gfp_t gfp_flags)
@@ -974,11 +1011,43 @@ int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize,
         if (!buf)
                 return 0;
         read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
-                                 buf, 0, 0, btree_get_extent, 0);
+                                 buf, 0, WAIT_NONE, btree_get_extent, 0);
         free_extent_buffer(buf);
         return ret;
 }
 
+int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize,
+                         int mirror_num, struct extent_buffer **eb)
+{
+        struct extent_buffer *buf = NULL;
+        struct inode *btree_inode = root->fs_info->btree_inode;
+        struct extent_io_tree *io_tree = &BTRFS_I(btree_inode)->io_tree;
+        int ret;
+
+        buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
+        if (!buf)
+                return 0;
+
+        set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags);
+
+        ret = read_extent_buffer_pages(io_tree, buf, 0, WAIT_PAGE_LOCK,
+                                       btree_get_extent, mirror_num);
+        if (ret) {
+                free_extent_buffer(buf);
+                return ret;
+        }
+
+        if (test_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags)) {
+                free_extent_buffer(buf);
+                return -EIO;
+        } else if (extent_buffer_uptodate(io_tree, buf, NULL)) {
+                *eb = buf;
+        } else {
+                free_extent_buffer(buf);
+        }
+        return 0;
+}
+
 struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
                                             u64 bytenr, u32 blocksize)
 {
@@ -1135,10 +1204,12 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
 
         generation = btrfs_root_generation(&root->root_item);
         blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
+        root->commit_root = NULL;
         root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
                                      blocksize, generation);
         if (!root->node || !btrfs_buffer_uptodate(root->node, generation)) {
                 free_extent_buffer(root->node);
+                root->node = NULL;
                 return -EIO;
         }
         root->commit_root = btrfs_root_node(root);
@@ -1577,6 +1648,235 @@ sleep:
         return 0;
 }
 
+/*
+ * this will find the highest generation in the array of
+ * root backups.  The index of the highest array is returned,
+ * or -1 if we can't find anything.
+ *
+ * We check to make sure the array is valid by comparing the
+ * generation of the latest  root in the array with the generation
+ * in the super block.  If they don't match we pitch it.
+ */
+static int find_newest_super_backup(struct btrfs_fs_info *info, u64 newest_gen)
+{
+        u64 cur;
+        int newest_index = -1;
+        struct btrfs_root_backup *root_backup;
+        int i;
+
+        for (i = 0; i < BTRFS_NUM_BACKUP_ROOTS; i++) {
+                root_backup = info->super_copy->super_roots + i;
+                cur = btrfs_backup_tree_root_gen(root_backup);
+                if (cur == newest_gen)
+                        newest_index = i;
+        }
+
+        /* check to see if we actually wrapped around */
+        if (newest_index == BTRFS_NUM_BACKUP_ROOTS - 1) {
+                root_backup = info->super_copy->super_roots;
+                cur = btrfs_backup_tree_root_gen(root_backup);
+                if (cur == newest_gen)
+                        newest_index = 0;
+        }
+        return newest_index;
+}
+
+
+/*
+ * find the oldest backup so we know where to store new entries
+ * in the backup array.  This will set the backup_root_index
+ * field in the fs_info struct
+ */
+static void find_oldest_super_backup(struct btrfs_fs_info *info,
+                                     u64 newest_gen)
+{
+        int newest_index = -1;
+
+        newest_index = find_newest_super_backup(info, newest_gen);
+        /* if there was garbage in there, just move along */
+        if (newest_index == -1) {
+                info->backup_root_index = 0;
+        } else {
+                info->backup_root_index = (newest_index + 1) % BTRFS_NUM_BACKUP_ROOTS;
+        }
+}
+
+/*
+ * copy all the root pointers into the super backup array.
+ * this will bump the backup pointer by one when it is
+ * done
+ */
+static void backup_super_roots(struct btrfs_fs_info *info)
+{
+        int next_backup;
+        struct btrfs_root_backup *root_backup;
+        int last_backup;
+
+        next_backup = info->backup_root_index;
+        last_backup = (next_backup + BTRFS_NUM_BACKUP_ROOTS - 1) %
+                BTRFS_NUM_BACKUP_ROOTS;
+
+        /*
+         * just overwrite the last backup if we're at the same generation
+         * this happens only at umount
+         */
+        root_backup = info->super_for_commit->super_roots + last_backup;
+        if (btrfs_backup_tree_root_gen(root_backup) ==
+            btrfs_header_generation(info->tree_root->node))
+                next_backup = last_backup;
+
+        root_backup = info->super_for_commit->super_roots + next_backup;
+
+        /*
+         * make sure all of our padding and empty slots get zero filled
+         * regardless of which ones we use today
+         */
+        memset(root_backup, 0, sizeof(*root_backup));
+
+        info->backup_root_index = (next_backup + 1) % BTRFS_NUM_BACKUP_ROOTS;
+
+        btrfs_set_backup_tree_root(root_backup, info->tree_root->node->start);
+        btrfs_set_backup_tree_root_gen(root_backup,
+                               btrfs_header_generation(info->tree_root->node));
+
+        btrfs_set_backup_tree_root_level(root_backup,
+                               btrfs_header_level(info->tree_root->node));
+
+        btrfs_set_backup_chunk_root(root_backup, info->chunk_root->node->start);
+        btrfs_set_backup_chunk_root_gen(root_backup,
+                               btrfs_header_generation(info->chunk_root->node));
+        btrfs_set_backup_chunk_root_level(root_backup,
+                               btrfs_header_level(info->chunk_root->node));
+
+        btrfs_set_backup_extent_root(root_backup, info->extent_root->node->start);
+        btrfs_set_backup_extent_root_gen(root_backup,
+                               btrfs_header_generation(info->extent_root->node));
+        btrfs_set_backup_extent_root_level(root_backup,
+                               btrfs_header_level(info->extent_root->node));
+
+        /*
+         * we might commit during log recovery, which happens before we set
+         * the fs_root.  Make sure it is valid before we fill it in.
+         */
+        if (info->fs_root && info->fs_root->node) {
+                btrfs_set_backup_fs_root(root_backup,
+                                         info->fs_root->node->start);
+                btrfs_set_backup_fs_root_gen(root_backup,
+                               btrfs_header_generation(info->fs_root->node));
+                btrfs_set_backup_fs_root_level(root_backup,
+                               btrfs_header_level(info->fs_root->node));
+        }
+
+        btrfs_set_backup_dev_root(root_backup, info->dev_root->node->start);
+        btrfs_set_backup_dev_root_gen(root_backup,
+                               btrfs_header_generation(info->dev_root->node));
+        btrfs_set_backup_dev_root_level(root_backup,
+                                       btrfs_header_level(info->dev_root->node));
+
+        btrfs_set_backup_csum_root(root_backup, info->csum_root->node->start);
+        btrfs_set_backup_csum_root_gen(root_backup,
+                               btrfs_header_generation(info->csum_root->node));
+        btrfs_set_backup_csum_root_level(root_backup,
+                               btrfs_header_level(info->csum_root->node));
+
+        btrfs_set_backup_total_bytes(root_backup,
+                             btrfs_super_total_bytes(info->super_copy));
+        btrfs_set_backup_bytes_used(root_backup,
+                             btrfs_super_bytes_used(info->super_copy));
+        btrfs_set_backup_num_devices(root_backup,
+                             btrfs_super_num_devices(info->super_copy));
+
+        /*
+         * if we don't copy this out to the super_copy, it won't get remembered
+         * for the next commit
+         */
+        memcpy(&info->super_copy->super_roots,
+               &info->super_for_commit->super_roots,
+               sizeof(*root_backup) * BTRFS_NUM_BACKUP_ROOTS);
+}
+
+/*
+ * this copies info out of the root backup array and back into
+ * the in-memory super block.  It is meant to help iterate through
+ * the array, so you send it the number of backups you've already
+ * tried and the last backup index you used.
+ *
+ * this returns -1 when it has tried all the backups
+ */
+static noinline int next_root_backup(struct btrfs_fs_info *info,
+                                     struct btrfs_super_block *super,
+                                     int *num_backups_tried, int *backup_index)
+{
+        struct btrfs_root_backup *root_backup;
+        int newest = *backup_index;
+
+        if (*num_backups_tried == 0) {
+                u64 gen = btrfs_super_generation(super);
+
+                newest = find_newest_super_backup(info, gen);
+                if (newest == -1)
+                        return -1;
+
+                *backup_index = newest;
+                *num_backups_tried = 1;
+        } else if (*num_backups_tried == BTRFS_NUM_BACKUP_ROOTS) {
+                /* we've tried all the backups, all done */
+                return -1;
+        } else {
+                /* jump to the next oldest backup */
+                newest = (*backup_index + BTRFS_NUM_BACKUP_ROOTS - 1) %
+                        BTRFS_NUM_BACKUP_ROOTS;
+                *backup_index = newest;
+                *num_backups_tried += 1;
+        }
+        root_backup = super->super_roots + newest;
+
+        btrfs_set_super_generation(super,
+                                   btrfs_backup_tree_root_gen(root_backup));
+        btrfs_set_super_root(super, btrfs_backup_tree_root(root_backup));
+        btrfs_set_super_root_level(super,
+                                   btrfs_backup_tree_root_level(root_backup));
+        btrfs_set_super_bytes_used(super, btrfs_backup_bytes_used(root_backup));
+
+        /*
+         * fixme: the total bytes and num_devices need to match or we should
+         * need a fsck
+         */
+        btrfs_set_super_total_bytes(super, btrfs_backup_total_bytes(root_backup));
+        btrfs_set_super_num_devices(super, btrfs_backup_num_devices(root_backup));
+        return 0;
+}
+
+/* helper to cleanup tree roots */
+static void free_root_pointers(struct btrfs_fs_info *info, int chunk_root)
+{
+        free_extent_buffer(info->tree_root->node);
+        free_extent_buffer(info->tree_root->commit_root);
+        free_extent_buffer(info->dev_root->node);
+        free_extent_buffer(info->dev_root->commit_root);
+        free_extent_buffer(info->extent_root->node);
+        free_extent_buffer(info->extent_root->commit_root);
+        free_extent_buffer(info->csum_root->node);
+        free_extent_buffer(info->csum_root->commit_root);
+
+        info->tree_root->node = NULL;
+        info->tree_root->commit_root = NULL;
+        info->dev_root->node = NULL;
+        info->dev_root->commit_root = NULL;
+        info->extent_root->node = NULL;
+        info->extent_root->commit_root = NULL;
+        info->csum_root->node = NULL;
+        info->csum_root->commit_root = NULL;
+
+        if (chunk_root) {
+                free_extent_buffer(info->chunk_root->node);
+                free_extent_buffer(info->chunk_root->commit_root);
+                info->chunk_root->node = NULL;
+                info->chunk_root->commit_root = NULL;
+        }
+}
+
+
 struct btrfs_root *open_ctree(struct super_block *sb,
                               struct btrfs_fs_devices *fs_devices,
                               char *options)
@@ -1604,6 +1904,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
         int ret;
         int err = -EINVAL;
+        int num_backups_tried = 0;
+        int backup_index = 0;
 
         struct btrfs_super_block *disk_super;
 
@@ -1648,6 +1950,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         spin_lock_init(&fs_info->fs_roots_radix_lock);
         spin_lock_init(&fs_info->delayed_iput_lock);
         spin_lock_init(&fs_info->defrag_inodes_lock);
+        spin_lock_init(&fs_info->free_chunk_lock);
         mutex_init(&fs_info->reloc_mutex);
 
         init_completion(&fs_info->kobj_unregister);
@@ -1665,8 +1968,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         btrfs_init_block_rsv(&fs_info->trans_block_rsv);
         btrfs_init_block_rsv(&fs_info->chunk_block_rsv);
         btrfs_init_block_rsv(&fs_info->empty_block_rsv);
-        INIT_LIST_HEAD(&fs_info->durable_block_rsv_list);
-        mutex_init(&fs_info->durable_block_rsv_mutex);
+        btrfs_init_block_rsv(&fs_info->delayed_block_rsv);
         atomic_set(&fs_info->nr_async_submits, 0);
         atomic_set(&fs_info->async_delalloc_pages, 0);
         atomic_set(&fs_info->async_submit_draining, 0);
@@ -1677,6 +1979,11 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         fs_info->metadata_ratio = 0;
         fs_info->defrag_inodes = RB_ROOT;
         fs_info->trans_no_join = 0;
+        fs_info->free_chunk_space = 0;
+
+        /* readahead state */
+        INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_WAIT);
+        spin_lock_init(&fs_info->reada_lock);
 
         fs_info->thread_pool_size = min_t(unsigned long,
                                           num_online_cpus() + 2, 8);
@@ -1766,14 +2073,14 @@ struct btrfs_root *open_ctree(struct super_block *sb,
                 goto fail_alloc;
         }
 
-        memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy));
-        memcpy(&fs_info->super_for_commit, &fs_info->super_copy,
-               sizeof(fs_info->super_for_commit));
+        memcpy(fs_info->super_copy, bh->b_data, sizeof(*fs_info->super_copy));
+        memcpy(fs_info->super_for_commit, fs_info->super_copy,
+               sizeof(*fs_info->super_for_commit));
         brelse(bh);
 
-        memcpy(fs_info->fsid, fs_info->super_copy.fsid, BTRFS_FSID_SIZE);
+        memcpy(fs_info->fsid, fs_info->super_copy->fsid, BTRFS_FSID_SIZE);
 
-        disk_super = &fs_info->super_copy;
+        disk_super = fs_info->super_copy;
         if (!btrfs_super_root(disk_super))
                 goto fail_alloc;
 
@@ -1783,6 +2090,13 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY);
 
         /*
+         * run through our array of backup supers and setup
+         * our ring pointer to the oldest one
+         */
+        generation = btrfs_super_generation(disk_super);
+        find_oldest_super_backup(fs_info, generation);
+
+        /*
          * In the long term, we'll store the compression type in the super
          * block, and it'll be used for per file compression control.
          */
@@ -1870,6 +2184,9 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta",
                            fs_info->thread_pool_size,
                            &fs_info->generic_worker);
+        btrfs_init_workers(&fs_info->readahead_workers, "readahead",
+                           fs_info->thread_pool_size,
+                           &fs_info->generic_worker);
 
         /*
          * endios are largely parallel and should have a very
@@ -1880,6 +2197,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
         fs_info->endio_write_workers.idle_thresh = 2;
         fs_info->endio_meta_write_workers.idle_thresh = 2;
+        fs_info->readahead_workers.idle_thresh = 2;
 
         btrfs_start_workers(&fs_info->workers, 1);
         btrfs_start_workers(&fs_info->generic_worker, 1);
@@ -1893,6 +2211,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         btrfs_start_workers(&fs_info->endio_freespace_worker, 1);
         btrfs_start_workers(&fs_info->delayed_workers, 1);
         btrfs_start_workers(&fs_info->caching_workers, 1);
+        btrfs_start_workers(&fs_info->readahead_workers, 1);
 
         fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super);
         fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages,
@@ -1939,7 +2258,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         if (!test_bit(EXTENT_BUFFER_UPTODATE, &chunk_root->node->bflags)) {
                 printk(KERN_WARNING "btrfs: failed to read chunk root on %s\n",
                        sb->s_id);
-                goto fail_chunk_root;
+                goto fail_tree_roots;
         }
         btrfs_set_root_node(&chunk_root->root_item, chunk_root->node);
         chunk_root->commit_root = btrfs_root_node(chunk_root);
@@ -1954,11 +2273,12 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         if (ret) {
                 printk(KERN_WARNING "btrfs: failed to read chunk tree on %s\n",
                        sb->s_id);
-                goto fail_chunk_root;
+                goto fail_tree_roots;
         }
 
         btrfs_close_extra_devices(fs_devices);
 
+retry_root_backup:
         blocksize = btrfs_level_size(tree_root,
                                      btrfs_super_root_level(disk_super));
         generation = btrfs_super_generation(disk_super);
@@ -1966,32 +2286,33 @@ struct btrfs_root *open_ctree(struct super_block *sb,
         tree_root->node = read_tree_block(tree_root,
                                           btrfs_super_root(disk_super),
                                           blocksize, generation);
-        if (!tree_root->node)
-                goto fail_chunk_root;
-        if (!test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) {
+        if (!tree_root->node ||
+            !test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) {
                 printk(KERN_WARNING "btrfs: failed to read tree root on %s\n",
                        sb->s_id);
-                goto fail_tree_root;
+
+                goto recovery_tree_root;
         }
+
         btrfs_set_root_node(&tree_root->root_item, tree_root->node);
         tree_root->commit_root = btrfs_root_node(tree_root);
 
         ret = find_and_setup_root(tree_root, fs_info,
                                   BTRFS_EXTENT_TREE_OBJECTID, extent_root);
         if (ret)
-                goto fail_tree_root;
+                goto recovery_tree_root;
         extent_root->track_dirty = 1;
 
         ret = find_and_setup_root(tree_root, fs_info,
                                   BTRFS_DEV_TREE_OBJECTID, dev_root);
         if (ret)
-                goto fail_extent_root;
+                goto recovery_tree_root;
         dev_root->track_dirty = 1;
 
         ret = find_and_setup_root(tree_root, fs_info,
                                   BTRFS_CSUM_TREE_OBJECTID, csum_root);
         if (ret)
-                goto fail_dev_root;
+                goto recovery_tree_root;
 
         csum_root->track_dirty = 1;
 
@@ -2124,22 +2445,13 @@ fail_cleaner:
 
 fail_block_groups:
         btrfs_free_block_groups(fs_info);
-        free_extent_buffer(csum_root->node);
-        free_extent_buffer(csum_root->commit_root);
-fail_dev_root:
-        free_extent_buffer(dev_root->node);
-        free_extent_buffer(dev_root->commit_root);
-fail_extent_root:
-        free_extent_buffer(extent_root->node);
-        free_extent_buffer(extent_root->commit_root);
-fail_tree_root:
-        free_extent_buffer(tree_root->node);
-        free_extent_buffer(tree_root->commit_root);
-fail_chunk_root:
-        free_extent_buffer(chunk_root->node);
-        free_extent_buffer(chunk_root->commit_root);
+
+fail_tree_roots:
+        free_root_pointers(fs_info, 1);
+
 fail_sb_buffer:
         btrfs_stop_workers(&fs_info->generic_worker);
+        btrfs_stop_workers(&fs_info->readahead_workers);
         btrfs_stop_workers(&fs_info->fixup_workers);
         btrfs_stop_workers(&fs_info->delalloc_workers);
         btrfs_stop_workers(&fs_info->workers);
@@ -2152,7 +2464,6 @@ fail_sb_buffer:
         btrfs_stop_workers(&fs_info->delayed_workers);
         btrfs_stop_workers(&fs_info->caching_workers);
 fail_alloc:
-        kfree(fs_info->delayed_root);
 fail_iput:
         invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
         iput(fs_info->btree_inode);
@@ -2164,13 +2475,27 @@ fail_bdi:
 fail_srcu:
         cleanup_srcu_struct(&fs_info->subvol_srcu);
 fail:
-        kfree(extent_root);
-        kfree(tree_root);
-        kfree(fs_info);
-        kfree(chunk_root);
-        kfree(dev_root);
-        kfree(csum_root);
+        free_fs_info(fs_info);
         return ERR_PTR(err);
+
+recovery_tree_root:
+
+        if (!btrfs_test_opt(tree_root, RECOVERY))
+                goto fail_tree_roots;
+
+        free_root_pointers(fs_info, 0);
+
+        /* don't use the log in recovery mode, it won't be valid */
+        btrfs_set_super_log_root(disk_super, 0);
+
+        /* we can't trust the free space cache either */
+        btrfs_set_opt(fs_info->mount_opt, CLEAR_CACHE);
+
+        ret = next_root_backup(fs_info, fs_info->super_copy,
+                               &num_backups_tried, &backup_index);
+        if (ret == -1)
+                goto fail_block_groups;
+        goto retry_root_backup;
 }
 
 static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate)
@@ -2338,10 +2663,11 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
         int total_errors = 0;
         u64 flags;
 
-        max_errors = btrfs_super_num_devices(&root->fs_info->super_copy) - 1;
+        max_errors = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
         do_barriers = !btrfs_test_opt(root, NOBARRIER);
+        backup_super_roots(root->fs_info);
 
-        sb = &root->fs_info->super_for_commit;
+        sb = root->fs_info->super_for_commit;
         dev_item = &sb->dev_item;
 
         mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
@@ -2545,8 +2871,6 @@ int close_ctree(struct btrfs_root *root)
         /* clear out the rbtree of defraggable inodes */
         btrfs_run_defrag_inodes(root->fs_info);
 
-        btrfs_put_block_group_cache(fs_info);
-
         /*
          * Here come 2 situations when btrfs is broken to flip readonly:
          *
@@ -2572,6 +2896,8 @@ int close_ctree(struct btrfs_root *root)
                         printk(KERN_ERR "btrfs: commit super ret %d\n", ret);
         }
 
+        btrfs_put_block_group_cache(fs_info);
+
         kthread_stop(root->fs_info->transaction_kthread);
         kthread_stop(root->fs_info->cleaner_kthread);
 
@@ -2603,7 +2929,6 @@ int close_ctree(struct btrfs_root *root)
         del_fs_roots(fs_info);
 
         iput(fs_info->btree_inode);
-        kfree(fs_info->delayed_root);
 
         btrfs_stop_workers(&fs_info->generic_worker);
         btrfs_stop_workers(&fs_info->fixup_workers);
@@ -2617,6 +2942,7 @@ int close_ctree(struct btrfs_root *root)
         btrfs_stop_workers(&fs_info->submit_workers);
         btrfs_stop_workers(&fs_info->delayed_workers);
         btrfs_stop_workers(&fs_info->caching_workers);
+        btrfs_stop_workers(&fs_info->readahead_workers);
 
         btrfs_close_devices(fs_info->fs_devices);
         btrfs_mapping_tree_free(&fs_info->mapping_tree);
@@ -2624,12 +2950,7 @@ int close_ctree(struct btrfs_root *root)
         bdi_destroy(&fs_info->bdi);
         cleanup_srcu_struct(&fs_info->subvol_srcu);
 
-        kfree(fs_info->extent_root);
-        kfree(fs_info->tree_root);
-        kfree(fs_info->chunk_root);
-        kfree(fs_info->dev_root);
-        kfree(fs_info->csum_root);
-        kfree(fs_info);
+        free_fs_info(fs_info);
 
         return 0;
 }
@@ -2735,7 +3056,8 @@ int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
         return ret;
 }
 
-int btree_lock_page_hook(struct page *page)
+static int btree_lock_page_hook(struct page *page, void *data,
+                                void (*flush_fn)(void *))
 {
         struct inode *inode = page->mapping->host;
         struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -2752,7 +3074,10 @@ int btree_lock_page_hook(struct page *page)
         if (!eb)
                 goto out;
 
-        btrfs_tree_lock(eb);
+        if (!btrfs_try_tree_write_lock(eb)) {
+                flush_fn(data);
+                btrfs_tree_lock(eb);
+        }
         btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
 
         if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
@@ -2767,7 +3092,10 @@ int btree_lock_page_hook(struct page *page)
         btrfs_tree_unlock(eb);
         free_extent_buffer(eb);
 out:
-        lock_page(page);
+        if (!trylock_page(page)) {
+                flush_fn(data);
+                lock_page(page);
+        }
         return 0;
 }
 
@@ -3123,6 +3451,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
 static struct extent_io_ops btree_extent_io_ops = {
         .write_cache_pages_lock_hook = btree_lock_page_hook,
         .readpage_end_io_hook = btree_readpage_end_io_hook,
+        .readpage_io_failed_hook = btree_io_failed_hook,
         .submit_bio_hook = btree_submit_bio_hook,
         /* note we're sharing with inode.c for the merge bio hook */
         .merge_bio_hook = btrfs_merge_bio_hook,
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index bec3ea4bd67f..c99d0a8f13fa 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -40,6 +40,8 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
                                       u32 blocksize, u64 parent_transid);
 int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize,
                          u64 parent_transid);
+int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize,
+                         int mirror_num, struct extent_buffer **eb);
 struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
                                                    u64 bytenr, u32 blocksize);
 int clean_tree_block(struct btrfs_trans_handle *trans,
@@ -83,8 +85,6 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
                              struct btrfs_fs_info *fs_info);
 int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
                        struct btrfs_root *root);
-int btree_lock_page_hook(struct page *page);
-
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 void btrfs_init_lockdep(void);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index f5be06a2462f..9879bd474632 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -23,6 +23,7 @@
 #include <linux/rcupdate.h>
 #include <linux/kthread.h>
 #include <linux/slab.h>
+#include <linux/ratelimit.h>
 #include "compat.h"
 #include "hash.h"
 #include "ctree.h"
@@ -52,6 +53,21 @@ enum {
         CHUNK_ALLOC_LIMITED = 2,
 };
 
+/*
+ * Control how reservations are dealt with.
+ *
+ * RESERVE_FREE - freeing a reservation.
+ * RESERVE_ALLOC - allocating space and we need to update bytes_may_use for
+ *   ENOSPC accounting
+ * RESERVE_ALLOC_NO_ACCOUNT - allocating space and we should not update
+ *   bytes_may_use as the ENOSPC accounting is done elsewhere
+ */
+enum {
+        RESERVE_FREE = 0,
+        RESERVE_ALLOC = 1,
+        RESERVE_ALLOC_NO_ACCOUNT = 2,
+};
+
 static int update_block_group(struct btrfs_trans_handle *trans,
                               struct btrfs_root *root,
                               u64 bytenr, u64 num_bytes, int alloc);
@@ -81,6 +97,8 @@ static int find_next_key(struct btrfs_path *path, int level,
                          struct btrfs_key *key);
 static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
                             int dump_block_groups);
+static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
+                                       u64 num_bytes, int reserve);
 
 static noinline int
 block_group_cache_done(struct btrfs_block_group_cache *cache)
@@ -104,7 +122,6 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
         if (atomic_dec_and_test(&cache->count)) {
                 WARN_ON(cache->pinned > 0);
                 WARN_ON(cache->reserved > 0);
-                WARN_ON(cache->reserved_pinned > 0);
                 kfree(cache->free_space_ctl);
                 kfree(cache);
         }
@@ -465,7 +482,8 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
          * we likely hold important locks.
          */
         if (trans && (!trans->transaction->in_commit) &&
-            (root && root != root->fs_info->tree_root)) {
+            (root && root != root->fs_info->tree_root) &&
+            btrfs_test_opt(root, SPACE_CACHE)) {
                 spin_lock(&cache->lock);
                 if (cache->cached != BTRFS_CACHE_NO) {
                         spin_unlock(&cache->lock);
@@ -1770,18 +1788,18 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 {
         int ret;
         u64 discarded_bytes = 0;
-        struct btrfs_multi_bio *multi = NULL;
+        struct btrfs_bio *bbio = NULL;
 
 
         /* Tell the block device(s) that the sectors can be discarded */
         ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD,
-                              bytenr, &num_bytes, &multi, 0);
+                              bytenr, &num_bytes, &bbio, 0);
         if (!ret) {
-                struct btrfs_bio_stripe *stripe = multi->stripes;
+                struct btrfs_bio_stripe *stripe = bbio->stripes;
                 int i;
 
 
-                for (i = 0; i < multi->num_stripes; i++, stripe++) {
+                for (i = 0; i < bbio->num_stripes; i++, stripe++) {
                         if (!stripe->dev->can_discard)
                                 continue;
 
@@ -1800,7 +1818,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
                          */
                         ret = 0;
                 }
-                kfree(multi);
+                kfree(bbio);
         }
 
         if (actual_bytes)
@@ -2700,6 +2718,13 @@ again:
                 goto again;
         }
 
+        /* We've already setup this transaction, go ahead and exit */
+        if (block_group->cache_generation == trans->transid &&
+            i_size_read(inode)) {
+                dcs = BTRFS_DC_SETUP;
+                goto out_put;
+        }
+
         /*
          * We want to set the generation to 0, that way if anything goes wrong
          * from here on out we know not to trust this cache when we load up next
@@ -2749,12 +2774,15 @@ again:
         if (!ret)
                 dcs = BTRFS_DC_SETUP;
         btrfs_free_reserved_data_space(inode, num_pages);
+
 out_put:
         iput(inode);
 out_free:
         btrfs_release_path(path);
 out:
         spin_lock(&block_group->lock);
+        if (!ret)
+                block_group->cache_generation = trans->transid;
         block_group->disk_cache_state = dcs;
         spin_unlock(&block_group->lock);
 
@@ -3122,16 +3150,13 @@ commit_trans:
                 return -ENOSPC;
         }
         data_sinfo->bytes_may_use += bytes;
-        BTRFS_I(inode)->reserved_bytes += bytes;
         spin_unlock(&data_sinfo->lock);
 
         return 0;
 }
 
 /*
- * called when we are clearing an delalloc extent from the
- * inode's io_tree or there was an error for whatever reason
- * after calling btrfs_check_data_free_space
+ * Called if we need to clear a data reservation for this inode.
  */
 void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
 {
@@ -3144,7 +3169,6 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
         data_sinfo = BTRFS_I(inode)->space_info;
         spin_lock(&data_sinfo->lock);
         data_sinfo->bytes_may_use -= bytes;
-        BTRFS_I(inode)->reserved_bytes -= bytes;
         spin_unlock(&data_sinfo->lock);
 }
 
@@ -3165,6 +3189,7 @@ static int should_alloc_chunk(struct btrfs_root *root,
                               struct btrfs_space_info *sinfo, u64 alloc_bytes,
                               int force)
 {
+        struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
         u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly;
         u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved;
         u64 thresh;
@@ -3173,11 +3198,18 @@ static int should_alloc_chunk(struct btrfs_root *root,
                 return 1;
 
         /*
+         * We need to take into account the global rsv because for all intents
+         * and purposes it's used space.  Don't worry about locking the
+         * global_rsv, it doesn't change except when the transaction commits.
+         */
+        num_allocated += global_rsv->size;
+
+        /*
          * in limited mode, we want to have some free space up to
          * about 1% of the FS size.
          */
         if (force == CHUNK_ALLOC_LIMITED) {
-                thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
+                thresh = btrfs_super_total_bytes(root->fs_info->super_copy);
                 thresh = max_t(u64, 64 * 1024 * 1024,
                                div_factor_fine(thresh, 1));
 
@@ -3199,7 +3231,7 @@ static int should_alloc_chunk(struct btrfs_root *root,
         if (num_allocated + alloc_bytes < div_factor(num_bytes, 8))
                 return 0;
 
-        thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
+        thresh = btrfs_super_total_bytes(root->fs_info->super_copy);
 
         /* 256MB or 5% of the FS */
         thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5));
@@ -3302,24 +3334,26 @@ out:
 /*
  * shrink metadata reservation for delalloc
  */
-static int shrink_delalloc(struct btrfs_trans_handle *trans,
-                           struct btrfs_root *root, u64 to_reclaim, int sync)
+static int shrink_delalloc(struct btrfs_root *root, u64 to_reclaim,
+                           bool wait_ordered)
 {
         struct btrfs_block_rsv *block_rsv;
         struct btrfs_space_info *space_info;
+        struct btrfs_trans_handle *trans;
         u64 reserved;
         u64 max_reclaim;
         u64 reclaimed = 0;
         long time_left;
-        int nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT;
+        unsigned long nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT;
         int loops = 0;
         unsigned long progress;
 
+        trans = (struct btrfs_trans_handle *)current->journal_info;
         block_rsv = &root->fs_info->delalloc_block_rsv;
         space_info = block_rsv->space_info;
 
         smp_mb();
-        reserved = space_info->bytes_reserved;
+        reserved = space_info->bytes_may_use;
         progress = space_info->reservation_progress;
 
         if (reserved == 0)
@@ -3334,18 +3368,20 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
         }
 
         max_reclaim = min(reserved, to_reclaim);
-
+        nr_pages = max_t(unsigned long, nr_pages,
+                         max_reclaim >> PAGE_CACHE_SHIFT);
         while (loops < 1024) {
                 /* have the flusher threads jump in and do some IO */
                 smp_mb();
                 nr_pages = min_t(unsigned long, nr_pages,
                        root->fs_info->delalloc_bytes >> PAGE_CACHE_SHIFT);
-                writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages);
+                writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages,
+                                                WB_REASON_FS_FREE_SPACE);
 
                 spin_lock(&space_info->lock);
-                if (reserved > space_info->bytes_reserved)
-                        reclaimed += reserved - space_info->bytes_reserved;
-                reserved = space_info->bytes_reserved;
+                if (reserved > space_info->bytes_may_use)
+                        reclaimed += reserved - space_info->bytes_may_use;
+                reserved = space_info->bytes_may_use;
                 spin_unlock(&space_info->lock);
 
                 loops++;
@@ -3356,11 +3392,15 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
                 if (trans && trans->transaction->blocked)
                         return -EAGAIN;
 
-                time_left = schedule_timeout_interruptible(1);
+                if (wait_ordered && !trans) {
+                        btrfs_wait_ordered_extents(root, 0, 0);
+                } else {
+                        time_left = schedule_timeout_interruptible(1);
 
-                /* We were interrupted, exit */
-                if (time_left)
-                        break;
+                        /* We were interrupted, exit */
+                        if (time_left)
+                                break;
+                }
 
                 /* we've kicked the IO a few times, if anything has been freed,
                  * exit.  There is no sense in looping here for a long time
@@ -3375,34 +3415,90 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
                 }
 
         }
-        if (reclaimed >= to_reclaim && !trans)
-                btrfs_wait_ordered_extents(root, 0, 0);
+
         return reclaimed >= to_reclaim;
 }
 
-/*
- * Retries tells us how many times we've called reserve_metadata_bytes.  The
- * idea is if this is the first call (retries == 0) then we will add to our
- * reserved count if we can't make the allocation in order to hold our place
- * while we go and try and free up space.  That way for retries > 1 we don't try
- * and add space, we just check to see if the amount of unused space is >= the
- * total space, meaning that our reservation is valid.
+/**
+ * maybe_commit_transaction - possibly commit the transaction if its ok to
+ * @root - the root we're allocating for
+ * @bytes - the number of bytes we want to reserve
+ * @force - force the commit
  *
- * However if we don't intend to retry this reservation, pass -1 as retries so
- * that it short circuits this logic.
+ * This will check to make sure that committing the transaction will actually
+ * get us somewhere and then commit the transaction if it does.  Otherwise it
+ * will return -ENOSPC.
  */
-static int reserve_metadata_bytes(struct btrfs_trans_handle *trans,
-                                  struct btrfs_root *root,
+static int may_commit_transaction(struct btrfs_root *root,
+                                  struct btrfs_space_info *space_info,
+                                  u64 bytes, int force)
+{
+        struct btrfs_block_rsv *delayed_rsv = &root->fs_info->delayed_block_rsv;
+        struct btrfs_trans_handle *trans;
+
+        trans = (struct btrfs_trans_handle *)current->journal_info;
+        if (trans)
+                return -EAGAIN;
+
+        if (force)
+                goto commit;
+
+        /* See if there is enough pinned space to make this reservation */
+        spin_lock(&space_info->lock);
+        if (space_info->bytes_pinned >= bytes) {
+                spin_unlock(&space_info->lock);
+                goto commit;
+        }
+        spin_unlock(&space_info->lock);
+
+        /*
+         * See if there is some space in the delayed insertion reservation for
+         * this reservation.
+         */
+        if (space_info != delayed_rsv->space_info)
+                return -ENOSPC;
+
+        spin_lock(&delayed_rsv->lock);
+        if (delayed_rsv->size < bytes) {
+                spin_unlock(&delayed_rsv->lock);
+                return -ENOSPC;
+        }
+        spin_unlock(&delayed_rsv->lock);
+
+commit:
+        trans = btrfs_join_transaction(root);
+        if (IS_ERR(trans))
+                return -ENOSPC;
+
+        return btrfs_commit_transaction(trans, root);
+}
+
+/**
+ * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space
+ * @root - the root we're allocating for
+ * @block_rsv - the block_rsv we're allocating for
+ * @orig_bytes - the number of bytes we want
+ * @flush - wether or not we can flush to make our reservation
+ *
+ * This will reserve orgi_bytes number of bytes from the space info associated
+ * with the block_rsv.  If there is not enough space it will make an attempt to
+ * flush out space to make room.  It will do this by flushing delalloc if
+ * possible or committing the transaction.  If flush is 0 then no attempts to
+ * regain reservations will be made and this will fail if there is not enough
+ * space already.
+ */
+static int reserve_metadata_bytes(struct btrfs_root *root,
                                   struct btrfs_block_rsv *block_rsv,
                                   u64 orig_bytes, int flush)
 {
         struct btrfs_space_info *space_info = block_rsv->space_info;
-        u64 unused;
+        u64 used;
         u64 num_bytes = orig_bytes;
         int retries = 0;
         int ret = 0;
         bool committed = false;
         bool flushing = false;
+        bool wait_ordered = false;
 
 again:
         ret = 0;
@@ -3419,7 +3515,7 @@ again:
                  * deadlock since we are waiting for the flusher to finish, but
                  * hold the current transaction open.
                  */
-                if (trans)
+                if (current->journal_info)
                         return -EAGAIN;
                 ret = wait_event_interruptible(space_info->wait,
                                                !space_info->flush);
@@ -3431,9 +3527,9 @@ again:
         }
 
         ret = -ENOSPC;
-        unused = space_info->bytes_used + space_info->bytes_reserved +
-                 space_info->bytes_pinned + space_info->bytes_readonly +
-                 space_info->bytes_may_use;
+        used = space_info->bytes_used + space_info->bytes_reserved +
+                space_info->bytes_pinned + space_info->bytes_readonly +
+                space_info->bytes_may_use;
 
         /*
          * The idea here is that we've not already over-reserved the block group
@@ -3442,10 +3538,9 @@ again:
          * lets start flushing stuff first and then come back and try to make
          * our reservation.
          */
-        if (unused <= space_info->total_bytes) {
-                unused = space_info->total_bytes - unused;
-                if (unused >= num_bytes) {
-                        space_info->bytes_reserved += orig_bytes;
+        if (used <= space_info->total_bytes) {
+                if (used + orig_bytes <= space_info->total_bytes) {
+                        space_info->bytes_may_use += orig_bytes;
                         ret = 0;
                 } else {
                         /*
@@ -3461,10 +3556,64 @@ again:
                  * amount plus the amount of bytes that we need for this
                  * reservation.
                  */
-                num_bytes = unused - space_info->total_bytes +
+                wait_ordered = true;
+                num_bytes = used - space_info->total_bytes +
                         (orig_bytes * (retries + 1));
         }
 
+        if (ret) {
+                u64 profile = btrfs_get_alloc_profile(root, 0);
+                u64 avail;
+
+                /*
+                 * If we have a lot of space that's pinned, don't bother doing
+                 * the overcommit dance yet and just commit the transaction.
+                 */
+                avail = (space_info->total_bytes - space_info->bytes_used) * 8;
+                do_div(avail, 10);
+                if (space_info->bytes_pinned >= avail && flush && !committed) {
+                        space_info->flush = 1;
+                        flushing = true;
+                        spin_unlock(&space_info->lock);
+                        ret = may_commit_transaction(root, space_info,
+                                                     orig_bytes, 1);
+                        if (ret)
+                                goto out;
+                        committed = true;
+                        goto again;
+                }
+
+                spin_lock(&root->fs_info->free_chunk_lock);
+                avail = root->fs_info->free_chunk_space;
+
+                /*
+                 * If we have dup, raid1 or raid10 then only half of the free
+                 * space is actually useable.
+                 */
+                if (profile & (BTRFS_BLOCK_GROUP_DUP |
+                               BTRFS_BLOCK_GROUP_RAID1 |
+                               BTRFS_BLOCK_GROUP_RAID10))
+                        avail >>= 1;
+
+                /*
+                 * If we aren't flushing don't let us overcommit too much, say
+                 * 1/8th of the space.  If we can flush, let it overcommit up to
+                 * 1/2 of the space.
+                 */
+                if (flush)
+                        avail >>= 3;
+                else
+                        avail >>= 1;
+                 spin_unlock(&root->fs_info->free_chunk_lock);
+
+                if (used + num_bytes < space_info->total_bytes + avail) {
+                        space_info->bytes_may_use += orig_bytes;
+                        ret = 0;
+                } else {
+                        wait_ordered = true;
+                }
+        }
+
         /*
          * Couldn't make our reservation, save our place so while we're trying
          * to reclaim space we can actually use it instead of somebody else
@@ -3484,7 +3633,7 @@ again:
          * We do synchronous shrinking since we don't actually unreserve
          * metadata until after the IO is completed.
          */
-        ret = shrink_delalloc(trans, root, num_bytes, 1);
+        ret = shrink_delalloc(root, num_bytes, wait_ordered);
         if (ret < 0)
                 goto out;
 
@@ -3496,35 +3645,17 @@ again:
          * so go back around and try again.
          */
         if (retries < 2) {
+                wait_ordered = true;
                 retries++;
                 goto again;
         }
 
-        /*
-         * Not enough space to be reclaimed, don't bother committing the
-         * transaction.
-         */
-        spin_lock(&space_info->lock);
-        if (space_info->bytes_pinned < orig_bytes)
-                ret = -ENOSPC;
-        spin_unlock(&space_info->lock);
-        if (ret)
-                goto out;
-
-        ret = -EAGAIN;
-        if (trans)
-                goto out;
-
         ret = -ENOSPC;
         if (committed)
                 goto out;
 
-        trans = btrfs_join_transaction(root);
-        if (IS_ERR(trans))
-                goto out;
-        ret = btrfs_commit_transaction(trans, root);
+        ret = may_commit_transaction(root, space_info, orig_bytes, 0);
         if (!ret) {
-                trans = NULL;
                 committed = true;
                 goto again;
         }
@@ -3542,10 +3673,12 @@ out:
 static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans,
                                              struct btrfs_root *root)
 {
-        struct btrfs_block_rsv *block_rsv;
-        if (root->ref_cows)
+        struct btrfs_block_rsv *block_rsv = NULL;
+
+        if (root->ref_cows || root == root->fs_info->csum_root)
                 block_rsv = trans->block_rsv;
-        else
+
+        if (!block_rsv)
                 block_rsv = root->block_rsv;
 
         if (!block_rsv)
@@ -3616,7 +3749,7 @@ static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv,
                 }
                 if (num_bytes) {
                         spin_lock(&space_info->lock);
-                        space_info->bytes_reserved -= num_bytes;
+                        space_info->bytes_may_use -= num_bytes;
                         space_info->reservation_progress++;
                         spin_unlock(&space_info->lock);
                 }
@@ -3640,9 +3773,6 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv)
 {
         memset(rsv, 0, sizeof(*rsv));
         spin_lock_init(&rsv->lock);
-        atomic_set(&rsv->usage, 1);
-        rsv->priority = 6;
-        INIT_LIST_HEAD(&rsv->list);
 }
 
 struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root)
@@ -3663,38 +3793,38 @@ struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root)
 void btrfs_free_block_rsv(struct btrfs_root *root,
                           struct btrfs_block_rsv *rsv)
 {
-        if (rsv && atomic_dec_and_test(&rsv->usage)) {
-                btrfs_block_rsv_release(root, rsv, (u64)-1);
-                if (!rsv->durable)
-                        kfree(rsv);
-        }
+        btrfs_block_rsv_release(root, rsv, (u64)-1);
+        kfree(rsv);
 }
 
-/*
- * make the block_rsv struct be able to capture freed space.
- * the captured space will re-add to the the block_rsv struct
- * after transaction commit
- */
-void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info,
-                                 struct btrfs_block_rsv *block_rsv)
+int btrfs_block_rsv_add(struct btrfs_root *root,
+                        struct btrfs_block_rsv *block_rsv,
+                        u64 num_bytes)
 {
-        block_rsv->durable = 1;
-        mutex_lock(&fs_info->durable_block_rsv_mutex);
-        list_add_tail(&block_rsv->list, &fs_info->durable_block_rsv_list);
-        mutex_unlock(&fs_info->durable_block_rsv_mutex);
+        int ret;
+
+        if (num_bytes == 0)
+                return 0;
+
+        ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 1);
+        if (!ret) {
+                block_rsv_add_bytes(block_rsv, num_bytes, 1);
+                return 0;
+        }
+
+        return ret;
 }
 
-int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
-                        struct btrfs_root *root,
-                        struct btrfs_block_rsv *block_rsv,
-                        u64 num_bytes)
+int btrfs_block_rsv_add_noflush(struct btrfs_root *root,
+                                struct btrfs_block_rsv *block_rsv,
+                                u64 num_bytes)
 {
         int ret;
 
         if (num_bytes == 0)
                 return 0;
 
-        ret = reserve_metadata_bytes(trans, root, block_rsv, num_bytes, 1);
+        ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 0);
         if (!ret) {
                 block_rsv_add_bytes(block_rsv, num_bytes, 1);
                 return 0;
@@ -3703,55 +3833,52 @@ int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
         return ret;
 }
 
-int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
-                          struct btrfs_root *root,
-                          struct btrfs_block_rsv *block_rsv,
-                          u64 min_reserved, int min_factor)
+int btrfs_block_rsv_check(struct btrfs_root *root,
+                          struct btrfs_block_rsv *block_rsv, int min_factor)
 {
         u64 num_bytes = 0;
-        int commit_trans = 0;
         int ret = -ENOSPC;
 
         if (!block_rsv)
                 return 0;
 
         spin_lock(&block_rsv->lock);
-        if (min_factor > 0)
-                num_bytes = div_factor(block_rsv->size, min_factor);
-        if (min_reserved > num_bytes)
-                num_bytes = min_reserved;
+        num_bytes = div_factor(block_rsv->size, min_factor);
+        if (block_rsv->reserved >= num_bytes)
+                ret = 0;
+        spin_unlock(&block_rsv->lock);
 
-        if (block_rsv->reserved >= num_bytes) {
+        return ret;
+}
+
+int btrfs_block_rsv_refill(struct btrfs_root *root,
+                          struct btrfs_block_rsv *block_rsv,
+                          u64 min_reserved)
+{
+        u64 num_bytes = 0;
+        int ret = -ENOSPC;
+
+        if (!block_rsv)
+                return 0;
+
+        spin_lock(&block_rsv->lock);
+        num_bytes = min_reserved;
+        if (block_rsv->reserved >= num_bytes)
                 ret = 0;
-        } else {
+        else
                 num_bytes -= block_rsv->reserved;
-                if (block_rsv->durable &&
-                    block_rsv->freed[0] + block_rsv->freed[1] >= num_bytes)
-                        commit_trans = 1;
-        }
         spin_unlock(&block_rsv->lock);
+
         if (!ret)
                 return 0;
 
-        if (block_rsv->refill_used) {
-                ret = reserve_metadata_bytes(trans, root, block_rsv,
-                                             num_bytes, 0);
-                if (!ret) {
-                        block_rsv_add_bytes(block_rsv, num_bytes, 0);
-                        return 0;
-                }
-        }
-
-        if (commit_trans) {
-                if (trans)
-                        return -EAGAIN;
-                trans = btrfs_join_transaction(root);
-                BUG_ON(IS_ERR(trans));
-                ret = btrfs_commit_transaction(trans, root);
+        ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 1);
+        if (!ret) {
+                block_rsv_add_bytes(block_rsv, num_bytes, 0);
                 return 0;
         }
 
-        return -ENOSPC;
+        return ret;
 }
 
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
@@ -3783,7 +3910,7 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info)
         u64 num_bytes;
         u64 meta_used;
         u64 data_used;
-        int csum_size = btrfs_super_csum_size(&fs_info->super_copy);
+        int csum_size = btrfs_super_csum_size(fs_info->super_copy);
 
         sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
         spin_lock(&sinfo->lock);
@@ -3827,12 +3954,12 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
         if (sinfo->total_bytes > num_bytes) {
                 num_bytes = sinfo->total_bytes - num_bytes;
                 block_rsv->reserved += num_bytes;
-                sinfo->bytes_reserved += num_bytes;
+                sinfo->bytes_may_use += num_bytes;
         }
 
         if (block_rsv->reserved >= block_rsv->size) {
                 num_bytes = block_rsv->reserved - block_rsv->size;
-                sinfo->bytes_reserved -= num_bytes;
+                sinfo->bytes_may_use -= num_bytes;
                 sinfo->reservation_progress++;
                 block_rsv->reserved = block_rsv->size;
                 block_rsv->full = 1;
@@ -3848,16 +3975,13 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
 
         space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM);
         fs_info->chunk_block_rsv.space_info = space_info;
-        fs_info->chunk_block_rsv.priority = 10;
 
         space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA);
         fs_info->global_block_rsv.space_info = space_info;
-        fs_info->global_block_rsv.priority = 10;
-        fs_info->global_block_rsv.refill_used = 1;
         fs_info->delalloc_block_rsv.space_info = space_info;
         fs_info->trans_block_rsv.space_info = space_info;
         fs_info->empty_block_rsv.space_info = space_info;
-        fs_info->empty_block_rsv.priority = 10;
+        fs_info->delayed_block_rsv.space_info = space_info;
 
         fs_info->extent_root->block_rsv = &fs_info->global_block_rsv;
         fs_info->csum_root->block_rsv = &fs_info->global_block_rsv;
@@ -3865,10 +3989,6 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
         fs_info->tree_root->block_rsv = &fs_info->global_block_rsv;
         fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv;
 
-        btrfs_add_durable_block_rsv(fs_info, &fs_info->global_block_rsv);
-
-        btrfs_add_durable_block_rsv(fs_info, &fs_info->delalloc_block_rsv);
-
         update_global_block_rsv(fs_info);
 }
 
@@ -3881,37 +4001,8 @@ static void release_global_block_rsv(struct btrfs_fs_info *fs_info)
         WARN_ON(fs_info->trans_block_rsv.reserved > 0);
         WARN_ON(fs_info->chunk_block_rsv.size > 0);
         WARN_ON(fs_info->chunk_block_rsv.reserved > 0);
-}
-
-int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans,
-                                    struct btrfs_root *root,
-                                    struct btrfs_block_rsv *rsv)
-{
-        struct btrfs_block_rsv *trans_rsv = &root->fs_info->trans_block_rsv;
-        u64 num_bytes;
-        int ret;
-
-        /*
-         * Truncate should be freeing data, but give us 2 items just in case it
-         * needs to use some space.  We may want to be smarter about this in the
-         * future.
-         */
-        num_bytes = btrfs_calc_trans_metadata_size(root, 2);
-
-        /* We already have enough bytes, just return */
-        if (rsv->reserved >= num_bytes)
-                return 0;
-
-        num_bytes -= rsv->reserved;
-
-        /*
-         * You should have reserved enough space before hand to do this, so this
-         * should not fail.
-         */
-        ret = block_rsv_migrate_bytes(trans_rsv, rsv, num_bytes);
-        BUG_ON(ret);
-
-        return 0;
+        WARN_ON(fs_info->delayed_block_rsv.size > 0);
+        WARN_ON(fs_info->delayed_block_rsv.reserved > 0);
 }
 
 void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
@@ -3920,9 +4011,7 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
         if (!trans->bytes_reserved)
                 return;
 
-        BUG_ON(trans->block_rsv != &root->fs_info->trans_block_rsv);
-        btrfs_block_rsv_release(root, trans->block_rsv,
-                                trans->bytes_reserved);
+        btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved);
         trans->bytes_reserved = 0;
 }
 
@@ -3964,11 +4053,19 @@ int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans,
         return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
 }
 
+/**
+ * drop_outstanding_extent - drop an outstanding extent
+ * @inode: the inode we're dropping the extent for
+ *
+ * This is called when we are freeing up an outstanding extent, either called
+ * after an error or after an extent is written.  This will return the number of
+ * reserved extents that need to be freed.  This must be called with
+ * BTRFS_I(inode)->lock held.
+ */
 static unsigned drop_outstanding_extent(struct inode *inode)
 {
         unsigned dropped_extents = 0;
 
-        spin_lock(&BTRFS_I(inode)->lock);
         BUG_ON(!BTRFS_I(inode)->outstanding_extents);
         BTRFS_I(inode)->outstanding_extents--;
 
@@ -3978,19 +4075,70 @@ static unsigned drop_outstanding_extent(struct inode *inode)
          */
         if (BTRFS_I(inode)->outstanding_extents >=
             BTRFS_I(inode)->reserved_extents)
-                goto out;
+                return 0;
 
         dropped_extents = BTRFS_I(inode)->reserved_extents -
                 BTRFS_I(inode)->outstanding_extents;
         BTRFS_I(inode)->reserved_extents -= dropped_extents;
-out:
-        spin_unlock(&BTRFS_I(inode)->lock);
         return dropped_extents;
 }
 
-static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes)
+/**
+ * calc_csum_metadata_size - return the amount of metada space that must be
+ *      reserved/free'd for the given bytes.
+ * @inode: the inode we're manipulating
+ * @num_bytes: the number of bytes in question
+ * @reserve: 1 if we are reserving space, 0 if we are freeing space
+ *
+ * This adjusts the number of csum_bytes in the inode and then returns the
+ * correct amount of metadata that must either be reserved or freed.  We
+ * calculate how many checksums we can fit into one leaf and then divide the
+ * number of bytes that will need to be checksumed by this value to figure out
+ * how many checksums will be required.  If we are adding bytes then the number
+ * may go up and we will return the number of additional bytes that must be
+ * reserved.  If it is going down we will return the number of bytes that must
+ * be freed.
+ *
+ * This must be called with BTRFS_I(inode)->lock held.
+ */
+static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes,
+                                   int reserve)
 {
-        return num_bytes >>= 3;
+        struct btrfs_root *root = BTRFS_I(inode)->root;
+        u64 csum_size;
+        int num_csums_per_leaf;
+        int num_csums;
+        int old_csums;
+
+        if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM &&
+            BTRFS_I(inode)->csum_bytes == 0)
+                return 0;
+
+        old_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize);
+        if (reserve)
+                BTRFS_I(inode)->csum_bytes += num_bytes;
+        else
+                BTRFS_I(inode)->csum_bytes -= num_bytes;
+        csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item);
+        num_csums_per_leaf = (int)div64_u64(csum_size,
+                                            sizeof(struct btrfs_csum_item) +
+                                            sizeof(struct btrfs_disk_key));
+        num_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize);
+        num_csums = num_csums + num_csums_per_leaf - 1;
+        num_csums = num_csums / num_csums_per_leaf;
+
+        old_csums = old_csums + num_csums_per_leaf - 1;
+        old_csums = old_csums / num_csums_per_leaf;
+
+        /* No change, no need to reserve more */
+        if (old_csums == num_csums)
+                return 0;
+
+        if (reserve)
+                return btrfs_calc_trans_metadata_size(root,
+                                                      num_csums - old_csums);
+
+        return btrfs_calc_trans_metadata_size(root, old_csums - num_csums);
 }
 
 int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
@@ -3999,9 +4147,13 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
         struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv;
         u64 to_reserve = 0;
         unsigned nr_extents = 0;
+        int flush = 1;
         int ret;
 
-        if (btrfs_transaction_in_commit(root->fs_info))
+        if (btrfs_is_free_space_inode(root, inode))
+                flush = 0;
+
+        if (flush && btrfs_transaction_in_commit(root->fs_info))
                 schedule_timeout(1);
 
         num_bytes = ALIGN(num_bytes, root->sectorsize);
@@ -4017,18 +4169,29 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
 
                 to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents);
         }
+        to_reserve += calc_csum_metadata_size(inode, num_bytes, 1);
         spin_unlock(&BTRFS_I(inode)->lock);
 
-        to_reserve += calc_csum_metadata_size(inode, num_bytes);
-        ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1);
+        ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush);
         if (ret) {
+                u64 to_free = 0;
                 unsigned dropped;
+
+                spin_lock(&BTRFS_I(inode)->lock);
+                dropped = drop_outstanding_extent(inode);
+                to_free = calc_csum_metadata_size(inode, num_bytes, 0);
+                spin_unlock(&BTRFS_I(inode)->lock);
+                to_free += btrfs_calc_trans_metadata_size(root, dropped);
+
                 /*
-                 * We don't need the return value since our reservation failed,
-                 * we just need to clean up our counter.
+                 * Somebody could have come in and twiddled with the
+                 * reservation, so if we have to free more than we would have
+                 * reserved from this reservation go ahead and release those
+                 * bytes.
                  */
-                dropped = drop_outstanding_extent(inode);
-                WARN_ON(dropped > 1);
+                to_free -= to_reserve;
+                if (to_free)
+                        btrfs_block_rsv_release(root, block_rsv, to_free);
                 return ret;
         }
 
@@ -4037,6 +4200,15 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
         return 0;
 }
 
+/**
+ * btrfs_delalloc_release_metadata - release a metadata reservation for an inode
+ * @inode: the inode to release the reservation for
+ * @num_bytes: the number of bytes we're releasing
+ *
+ * This will release the metadata reservation for an inode.  This can be called
+ * once we complete IO for a given set of bytes to release their metadata
+ * reservations.
+ */
 void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
 {
         struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -4044,9 +4216,11 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
         unsigned dropped;
 
         num_bytes = ALIGN(num_bytes, root->sectorsize);
+        spin_lock(&BTRFS_I(inode)->lock);
         dropped = drop_outstanding_extent(inode);
 
-        to_free = calc_csum_metadata_size(inode, num_bytes);
+        to_free = calc_csum_metadata_size(inode, num_bytes, 0);
+        spin_unlock(&BTRFS_I(inode)->lock);
         if (dropped > 0)
                 to_free += btrfs_calc_trans_metadata_size(root, dropped);
 
@@ -4054,6 +4228,21 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
                                 to_free);
 }
 
+/**
+ * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc
+ * @inode: inode we're writing to
+ * @num_bytes: the number of bytes we want to allocate
+ *
+ * This will do the following things
+ *
+ * o reserve space in the data space info for num_bytes
+ * o reserve space in the metadata space info based on number of outstanding
+ *   extents and how much csums will be needed
+ * o add to the inodes ->delalloc_bytes
+ * o add it to the fs_info's delalloc inodes list.
+ *
+ * This will return 0 for success and -ENOSPC if there is no space left.
+ */
 int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
 {
         int ret;
@@ -4071,6 +4260,19 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
         return 0;
 }
 
+/**
+ * btrfs_delalloc_release_space - release data and metadata space for delalloc
+ * @inode: inode we're releasing space for
+ * @num_bytes: the number of bytes we want to free up
+ *
+ * This must be matched with a call to btrfs_delalloc_reserve_space.  This is
+ * called in the case that we don't need the metadata AND data reservations
+ * anymore.  So if there is an error or we insert an inline extent.
+ *
+ * This function will release the metadata space that was not used and will
+ * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes
+ * list if there are no delalloc bytes left.
+ */
 void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes)
 {
         btrfs_delalloc_release_metadata(inode, num_bytes);
@@ -4090,12 +4292,12 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 
         /* block accounting for super block */
         spin_lock(&info->delalloc_lock);
-        old_val = btrfs_super_bytes_used(&info->super_copy);
+        old_val = btrfs_super_bytes_used(info->super_copy);
         if (alloc)
                 old_val += num_bytes;
         else
                 old_val -= num_bytes;
-        btrfs_set_super_bytes_used(&info->super_copy, old_val);
+        btrfs_set_super_bytes_used(info->super_copy, old_val);
         spin_unlock(&info->delalloc_lock);
 
         while (total) {
@@ -4123,7 +4325,7 @@ static int update_block_group(struct btrfs_trans_handle *trans,
                 spin_lock(&cache->space_info->lock);
                 spin_lock(&cache->lock);
 
-                if (btrfs_super_cache_generation(&info->super_copy) != 0 &&
+                if (btrfs_test_opt(root, SPACE_CACHE) &&
                     cache->disk_cache_state < BTRFS_DC_CLEAR)
                         cache->disk_cache_state = BTRFS_DC_CLEAR;
 
@@ -4135,7 +4337,6 @@ static int update_block_group(struct btrfs_trans_handle *trans,
                         btrfs_set_block_group_used(&cache->item, old_val);
                         cache->reserved -= num_bytes;
                         cache->space_info->bytes_reserved -= num_bytes;
-                        cache->space_info->reservation_progress++;
                         cache->space_info->bytes_used += num_bytes;
                         cache->space_info->disk_used += num_bytes * factor;
                         spin_unlock(&cache->lock);
@@ -4187,7 +4388,6 @@ static int pin_down_extent(struct btrfs_root *root,
         if (reserved) {
                 cache->reserved -= num_bytes;
                 cache->space_info->bytes_reserved -= num_bytes;
-                cache->space_info->reservation_progress++;
         }
         spin_unlock(&cache->lock);
         spin_unlock(&cache->space_info->lock);
@@ -4215,45 +4415,82 @@ int btrfs_pin_extent(struct btrfs_root *root,
 }
 
 /*
- * update size of reserved extents. this function may return -EAGAIN
- * if 'reserve' is true or 'sinfo' is false.
+ * this function must be called within transaction
+ */
+int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans,
+                                    struct btrfs_root *root,
+                                    u64 bytenr, u64 num_bytes)
+{
+        struct btrfs_block_group_cache *cache;
+
+        cache = btrfs_lookup_block_group(root->fs_info, bytenr);
+        BUG_ON(!cache);
+
+        /*
+         * pull in the free space cache (if any) so that our pin
+         * removes the free space from the cache.  We have load_only set
+         * to one because the slow code to read in the free extents does check
+         * the pinned extents.
+         */
+        cache_block_group(cache, trans, root, 1);
+
+        pin_down_extent(root, cache, bytenr, num_bytes, 0);
+
+        /* remove us from the free space cache (if we're there at all) */
+        btrfs_remove_free_space(cache, bytenr, num_bytes);
+        btrfs_put_block_group(cache);
+        return 0;
+}
+
+/**
+ * btrfs_update_reserved_bytes - update the block_group and space info counters
+ * @cache:      The cache we are manipulating
+ * @num_bytes:  The number of bytes in question
+ * @reserve:    One of the reservation enums
+ *
+ * This is called by the allocator when it reserves space, or by somebody who is
+ * freeing space that was never actually used on disk.  For example if you
+ * reserve some space for a new leaf in transaction A and before transaction A
+ * commits you free that leaf, you call this with reserve set to 0 in order to
+ * clear the reservation.
+ *
+ * Metadata reservations should be called with RESERVE_ALLOC so we do the proper
+ * ENOSPC accounting.  For data we handle the reservation through clearing the
+ * delalloc bits in the io_tree.  We have to do this since we could end up
+ * allocating less disk space for the amount of data we have reserved in the
+ * case of compression.
+ *
+ * If this is a reservation and the block group has become read only we cannot
+ * make the reservation and return -EAGAIN, otherwise this function always
+ * succeeds.
  */
-int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
-                                u64 num_bytes, int reserve, int sinfo)
+static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
+                                       u64 num_bytes, int reserve)
 {
+        struct btrfs_space_info *space_info = cache->space_info;
         int ret = 0;
-        if (sinfo) {
-                struct btrfs_space_info *space_info = cache->space_info;
-                spin_lock(&space_info->lock);
-                spin_lock(&cache->lock);
-                if (reserve) {
-                        if (cache->ro) {
-                                ret = -EAGAIN;
-                        } else {
-                                cache->reserved += num_bytes;
-                                space_info->bytes_reserved += num_bytes;
-                        }
-                } else {
-                        if (cache->ro)
-                                space_info->bytes_readonly += num_bytes;
-                        cache->reserved -= num_bytes;
-                        space_info->bytes_reserved -= num_bytes;
-                        space_info->reservation_progress++;
-                }
-                spin_unlock(&cache->lock);
-                spin_unlock(&space_info->lock);
-        } else {
-                spin_lock(&cache->lock);
+        spin_lock(&space_info->lock);
+        spin_lock(&cache->lock);
+        if (reserve != RESERVE_FREE) {
                 if (cache->ro) {
                         ret = -EAGAIN;
                 } else {
-                        if (reserve)
-                                cache->reserved += num_bytes;
-                        else
-                                cache->reserved -= num_bytes;
+                        cache->reserved += num_bytes;
+                        space_info->bytes_reserved += num_bytes;
+                        if (reserve == RESERVE_ALLOC) {
+                                BUG_ON(space_info->bytes_may_use < num_bytes);
+                                space_info->bytes_may_use -= num_bytes;
+                        }
                 }
-                spin_unlock(&cache->lock);
+        } else {
+                if (cache->ro)
+                        space_info->bytes_readonly += num_bytes;
+                cache->reserved -= num_bytes;
+                space_info->bytes_reserved -= num_bytes;
+                space_info->reservation_progress++;
         }
+        spin_unlock(&cache->lock);
+        spin_unlock(&space_info->lock);
         return ret;
 }
 
@@ -4319,13 +4556,8 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
                 spin_lock(&cache->lock);
                 cache->pinned -= len;
                 cache->space_info->bytes_pinned -= len;
-                if (cache->ro) {
+                if (cache->ro)
                         cache->space_info->bytes_readonly += len;
-                } else if (cache->reserved_pinned > 0) {
-                        len = min(len, cache->reserved_pinned);
-                        cache->reserved_pinned -= len;
-                        cache->space_info->bytes_reserved += len;
-                }
                 spin_unlock(&cache->lock);
                 spin_unlock(&cache->space_info->lock);
         }
@@ -4340,11 +4572,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 {
         struct btrfs_fs_info *fs_info = root->fs_info;
         struct extent_io_tree *unpin;
-        struct btrfs_block_rsv *block_rsv;
-        struct btrfs_block_rsv *next_rsv;
         u64 start;
         u64 end;
-        int idx;
         int ret;
 
         if (fs_info->pinned_extents == &fs_info->freed_extents[0])
@@ -4367,30 +4596,6 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
                 cond_resched();
         }
 
-        mutex_lock(&fs_info->durable_block_rsv_mutex);
-        list_for_each_entry_safe(block_rsv, next_rsv,
-                                 &fs_info->durable_block_rsv_list, list) {
-
-                idx = trans->transid & 0x1;
-                if (block_rsv->freed[idx] > 0) {
-                        block_rsv_add_bytes(block_rsv,
-                                            block_rsv->freed[idx], 0);
-                        block_rsv->freed[idx] = 0;
-                }
-                if (atomic_read(&block_rsv->usage) == 0) {
-                        btrfs_block_rsv_release(root, block_rsv, (u64)-1);
-
-                        if (block_rsv->freed[0] == 0 &&
-                            block_rsv->freed[1] == 0) {
-                                list_del_init(&block_rsv->list);
-                                kfree(block_rsv);
-                        }
-                } else {
-                        btrfs_block_rsv_release(root, block_rsv, 0);
-                }
-        }
-        mutex_unlock(&fs_info->durable_block_rsv_mutex);
-
         return 0;
 }
 
@@ -4668,7 +4873,6 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
                            struct extent_buffer *buf,
                            u64 parent, int last_ref)
 {
-        struct btrfs_block_rsv *block_rsv;
         struct btrfs_block_group_cache *cache = NULL;
         int ret;
 
@@ -4683,64 +4887,24 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
         if (!last_ref)
                 return;
 
-        block_rsv = get_block_rsv(trans, root);
         cache = btrfs_lookup_block_group(root->fs_info, buf->start);
-        if (block_rsv->space_info != cache->space_info)
-                goto out;
 
         if (btrfs_header_generation(buf) == trans->transid) {
                 if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
                         ret = check_ref_cleanup(trans, root, buf->start);
                         if (!ret)
-                                goto pin;
+                                goto out;
                 }
 
                 if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
                         pin_down_extent(root, cache, buf->start, buf->len, 1);
-                        goto pin;
+                        goto out;
                 }
 
                 WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags));
 
                 btrfs_add_free_space(cache, buf->start, buf->len);
-                ret = btrfs_update_reserved_bytes(cache, buf->len, 0, 0);
-                if (ret == -EAGAIN) {
-                        /* block group became read-only */
-                        btrfs_update_reserved_bytes(cache, buf->len, 0, 1);
-                        goto out;
-                }
-
-                ret = 1;
-                spin_lock(&block_rsv->lock);
-                if (block_rsv->reserved < block_rsv->size) {
-                        block_rsv->reserved += buf->len;
-                        ret = 0;
-                }
-                spin_unlock(&block_rsv->lock);
-
-                if (ret) {
-                        spin_lock(&cache->space_info->lock);
-                        cache->space_info->bytes_reserved -= buf->len;
-                        cache->space_info->reservation_progress++;
-                        spin_unlock(&cache->space_info->lock);
-                }
-                goto out;
-        }
-pin:
-        if (block_rsv->durable && !cache->ro) {
-                ret = 0;
-                spin_lock(&cache->lock);
-                if (!cache->ro) {
-                        cache->reserved_pinned += buf->len;
-                        ret = 1;
-                }
-                spin_unlock(&cache->lock);
-
-                if (ret) {
-                        spin_lock(&block_rsv->lock);
-                        block_rsv->freed[trans->transid & 0x1] += buf->len;
-                        spin_unlock(&block_rsv->lock);
-                }
+                btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE);
         }
 out:
         /*
@@ -4883,10 +5047,13 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
         int last_ptr_loop = 0;
         int loop = 0;
         int index = 0;
+        int alloc_type = (data & BTRFS_BLOCK_GROUP_DATA) ?
+                RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC;
         bool found_uncached_bg = false;
         bool failed_cluster_refill = false;
         bool failed_alloc = false;
         bool use_cluster = true;
+        bool have_caching_bg = false;
         u64 ideal_cache_percent = 0;
         u64 ideal_cache_offset = 0;
 
@@ -4969,6 +5136,7 @@ ideal_cache:
                 }
         }
 search:
+        have_caching_bg = false;
         down_read(&space_info->groups_sem);
         list_for_each_entry(block_group, &space_info->block_groups[index],
                             list) {
@@ -5177,6 +5345,8 @@ refill_cluster:
                         failed_alloc = true;
                         goto have_block_group;
                 } else if (!offset) {
+                        if (!cached)
+                                have_caching_bg = true;
                         goto loop;
                 }
 checks:
@@ -5202,8 +5372,8 @@ checks:
                                              search_start - offset);
                 BUG_ON(offset > search_start);
 
-                ret = btrfs_update_reserved_bytes(block_group, num_bytes, 1,
-                                            (data & BTRFS_BLOCK_GROUP_DATA));
+                ret = btrfs_update_reserved_bytes(block_group, num_bytes,
+                                                  alloc_type);
                 if (ret == -EAGAIN) {
                         btrfs_add_free_space(block_group, offset, num_bytes);
                         goto loop;
@@ -5227,6 +5397,9 @@ loop:
         }
         up_read(&space_info->groups_sem);
 
+        if (!ins->objectid && loop >= LOOP_CACHING_WAIT && have_caching_bg)
+                goto search;
+
         if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES)
                 goto search;
 
@@ -5325,7 +5498,8 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
         int index = 0;
 
         spin_lock(&info->lock);
-        printk(KERN_INFO "space_info has %llu free, is %sfull\n",
+        printk(KERN_INFO "space_info %llu has %llu free, is %sfull\n",
+               (unsigned long long)info->flags,
                (unsigned long long)(info->total_bytes - info->bytes_used -
                                     info->bytes_pinned - info->bytes_reserved -
                                     info->bytes_readonly),
@@ -5411,7 +5585,8 @@ again:
         return ret;
 }
 
-int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
+static int __btrfs_free_reserved_extent(struct btrfs_root *root,
+                                        u64 start, u64 len, int pin)
 {
         struct btrfs_block_group_cache *cache;
         int ret = 0;
@@ -5426,8 +5601,12 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
         if (btrfs_test_opt(root, DISCARD))
                 ret = btrfs_discard_extent(root, start, len, NULL);
 
-        btrfs_add_free_space(cache, start, len);
-        btrfs_update_reserved_bytes(cache, len, 0, 1);
+        if (pin)
+                pin_down_extent(root, cache, start, len, 1);
+        else {
+                btrfs_add_free_space(cache, start, len);
+                btrfs_update_reserved_bytes(cache, len, RESERVE_FREE);
+        }
         btrfs_put_block_group(cache);
 
         trace_btrfs_reserved_extent_free(root, start, len);
@@ -5435,6 +5614,18 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
         return ret;
 }
 
+int btrfs_free_reserved_extent(struct btrfs_root *root,
+                                        u64 start, u64 len)
+{
+        return __btrfs_free_reserved_extent(root, start, len, 0);
+}
+
+int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root,
+                                       u64 start, u64 len)
+{
+        return __btrfs_free_reserved_extent(root, start, len, 1);
+}
+
 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
                                       struct btrfs_root *root,
                                       u64 parent, u64 root_objectid,
@@ -5630,7 +5821,8 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
                 put_caching_control(caching_ctl);
         }
 
-        ret = btrfs_update_reserved_bytes(block_group, ins->offset, 1, 1);
+        ret = btrfs_update_reserved_bytes(block_group, ins->offset,
+                                          RESERVE_ALLOC_NO_ACCOUNT);
         BUG_ON(ret);
         btrfs_put_block_group(block_group);
         ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
@@ -5687,8 +5879,7 @@ use_block_rsv(struct btrfs_trans_handle *trans,
         block_rsv = get_block_rsv(trans, root);
 
         if (block_rsv->size == 0) {
-                ret = reserve_metadata_bytes(trans, root, block_rsv,
-                                             blocksize, 0);
+                ret = reserve_metadata_bytes(root, block_rsv, blocksize, 0);
                 /*
                  * If we couldn't reserve metadata bytes try and use some from
                  * the global reserve.
@@ -5708,13 +5899,15 @@ use_block_rsv(struct btrfs_trans_handle *trans,
         if (!ret)
                 return block_rsv;
         if (ret) {
-                WARN_ON(1);
-                ret = reserve_metadata_bytes(trans, root, block_rsv, blocksize,
-                                             0);
+                static DEFINE_RATELIMIT_STATE(_rs,
+                                DEFAULT_RATELIMIT_INTERVAL,
+                                /*DEFAULT_RATELIMIT_BURST*/ 2);
+                if (__ratelimit(&_rs)) {
+                        printk(KERN_DEBUG "btrfs: block rsv returned %d\n", ret);
+                        WARN_ON(1);
+                }
+                ret = reserve_metadata_bytes(root, block_rsv, blocksize, 0);
                 if (!ret) {
-                        spin_lock(&block_rsv->lock);
-                        block_rsv->size += blocksize;
-                        spin_unlock(&block_rsv->lock);
                         return block_rsv;
                 } else if (ret && block_rsv != global_rsv) {
                         ret = block_rsv_use_bytes(global_rsv, blocksize);
@@ -6592,12 +6785,9 @@ static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force)
                     cache->bytes_super - btrfs_block_group_used(&cache->item);
 
         if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned +
-            sinfo->bytes_may_use + sinfo->bytes_readonly +
-            cache->reserved_pinned + num_bytes + min_allocable_bytes <=
-            sinfo->total_bytes) {
+            sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes +
+            min_allocable_bytes <= sinfo->total_bytes) {
                 sinfo->bytes_readonly += num_bytes;
-                sinfo->bytes_reserved += cache->reserved_pinned;
-                cache->reserved_pinned = 0;
                 cache->ro = 1;
                 ret = 0;
         }
@@ -6964,7 +7154,8 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
                                         struct btrfs_space_info,
                                         list);
                 if (space_info->bytes_pinned > 0 ||
-                    space_info->bytes_reserved > 0) {
+                    space_info->bytes_reserved > 0 ||
+                    space_info->bytes_may_use > 0) {
                         WARN_ON(1);
                         dump_space_info(space_info, 0, 0);
                 }
@@ -7006,14 +7197,12 @@ int btrfs_read_block_groups(struct btrfs_root *root)
                 return -ENOMEM;
         path->reada = 1;
 
-        cache_gen = btrfs_super_cache_generation(&root->fs_info->super_copy);
-        if (cache_gen != 0 &&
-            btrfs_super_generation(&root->fs_info->super_copy) != cache_gen)
+        cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy);
+        if (btrfs_test_opt(root, SPACE_CACHE) &&
+            btrfs_super_generation(root->fs_info->super_copy) != cache_gen)
                 need_clear = 1;
         if (btrfs_test_opt(root, CLEAR_CACHE))
                 need_clear = 1;
-        if (!btrfs_test_opt(root, SPACE_CACHE) && cache_gen)
-                printk(KERN_INFO "btrfs: disk space caching is enabled\n");
 
         while (1) {
                 ret = find_first_block_group(root, path, &key);
@@ -7252,7 +7441,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
                 goto out;
         }
 
-        inode = lookup_free_space_inode(root, block_group, path);
+        inode = lookup_free_space_inode(tree_root, block_group, path);
         if (!IS_ERR(inode)) {
                 ret = btrfs_orphan_add(trans, inode);
                 BUG_ON(ret);
@@ -7268,7 +7457,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
                         spin_unlock(&block_group->lock);
                 }
                 /* One for our lookup ref */
-                iput(inode);
+                btrfs_add_delayed_iput(inode);
         }
 
         key.objectid = BTRFS_FREE_SPACE_OBJECTID;
@@ -7339,7 +7528,7 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info)
         int mixed = 0;
         int ret;
 
-        disk_super = &fs_info->super_copy;
+        disk_super = fs_info->super_copy;
         if (!btrfs_super_root(disk_super))
                 return 1;
 
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index d418164a35f1..1f87c4d0e7a0 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -17,6 +17,7 @@
 #include "compat.h"
 #include "ctree.h"
 #include "btrfs_inode.h"
+#include "volumes.h"
 
 static struct kmem_cache *extent_state_cache;
 static struct kmem_cache *extent_buffer_cache;
@@ -894,6 +895,194 @@ search_again:
         goto again;
 }
 
+/**
+ * convert_extent - convert all bits in a given range from one bit to another
+ * @tree:       the io tree to search
+ * @start:      the start offset in bytes
+ * @end:        the end offset in bytes (inclusive)
+ * @bits:       the bits to set in this range
+ * @clear_bits: the bits to clear in this range
+ * @mask:       the allocation mask
+ *
+ * This will go through and set bits for the given range.  If any states exist
+ * already in this range they are set with the given bit and cleared of the
+ * clear_bits.  This is only meant to be used by things that are mergeable, ie
+ * converting from say DELALLOC to DIRTY.  This is not meant to be used with
+ * boundary bits like LOCK.
+ */
+int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+                       int bits, int clear_bits, gfp_t mask)
+{
+        struct extent_state *state;
+        struct extent_state *prealloc = NULL;
+        struct rb_node *node;
+        int err = 0;
+        u64 last_start;
+        u64 last_end;
+
+again:
+        if (!prealloc && (mask & __GFP_WAIT)) {
+                prealloc = alloc_extent_state(mask);
+                if (!prealloc)
+                        return -ENOMEM;
+        }
+
+        spin_lock(&tree->lock);
+        /*
+         * this search will find all the extents that end after
+         * our range starts.
+         */
+        node = tree_search(tree, start);
+        if (!node) {
+                prealloc = alloc_extent_state_atomic(prealloc);
+                if (!prealloc)
+                        return -ENOMEM;
+                err = insert_state(tree, prealloc, start, end, &bits);
+                prealloc = NULL;
+                BUG_ON(err == -EEXIST);
+                goto out;
+        }
+        state = rb_entry(node, struct extent_state, rb_node);
+hit_next:
+        last_start = state->start;
+        last_end = state->end;
+
+        /*
+         * | ---- desired range ---- |
+         * | state |
+         *
+         * Just lock what we found and keep going
+         */
+        if (state->start == start && state->end <= end) {
+                struct rb_node *next_node;
+
+                set_state_bits(tree, state, &bits);
+                clear_state_bit(tree, state, &clear_bits, 0);
+
+                merge_state(tree, state);
+                if (last_end == (u64)-1)
+                        goto out;
+
+                start = last_end + 1;
+                next_node = rb_next(&state->rb_node);
+                if (next_node && start < end && prealloc && !need_resched()) {
+                        state = rb_entry(next_node, struct extent_state,
+                                         rb_node);
+                        if (state->start == start)
+                                goto hit_next;
+                }
+                goto search_again;
+        }
+
+        /*
+         *     | ---- desired range ---- |
+         * | state |
+         *   or
+         * | ------------- state -------------- |
+         *
+         * We need to split the extent we found, and may flip bits on
+         * second half.
+         *
+         * If the extent we found extends past our
+         * range, we just split and search again.  It'll get split
+         * again the next time though.
+         *
+         * If the extent we found is inside our range, we set the
+         * desired bit on it.
+         */
+        if (state->start < start) {
+                prealloc = alloc_extent_state_atomic(prealloc);
+                if (!prealloc)
+                        return -ENOMEM;
+                err = split_state(tree, state, prealloc, start);
+                BUG_ON(err == -EEXIST);
+                prealloc = NULL;
+                if (err)
+                        goto out;
+                if (state->end <= end) {
+                        set_state_bits(tree, state, &bits);
+                        clear_state_bit(tree, state, &clear_bits, 0);
+                        merge_state(tree, state);
+                        if (last_end == (u64)-1)
+                                goto out;
+                        start = last_end + 1;
+                }
+                goto search_again;
+        }
+        /*
+         * | ---- desired range ---- |
+         *     | state | or               | state |
+         *
+         * There's a hole, we need to insert something in it and
+         * ignore the extent we found.
+         */
+        if (state->start > start) {
+                u64 this_end;
+                if (end < last_start)
+                        this_end = end;
+                else
+                        this_end = last_start - 1;
+
+                prealloc = alloc_extent_state_atomic(prealloc);
+                if (!prealloc)
+                        return -ENOMEM;
+
+                /*
+                 * Avoid to free 'prealloc' if it can be merged with
+                 * the later extent.
+                 */
+                err = insert_state(tree, prealloc, start, this_end,
+                                   &bits);
+                BUG_ON(err == -EEXIST);
+                if (err) {
+                        free_extent_state(prealloc);
+                        prealloc = NULL;
+                        goto out;
+                }
+                prealloc = NULL;
+                start = this_end + 1;
+                goto search_again;
+        }
+        /*
+         * | ---- desired range ---- |
+         *                        | state |
+         * We need to split the extent, and set the bit
+         * on the first half
+         */
+        if (state->start <= end && state->end > end) {
+                prealloc = alloc_extent_state_atomic(prealloc);
+                if (!prealloc)
+                        return -ENOMEM;
+
+                err = split_state(tree, state, prealloc, end + 1);
+                BUG_ON(err == -EEXIST);
+
+                set_state_bits(tree, prealloc, &bits);
+                clear_state_bit(tree, prealloc, &clear_bits, 0);
+
+                merge_state(tree, prealloc);
+                prealloc = NULL;
+                goto out;
+        }
+
+        goto search_again;
+
+out:
+        spin_unlock(&tree->lock);
+        if (prealloc)
+                free_extent_state(prealloc);
+
+        return err;
+
+search_again:
+        if (start > end)
+                goto out;
+        spin_unlock(&tree->lock);
+        if (mask & __GFP_WAIT)
+                cond_resched();
+        goto again;
+}
+
 /* wrappers around set/clear extent bit */
 int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
                      gfp_t mask)
@@ -919,7 +1108,7 @@ int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
                         struct extent_state **cached_state, gfp_t mask)
 {
         return set_extent_bit(tree, start, end,
-                              EXTENT_DELALLOC | EXTENT_DIRTY | EXTENT_UPTODATE,
+                              EXTENT_DELALLOC | EXTENT_UPTODATE,
                               0, NULL, cached_state, mask);
 }
 
@@ -1599,6 +1788,368 @@ static int check_page_writeback(struct extent_io_tree *tree,
         return 0;
 }
 
+/*
+ * When IO fails, either with EIO or csum verification fails, we
+ * try other mirrors that might have a good copy of the data.  This
+ * io_failure_record is used to record state as we go through all the
+ * mirrors.  If another mirror has good data, the page is set up to date
+ * and things continue.  If a good mirror can't be found, the original
+ * bio end_io callback is called to indicate things have failed.
+ */
+struct io_failure_record {
+        struct page *page;
+        u64 start;
+        u64 len;
+        u64 logical;
+        unsigned long bio_flags;
+        int this_mirror;
+        int failed_mirror;
+        int in_validation;
+};
+
+static int free_io_failure(struct inode *inode, struct io_failure_record *rec,
+                                int did_repair)
+{
+        int ret;
+        int err = 0;
+        struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree;
+
+        set_state_private(failure_tree, rec->start, 0);
+        ret = clear_extent_bits(failure_tree, rec->start,
+                                rec->start + rec->len - 1,
+                                EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS);
+        if (ret)
+                err = ret;
+
+        if (did_repair) {
+                ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start,
+                                        rec->start + rec->len - 1,
+                                        EXTENT_DAMAGED, GFP_NOFS);
+                if (ret && !err)
+                        err = ret;
+        }
+
+        kfree(rec);
+        return err;
+}
+
+static void repair_io_failure_callback(struct bio *bio, int err)
+{
+        complete(bio->bi_private);
+}
+
+/*
+ * this bypasses the standard btrfs submit functions deliberately, as
+ * the standard behavior is to write all copies in a raid setup. here we only
+ * want to write the one bad copy. so we do the mapping for ourselves and issue
+ * submit_bio directly.
+ * to avoid any synchonization issues, wait for the data after writing, which
+ * actually prevents the read that triggered the error from finishing.
+ * currently, there can be no more than two copies of every data bit. thus,
+ * exactly one rewrite is required.
+ */
+int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start,
+                        u64 length, u64 logical, struct page *page,
+                        int mirror_num)
+{
+        struct bio *bio;
+        struct btrfs_device *dev;
+        DECLARE_COMPLETION_ONSTACK(compl);
+        u64 map_length = 0;
+        u64 sector;
+        struct btrfs_bio *bbio = NULL;
+        int ret;
+
+        BUG_ON(!mirror_num);
+
+        bio = bio_alloc(GFP_NOFS, 1);
+        if (!bio)
+                return -EIO;
+        bio->bi_private = &compl;
+        bio->bi_end_io = repair_io_failure_callback;
+        bio->bi_size = 0;
+        map_length = length;
+
+        ret = btrfs_map_block(map_tree, WRITE, logical,
+                              &map_length, &bbio, mirror_num);
+        if (ret) {
+                bio_put(bio);
+                return -EIO;
+        }
+        BUG_ON(mirror_num != bbio->mirror_num);
+        sector = bbio->stripes[mirror_num-1].physical >> 9;
+        bio->bi_sector = sector;
+        dev = bbio->stripes[mirror_num-1].dev;
+        kfree(bbio);
+        if (!dev || !dev->bdev || !dev->writeable) {
+                bio_put(bio);
+                return -EIO;
+        }
+        bio->bi_bdev = dev->bdev;
+        bio_add_page(bio, page, length, start-page_offset(page));
+        submit_bio(WRITE_SYNC, bio);
+        wait_for_completion(&compl);
+
+        if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) {
+                /* try to remap that extent elsewhere? */
+                bio_put(bio);
+                return -EIO;
+        }
+
+        printk(KERN_INFO "btrfs read error corrected: ino %lu off %llu (dev %s "
+                        "sector %llu)\n", page->mapping->host->i_ino, start,
+                        dev->name, sector);
+
+        bio_put(bio);
+        return 0;
+}
+
+/*
+ * each time an IO finishes, we do a fast check in the IO failure tree
+ * to see if we need to process or clean up an io_failure_record
+ */
+static int clean_io_failure(u64 start, struct page *page)
+{
+        u64 private;
+        u64 private_failure;
+        struct io_failure_record *failrec;
+        struct btrfs_mapping_tree *map_tree;
+        struct extent_state *state;
+        int num_copies;
+        int did_repair = 0;
+        int ret;
+        struct inode *inode = page->mapping->host;
+
+        private = 0;
+        ret = count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private,
+                                (u64)-1, 1, EXTENT_DIRTY, 0);
+        if (!ret)
+                return 0;
+
+        ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, start,
+                                &private_failure);
+        if (ret)
+                return 0;
+
+        failrec = (struct io_failure_record *)(unsigned long) private_failure;
+        BUG_ON(!failrec->this_mirror);
+
+        if (failrec->in_validation) {
+                /* there was no real error, just free the record */
+                pr_debug("clean_io_failure: freeing dummy error at %llu\n",
+                         failrec->start);
+                did_repair = 1;
+                goto out;
+        }
+
+        spin_lock(&BTRFS_I(inode)->io_tree.lock);
+        state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree,
+                                            failrec->start,
+                                            EXTENT_LOCKED);
+        spin_unlock(&BTRFS_I(inode)->io_tree.lock);
+
+        if (state && state->start == failrec->start) {
+                map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree;
+                num_copies = btrfs_num_copies(map_tree, failrec->logical,
+                                                failrec->len);
+                if (num_copies > 1)  {
+                        ret = repair_io_failure(map_tree, start, failrec->len,
+                                                failrec->logical, page,
+                                                failrec->failed_mirror);
+                        did_repair = !ret;
+                }
+        }
+
+out:
+        if (!ret)
+                ret = free_io_failure(inode, failrec, did_repair);
+
+        return ret;
+}
+
+/*
+ * this is a generic handler for readpage errors (default
+ * readpage_io_failed_hook). if other copies exist, read those and write back
+ * good data to the failed position. does not investigate in remapping the
+ * failed extent elsewhere, hoping the device will be smart enough to do this as
+ * needed
+ */
+
+static int bio_readpage_error(struct bio *failed_bio, struct page *page,
+                                u64 start, u64 end, int failed_mirror,
+                                struct extent_state *state)
+{
+        struct io_failure_record *failrec = NULL;
+        u64 private;
+        struct extent_map *em;
+        struct inode *inode = page->mapping->host;
+        struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree;
+        struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
+        struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+        struct bio *bio;
+        int num_copies;
+        int ret;
+        int read_mode;
+        u64 logical;
+
+        BUG_ON(failed_bio->bi_rw & REQ_WRITE);
+
+        ret = get_state_private(failure_tree, start, &private);
+        if (ret) {
+                failrec = kzalloc(sizeof(*failrec), GFP_NOFS);
+                if (!failrec)
+                        return -ENOMEM;
+                failrec->start = start;
+                failrec->len = end - start + 1;
+                failrec->this_mirror = 0;
+                failrec->bio_flags = 0;
+                failrec->in_validation = 0;
+
+                read_lock(&em_tree->lock);
+                em = lookup_extent_mapping(em_tree, start, failrec->len);
+                if (!em) {
+                        read_unlock(&em_tree->lock);
+                        kfree(failrec);
+                        return -EIO;
+                }
+
+                if (em->start > start || em->start + em->len < start) {
+                        free_extent_map(em);
+                        em = NULL;
+                }
+                read_unlock(&em_tree->lock);
+
+                if (!em || IS_ERR(em)) {
+                        kfree(failrec);
+                        return -EIO;
+                }
+                logical = start - em->start;
+                logical = em->block_start + logical;
+                if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
+                        logical = em->block_start;
+                        failrec->bio_flags = EXTENT_BIO_COMPRESSED;
+                        extent_set_compress_type(&failrec->bio_flags,
+                                                 em->compress_type);
+                }
+                pr_debug("bio_readpage_error: (new) logical=%llu, start=%llu, "
+                         "len=%llu\n", logical, start, failrec->len);
+                failrec->logical = logical;
+                free_extent_map(em);
+
+                /* set the bits in the private failure tree */
+                ret = set_extent_bits(failure_tree, start, end,
+                                        EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS);
+                if (ret >= 0)
+                        ret = set_state_private(failure_tree, start,
+                                                (u64)(unsigned long)failrec);
+                /* set the bits in the inode's tree */
+                if (ret >= 0)
+                        ret = set_extent_bits(tree, start, end, EXTENT_DAMAGED,
+                                                GFP_NOFS);
+                if (ret < 0) {
+                        kfree(failrec);
+                        return ret;
+                }
+        } else {
+                failrec = (struct io_failure_record *)(unsigned long)private;
+                pr_debug("bio_readpage_error: (found) logical=%llu, "
+                         "start=%llu, len=%llu, validation=%d\n",
+                         failrec->logical, failrec->start, failrec->len,
+                         failrec->in_validation);
+                /*
+                 * when data can be on disk more than twice, add to failrec here
+                 * (e.g. with a list for failed_mirror) to make
+                 * clean_io_failure() clean all those errors at once.
+                 */
+        }
+        num_copies = btrfs_num_copies(
+                              &BTRFS_I(inode)->root->fs_info->mapping_tree,
+                              failrec->logical, failrec->len);
+        if (num_copies == 1) {
+                /*
+                 * we only have a single copy of the data, so don't bother with
+                 * all the retry and error correction code that follows. no
+                 * matter what the error is, it is very likely to persist.
+                 */
+                pr_debug("bio_readpage_error: cannot repair, num_copies == 1. "
+                         "state=%p, num_copies=%d, next_mirror %d, "
+                         "failed_mirror %d\n", state, num_copies,
+                         failrec->this_mirror, failed_mirror);
+                free_io_failure(inode, failrec, 0);
+                return -EIO;
+        }
+
+        if (!state) {
+                spin_lock(&tree->lock);
+                state = find_first_extent_bit_state(tree, failrec->start,
+                                                    EXTENT_LOCKED);
+                if (state && state->start != failrec->start)
+                        state = NULL;
+                spin_unlock(&tree->lock);
+        }
+
+        /*
+         * there are two premises:
+         *      a) deliver good data to the caller
+         *      b) correct the bad sectors on disk
+         */
+        if (failed_bio->bi_vcnt > 1) {
+                /*
+                 * to fulfill b), we need to know the exact failing sectors, as
+                 * we don't want to rewrite any more than the failed ones. thus,
+                 * we need separate read requests for the failed bio
+                 *
+                 * if the following BUG_ON triggers, our validation request got
+                 * merged. we need separate requests for our algorithm to work.
+                 */
+                BUG_ON(failrec->in_validation);
+                failrec->in_validation = 1;
+                failrec->this_mirror = failed_mirror;
+                read_mode = READ_SYNC | REQ_FAILFAST_DEV;
+        } else {
+                /*
+                 * we're ready to fulfill a) and b) alongside. get a good copy
+                 * of the failed sector and if we succeed, we have setup
+                 * everything for repair_io_failure to do the rest for us.
+                 */
+                if (failrec->in_validation) {
+                        BUG_ON(failrec->this_mirror != failed_mirror);
+                        failrec->in_validation = 0;
+                        failrec->this_mirror = 0;
+                }
+                failrec->failed_mirror = failed_mirror;
+                failrec->this_mirror++;
+                if (failrec->this_mirror == failed_mirror)
+                        failrec->this_mirror++;
+                read_mode = READ_SYNC;
+        }
+
+        if (!state || failrec->this_mirror > num_copies) {
+                pr_debug("bio_readpage_error: (fail) state=%p, num_copies=%d, "
+                         "next_mirror %d, failed_mirror %d\n", state,
+                         num_copies, failrec->this_mirror, failed_mirror);
+                free_io_failure(inode, failrec, 0);
+                return -EIO;
+        }
+
+        bio = bio_alloc(GFP_NOFS, 1);
+        bio->bi_private = state;
+        bio->bi_end_io = failed_bio->bi_end_io;
+        bio->bi_sector = failrec->logical >> 9;
+        bio->bi_bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
+        bio->bi_size = 0;
+
+        bio_add_page(bio, page, failrec->len, start - page_offset(page));
+
+        pr_debug("bio_readpage_error: submitting new read[%#x] to "
+                 "this_mirror=%d, num_copies=%d, in_validation=%d\n", read_mode,
+                 failrec->this_mirror, num_copies, failrec->in_validation);
+
+        tree->ops->submit_bio_hook(inode, read_mode, bio, failrec->this_mirror,
+                                        failrec->bio_flags, 0);
+        return 0;
+}
+
 /* lots and lots of room for performance fixes in the end_bio funcs */
 
 /*
@@ -1697,6 +2248,9 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
                 struct extent_state *cached = NULL;
                 struct extent_state *state;
 
+                pr_debug("end_bio_extent_readpage: bi_vcnt=%d, idx=%d, err=%d, "
+                         "mirror=%ld\n", bio->bi_vcnt, bio->bi_idx, err,
+                         (long int)bio->bi_bdev);
                 tree = &BTRFS_I(page->mapping->host)->io_tree;
 
                 start = ((u64)page->index << PAGE_CACHE_SHIFT) +
@@ -1727,11 +2281,19 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
                                                               state);
                         if (ret)
                                 uptodate = 0;
+                        else
+                                clean_io_failure(start, page);
                 }
-                if (!uptodate && tree->ops &&
-                    tree->ops->readpage_io_failed_hook) {
-                        ret = tree->ops->readpage_io_failed_hook(bio, page,
-                                                         start, end, NULL);
+                if (!uptodate) {
+                        u64 failed_mirror;
+                        failed_mirror = (u64)bio->bi_bdev;
+                        if (tree->ops && tree->ops->readpage_io_failed_hook)
+                                ret = tree->ops->readpage_io_failed_hook(
+                                                bio, page, start, end,
+                                                failed_mirror, state);
+                        else
+                                ret = bio_readpage_error(bio, page, start, end,
+                                                         failed_mirror, NULL);
                         if (ret == 0) {
                                 uptodate =
                                         test_bit(BIO_UPTODATE, &bio->bi_flags);
@@ -1811,6 +2373,7 @@ static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
                                            mirror_num, bio_flags, start);
         else
                 submit_bio(rw, bio);
+
         if (bio_flagged(bio, BIO_EOPNOTSUPP))
                 ret = -EOPNOTSUPP;
         bio_put(bio);
@@ -2076,16 +2639,16 @@ out:
 }
 
 int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
-                            get_extent_t *get_extent)
+                            get_extent_t *get_extent, int mirror_num)
 {
         struct bio *bio = NULL;
         unsigned long bio_flags = 0;
         int ret;
 
-        ret = __extent_read_full_page(tree, page, get_extent, &bio, 0,
+        ret = __extent_read_full_page(tree, page, get_extent, &bio, mirror_num,
                                       &bio_flags);
         if (bio)
-                ret = submit_one_bio(READ, bio, 0, bio_flags);
+                ret = submit_one_bio(READ, bio, mirror_num, bio_flags);
         return ret;
 }
 
@@ -2136,6 +2699,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
         int compressed;
         int write_flags;
         unsigned long nr_written = 0;
+        bool fill_delalloc = true;
 
         if (wbc->sync_mode == WB_SYNC_ALL)
                 write_flags = WRITE_SYNC;
@@ -2145,6 +2709,9 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
         trace___extent_writepage(page, inode, wbc);
 
         WARN_ON(!PageLocked(page));
+
+        ClearPageError(page);
+
         pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
         if (page->index > end_index ||
            (page->index == end_index && !pg_offset)) {
@@ -2166,10 +2733,13 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 
         set_page_extent_mapped(page);
 
+        if (!tree->ops || !tree->ops->fill_delalloc)
+                fill_delalloc = false;
+
         delalloc_start = start;
         delalloc_end = 0;
         page_started = 0;
-        if (!epd->extent_locked) {
+        if (!epd->extent_locked && fill_delalloc) {
                 u64 delalloc_to_write = 0;
                 /*
                  * make sure the wbc mapping index is at least updated
@@ -2421,10 +2991,16 @@ retry:
                          * swizzled back from swapper_space to tmpfs file
                          * mapping
                          */
-                        if (tree->ops && tree->ops->write_cache_pages_lock_hook)
-                                tree->ops->write_cache_pages_lock_hook(page);
-                        else
-                                lock_page(page);
+                        if (tree->ops &&
+                            tree->ops->write_cache_pages_lock_hook) {
+                                tree->ops->write_cache_pages_lock_hook(page,
+                                                               data, flush_fn);
+                        } else {
+                                if (!trylock_page(page)) {
+                                        flush_fn(data);
+                                        lock_page(page);
+                                }
+                        }
 
                         if (unlikely(page->mapping != mapping)) {
                                 unlock_page(page);
@@ -2926,7 +3502,7 @@ out:
         return ret;
 }
 
-static inline struct page *extent_buffer_page(struct extent_buffer *eb,
+inline struct page *extent_buffer_page(struct extent_buffer *eb,
                                               unsigned long i)
 {
         struct page *p;
@@ -2951,7 +3527,7 @@ static inline struct page *extent_buffer_page(struct extent_buffer *eb,
         return p;
 }
 
-static inline unsigned long num_extent_pages(u64 start, u64 len)
+inline unsigned long num_extent_pages(u64 start, u64 len)
 {
         return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
                 (start >> PAGE_CACHE_SHIFT);
@@ -3204,6 +3780,7 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree,
                                                 PAGECACHE_TAG_DIRTY);
                 }
                 spin_unlock_irq(&page->mapping->tree_lock);
+                ClearPageError(page);
                 unlock_page(page);
         }
         return 0;
@@ -3349,8 +3926,7 @@ int extent_buffer_uptodate(struct extent_io_tree *tree,
 }
 
 int read_extent_buffer_pages(struct extent_io_tree *tree,
-                             struct extent_buffer *eb,
-                             u64 start, int wait,
+                             struct extent_buffer *eb, u64 start, int wait,
                              get_extent_t *get_extent, int mirror_num)
 {
         unsigned long i;
@@ -3386,7 +3962,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
         num_pages = num_extent_pages(eb->start, eb->len);
         for (i = start_i; i < num_pages; i++) {
                 page = extent_buffer_page(eb, i);
-                if (!wait) {
+                if (wait == WAIT_NONE) {
                         if (!trylock_page(page))
                                 goto unlock_exit;
                 } else {
@@ -3430,7 +4006,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
         if (bio)
                 submit_one_bio(READ, bio, mirror_num, bio_flags);
 
-        if (ret || !wait)
+        if (ret || wait != WAIT_COMPLETE)
                 return ret;
 
         for (i = start_i; i < num_pages; i++) {
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index 7b2f0c3e7929..feb9be0e23bc 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -17,6 +17,8 @@
 #define EXTENT_NODATASUM (1 << 10)
 #define EXTENT_DO_ACCOUNTING (1 << 11)
 #define EXTENT_FIRST_DELALLOC (1 << 12)
+#define EXTENT_NEED_WAIT (1 << 13)
+#define EXTENT_DAMAGED (1 << 14)
 #define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK)
 #define EXTENT_CTLBITS (EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC)
 
@@ -32,6 +34,7 @@
 #define EXTENT_BUFFER_BLOCKING 1
 #define EXTENT_BUFFER_DIRTY 2
 #define EXTENT_BUFFER_CORRUPT 3
+#define EXTENT_BUFFER_READAHEAD 4       /* this got triggered by readahead */
 
 /* these are flags for extent_clear_unlock_delalloc */
 #define EXTENT_CLEAR_UNLOCK_PAGE 0x1
@@ -67,7 +70,7 @@ struct extent_io_ops {
                               unsigned long bio_flags);
         int (*readpage_io_hook)(struct page *page, u64 start, u64 end);
         int (*readpage_io_failed_hook)(struct bio *bio, struct page *page,
-                                       u64 start, u64 end,
+                                       u64 start, u64 end, u64 failed_mirror,
                                        struct extent_state *state);
         int (*writepage_io_failed_hook)(struct bio *bio, struct page *page,
                                         u64 start, u64 end,
@@ -85,7 +88,8 @@ struct extent_io_ops {
                                   struct extent_state *other);
         void (*split_extent_hook)(struct inode *inode,
                                   struct extent_state *orig, u64 split);
-        int (*write_cache_pages_lock_hook)(struct page *page);
+        int (*write_cache_pages_lock_hook)(struct page *page, void *data,
+                                           void (*flush_fn)(void *));
 };
 
 struct extent_io_tree {
@@ -185,7 +189,7 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
 int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
                     gfp_t mask);
 int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
-                          get_extent_t *get_extent);
+                          get_extent_t *get_extent, int mirror_num);
 int __init extent_io_init(void);
 void extent_io_exit(void);
 
@@ -214,6 +218,8 @@ int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
                      gfp_t mask);
 int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
                        gfp_t mask);
+int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+                       int bits, int clear_bits, gfp_t mask);
 int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
                         struct extent_state **cached_state, gfp_t mask);
 int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
@@ -248,9 +254,14 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
                                          u64 start, unsigned long len);
 void free_extent_buffer(struct extent_buffer *eb);
+#define WAIT_NONE       0
+#define WAIT_COMPLETE   1
+#define WAIT_PAGE_LOCK  2
 int read_extent_buffer_pages(struct extent_io_tree *tree,
                              struct extent_buffer *eb, u64 start, int wait,
                              get_extent_t *get_extent, int mirror_num);
+unsigned long num_extent_pages(u64 start, u64 len);
+struct page *extent_buffer_page(struct extent_buffer *eb, unsigned long i);
 
 static inline void extent_buffer_get(struct extent_buffer *eb)
 {
@@ -300,4 +311,10 @@ int extent_clear_unlock_delalloc(struct inode *inode,
 struct bio *
 btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
                 gfp_t gfp_flags);
+
+struct btrfs_mapping_tree;
+
+int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start,
+                        u64 length, u64 logical, struct page *page,
+                        int mirror_num);
 #endif
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index a1cb7821becd..c7fb3a4247d3 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -91,8 +91,7 @@ struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans,
         struct btrfs_csum_item *item;
         struct extent_buffer *leaf;
         u64 csum_offset = 0;
-        u16 csum_size =
-                btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
         int csums_in_item;
 
         file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
@@ -162,8 +161,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
         u64 item_last_offset = 0;
         u64 disk_bytenr;
         u32 diff;
-        u16 csum_size =
-                btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
         int ret;
         struct btrfs_path *path;
         struct btrfs_csum_item *item = NULL;
@@ -290,7 +288,7 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
         int ret;
         size_t size;
         u64 csum_end;
-        u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
 
         path = btrfs_alloc_path();
         if (!path)
@@ -492,8 +490,7 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
                                       u64 bytenr, u64 len)
 {
         struct extent_buffer *leaf;
-        u16 csum_size =
-                btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
         u64 csum_end;
         u64 end_byte = bytenr + len;
         u32 blocksize_bits = root->fs_info->sb->s_blocksize_bits;
@@ -549,8 +546,7 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
         u64 csum_end;
         struct extent_buffer *leaf;
         int ret;
-        u16 csum_size =
-                btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
         int blocksize_bits = root->fs_info->sb->s_blocksize_bits;
 
         root = root->fs_info->csum_root;
@@ -676,8 +672,7 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
         struct btrfs_sector_sum *sector_sum;
         u32 nritems;
         u32 ins_size;
-        u16 csum_size =
-                btrfs_super_csum_size(&root->fs_info->super_copy);
+        u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
 
         path = btrfs_alloc_path();
         if (!path)
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 1266f6e9cdb2..dafdfa059bf6 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -1069,6 +1069,7 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file,
         int i;
         unsigned long index = pos >> PAGE_CACHE_SHIFT;
         struct inode *inode = fdentry(file)->d_inode;
+        gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
         int err = 0;
         int faili = 0;
         u64 start_pos;
@@ -1080,7 +1081,7 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file,
 again:
         for (i = 0; i < num_pages; i++) {
                 pages[i] = find_or_create_page(inode->i_mapping, index + i,
-                                               GFP_NOFS);
+                                               mask);
                 if (!pages[i]) {
                         faili = i - 1;
                         err = -ENOMEM;
@@ -1615,10 +1616,6 @@ static long btrfs_fallocate(struct file *file, int mode,
                         goto out;
         }
 
-        ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start);
-        if (ret)
-                goto out;
-
         locked_end = alloc_end - 1;
         while (1) {
                 struct btrfs_ordered_extent *ordered;
@@ -1664,11 +1661,27 @@ static long btrfs_fallocate(struct file *file, int mode,
                 if (em->block_start == EXTENT_MAP_HOLE ||
                     (cur_offset >= inode->i_size &&
                      !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
+
+                        /*
+                         * Make sure we have enough space before we do the
+                         * allocation.
+                         */
+                        ret = btrfs_check_data_free_space(inode, last_byte -
+                                                          cur_offset);
+                        if (ret) {
+                                free_extent_map(em);
+                                break;
+                        }
+
                         ret = btrfs_prealloc_file_range(inode, mode, cur_offset,
                                                         last_byte - cur_offset,
                                                         1 << inode->i_blkbits,
                                                         offset + len,
                                                         &alloc_hint);
+
+                        /* Let go of our reservation. */
+                        btrfs_free_reserved_data_space(inode, last_byte -
+                                                       cur_offset);
                         if (ret < 0) {
                                 free_extent_map(em);
                                 break;
@@ -1694,8 +1707,6 @@ static long btrfs_fallocate(struct file *file, int mode,
         }
         unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
                              &cached_state, GFP_NOFS);
-
-        btrfs_free_reserved_data_space(inode, alloc_end - alloc_start);
 out:
         mutex_unlock(&inode->i_mutex);
         return ret;
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 41ac927401d0..7a15fcfb3e1f 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -20,6 +20,7 @@
 #include <linux/sched.h>
 #include <linux/slab.h>
 #include <linux/math64.h>
+#include <linux/ratelimit.h>
 #include "ctree.h"
 #include "free-space-cache.h"
 #include "transaction.h"
@@ -84,6 +85,7 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
                                       *block_group, struct btrfs_path *path)
 {
         struct inode *inode = NULL;
+        u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
 
         spin_lock(&block_group->lock);
         if (block_group->inode)
@@ -98,13 +100,14 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
                 return inode;
 
         spin_lock(&block_group->lock);
-        if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) {
+        if (!((BTRFS_I(inode)->flags & flags) == flags)) {
                 printk(KERN_INFO "Old style space inode found, converting.\n");
-                BTRFS_I(inode)->flags &= ~BTRFS_INODE_NODATASUM;
+                BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM |
+                        BTRFS_INODE_NODATACOW;
                 block_group->disk_cache_state = BTRFS_DC_CLEAR;
         }
 
-        if (!btrfs_fs_closing(root->fs_info)) {
+        if (!block_group->iref) {
                 block_group->inode = igrab(inode);
                 block_group->iref = 1;
         }
@@ -122,12 +125,17 @@ int __create_free_space_inode(struct btrfs_root *root,
         struct btrfs_free_space_header *header;
         struct btrfs_inode_item *inode_item;
         struct extent_buffer *leaf;
+        u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC;
         int ret;
 
         ret = btrfs_insert_empty_inode(trans, root, path, ino);
         if (ret)
                 return ret;
 
+        /* We inline crc's for the free disk space cache */
+        if (ino != BTRFS_FREE_INO_OBJECTID)
+                flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
+
         leaf = path->nodes[0];
         inode_item = btrfs_item_ptr(leaf, path->slots[0],
                                     struct btrfs_inode_item);
@@ -140,8 +148,7 @@ int __create_free_space_inode(struct btrfs_root *root,
         btrfs_set_inode_uid(leaf, inode_item, 0);
         btrfs_set_inode_gid(leaf, inode_item, 0);
         btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600);
-        btrfs_set_inode_flags(leaf, inode_item, BTRFS_INODE_NOCOMPRESS |
-                              BTRFS_INODE_PREALLOC);
+        btrfs_set_inode_flags(leaf, inode_item, flags);
         btrfs_set_inode_nlink(leaf, inode_item, 1);
         btrfs_set_inode_transid(leaf, inode_item, trans->transid);
         btrfs_set_inode_block_group(leaf, inode_item, offset);
@@ -191,16 +198,24 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
                                     struct inode *inode)
 {
         struct btrfs_block_rsv *rsv;
+        u64 needed_bytes;
         loff_t oldsize;
         int ret = 0;
 
         rsv = trans->block_rsv;
-        trans->block_rsv = root->orphan_block_rsv;
-        ret = btrfs_block_rsv_check(trans, root,
-                                    root->orphan_block_rsv,
-                                    0, 5);
-        if (ret)
-                return ret;
+        trans->block_rsv = &root->fs_info->global_block_rsv;
+
+        /* 1 for slack space, 1 for updating the inode */
+        needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) +
+                btrfs_calc_trans_metadata_size(root, 1);
+
+        spin_lock(&trans->block_rsv->lock);
+        if (trans->block_rsv->reserved < needed_bytes) {
+                spin_unlock(&trans->block_rsv->lock);
+                trans->block_rsv = rsv;
+                return -ENOSPC;
+        }
+        spin_unlock(&trans->block_rsv->lock);
 
         oldsize = i_size_read(inode);
         btrfs_i_size_write(inode, 0);
@@ -213,13 +228,15 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
         ret = btrfs_truncate_inode_items(trans, root, inode,
                                          0, BTRFS_EXTENT_DATA_KEY);
 
-        trans->block_rsv = rsv;
         if (ret) {
+                trans->block_rsv = rsv;
                 WARN_ON(1);
                 return ret;
         }
 
         ret = btrfs_update_inode(trans, root, inode);
+        trans->block_rsv = rsv;
+
         return ret;
 }
 
@@ -242,26 +259,342 @@ static int readahead_cache(struct inode *inode)
         return 0;
 }
 
+struct io_ctl {
+        void *cur, *orig;
+        struct page *page;
+        struct page **pages;
+        struct btrfs_root *root;
+        unsigned long size;
+        int index;
+        int num_pages;
+        unsigned check_crcs:1;
+};
+
+static int io_ctl_init(struct io_ctl *io_ctl, struct inode *inode,
+                       struct btrfs_root *root)
+{
+        memset(io_ctl, 0, sizeof(struct io_ctl));
+        io_ctl->num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
+                PAGE_CACHE_SHIFT;
+        io_ctl->pages = kzalloc(sizeof(struct page *) * io_ctl->num_pages,
+                                GFP_NOFS);
+        if (!io_ctl->pages)
+                return -ENOMEM;
+        io_ctl->root = root;
+        if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID)
+                io_ctl->check_crcs = 1;
+        return 0;
+}
+
+static void io_ctl_free(struct io_ctl *io_ctl)
+{
+        kfree(io_ctl->pages);
+}
+
+static void io_ctl_unmap_page(struct io_ctl *io_ctl)
+{
+        if (io_ctl->cur) {
+                kunmap(io_ctl->page);
+                io_ctl->cur = NULL;
+                io_ctl->orig = NULL;
+        }
+}
+
+static void io_ctl_map_page(struct io_ctl *io_ctl, int clear)
+{
+        WARN_ON(io_ctl->cur);
+        BUG_ON(io_ctl->index >= io_ctl->num_pages);
+        io_ctl->page = io_ctl->pages[io_ctl->index++];
+        io_ctl->cur = kmap(io_ctl->page);
+        io_ctl->orig = io_ctl->cur;
+        io_ctl->size = PAGE_CACHE_SIZE;
+        if (clear)
+                memset(io_ctl->cur, 0, PAGE_CACHE_SIZE);
+}
+
+static void io_ctl_drop_pages(struct io_ctl *io_ctl)
+{
+        int i;
+
+        io_ctl_unmap_page(io_ctl);
+
+        for (i = 0; i < io_ctl->num_pages; i++) {
+                ClearPageChecked(io_ctl->pages[i]);
+                unlock_page(io_ctl->pages[i]);
+                page_cache_release(io_ctl->pages[i]);
+        }
+}
+
+static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode,
+                                int uptodate)
+{
+        struct page *page;
+        gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
+        int i;
+
+        for (i = 0; i < io_ctl->num_pages; i++) {
+                page = find_or_create_page(inode->i_mapping, i, mask);
+                if (!page) {
+                        io_ctl_drop_pages(io_ctl);
+                        return -ENOMEM;
+                }
+                io_ctl->pages[i] = page;
+                if (uptodate && !PageUptodate(page)) {
+                        btrfs_readpage(NULL, page);
+                        lock_page(page);
+                        if (!PageUptodate(page)) {
+                                printk(KERN_ERR "btrfs: error reading free "
+                                       "space cache\n");
+                                io_ctl_drop_pages(io_ctl);
+                                return -EIO;
+                        }
+                }
+        }
+
+        return 0;
+}
+
+static void io_ctl_set_generation(struct io_ctl *io_ctl, u64 generation)
+{
+        u64 *val;
+
+        io_ctl_map_page(io_ctl, 1);
+
+        /*
+         * Skip the csum areas.  If we don't check crcs then we just have a
+         * 64bit chunk at the front of the first page.
+         */
+        if (io_ctl->check_crcs) {
+                io_ctl->cur += (sizeof(u32) * io_ctl->num_pages);
+                io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages);
+        } else {
+                io_ctl->cur += sizeof(u64);
+                io_ctl->size -= sizeof(u64) * 2;
+        }
+
+        val = io_ctl->cur;
+        *val = cpu_to_le64(generation);
+        io_ctl->cur += sizeof(u64);
+}
+
+static int io_ctl_check_generation(struct io_ctl *io_ctl, u64 generation)
+{
+        u64 *gen;
+
+        /*
+         * Skip the crc area.  If we don't check crcs then we just have a 64bit
+         * chunk at the front of the first page.
+         */
+        if (io_ctl->check_crcs) {
+                io_ctl->cur += sizeof(u32) * io_ctl->num_pages;
+                io_ctl->size -= sizeof(u64) +
+                        (sizeof(u32) * io_ctl->num_pages);
+        } else {
+                io_ctl->cur += sizeof(u64);
+                io_ctl->size -= sizeof(u64) * 2;
+        }
+
+        gen = io_ctl->cur;
+        if (le64_to_cpu(*gen) != generation) {
+                printk_ratelimited(KERN_ERR "btrfs: space cache generation "
+                                   "(%Lu) does not match inode (%Lu)\n", *gen,
+                                   generation);
+                io_ctl_unmap_page(io_ctl);
+                return -EIO;
+        }
+        io_ctl->cur += sizeof(u64);
+        return 0;
+}
+
+static void io_ctl_set_crc(struct io_ctl *io_ctl, int index)
+{
+        u32 *tmp;
+        u32 crc = ~(u32)0;
+        unsigned offset = 0;
+
+        if (!io_ctl->check_crcs) {
+                io_ctl_unmap_page(io_ctl);
+                return;
+        }
+
+        if (index == 0)
+                offset = sizeof(u32) * io_ctl->num_pages;;
+
+        crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc,
+                              PAGE_CACHE_SIZE - offset);
+        btrfs_csum_final(crc, (char *)&crc);
+        io_ctl_unmap_page(io_ctl);
+        tmp = kmap(io_ctl->pages[0]);
+        tmp += index;
+        *tmp = crc;
+        kunmap(io_ctl->pages[0]);
+}
+
+static int io_ctl_check_crc(struct io_ctl *io_ctl, int index)
+{
+        u32 *tmp, val;
+        u32 crc = ~(u32)0;
+        unsigned offset = 0;
+
+        if (!io_ctl->check_crcs) {
+                io_ctl_map_page(io_ctl, 0);
+                return 0;
+        }
+
+        if (index == 0)
+                offset = sizeof(u32) * io_ctl->num_pages;
+
+        tmp = kmap(io_ctl->pages[0]);
+        tmp += index;
+        val = *tmp;
+        kunmap(io_ctl->pages[0]);
+
+        io_ctl_map_page(io_ctl, 0);
+        crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc,
+                              PAGE_CACHE_SIZE - offset);
+        btrfs_csum_final(crc, (char *)&crc);
+        if (val != crc) {
+                printk_ratelimited(KERN_ERR "btrfs: csum mismatch on free "
+                                   "space cache\n");
+                io_ctl_unmap_page(io_ctl);
+                return -EIO;
+        }
+
+        return 0;
+}
+
+static int io_ctl_add_entry(struct io_ctl *io_ctl, u64 offset, u64 bytes,
+                            void *bitmap)
+{
+        struct btrfs_free_space_entry *entry;
+
+        if (!io_ctl->cur)
+                return -ENOSPC;
+
+        entry = io_ctl->cur;
+        entry->offset = cpu_to_le64(offset);
+        entry->bytes = cpu_to_le64(bytes);
+        entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP :
+                BTRFS_FREE_SPACE_EXTENT;
+        io_ctl->cur += sizeof(struct btrfs_free_space_entry);
+        io_ctl->size -= sizeof(struct btrfs_free_space_entry);
+
+        if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
+                return 0;
+
+        io_ctl_set_crc(io_ctl, io_ctl->index - 1);
+
+        /* No more pages to map */
+        if (io_ctl->index >= io_ctl->num_pages)
+                return 0;
+
+        /* map the next page */
+        io_ctl_map_page(io_ctl, 1);
+        return 0;
+}
+
+static int io_ctl_add_bitmap(struct io_ctl *io_ctl, void *bitmap)
+{
+        if (!io_ctl->cur)
+                return -ENOSPC;
+
+        /*
+         * If we aren't at the start of the current page, unmap this one and
+         * map the next one if there is any left.
+         */
+        if (io_ctl->cur != io_ctl->orig) {
+                io_ctl_set_crc(io_ctl, io_ctl->index - 1);
+                if (io_ctl->index >= io_ctl->num_pages)
+                        return -ENOSPC;
+                io_ctl_map_page(io_ctl, 0);
+        }
+
+        memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE);
+        io_ctl_set_crc(io_ctl, io_ctl->index - 1);
+        if (io_ctl->index < io_ctl->num_pages)
+                io_ctl_map_page(io_ctl, 0);
+        return 0;
+}
+
+static void io_ctl_zero_remaining_pages(struct io_ctl *io_ctl)
+{
+        /*
+         * If we're not on the boundary we know we've modified the page and we
+         * need to crc the page.
+         */
+        if (io_ctl->cur != io_ctl->orig)
+                io_ctl_set_crc(io_ctl, io_ctl->index - 1);
+        else
+                io_ctl_unmap_page(io_ctl);
+
+        while (io_ctl->index < io_ctl->num_pages) {
+                io_ctl_map_page(io_ctl, 1);
+                io_ctl_set_crc(io_ctl, io_ctl->index - 1);
+        }
+}
+
+static int io_ctl_read_entry(struct io_ctl *io_ctl,
+                            struct btrfs_free_space *entry, u8 *type)
+{
+        struct btrfs_free_space_entry *e;
+
+        e = io_ctl->cur;
+        entry->offset = le64_to_cpu(e->offset);
+        entry->bytes = le64_to_cpu(e->bytes);
+        *type = e->type;
+        io_ctl->cur += sizeof(struct btrfs_free_space_entry);
+        io_ctl->size -= sizeof(struct btrfs_free_space_entry);
+
+        if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
+                return 0;
+
+        io_ctl_unmap_page(io_ctl);
+
+        if (io_ctl->index >= io_ctl->num_pages)
+                return 0;
+
+        return io_ctl_check_crc(io_ctl, io_ctl->index);
+}
+
+static int io_ctl_read_bitmap(struct io_ctl *io_ctl,
+                              struct btrfs_free_space *entry)
+{
+        int ret;
+
+        if (io_ctl->cur && io_ctl->cur != io_ctl->orig)
+                io_ctl_unmap_page(io_ctl);
+
+        ret = io_ctl_check_crc(io_ctl, io_ctl->index);
+        if (ret)
+                return ret;
+
+        memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE);
+        io_ctl_unmap_page(io_ctl);
+
+        return 0;
+}
+
 int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
                             struct btrfs_free_space_ctl *ctl,
                             struct btrfs_path *path, u64 offset)
 {
         struct btrfs_free_space_header *header;
         struct extent_buffer *leaf;
-        struct page *page;
+        struct io_ctl io_ctl;
         struct btrfs_key key;
+        struct btrfs_free_space *e, *n;
         struct list_head bitmaps;
         u64 num_entries;
         u64 num_bitmaps;
         u64 generation;
-        pgoff_t index = 0;
+        u8 type;
         int ret = 0;
 
         INIT_LIST_HEAD(&bitmaps);
 
         /* Nothing in the space cache, goodbye */
         if (!i_size_read(inode))
-                goto out;
+                return 0;
 
         key.objectid = BTRFS_FREE_SPACE_OBJECTID;
         key.offset = offset;
@@ -269,11 +602,10 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
 
         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
         if (ret < 0)
-                goto out;
+                return 0;
         else if (ret > 0) {
                 btrfs_release_path(path);
-                ret = 0;
-                goto out;
+                return 0;
         }
 
         ret = -1;
@@ -291,169 +623,100 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
                        " not match free space cache generation (%llu)\n",
                        (unsigned long long)BTRFS_I(inode)->generation,
                        (unsigned long long)generation);
-                goto out;
+                return 0;
         }
 
         if (!num_entries)
-                goto out;
+                return 0;
 
+        io_ctl_init(&io_ctl, inode, root);
         ret = readahead_cache(inode);
         if (ret)
                 goto out;
 
-        while (1) {
-                struct btrfs_free_space_entry *entry;
-                struct btrfs_free_space *e;
-                void *addr;
-                unsigned long offset = 0;
-                int need_loop = 0;
+        ret = io_ctl_prepare_pages(&io_ctl, inode, 1);
+        if (ret)
+                goto out;
 
-                if (!num_entries && !num_bitmaps)
-                        break;
+        ret = io_ctl_check_crc(&io_ctl, 0);
+        if (ret)
+                goto free_cache;
 
-                page = find_or_create_page(inode->i_mapping, index, GFP_NOFS);
-                if (!page)
+        ret = io_ctl_check_generation(&io_ctl, generation);
+        if (ret)
+                goto free_cache;
+
+        while (num_entries) {
+                e = kmem_cache_zalloc(btrfs_free_space_cachep,
+                                      GFP_NOFS);
+                if (!e)
                         goto free_cache;
 
-                if (!PageUptodate(page)) {
-                        btrfs_readpage(NULL, page);
-                        lock_page(page);
-                        if (!PageUptodate(page)) {
-                                unlock_page(page);
-                                page_cache_release(page);
-                                printk(KERN_ERR "btrfs: error reading free "
-                                       "space cache\n");
-                                goto free_cache;
-                        }
+                ret = io_ctl_read_entry(&io_ctl, e, &type);
+                if (ret) {
+                        kmem_cache_free(btrfs_free_space_cachep, e);
+                        goto free_cache;
                 }
-                addr = kmap(page);
 
-                if (index == 0) {
-                        u64 *gen;
+                if (!e->bytes) {
+                        kmem_cache_free(btrfs_free_space_cachep, e);
+                        goto free_cache;
+                }
 
-                        /*
-                         * We put a bogus crc in the front of the first page in
-                         * case old kernels try to mount a fs with the new
-                         * format to make sure they discard the cache.
-                         */
-                        addr += sizeof(u64);
-                        offset += sizeof(u64);
-
-                        gen = addr;
-                        if (*gen != BTRFS_I(inode)->generation) {
-                                printk(KERN_ERR "btrfs: space cache generation"
-                                       " (%llu) does not match inode (%llu)\n",
-                                       (unsigned long long)*gen,
-                                       (unsigned long long)
-                                       BTRFS_I(inode)->generation);
-                                kunmap(page);
-                                unlock_page(page);
-                                page_cache_release(page);
+                if (type == BTRFS_FREE_SPACE_EXTENT) {
+                        spin_lock(&ctl->tree_lock);
+                        ret = link_free_space(ctl, e);
+                        spin_unlock(&ctl->tree_lock);
+                        if (ret) {
+                                printk(KERN_ERR "Duplicate entries in "
+                                       "free space cache, dumping\n");
+                                kmem_cache_free(btrfs_free_space_cachep, e);
                                 goto free_cache;
                         }
-                        addr += sizeof(u64);
-                        offset += sizeof(u64);
-                }
-                entry = addr;
-
-                while (1) {
-                        if (!num_entries)
-                                break;
-
-                        need_loop = 1;
-                        e = kmem_cache_zalloc(btrfs_free_space_cachep,
-                                              GFP_NOFS);
-                        if (!e) {
-                                kunmap(page);
-                                unlock_page(page);
-                                page_cache_release(page);
+                } else {
+                        BUG_ON(!num_bitmaps);
+                        num_bitmaps--;
+                        e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
+                        if (!e->bitmap) {
+                                kmem_cache_free(
+                                        btrfs_free_space_cachep, e);
                                 goto free_cache;
                         }
-
-                        e->offset = le64_to_cpu(entry->offset);
-                        e->bytes = le64_to_cpu(entry->bytes);
-                        if (!e->bytes) {
-                                kunmap(page);
+                        spin_lock(&ctl->tree_lock);
+                        ret = link_free_space(ctl, e);
+                        ctl->total_bitmaps++;
+                        ctl->op->recalc_thresholds(ctl);
+                        spin_unlock(&ctl->tree_lock);
+                        if (ret) {
+                                printk(KERN_ERR "Duplicate entries in "
+                                       "free space cache, dumping\n");
                                 kmem_cache_free(btrfs_free_space_cachep, e);
-                                unlock_page(page);
-                                page_cache_release(page);
                                 goto free_cache;
                         }
-
-                        if (entry->type == BTRFS_FREE_SPACE_EXTENT) {
-                                spin_lock(&ctl->tree_lock);
-                                ret = link_free_space(ctl, e);
-                                spin_unlock(&ctl->tree_lock);
-                                if (ret) {
-                                        printk(KERN_ERR "Duplicate entries in "
-                                               "free space cache, dumping\n");
-                                        kunmap(page);
-                                        unlock_page(page);
-                                        page_cache_release(page);
-                                        goto free_cache;
-                                }
-                        } else {
-                                e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
-                                if (!e->bitmap) {
-                                        kunmap(page);
-                                        kmem_cache_free(
-                                                btrfs_free_space_cachep, e);
-                                        unlock_page(page);
-                                        page_cache_release(page);
-                                        goto free_cache;
-                                }
-                                spin_lock(&ctl->tree_lock);
-                                ret = link_free_space(ctl, e);
-                                ctl->total_bitmaps++;
-                                ctl->op->recalc_thresholds(ctl);
-                                spin_unlock(&ctl->tree_lock);
-                                if (ret) {
-                                        printk(KERN_ERR "Duplicate entries in "
-                                               "free space cache, dumping\n");
-                                        kunmap(page);
-                                        unlock_page(page);
-                                        page_cache_release(page);
-                                        goto free_cache;
-                                }
-                                list_add_tail(&e->list, &bitmaps);
-                        }
-
-                        num_entries--;
-                        offset += sizeof(struct btrfs_free_space_entry);
-                        if (offset + sizeof(struct btrfs_free_space_entry) >=
-                            PAGE_CACHE_SIZE)
-                                break;
-                        entry++;
+                        list_add_tail(&e->list, &bitmaps);
                 }
 
-                /*
-                 * We read an entry out of this page, we need to move on to the
-                 * next page.
-                 */
-                if (need_loop) {
-                        kunmap(page);
-                        goto next;
-                }
+                num_entries--;
+        }
 
-                /*
-                 * We add the bitmaps at the end of the entries in order that
-                 * the bitmap entries are added to the cache.
-                 */
-                e = list_entry(bitmaps.next, struct btrfs_free_space, list);
+        /*
+         * We add the bitmaps at the end of the entries in order that
+         * the bitmap entries are added to the cache.
+         */
+        list_for_each_entry_safe(e, n, &bitmaps, list) {
                 list_del_init(&e->list);
-                memcpy(e->bitmap, addr, PAGE_CACHE_SIZE);
-                kunmap(page);
-                num_bitmaps--;
-next:
-                unlock_page(page);
-                page_cache_release(page);
-                index++;
+                ret = io_ctl_read_bitmap(&io_ctl, e);
+                if (ret)
+                        goto free_cache;
         }
 
+        io_ctl_drop_pages(&io_ctl);
         ret = 1;
 out:
+        io_ctl_free(&io_ctl);
         return ret;
 free_cache:
+        io_ctl_drop_pages(&io_ctl);
         __btrfs_remove_free_space_cache(ctl);
         goto out;
 }
@@ -465,7 +728,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
         struct btrfs_root *root = fs_info->tree_root;
         struct inode *inode;
         struct btrfs_path *path;
-        int ret;
+        int ret = 0;
         bool matched;
         u64 used = btrfs_block_group_used(&block_group->item);
 
@@ -497,6 +760,14 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
                 return 0;
         }
 
+        /* We may have converted the inode and made the cache invalid. */
+        spin_lock(&block_group->lock);
+        if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
+                spin_unlock(&block_group->lock);
+                goto out;
+        }
+        spin_unlock(&block_group->lock);
+
         ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
                                       path, block_group->key.objectid);
         btrfs_free_path(path);
@@ -530,6 +801,19 @@ out:
         return ret;
 }
 
+/**
+ * __btrfs_write_out_cache - write out cached info to an inode
+ * @root - the root the inode belongs to
+ * @ctl - the free space cache we are going to write out
+ * @block_group - the block_group for this cache if it belongs to a block_group
+ * @trans - the trans handle
+ * @path - the path to use
+ * @offset - the offset for the key we'll insert
+ *
+ * This function writes out a free space cache struct to disk for quick recovery
+ * on mount.  This will return 0 if it was successfull in writing the cache out,
+ * and -1 if it was not.
+ */
 int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
                             struct btrfs_free_space_ctl *ctl,
                             struct btrfs_block_group_cache *block_group,
@@ -540,42 +824,24 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
         struct extent_buffer *leaf;
         struct rb_node *node;
         struct list_head *pos, *n;
-        struct page **pages;
-        struct page *page;
         struct extent_state *cached_state = NULL;
         struct btrfs_free_cluster *cluster = NULL;
         struct extent_io_tree *unpin = NULL;
+        struct io_ctl io_ctl;
         struct list_head bitmap_list;
         struct btrfs_key key;
         u64 start, end, len;
-        u64 bytes = 0;
-        u32 crc = ~(u32)0;
-        int index = 0, num_pages = 0;
         int entries = 0;
         int bitmaps = 0;
-        int ret = -1;
-        bool next_page = false;
-        bool out_of_space = false;
+        int ret;
+        int err = -1;
 
         INIT_LIST_HEAD(&bitmap_list);
 
-        node = rb_first(&ctl->free_space_offset);
-        if (!node)
-                return 0;
-
         if (!i_size_read(inode))
                 return -1;
 
-        num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
-                PAGE_CACHE_SHIFT;
-
-        filemap_write_and_wait(inode->i_mapping);
-        btrfs_wait_ordered_range(inode, inode->i_size &
-                                 ~(root->sectorsize - 1), (u64)-1);
-
-        pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS);
-        if (!pages)
-                return -1;
+        io_ctl_init(&io_ctl, inode, root);
 
         /* Get the cluster for this block_group if it exists */
         if (block_group && !list_empty(&block_group->cluster_list))
@@ -589,30 +855,9 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
          */
         unpin = root->fs_info->pinned_extents;
 
-        /*
-         * Lock all pages first so we can lock the extent safely.
-         *
-         * NOTE: Because we hold the ref the entire time we're going to write to
-         * the page find_get_page should never fail, so we don't do a check
-         * after find_get_page at this point.  Just putting this here so people
-         * know and don't freak out.
-         */
-        while (index < num_pages) {
-                page = find_or_create_page(inode->i_mapping, index, GFP_NOFS);
-                if (!page) {
-                        int i;
-
-                        for (i = 0; i < num_pages; i++) {
-                                unlock_page(pages[i]);
-                                page_cache_release(pages[i]);
-                        }
-                        goto out;
-                }
-                pages[index] = page;
-                index++;
-        }
+        /* Lock all pages first so we can lock the extent safely. */
+        io_ctl_prepare_pages(&io_ctl, inode, 0);
 
-        index = 0;
         lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
                          0, &cached_state, GFP_NOFS);
 
@@ -623,189 +868,111 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
         if (block_group)
                 start = block_group->key.objectid;
 
-        /* Write out the extent entries */
-        do {
-                struct btrfs_free_space_entry *entry;
-                void *addr, *orig;
-                unsigned long offset = 0;
+        node = rb_first(&ctl->free_space_offset);
+        if (!node && cluster) {
+                node = rb_first(&cluster->root);
+                cluster = NULL;
+        }
 
-                next_page = false;
+        /* Make sure we can fit our crcs into the first page */
+        if (io_ctl.check_crcs &&
+            (io_ctl.num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE) {
+                WARN_ON(1);
+                goto out_nospc;
+        }
 
-                if (index >= num_pages) {
-                        out_of_space = true;
-                        break;
-                }
+        io_ctl_set_generation(&io_ctl, trans->transid);
 
-                page = pages[index];
+        /* Write out the extent entries */
+        while (node) {
+                struct btrfs_free_space *e;
 
-                orig = addr = kmap(page);
-                if (index == 0) {
-                        u64 *gen;
+                e = rb_entry(node, struct btrfs_free_space, offset_index);
+                entries++;
 
-                        /*
-                         * We're going to put in a bogus crc for this page to
-                         * make sure that old kernels who aren't aware of this
-                         * format will be sure to discard the cache.
-                         */
-                        addr += sizeof(u64);
-                        offset += sizeof(u64);
+                ret = io_ctl_add_entry(&io_ctl, e->offset, e->bytes,
+                                       e->bitmap);
+                if (ret)
+                        goto out_nospc;
 
-                        gen = addr;
-                        *gen = trans->transid;
-                        addr += sizeof(u64);
-                        offset += sizeof(u64);
+                if (e->bitmap) {
+                        list_add_tail(&e->list, &bitmap_list);
+                        bitmaps++;
                 }
-                entry = addr;
-
-                memset(addr, 0, PAGE_CACHE_SIZE - offset);
-                while (node && !next_page) {
-                        struct btrfs_free_space *e;
-
-                        e = rb_entry(node, struct btrfs_free_space, offset_index);
-                        entries++;
-
-                        entry->offset = cpu_to_le64(e->offset);
-                        entry->bytes = cpu_to_le64(e->bytes);
-                        if (e->bitmap) {
-                                entry->type = BTRFS_FREE_SPACE_BITMAP;
-                                list_add_tail(&e->list, &bitmap_list);
-                                bitmaps++;
-                        } else {
-                                entry->type = BTRFS_FREE_SPACE_EXTENT;
-                        }
-                        node = rb_next(node);
-                        if (!node && cluster) {
-                                node = rb_first(&cluster->root);
-                                cluster = NULL;
-                        }
-                        offset += sizeof(struct btrfs_free_space_entry);
-                        if (offset + sizeof(struct btrfs_free_space_entry) >=
-                            PAGE_CACHE_SIZE)
-                                next_page = true;
-                        entry++;
+                node = rb_next(node);
+                if (!node && cluster) {
+                        node = rb_first(&cluster->root);
+                        cluster = NULL;
                 }
+        }
 
-                /*
-                 * We want to add any pinned extents to our free space cache
-                 * so we don't leak the space
-                 */
-                while (block_group && !next_page &&
-                       (start < block_group->key.objectid +
-                        block_group->key.offset)) {
-                        ret = find_first_extent_bit(unpin, start, &start, &end,
-                                                    EXTENT_DIRTY);
-                        if (ret) {
-                                ret = 0;
-                                break;
-                        }
-
-                        /* This pinned extent is out of our range */
-                        if (start >= block_group->key.objectid +
-                            block_group->key.offset)
-                                break;
-
-                        len = block_group->key.objectid +
-                                block_group->key.offset - start;
-                        len = min(len, end + 1 - start);
-
-                        entries++;
-                        entry->offset = cpu_to_le64(start);
-                        entry->bytes = cpu_to_le64(len);
-                        entry->type = BTRFS_FREE_SPACE_EXTENT;
-
-                        start = end + 1;
-                        offset += sizeof(struct btrfs_free_space_entry);
-                        if (offset + sizeof(struct btrfs_free_space_entry) >=
-                            PAGE_CACHE_SIZE)
-                                next_page = true;
-                        entry++;
+        /*
+         * We want to add any pinned extents to our free space cache
+         * so we don't leak the space
+         */
+        while (block_group && (start < block_group->key.objectid +
+                               block_group->key.offset)) {
+                ret = find_first_extent_bit(unpin, start, &start, &end,
+                                            EXTENT_DIRTY);
+                if (ret) {
+                        ret = 0;
+                        break;
                 }
 
-                /* Generate bogus crc value */
-                if (index == 0) {
-                        u32 *tmp;
-                        crc = btrfs_csum_data(root, orig + sizeof(u64), crc,
-                                              PAGE_CACHE_SIZE - sizeof(u64));
-                        btrfs_csum_final(crc, (char *)&crc);
-                        crc++;
-                        tmp = orig;
-                        *tmp = crc;
-                }
+                /* This pinned extent is out of our range */
+                if (start >= block_group->key.objectid +
+                    block_group->key.offset)
+                        break;
 
-                kunmap(page);
+                len = block_group->key.objectid +
+                        block_group->key.offset - start;
+                len = min(len, end + 1 - start);
 
-                bytes += PAGE_CACHE_SIZE;
+                entries++;
+                ret = io_ctl_add_entry(&io_ctl, start, len, NULL);
+                if (ret)
+                        goto out_nospc;
 
-                index++;
-        } while (node || next_page);
+                start = end + 1;
+        }
 
         /* Write out the bitmaps */
         list_for_each_safe(pos, n, &bitmap_list) {
-                void *addr;
                 struct btrfs_free_space *entry =
                         list_entry(pos, struct btrfs_free_space, list);
 
-                if (index >= num_pages) {
-                        out_of_space = true;
-                        break;
-                }
-                page = pages[index];
-
-                addr = kmap(page);
-                memcpy(addr, entry->bitmap, PAGE_CACHE_SIZE);
-                kunmap(page);
-                bytes += PAGE_CACHE_SIZE;
-
+                ret = io_ctl_add_bitmap(&io_ctl, entry->bitmap);
+                if (ret)
+                        goto out_nospc;
                 list_del_init(&entry->list);
-                index++;
-        }
-
-        if (out_of_space) {
-                btrfs_drop_pages(pages, num_pages);
-                unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
-                                     i_size_read(inode) - 1, &cached_state,
-                                     GFP_NOFS);
-                ret = 0;
-                goto out;
         }
 
         /* Zero out the rest of the pages just to make sure */
-        while (index < num_pages) {
-                void *addr;
-
-                page = pages[index];
-                addr = kmap(page);
-                memset(addr, 0, PAGE_CACHE_SIZE);
-                kunmap(page);
-                bytes += PAGE_CACHE_SIZE;
-                index++;
-        }
+        io_ctl_zero_remaining_pages(&io_ctl);
 
-        ret = btrfs_dirty_pages(root, inode, pages, num_pages, 0,
-                                            bytes, &cached_state);
-        btrfs_drop_pages(pages, num_pages);
+        ret = btrfs_dirty_pages(root, inode, io_ctl.pages, io_ctl.num_pages,
+                                0, i_size_read(inode), &cached_state);
+        io_ctl_drop_pages(&io_ctl);
         unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
                              i_size_read(inode) - 1, &cached_state, GFP_NOFS);
 
-        if (ret) {
-                ret = 0;
+        if (ret)
                 goto out;
-        }
 
-        BTRFS_I(inode)->generation = trans->transid;
 
-        filemap_write_and_wait(inode->i_mapping);
+        ret = filemap_write_and_wait(inode->i_mapping);
+        if (ret)
+                goto out;
 
         key.objectid = BTRFS_FREE_SPACE_OBJECTID;
         key.offset = offset;
         key.type = 0;
 
-        ret = btrfs_search_slot(trans, root, &key, path, 1, 1);
+        ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
         if (ret < 0) {
-                ret = -1;
-                clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
-                                 EXTENT_DIRTY | EXTENT_DELALLOC |
-                                 EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS);
+                clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
+                                 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
+                                 GFP_NOFS);
                 goto out;
         }
         leaf = path->nodes[0];
@@ -816,15 +983,16 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
                 if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
                     found_key.offset != offset) {
-                        ret = -1;
-                        clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
-                                         EXTENT_DIRTY | EXTENT_DELALLOC |
-                                         EXTENT_DO_ACCOUNTING, 0, 0, NULL,
-                                         GFP_NOFS);
+                        clear_extent_bit(&BTRFS_I(inode)->io_tree, 0,
+                                         inode->i_size - 1,
+                                         EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0,
+                                         NULL, GFP_NOFS);
                         btrfs_release_path(path);
                         goto out;
                 }
         }
+
+        BTRFS_I(inode)->generation = trans->transid;
         header = btrfs_item_ptr(leaf, path->slots[0],
                                 struct btrfs_free_space_header);
         btrfs_set_free_space_entries(leaf, header, entries);
@@ -833,16 +1001,26 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
         btrfs_mark_buffer_dirty(leaf);
         btrfs_release_path(path);
 
-        ret = 1;
-
+        err = 0;
 out:
-        kfree(pages);
-        if (ret != 1) {
-                invalidate_inode_pages2_range(inode->i_mapping, 0, index);
+        io_ctl_free(&io_ctl);
+        if (err) {
+                invalidate_inode_pages2(inode->i_mapping);
                 BTRFS_I(inode)->generation = 0;
         }
         btrfs_update_inode(trans, root, inode);
-        return ret;
+        return err;
+
+out_nospc:
+        list_for_each_safe(pos, n, &bitmap_list) {
+                struct btrfs_free_space *entry =
+                        list_entry(pos, struct btrfs_free_space, list);
+                list_del_init(&entry->list);
+        }
+        io_ctl_drop_pages(&io_ctl);
+        unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
+                             i_size_read(inode) - 1, &cached_state, GFP_NOFS);
+        goto out;
 }
 
 int btrfs_write_out_cache(struct btrfs_root *root,
@@ -869,14 +1047,15 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 
         ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans,
                                       path, block_group->key.objectid);
-        if (ret < 0) {
+        if (ret) {
                 spin_lock(&block_group->lock);
                 block_group->disk_cache_state = BTRFS_DC_ERROR;
                 spin_unlock(&block_group->lock);
                 ret = 0;
-
+#ifdef DEBUG
                 printk(KERN_ERR "btrfs: failed to write free space cace "
                        "for block group %llu\n", block_group->key.objectid);
+#endif
         }
 
         iput(inode);
@@ -1701,6 +1880,7 @@ again:
                         ctl->total_bitmaps--;
                 }
                 kmem_cache_free(btrfs_free_space_cachep, info);
+                ret = 0;
                 goto out_lock;
         }
 
@@ -1708,7 +1888,8 @@ again:
                 unlink_free_space(ctl, info);
                 info->offset += bytes;
                 info->bytes -= bytes;
-                link_free_space(ctl, info);
+                ret = link_free_space(ctl, info);
+                WARN_ON(ret);
                 goto out_lock;
         }
 
@@ -2472,9 +2653,19 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
                 spin_unlock(&ctl->tree_lock);
 
                 if (bytes >= minlen) {
-                        int update_ret;
-                        update_ret = btrfs_update_reserved_bytes(block_group,
-                                                                 bytes, 1, 1);
+                        struct btrfs_space_info *space_info;
+                        int update = 0;
+
+                        space_info = block_group->space_info;
+                        spin_lock(&space_info->lock);
+                        spin_lock(&block_group->lock);
+                        if (!block_group->ro) {
+                                block_group->reserved += bytes;
+                                space_info->bytes_reserved += bytes;
+                                update = 1;
+                        }
+                        spin_unlock(&block_group->lock);
+                        spin_unlock(&space_info->lock);
 
                         ret = btrfs_error_discard_extent(fs_info->extent_root,
                                                          start,
@@ -2482,9 +2673,16 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
                                                          &actually_trimmed);
 
                         btrfs_add_free_space(block_group, start, bytes);
-                        if (!update_ret)
-                                btrfs_update_reserved_bytes(block_group,
-                                                            bytes, 0, 1);
+                        if (update) {
+                                spin_lock(&space_info->lock);
+                                spin_lock(&block_group->lock);
+                                if (block_group->ro)
+                                        space_info->bytes_readonly += bytes;
+                                block_group->reserved -= bytes;
+                                space_info->bytes_reserved -= bytes;
+                                spin_unlock(&space_info->lock);
+                                spin_unlock(&block_group->lock);
+                        }
 
                         if (ret)
                                 break;
@@ -2643,9 +2841,13 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root,
                 return 0;
 
         ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0);
-        if (ret < 0)
+        if (ret) {
+                btrfs_delalloc_release_metadata(inode, inode->i_size);
+#ifdef DEBUG
                 printk(KERN_ERR "btrfs: failed to write free ino cache "
                        "for root %llu\n", root->root_key.objectid);
+#endif
+        }
 
         iput(inode);
         return ret;
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index b4087e0fa871..53dcbdf446cd 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -465,14 +465,16 @@ again:
         /* Just to make sure we have enough space */
         prealloc += 8 * PAGE_CACHE_SIZE;
 
-        ret = btrfs_check_data_free_space(inode, prealloc);
+        ret = btrfs_delalloc_reserve_space(inode, prealloc);
         if (ret)
                 goto out_put;
 
         ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc,
                                               prealloc, prealloc, &alloc_hint);
-        if (ret)
+        if (ret) {
+                btrfs_delalloc_release_space(inode, prealloc);
                 goto out_put;
+        }
         btrfs_free_reserved_data_space(inode, prealloc);
 
 out_put:
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 75686a61bd45..966ddcc4c63d 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -45,10 +45,10 @@
 #include "btrfs_inode.h"
 #include "ioctl.h"
 #include "print-tree.h"
-#include "volumes.h"
 #include "ordered-data.h"
 #include "xattr.h"
 #include "tree-log.h"
+#include "volumes.h"
 #include "compression.h"
 #include "locking.h"
 #include "free-space-cache.h"
@@ -393,7 +393,10 @@ again:
              (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) {
                 WARN_ON(pages);
                 pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS);
-                BUG_ON(!pages);
+                if (!pages) {
+                        /* just bail out to the uncompressed code */
+                        goto cont;
+                }
 
                 if (BTRFS_I(inode)->force_compress)
                         compress_type = BTRFS_I(inode)->force_compress;
@@ -424,6 +427,7 @@ again:
                         will_compress = 1;
                 }
         }
+cont:
         if (start == 0) {
                 trans = btrfs_join_transaction(root);
                 BUG_ON(IS_ERR(trans));
@@ -820,7 +824,7 @@ static noinline int cow_file_range(struct inode *inode,
         }
 
         BUG_ON(disk_num_bytes >
-               btrfs_super_total_bytes(&root->fs_info->super_copy));
+               btrfs_super_total_bytes(root->fs_info->super_copy));
 
         alloc_hint = get_extent_allocation_hint(inode, start, num_bytes);
         btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
@@ -1792,12 +1796,12 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
         }
         ret = 0;
 out:
-        if (nolock) {
-                if (trans)
-                        btrfs_end_transaction_nolock(trans, root);
-        } else {
+        if (root != root->fs_info->tree_root)
                 btrfs_delalloc_release_metadata(inode, ordered_extent->len);
-                if (trans)
+        if (trans) {
+                if (nolock)
+                        btrfs_end_transaction_nolock(trans, root);
+                else
                         btrfs_end_transaction(trans, root);
         }
 
@@ -1819,153 +1823,9 @@ static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
 }
 
 /*
- * When IO fails, either with EIO or csum verification fails, we
- * try other mirrors that might have a good copy of the data.  This
- * io_failure_record is used to record state as we go through all the
- * mirrors.  If another mirror has good data, the page is set up to date
- * and things continue.  If a good mirror can't be found, the original
- * bio end_io callback is called to indicate things have failed.
- */
-struct io_failure_record {
-        struct page *page;
-        u64 start;
-        u64 len;
-        u64 logical;
-        unsigned long bio_flags;
-        int last_mirror;
-};
-
-static int btrfs_io_failed_hook(struct bio *failed_bio,
-                         struct page *page, u64 start, u64 end,
-                         struct extent_state *state)
-{
-        struct io_failure_record *failrec = NULL;
-        u64 private;
-        struct extent_map *em;
-        struct inode *inode = page->mapping->host;
-        struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree;
-        struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
-        struct bio *bio;
-        int num_copies;
-        int ret;
-        int rw;
-        u64 logical;
-
-        ret = get_state_private(failure_tree, start, &private);
-        if (ret) {
-                failrec = kmalloc(sizeof(*failrec), GFP_NOFS);
-                if (!failrec)
-                        return -ENOMEM;
-                failrec->start = start;
-                failrec->len = end - start + 1;
-                failrec->last_mirror = 0;
-                failrec->bio_flags = 0;
-
-                read_lock(&em_tree->lock);
-                em = lookup_extent_mapping(em_tree, start, failrec->len);
-                if (em->start > start || em->start + em->len < start) {
-                        free_extent_map(em);
-                        em = NULL;
-                }
-                read_unlock(&em_tree->lock);
-
-                if (IS_ERR_OR_NULL(em)) {
-                        kfree(failrec);
-                        return -EIO;
-                }
-                logical = start - em->start;
-                logical = em->block_start + logical;
-                if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
-                        logical = em->block_start;
-                        failrec->bio_flags = EXTENT_BIO_COMPRESSED;
-                        extent_set_compress_type(&failrec->bio_flags,
-                                                 em->compress_type);
-                }
-                failrec->logical = logical;
-                free_extent_map(em);
-                set_extent_bits(failure_tree, start, end, EXTENT_LOCKED |
-                                EXTENT_DIRTY, GFP_NOFS);
-                set_state_private(failure_tree, start,
-                                 (u64)(unsigned long)failrec);
-        } else {
-                failrec = (struct io_failure_record *)(unsigned long)private;
-        }
-        num_copies = btrfs_num_copies(
-                              &BTRFS_I(inode)->root->fs_info->mapping_tree,
-                              failrec->logical, failrec->len);
-        failrec->last_mirror++;
-        if (!state) {
-                spin_lock(&BTRFS_I(inode)->io_tree.lock);
-                state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree,
-                                                    failrec->start,
-                                                    EXTENT_LOCKED);
-                if (state && state->start != failrec->start)
-                        state = NULL;
-                spin_unlock(&BTRFS_I(inode)->io_tree.lock);
-        }
-        if (!state || failrec->last_mirror > num_copies) {
-                set_state_private(failure_tree, failrec->start, 0);
-                clear_extent_bits(failure_tree, failrec->start,
-                                  failrec->start + failrec->len - 1,
-                                  EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS);
-                kfree(failrec);
-                return -EIO;
-        }
-        bio = bio_alloc(GFP_NOFS, 1);
-        bio->bi_private = state;
-        bio->bi_end_io = failed_bio->bi_end_io;
-        bio->bi_sector = failrec->logical >> 9;
-        bio->bi_bdev = failed_bio->bi_bdev;
-        bio->bi_size = 0;
-
-        bio_add_page(bio, page, failrec->len, start - page_offset(page));
-        if (failed_bio->bi_rw & REQ_WRITE)
-                rw = WRITE;
-        else
-                rw = READ;
-
-        ret = BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio,
-                                                      failrec->last_mirror,
-                                                      failrec->bio_flags, 0);
-        return ret;
-}
-
-/*
- * each time an IO finishes, we do a fast check in the IO failure tree
- * to see if we need to process or clean up an io_failure_record
- */
-static int btrfs_clean_io_failures(struct inode *inode, u64 start)
-{
-        u64 private;
-        u64 private_failure;
-        struct io_failure_record *failure;
-        int ret;
-
-        private = 0;
-        if (count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private,
-                             (u64)-1, 1, EXTENT_DIRTY, 0)) {
-                ret = get_state_private(&BTRFS_I(inode)->io_failure_tree,
-                                        start, &private_failure);
-                if (ret == 0) {
-                        failure = (struct io_failure_record *)(unsigned long)
-                                   private_failure;
-                        set_state_private(&BTRFS_I(inode)->io_failure_tree,
-                                          failure->start, 0);
-                        clear_extent_bits(&BTRFS_I(inode)->io_failure_tree,
-                                          failure->start,
-                                          failure->start + failure->len - 1,
-                                          EXTENT_DIRTY | EXTENT_LOCKED,
-                                          GFP_NOFS);
-                        kfree(failure);
-                }
-        }
-        return 0;
-}
-
-/*
  * when reads are done, we need to check csums to verify the data is correct
- * if there's a match, we allow the bio to finish.  If not, we go through
- * the io_failure_record routines to find good copies
+ * if there's a match, we allow the bio to finish.  If not, the code in
+ * extent_io.c will try to find good copies for us.
  */
 static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
                                struct extent_state *state)
@@ -2011,10 +1871,6 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 
         kunmap_atomic(kaddr, KM_USER0);
 good:
-        /* if the io failure tree for this inode is non-empty,
-         * check to see if we've recovered from a failed IO
-         */
-        btrfs_clean_io_failures(inode, start);
         return 0;
 
 zeroit:
@@ -2079,89 +1935,6 @@ void btrfs_run_delayed_iputs(struct btrfs_root *root)
         up_read(&root->fs_info->cleanup_work_sem);
 }
 
-/*
- * calculate extra metadata reservation when snapshotting a subvolume
- * contains orphan files.
- */
-void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans,
-                                struct btrfs_pending_snapshot *pending,
-                                u64 *bytes_to_reserve)
-{
-        struct btrfs_root *root;
-        struct btrfs_block_rsv *block_rsv;
-        u64 num_bytes;
-        int index;
-
-        root = pending->root;
-        if (!root->orphan_block_rsv || list_empty(&root->orphan_list))
-                return;
-
-        block_rsv = root->orphan_block_rsv;
-
-        /* orphan block reservation for the snapshot */
-        num_bytes = block_rsv->size;
-
-        /*
-         * after the snapshot is created, COWing tree blocks may use more
-         * space than it frees. So we should make sure there is enough
-         * reserved space.
-         */
-        index = trans->transid & 0x1;
-        if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) {
-                num_bytes += block_rsv->size -
-                             (block_rsv->reserved + block_rsv->freed[index]);
-        }
-
-        *bytes_to_reserve += num_bytes;
-}
-
-void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans,
-                                struct btrfs_pending_snapshot *pending)
-{
-        struct btrfs_root *root = pending->root;
-        struct btrfs_root *snap = pending->snap;
-        struct btrfs_block_rsv *block_rsv;
-        u64 num_bytes;
-        int index;
-        int ret;
-
-        if (!root->orphan_block_rsv || list_empty(&root->orphan_list))
-                return;
-
-        /* refill source subvolume's orphan block reservation */
-        block_rsv = root->orphan_block_rsv;
-        index = trans->transid & 0x1;
-        if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) {
-                num_bytes = block_rsv->size -
-                            (block_rsv->reserved + block_rsv->freed[index]);
-                ret = btrfs_block_rsv_migrate(&pending->block_rsv,
-                                              root->orphan_block_rsv,
-                                              num_bytes);
-                BUG_ON(ret);
-        }
-
-        /* setup orphan block reservation for the snapshot */
-        block_rsv = btrfs_alloc_block_rsv(snap);
-        BUG_ON(!block_rsv);
-
-        btrfs_add_durable_block_rsv(root->fs_info, block_rsv);
-        snap->orphan_block_rsv = block_rsv;
-
-        num_bytes = root->orphan_block_rsv->size;
-        ret = btrfs_block_rsv_migrate(&pending->block_rsv,
-                                      block_rsv, num_bytes);
-        BUG_ON(ret);
-
-#if 0
-        /* insert orphan item for the snapshot */
-        WARN_ON(!root->orphan_item_inserted);
-        ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root,
-                                       snap->root_key.objectid);
-        BUG_ON(ret);
-        snap->orphan_item_inserted = 1;
-#endif
-}
-
 enum btrfs_orphan_cleanup_state {
         ORPHAN_CLEANUP_STARTED  = 1,
         ORPHAN_CLEANUP_DONE     = 2,
@@ -2247,9 +2020,6 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
         }
         spin_unlock(&root->orphan_lock);
 
-        if (block_rsv)
-                btrfs_add_durable_block_rsv(root->fs_info, block_rsv);
-
         /* grab metadata reservation from transaction handle */
         if (reserve) {
                 ret = btrfs_orphan_reserve_metadata(trans, inode);
@@ -2316,6 +2086,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
         struct btrfs_key key, found_key;
         struct btrfs_trans_handle *trans;
         struct inode *inode;
+        u64 last_objectid = 0;
         int ret = 0, nr_unlink = 0, nr_truncate = 0;
 
         if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED))
@@ -2367,41 +2138,49 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
                  * crossing root thing.  we store the inode number in the
                  * offset of the orphan item.
                  */
+
+                if (found_key.offset == last_objectid) {
+                        printk(KERN_ERR "btrfs: Error removing orphan entry, "
+                               "stopping orphan cleanup\n");
+                        ret = -EINVAL;
+                        goto out;
+                }
+
+                last_objectid = found_key.offset;
+
                 found_key.objectid = found_key.offset;
                 found_key.type = BTRFS_INODE_ITEM_KEY;
                 found_key.offset = 0;
                 inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL);
-                if (IS_ERR(inode)) {
-                        ret = PTR_ERR(inode);
+                ret = PTR_RET(inode);
+                if (ret && ret != -ESTALE)
                         goto out;
-                }
 
                 /*
-                 * add this inode to the orphan list so btrfs_orphan_del does
-                 * the proper thing when we hit it
+                 * Inode is already gone but the orphan item is still there,
+                 * kill the orphan item.
                  */
-                spin_lock(&root->orphan_lock);
-                list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
-                spin_unlock(&root->orphan_lock);
-
-                /*
-                 * if this is a bad inode, means we actually succeeded in
-                 * removing the inode, but not the orphan record, which means
-                 * we need to manually delete the orphan since iput will just
-                 * do a destroy_inode
-                 */
-                if (is_bad_inode(inode)) {
-                        trans = btrfs_start_transaction(root, 0);
+                if (ret == -ESTALE) {
+                        trans = btrfs_start_transaction(root, 1);
                         if (IS_ERR(trans)) {
                                 ret = PTR_ERR(trans);
                                 goto out;
                         }
-                        btrfs_orphan_del(trans, inode);
+                        ret = btrfs_del_orphan_item(trans, root,
+                                                    found_key.objectid);
+                        BUG_ON(ret);
                         btrfs_end_transaction(trans, root);
-                        iput(inode);
                         continue;
                 }
 
+                /*
+                 * add this inode to the orphan list so btrfs_orphan_del does
+                 * the proper thing when we hit it
+                 */
+                spin_lock(&root->orphan_lock);
+                list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+                spin_unlock(&root->orphan_lock);
+
                 /* if we have links, this was a truncate, lets do that */
                 if (inode->i_nlink) {
                         if (!S_ISREG(inode->i_mode)) {
@@ -2835,7 +2614,16 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
         u64 ino = btrfs_ino(inode);
         u64 dir_ino = btrfs_ino(dir);
 
-        trans = btrfs_start_transaction(root, 10);
+        /*
+         * 1 for the possible orphan item
+         * 1 for the dir item
+         * 1 for the dir index
+         * 1 for the inode ref
+         * 1 for the inode ref in the tree log
+         * 2 for the dir entries in the log
+         * 1 for the inode
+         */
+        trans = btrfs_start_transaction(root, 8);
         if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
                 return trans;
 
@@ -2858,7 +2646,8 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
                 return ERR_PTR(-ENOMEM);
         }
 
-        trans = btrfs_start_transaction(root, 0);
+        /* 1 for the orphan item */
+        trans = btrfs_start_transaction(root, 1);
         if (IS_ERR(trans)) {
                 btrfs_free_path(path);
                 root->fs_info->enospc_unlink = 0;
@@ -2963,6 +2752,12 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
         err = 0;
 out:
         btrfs_free_path(path);
+        /* Migrate the orphan reservation over */
+        if (!err)
+                err = btrfs_block_rsv_migrate(trans->block_rsv,
+                                &root->fs_info->global_block_rsv,
+                                trans->bytes_reserved);
+
         if (err) {
                 btrfs_end_transaction(trans, root);
                 root->fs_info->enospc_unlink = 0;
@@ -2977,6 +2772,9 @@ static void __unlink_end_trans(struct btrfs_trans_handle *trans,
                                struct btrfs_root *root)
 {
         if (trans->block_rsv == &root->fs_info->global_block_rsv) {
+                btrfs_block_rsv_release(root, trans->block_rsv,
+                                        trans->bytes_reserved);
+                trans->block_rsv = &root->fs_info->trans_block_rsv;
                 BUG_ON(!root->fs_info->enospc_unlink);
                 root->fs_info->enospc_unlink = 0;
         }
@@ -3368,6 +3166,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
         pgoff_t index = from >> PAGE_CACHE_SHIFT;
         unsigned offset = from & (PAGE_CACHE_SIZE-1);
         struct page *page;
+        gfp_t mask = btrfs_alloc_write_mask(mapping);
         int ret = 0;
         u64 page_start;
         u64 page_end;
@@ -3380,7 +3179,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
 
         ret = -ENOMEM;
 again:
-        page = find_or_create_page(mapping, index, GFP_NOFS);
+        page = find_or_create_page(mapping, index, mask);
         if (!page) {
                 btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
                 goto out;
@@ -3613,6 +3412,8 @@ void btrfs_evict_inode(struct inode *inode)
 {
         struct btrfs_trans_handle *trans;
         struct btrfs_root *root = BTRFS_I(inode)->root;
+        struct btrfs_block_rsv *rsv, *global_rsv;
+        u64 min_size = btrfs_calc_trunc_metadata_size(root, 1);
         unsigned long nr;
         int ret;
 
@@ -3640,22 +3441,55 @@ void btrfs_evict_inode(struct inode *inode)
                 goto no_delete;
         }
 
+        rsv = btrfs_alloc_block_rsv(root);
+        if (!rsv) {
+                btrfs_orphan_del(NULL, inode);
+                goto no_delete;
+        }
+        rsv->size = min_size;
+        global_rsv = &root->fs_info->global_block_rsv;
+
         btrfs_i_size_write(inode, 0);
 
+        /*
+         * This is a bit simpler than btrfs_truncate since
+         *
+         * 1) We've already reserved our space for our orphan item in the
+         *    unlink.
+         * 2) We're going to delete the inode item, so we don't need to update
+         *    it at all.
+         *
+         * So we just need to reserve some slack space in case we add bytes when
+         * doing the truncate.
+         */
         while (1) {
-                trans = btrfs_join_transaction(root);
-                BUG_ON(IS_ERR(trans));
-                trans->block_rsv = root->orphan_block_rsv;
+                ret = btrfs_block_rsv_refill(root, rsv, min_size);
+
+                /*
+                 * Try and steal from the global reserve since we will
+                 * likely not use this space anyway, we want to try as
+                 * hard as possible to get this to work.
+                 */
+                if (ret)
+                        ret = btrfs_block_rsv_migrate(global_rsv, rsv, min_size);
 
-                ret = btrfs_block_rsv_check(trans, root,
-                                            root->orphan_block_rsv, 0, 5);
                 if (ret) {
-                        BUG_ON(ret != -EAGAIN);
-                        ret = btrfs_commit_transaction(trans, root);
-                        BUG_ON(ret);
-                        continue;
+                        printk(KERN_WARNING "Could not get space for a "
+                               "delete, will truncate on mount %d\n", ret);
+                        btrfs_orphan_del(NULL, inode);
+                        btrfs_free_block_rsv(root, rsv);
+                        goto no_delete;
                 }
 
+                trans = btrfs_start_transaction(root, 0);
+                if (IS_ERR(trans)) {
+                        btrfs_orphan_del(NULL, inode);
+                        btrfs_free_block_rsv(root, rsv);
+                        goto no_delete;
+                }
+
+                trans->block_rsv = rsv;
+
                 ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
                 if (ret != -EAGAIN)
                         break;
@@ -3664,14 +3498,17 @@ void btrfs_evict_inode(struct inode *inode)
                 btrfs_end_transaction(trans, root);
                 trans = NULL;
                 btrfs_btree_balance_dirty(root, nr);
-
         }
 
+        btrfs_free_block_rsv(root, rsv);
+
         if (ret == 0) {
+                trans->block_rsv = root->orphan_block_rsv;
                 ret = btrfs_orphan_del(trans, inode);
                 BUG_ON(ret);
         }
 
+        trans->block_rsv = &root->fs_info->trans_block_rsv;
         if (!(root == root->fs_info->tree_root ||
               root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID))
                 btrfs_return_ino(root, btrfs_ino(inode));
@@ -5795,8 +5632,7 @@ again:
         if (test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) {
                 ret = btrfs_ordered_update_i_size(inode, 0, ordered);
                 if (!ret)
-                        ret = btrfs_update_inode(trans, root, inode);
-                err = ret;
+                        err = btrfs_update_inode(trans, root, inode);
                 goto out;
         }
 
@@ -6289,7 +6125,7 @@ int btrfs_readpage(struct file *file, struct page *page)
 {
         struct extent_io_tree *tree;
         tree = &BTRFS_I(page->mapping->host)->io_tree;
-        return extent_read_full_page(tree, page, btrfs_get_extent);
+        return extent_read_full_page(tree, page, btrfs_get_extent, 0);
 }
 
 static int btrfs_writepage(struct page *page, struct writeback_control *wbc)
@@ -6541,6 +6377,7 @@ static int btrfs_truncate(struct inode *inode)
         struct btrfs_trans_handle *trans;
         unsigned long nr;
         u64 mask = root->sectorsize - 1;
+        u64 min_size = btrfs_calc_trunc_metadata_size(root, 1);
 
         ret = btrfs_truncate_page(inode->i_mapping, inode->i_size);
         if (ret)
@@ -6588,19 +6425,23 @@ static int btrfs_truncate(struct inode *inode)
         rsv = btrfs_alloc_block_rsv(root);
         if (!rsv)
                 return -ENOMEM;
-        btrfs_add_durable_block_rsv(root->fs_info, rsv);
+        rsv->size = min_size;
 
+        /*
+         * 1 for the truncate slack space
+         * 1 for the orphan item we're going to add
+         * 1 for the orphan item deletion
+         * 1 for updating the inode.
+         */
         trans = btrfs_start_transaction(root, 4);
         if (IS_ERR(trans)) {
                 err = PTR_ERR(trans);
                 goto out;
         }
 
-        /*
-         * Reserve space for the truncate process.  Truncate should be adding
-         * space, but if there are snapshots it may end up using space.
-         */
-        ret = btrfs_truncate_reserve_metadata(trans, root, rsv);
+        /* Migrate the slack space for the truncate to our reserve */
+        ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv,
+                                      min_size);
         BUG_ON(ret);
 
         ret = btrfs_orphan_add(trans, inode);
@@ -6609,21 +6450,6 @@ static int btrfs_truncate(struct inode *inode)
                 goto out;
         }
 
-        nr = trans->blocks_used;
-        btrfs_end_transaction(trans, root);
-        btrfs_btree_balance_dirty(root, nr);
-
-        /*
-         * Ok so we've already migrated our bytes over for the truncate, so here
-         * just reserve the one slot we need for updating the inode.
-         */
-        trans = btrfs_start_transaction(root, 1);
-        if (IS_ERR(trans)) {
-                err = PTR_ERR(trans);
-                goto out;
-        }
-        trans->block_rsv = rsv;
-
         /*
          * setattr is responsible for setting the ordered_data_close flag,
          * but that is only tested during the last file release.  That
@@ -6645,20 +6471,30 @@ static int btrfs_truncate(struct inode *inode)
                 btrfs_add_ordered_operation(trans, root, inode);
 
         while (1) {
+                ret = btrfs_block_rsv_refill(root, rsv, min_size);
+                if (ret) {
+                        /*
+                         * This can only happen with the original transaction we
+                         * started above, every other time we shouldn't have a
+                         * transaction started yet.
+                         */
+                        if (ret == -EAGAIN)
+                                goto end_trans;
+                        err = ret;
+                        break;
+                }
+
                 if (!trans) {
-                        trans = btrfs_start_transaction(root, 3);
+                        /* Just need the 1 for updating the inode */
+                        trans = btrfs_start_transaction(root, 1);
                         if (IS_ERR(trans)) {
                                 err = PTR_ERR(trans);
                                 goto out;
                         }
-
-                        ret = btrfs_truncate_reserve_metadata(trans, root,
-                                                              rsv);
-                        BUG_ON(ret);
-
-                        trans->block_rsv = rsv;
                 }
 
+                trans->block_rsv = rsv;
+
                 ret = btrfs_truncate_inode_items(trans, root, inode,
                                                  inode->i_size,
                                                  BTRFS_EXTENT_DATA_KEY);
@@ -6673,7 +6509,7 @@ static int btrfs_truncate(struct inode *inode)
                         err = ret;
                         break;
                 }
-
+end_trans:
                 nr = trans->blocks_used;
                 btrfs_end_transaction(trans, root);
                 trans = NULL;
@@ -6755,9 +6591,9 @@ struct inode *btrfs_alloc_inode(struct super_block *sb)
         ei->last_sub_trans = 0;
         ei->logged_trans = 0;
         ei->delalloc_bytes = 0;
-        ei->reserved_bytes = 0;
         ei->disk_i_size = 0;
         ei->flags = 0;
+        ei->csum_bytes = 0;
         ei->index_cnt = (u64)-1;
         ei->last_unlink_trans = 0;
 
@@ -6803,6 +6639,8 @@ void btrfs_destroy_inode(struct inode *inode)
         WARN_ON(inode->i_data.nrpages);
         WARN_ON(BTRFS_I(inode)->outstanding_extents);
         WARN_ON(BTRFS_I(inode)->reserved_extents);
+        WARN_ON(BTRFS_I(inode)->delalloc_bytes);
+        WARN_ON(BTRFS_I(inode)->csum_bytes);
 
         /*
          * This can happen where we create an inode, but somebody else also
@@ -7420,7 +7258,6 @@ static struct extent_io_ops btrfs_extent_io_ops = {
         .readpage_end_io_hook = btrfs_readpage_end_io_hook,
         .writepage_end_io_hook = btrfs_writepage_end_io_hook,
         .writepage_start_hook = btrfs_writepage_start_hook,
-        .readpage_io_failed_hook = btrfs_io_failed_hook,
         .set_bit_hook = btrfs_set_bit_hook,
         .clear_bit_hook = btrfs_clear_bit_hook,
         .merge_extent_hook = btrfs_merge_extent_hook,
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index dae5dfe41ba5..4a34c472f126 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -51,6 +51,7 @@
 #include "volumes.h"
 #include "locking.h"
 #include "inode-map.h"
+#include "backref.h"
 
 /* Mask out flags that are inappropriate for the given type of inode. */
 static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
@@ -117,7 +118,7 @@ void btrfs_update_iflags(struct inode *inode)
 /*
  * Inherit flags from the parent inode.
  *
- * Unlike extN we don't have any flags we don't want to inherit currently.
+ * Currently only the compression flags and the cow flags are inherited.
  */
 void btrfs_inherit_iflags(struct inode *inode, struct inode *dir)
 {
@@ -128,12 +129,17 @@ void btrfs_inherit_iflags(struct inode *inode, struct inode *dir)
 
         flags = BTRFS_I(dir)->flags;
 
-        if (S_ISREG(inode->i_mode))
-                flags &= ~BTRFS_INODE_DIRSYNC;
-        else if (!S_ISDIR(inode->i_mode))
-                flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME);
+        if (flags & BTRFS_INODE_NOCOMPRESS) {
+                BTRFS_I(inode)->flags &= ~BTRFS_INODE_COMPRESS;
+                BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
+        } else if (flags & BTRFS_INODE_COMPRESS) {
+                BTRFS_I(inode)->flags &= ~BTRFS_INODE_NOCOMPRESS;
+                BTRFS_I(inode)->flags |= BTRFS_INODE_COMPRESS;
+        }
+
+        if (flags & BTRFS_INODE_NODATACOW)
+                BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW;
 
-        BTRFS_I(inode)->flags = flags;
         btrfs_update_iflags(inode);
 }
 
@@ -277,6 +283,7 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg)
         struct fstrim_range range;
         u64 minlen = ULLONG_MAX;
         u64 num_devices = 0;
+        u64 total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy);
         int ret;
 
         if (!capable(CAP_SYS_ADMIN))
@@ -295,12 +302,15 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg)
                 }
         }
         rcu_read_unlock();
+
         if (!num_devices)
                 return -EOPNOTSUPP;
-
         if (copy_from_user(&range, arg, sizeof(range)))
                 return -EFAULT;
+        if (range.start > total_bytes)
+                return -EINVAL;
 
+        range.len = min(range.len, total_bytes - range.start);
         range.minlen = max(range.minlen, minlen);
         ret = btrfs_trim_fs(root, &range);
         if (ret < 0)
@@ -760,7 +770,7 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len,
         int ret = 1;
 
         /*
-         * make sure that once we start defragging and extent, we keep on
+         * make sure that once we start defragging an extent, we keep on
          * defragging it
          */
         if (start < *defrag_end)
@@ -805,7 +815,6 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len,
          * extent will force at least part of that big extent to be defragged.
          */
         if (ret) {
-                *last_len += len;
                 *defrag_end = extent_map_end(em);
         } else {
                 *last_len = 0;
@@ -843,6 +852,7 @@ static int cluster_pages_for_defrag(struct inode *inode,
         int i_done;
         struct btrfs_ordered_extent *ordered;
         struct extent_state *cached_state = NULL;
+        gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
 
         if (isize == 0)
                 return 0;
@@ -860,7 +870,7 @@ again:
         for (i = 0; i < num_pages; i++) {
                 struct page *page;
                 page = find_or_create_page(inode->i_mapping,
-                                            start_index + i, GFP_NOFS);
+                                            start_index + i, mask);
                 if (!page)
                         break;
 
@@ -972,18 +982,20 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
         struct btrfs_super_block *disk_super;
         struct file_ra_state *ra = NULL;
         unsigned long last_index;
+        u64 isize = i_size_read(inode);
         u64 features;
         u64 last_len = 0;
         u64 skip = 0;
         u64 defrag_end = 0;
         u64 newer_off = range->start;
-        int newer_left = 0;
         unsigned long i;
+        unsigned long ra_index = 0;
         int ret;
         int defrag_count = 0;
         int compress_type = BTRFS_COMPRESS_ZLIB;
         int extent_thresh = range->extent_thresh;
-        int newer_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT;
+        int max_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT;
+        int cluster = max_cluster;
         u64 new_align = ~((u64)128 * 1024 - 1);
         struct page **pages = NULL;
 
@@ -997,7 +1009,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
                         compress_type = range->compress_type;
         }
 
-        if (inode->i_size == 0)
+        if (isize == 0)
                 return 0;
 
         /*
@@ -1013,7 +1025,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
                 ra = &file->f_ra;
         }
 
-        pages = kmalloc(sizeof(struct page *) * newer_cluster,
+        pages = kmalloc(sizeof(struct page *) * max_cluster,
                         GFP_NOFS);
         if (!pages) {
                 ret = -ENOMEM;
@@ -1022,10 +1034,10 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 
         /* find the last page to defrag */
         if (range->start + range->len > range->start) {
-                last_index = min_t(u64, inode->i_size - 1,
+                last_index = min_t(u64, isize - 1,
                          range->start + range->len - 1) >> PAGE_CACHE_SHIFT;
         } else {
-                last_index = (inode->i_size - 1) >> PAGE_CACHE_SHIFT;
+                last_index = (isize - 1) >> PAGE_CACHE_SHIFT;
         }
 
         if (newer_than) {
@@ -1038,14 +1050,13 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
                          * the extents in the file evenly spaced
                          */
                         i = (newer_off & new_align) >> PAGE_CACHE_SHIFT;
-                        newer_left = newer_cluster;
                 } else
                         goto out_ra;
         } else {
                 i = range->start >> PAGE_CACHE_SHIFT;
         }
         if (!max_to_defrag)
-                max_to_defrag = last_index - 1;
+                max_to_defrag = last_index;
 
         /*
          * make writeback starts from i, so the defrag range can be
@@ -1079,18 +1090,31 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
                         i = max(i + 1, next);
                         continue;
                 }
+
+                if (!newer_than) {
+                        cluster = (PAGE_CACHE_ALIGN(defrag_end) >>
+                                   PAGE_CACHE_SHIFT) - i;
+                        cluster = min(cluster, max_cluster);
+                } else {
+                        cluster = max_cluster;
+                }
+
                 if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
                         BTRFS_I(inode)->force_compress = compress_type;
 
-                btrfs_force_ra(inode->i_mapping, ra, file, i, newer_cluster);
+                if (i + cluster > ra_index) {
+                        ra_index = max(i, ra_index);
+                        btrfs_force_ra(inode->i_mapping, ra, file, ra_index,
+                                       cluster);
+                        ra_index += max_cluster;
+                }
 
-                ret = cluster_pages_for_defrag(inode, pages, i, newer_cluster);
+                ret = cluster_pages_for_defrag(inode, pages, i, cluster);
                 if (ret < 0)
                         goto out_ra;
 
                 defrag_count += ret;
                 balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret);
-                i += ret;
 
                 if (newer_than) {
                         if (newer_off == (u64)-1)
@@ -1105,12 +1129,17 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
                         if (!ret) {
                                 range->start = newer_off;
                                 i = (newer_off & new_align) >> PAGE_CACHE_SHIFT;
-                                newer_left = newer_cluster;
                         } else {
                                 break;
                         }
                 } else {
-                        i++;
+                        if (ret > 0) {
+                                i += ret;
+                                last_len += ret << PAGE_CACHE_SHIFT;
+                        } else {
+                                i++;
+                                last_len = 0;
+                        }
                 }
         }
 
@@ -1136,16 +1165,14 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
                 mutex_unlock(&inode->i_mutex);
         }
 
-        disk_super = &root->fs_info->super_copy;
+        disk_super = root->fs_info->super_copy;
         features = btrfs_super_incompat_flags(disk_super);
         if (range->compress_type == BTRFS_COMPRESS_LZO) {
                 features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO;
                 btrfs_set_super_incompat_flags(disk_super, features);
         }
 
-        if (!file)
-                kfree(ra);
-        return defrag_count;
+        ret = defrag_count;
 
 out_ra:
         if (!file)
@@ -2587,7 +2614,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp)
                 return PTR_ERR(trans);
         }
 
-        dir_id = btrfs_super_root_dir(&root->fs_info->super_copy);
+        dir_id = btrfs_super_root_dir(root->fs_info->super_copy);
         di = btrfs_lookup_dir_item(trans, root->fs_info->tree_root, path,
                                    dir_id, "default", 7, 1);
         if (IS_ERR_OR_NULL(di)) {
@@ -2603,7 +2630,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp)
         btrfs_mark_buffer_dirty(path->nodes[0]);
         btrfs_free_path(path);
 
-        disk_super = &root->fs_info->super_copy;
+        disk_super = root->fs_info->super_copy;
         features = btrfs_super_incompat_flags(disk_super);
         if (!(features & BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL)) {
                 features |= BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL;
@@ -2864,6 +2891,144 @@ static long btrfs_ioctl_scrub_progress(struct btrfs_root *root,
         return ret;
 }
 
+static long btrfs_ioctl_ino_to_path(struct btrfs_root *root, void __user *arg)
+{
+        int ret = 0;
+        int i;
+        u64 rel_ptr;
+        int size;
+        struct btrfs_ioctl_ino_path_args *ipa = NULL;
+        struct inode_fs_paths *ipath = NULL;
+        struct btrfs_path *path;
+
+        if (!capable(CAP_SYS_ADMIN))
+                return -EPERM;
+
+        path = btrfs_alloc_path();
+        if (!path) {
+                ret = -ENOMEM;
+                goto out;
+        }
+
+        ipa = memdup_user(arg, sizeof(*ipa));
+        if (IS_ERR(ipa)) {
+                ret = PTR_ERR(ipa);
+                ipa = NULL;
+                goto out;
+        }
+
+        size = min_t(u32, ipa->size, 4096);
+        ipath = init_ipath(size, root, path);
+        if (IS_ERR(ipath)) {
+                ret = PTR_ERR(ipath);
+                ipath = NULL;
+                goto out;
+        }
+
+        ret = paths_from_inode(ipa->inum, ipath);
+        if (ret < 0)
+                goto out;
+
+        for (i = 0; i < ipath->fspath->elem_cnt; ++i) {
+                rel_ptr = ipath->fspath->val[i] - (u64)ipath->fspath->val;
+                ipath->fspath->val[i] = rel_ptr;
+        }
+
+        ret = copy_to_user((void *)ipa->fspath, (void *)ipath->fspath, size);
+        if (ret) {
+                ret = -EFAULT;
+                goto out;
+        }
+
+out:
+        btrfs_free_path(path);
+        free_ipath(ipath);
+        kfree(ipa);
+
+        return ret;
+}
+
+static int build_ino_list(u64 inum, u64 offset, u64 root, void *ctx)
+{
+        struct btrfs_data_container *inodes = ctx;
+        const size_t c = 3 * sizeof(u64);
+
+        if (inodes->bytes_left >= c) {
+                inodes->bytes_left -= c;
+                inodes->val[inodes->elem_cnt] = inum;
+                inodes->val[inodes->elem_cnt + 1] = offset;
+                inodes->val[inodes->elem_cnt + 2] = root;
+                inodes->elem_cnt += 3;
+        } else {
+                inodes->bytes_missing += c - inodes->bytes_left;
+                inodes->bytes_left = 0;
+                inodes->elem_missed += 3;
+        }
+
+        return 0;
+}
+
+static long btrfs_ioctl_logical_to_ino(struct btrfs_root *root,
+                                        void __user *arg)
+{
+        int ret = 0;
+        int size;
+        u64 extent_offset;
+        struct btrfs_ioctl_logical_ino_args *loi;
+        struct btrfs_data_container *inodes = NULL;
+        struct btrfs_path *path = NULL;
+        struct btrfs_key key;
+
+        if (!capable(CAP_SYS_ADMIN))
+                return -EPERM;
+
+        loi = memdup_user(arg, sizeof(*loi));
+        if (IS_ERR(loi)) {
+                ret = PTR_ERR(loi);
+                loi = NULL;
+                goto out;
+        }
+
+        path = btrfs_alloc_path();
+        if (!path) {
+                ret = -ENOMEM;
+                goto out;
+        }
+
+        size = min_t(u32, loi->size, 4096);
+        inodes = init_data_container(size);
+        if (IS_ERR(inodes)) {
+                ret = PTR_ERR(inodes);
+                inodes = NULL;
+                goto out;
+        }
+
+        ret = extent_from_logical(root->fs_info, loi->logical, path, &key);
+
+        if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK)
+                ret = -ENOENT;
+        if (ret < 0)
+                goto out;
+
+        extent_offset = loi->logical - key.objectid;
+        ret = iterate_extent_inodes(root->fs_info, path, key.objectid,
+                                        extent_offset, build_ino_list, inodes);
+
+        if (ret < 0)
+                goto out;
+
+        ret = copy_to_user((void *)loi->inodes, (void *)inodes, size);
+        if (ret)
+                ret = -EFAULT;
+
+out:
+        btrfs_free_path(path);
+        kfree(inodes);
+        kfree(loi);
+
+        return ret;
+}
+
 long btrfs_ioctl(struct file *file, unsigned int
                 cmd, unsigned long arg)
 {
@@ -2921,6 +3086,10 @@ long btrfs_ioctl(struct file *file, unsigned int
                 return btrfs_ioctl_tree_search(file, argp);
         case BTRFS_IOC_INO_LOOKUP:
                 return btrfs_ioctl_ino_lookup(file, argp);
+        case BTRFS_IOC_INO_PATHS:
+                return btrfs_ioctl_ino_to_path(root, argp);
+        case BTRFS_IOC_LOGICAL_INO:
+                return btrfs_ioctl_logical_to_ino(root, argp);
         case BTRFS_IOC_SPACE_INFO:
                 return btrfs_ioctl_space_info(root, argp);
         case BTRFS_IOC_SYNC:
diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h
index ad1ea789fcb4..252ae9915de8 100644
--- a/fs/btrfs/ioctl.h
+++ b/fs/btrfs/ioctl.h
@@ -193,6 +193,30 @@ struct btrfs_ioctl_space_args {
         struct btrfs_ioctl_space_info spaces[0];
 };
 
+struct btrfs_data_container {
+        __u32   bytes_left;     /* out -- bytes not needed to deliver output */
+        __u32   bytes_missing;  /* out -- additional bytes needed for result */
+        __u32   elem_cnt;       /* out */
+        __u32   elem_missed;    /* out */
+        __u64   val[0];         /* out */
+};
+
+struct btrfs_ioctl_ino_path_args {
+        __u64                           inum;           /* in */
+        __u32                           size;           /* in */
+        __u64                           reserved[4];
+        /* struct btrfs_data_container  *fspath;           out */
+        __u64                           fspath;         /* out */
+};
+
+struct btrfs_ioctl_logical_ino_args {
+        __u64                           logical;        /* in */
+        __u32                           size;           /* in */
+        __u64                           reserved[4];
+        /* struct btrfs_data_container  *inodes;        out   */
+        __u64                           inodes;
+};
+
 #define BTRFS_IOC_SNAP_CREATE _IOW(BTRFS_IOCTL_MAGIC, 1, \
                                    struct btrfs_ioctl_vol_args)
 #define BTRFS_IOC_DEFRAG _IOW(BTRFS_IOCTL_MAGIC, 2, \
@@ -248,4 +272,9 @@ struct btrfs_ioctl_space_args {
                                  struct btrfs_ioctl_dev_info_args)
 #define BTRFS_IOC_FS_INFO _IOR(BTRFS_IOCTL_MAGIC, 31, \
                                struct btrfs_ioctl_fs_info_args)
+#define BTRFS_IOC_INO_PATHS _IOWR(BTRFS_IOCTL_MAGIC, 35, \
+                                        struct btrfs_ioctl_ino_path_args)
+#define BTRFS_IOC_LOGICAL_INO _IOWR(BTRFS_IOCTL_MAGIC, 36, \
+                                        struct btrfs_ioctl_ino_path_args)
+
 #endif
diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c
index fb2605d998e9..f38e452486b8 100644
--- a/fs/btrfs/print-tree.c
+++ b/fs/btrfs/print-tree.c
@@ -158,8 +158,7 @@ static void print_extent_ref_v0(struct extent_buffer *eb, int slot)
 void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 {
         int i;
-        u32 type;
-        u32 nr = btrfs_header_nritems(l);
+        u32 type, nr;
         struct btrfs_item *item;
         struct btrfs_root_item *ri;
         struct btrfs_dir_item *di;
@@ -172,6 +171,11 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
         struct btrfs_key key;
         struct btrfs_key found_key;
 
+        if (!l)
+                return;
+
+        nr = btrfs_header_nritems(l);
+
         printk(KERN_INFO "leaf %llu total ptrs %d free space %d\n",
                 (unsigned long long)btrfs_header_bytenr(l), nr,
                 btrfs_leaf_free_space(root, l));
diff --git a/fs/btrfs/reada.c b/fs/btrfs/reada.c
new file mode 100644
index 000000000000..2373b39a132b
--- /dev/null
+++ b/fs/btrfs/reada.c
@@ -0,0 +1,951 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include <linux/pagemap.h>
+#include <linux/writeback.h>
+#include <linux/blkdev.h>
+#include <linux/rbtree.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+#include "ctree.h"
+#include "volumes.h"
+#include "disk-io.h"
+#include "transaction.h"
+
+#undef DEBUG
+
+/*
+ * This is the implementation for the generic read ahead framework.
+ *
+ * To trigger a readahead, btrfs_reada_add must be called. It will start
+ * a read ahead for the given range [start, end) on tree root. The returned
+ * handle can either be used to wait on the readahead to finish
+ * (btrfs_reada_wait), or to send it to the background (btrfs_reada_detach).
+ *
+ * The read ahead works as follows:
+ * On btrfs_reada_add, the root of the tree is inserted into a radix_tree.
+ * reada_start_machine will then search for extents to prefetch and trigger
+ * some reads. When a read finishes for a node, all contained node/leaf
+ * pointers that lie in the given range will also be enqueued. The reads will
+ * be triggered in sequential order, thus giving a big win over a naive
+ * enumeration. It will also make use of multi-device layouts. Each disk
+ * will have its on read pointer and all disks will by utilized in parallel.
+ * Also will no two disks read both sides of a mirror simultaneously, as this
+ * would waste seeking capacity. Instead both disks will read different parts
+ * of the filesystem.
+ * Any number of readaheads can be started in parallel. The read order will be
+ * determined globally, i.e. 2 parallel readaheads will normally finish faster
+ * than the 2 started one after another.
+ */
+
+#define MAX_MIRRORS 2
+#define MAX_IN_FLIGHT 6
+
+struct reada_extctl {
+        struct list_head        list;
+        struct reada_control    *rc;
+        u64                     generation;
+};
+
+struct reada_extent {
+        u64                     logical;
+        struct btrfs_key        top;
+        u32                     blocksize;
+        int                     err;
+        struct list_head        extctl;
+        struct kref             refcnt;
+        spinlock_t              lock;
+        struct reada_zone       *zones[MAX_MIRRORS];
+        int                     nzones;
+        struct btrfs_device     *scheduled_for;
+};
+
+struct reada_zone {
+        u64                     start;
+        u64                     end;
+        u64                     elems;
+        struct list_head        list;
+        spinlock_t              lock;
+        int                     locked;
+        struct btrfs_device     *device;
+        struct btrfs_device     *devs[MAX_MIRRORS]; /* full list, incl self */
+        int                     ndevs;
+        struct kref             refcnt;
+};
+
+struct reada_machine_work {
+        struct btrfs_work       work;
+        struct btrfs_fs_info    *fs_info;
+};
+
+static void reada_extent_put(struct btrfs_fs_info *, struct reada_extent *);
+static void reada_control_release(struct kref *kref);
+static void reada_zone_release(struct kref *kref);
+static void reada_start_machine(struct btrfs_fs_info *fs_info);
+static void __reada_start_machine(struct btrfs_fs_info *fs_info);
+
+static int reada_add_block(struct reada_control *rc, u64 logical,
+                           struct btrfs_key *top, int level, u64 generation);
+
+/* recurses */
+/* in case of err, eb might be NULL */
+static int __readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
+                            u64 start, int err)
+{
+        int level = 0;
+        int nritems;
+        int i;
+        u64 bytenr;
+        u64 generation;
+        struct reada_extent *re;
+        struct btrfs_fs_info *fs_info = root->fs_info;
+        struct list_head list;
+        unsigned long index = start >> PAGE_CACHE_SHIFT;
+        struct btrfs_device *for_dev;
+
+        if (eb)
+                level = btrfs_header_level(eb);
+
+        /* find extent */
+        spin_lock(&fs_info->reada_lock);
+        re = radix_tree_lookup(&fs_info->reada_tree, index);
+        if (re)
+                kref_get(&re->refcnt);
+        spin_unlock(&fs_info->reada_lock);
+
+        if (!re)
+                return -1;
+
+        spin_lock(&re->lock);
+        /*
+         * just take the full list from the extent. afterwards we
+         * don't need the lock anymore
+         */
+        list_replace_init(&re->extctl, &list);
+        for_dev = re->scheduled_for;
+        re->scheduled_for = NULL;
+        spin_unlock(&re->lock);
+
+        if (err == 0) {
+                nritems = level ? btrfs_header_nritems(eb) : 0;
+                generation = btrfs_header_generation(eb);
+                /*
+                 * FIXME: currently we just set nritems to 0 if this is a leaf,
+                 * effectively ignoring the content. In a next step we could
+                 * trigger more readahead depending from the content, e.g.
+                 * fetch the checksums for the extents in the leaf.
+                 */
+        } else {
+                /*
+                 * this is the error case, the extent buffer has not been
+                 * read correctly. We won't access anything from it and
+                 * just cleanup our data structures. Effectively this will
+                 * cut the branch below this node from read ahead.
+                 */
+                nritems = 0;
+                generation = 0;
+        }
+
+        for (i = 0; i < nritems; i++) {
+                struct reada_extctl *rec;
+                u64 n_gen;
+                struct btrfs_key key;
+                struct btrfs_key next_key;
+
+                btrfs_node_key_to_cpu(eb, &key, i);
+                if (i + 1 < nritems)
+                        btrfs_node_key_to_cpu(eb, &next_key, i + 1);
+                else
+                        next_key = re->top;
+                bytenr = btrfs_node_blockptr(eb, i);
+                n_gen = btrfs_node_ptr_generation(eb, i);
+
+                list_for_each_entry(rec, &list, list) {
+                        struct reada_control *rc = rec->rc;
+
+                        /*
+                         * if the generation doesn't match, just ignore this
+                         * extctl. This will probably cut off a branch from
+                         * prefetch. Alternatively one could start a new (sub-)
+                         * prefetch for this branch, starting again from root.
+                         * FIXME: move the generation check out of this loop
+                         */
+#ifdef DEBUG
+                        if (rec->generation != generation) {
+                                printk(KERN_DEBUG "generation mismatch for "
+                                                "(%llu,%d,%llu) %llu != %llu\n",
+                                       key.objectid, key.type, key.offset,
+                                       rec->generation, generation);
+                        }
+#endif
+                        if (rec->generation == generation &&
+                            btrfs_comp_cpu_keys(&key, &rc->key_end) < 0 &&
+                            btrfs_comp_cpu_keys(&next_key, &rc->key_start) > 0)
+                                reada_add_block(rc, bytenr, &next_key,
+                                                level - 1, n_gen);
+                }
+        }
+        /*
+         * free extctl records
+         */
+        while (!list_empty(&list)) {
+                struct reada_control *rc;
+                struct reada_extctl *rec;
+
+                rec = list_first_entry(&list, struct reada_extctl, list);
+                list_del(&rec->list);
+                rc = rec->rc;
+                kfree(rec);
+
+                kref_get(&rc->refcnt);
+                if (atomic_dec_and_test(&rc->elems)) {
+                        kref_put(&rc->refcnt, reada_control_release);
+                        wake_up(&rc->wait);
+                }
+                kref_put(&rc->refcnt, reada_control_release);
+
+                reada_extent_put(fs_info, re);  /* one ref for each entry */
+        }
+        reada_extent_put(fs_info, re);  /* our ref */
+        if (for_dev)
+                atomic_dec(&for_dev->reada_in_flight);
+
+        return 0;
+}
+
+/*
+ * start is passed separately in case eb in NULL, which may be the case with
+ * failed I/O
+ */
+int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
+                         u64 start, int err)
+{
+        int ret;
+
+        ret = __readahead_hook(root, eb, start, err);
+
+        reada_start_machine(root->fs_info);
+
+        return ret;
+}
+
+static struct reada_zone *reada_find_zone(struct btrfs_fs_info *fs_info,
+                                          struct btrfs_device *dev, u64 logical,
+                                          struct btrfs_bio *bbio)
+{
+        int ret;
+        int looped = 0;
+        struct reada_zone *zone;
+        struct btrfs_block_group_cache *cache = NULL;
+        u64 start;
+        u64 end;
+        int i;
+
+again:
+        zone = NULL;
+        spin_lock(&fs_info->reada_lock);
+        ret = radix_tree_gang_lookup(&dev->reada_zones, (void **)&zone,
+                                     logical >> PAGE_CACHE_SHIFT, 1);
+        if (ret == 1)
+                kref_get(&zone->refcnt);
+        spin_unlock(&fs_info->reada_lock);
+
+        if (ret == 1) {
+                if (logical >= zone->start && logical < zone->end)
+                        return zone;
+                spin_lock(&fs_info->reada_lock);
+                kref_put(&zone->refcnt, reada_zone_release);
+                spin_unlock(&fs_info->reada_lock);
+        }
+
+        if (looped)
+                return NULL;
+
+        cache = btrfs_lookup_block_group(fs_info, logical);
+        if (!cache)
+                return NULL;
+
+        start = cache->key.objectid;
+        end = start + cache->key.offset - 1;
+        btrfs_put_block_group(cache);
+
+        zone = kzalloc(sizeof(*zone), GFP_NOFS);
+        if (!zone)
+                return NULL;
+
+        zone->start = start;
+        zone->end = end;
+        INIT_LIST_HEAD(&zone->list);
+        spin_lock_init(&zone->lock);
+        zone->locked = 0;
+        kref_init(&zone->refcnt);
+        zone->elems = 0;
+        zone->device = dev; /* our device always sits at index 0 */
+        for (i = 0; i < bbio->num_stripes; ++i) {
+                /* bounds have already been checked */
+                zone->devs[i] = bbio->stripes[i].dev;
+        }
+        zone->ndevs = bbio->num_stripes;
+
+        spin_lock(&fs_info->reada_lock);
+        ret = radix_tree_insert(&dev->reada_zones,
+                                (unsigned long)zone->end >> PAGE_CACHE_SHIFT,
+                                zone);
+        spin_unlock(&fs_info->reada_lock);
+
+        if (ret) {
+                kfree(zone);
+                looped = 1;
+                goto again;
+        }
+
+        return zone;
+}
+
+static struct reada_extent *reada_find_extent(struct btrfs_root *root,
+                                              u64 logical,
+                                              struct btrfs_key *top, int level)
+{
+        int ret;
+        int looped = 0;
+        struct reada_extent *re = NULL;
+        struct btrfs_fs_info *fs_info = root->fs_info;
+        struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
+        struct btrfs_bio *bbio = NULL;
+        struct btrfs_device *dev;
+        u32 blocksize;
+        u64 length;
+        int nzones = 0;
+        int i;
+        unsigned long index = logical >> PAGE_CACHE_SHIFT;
+
+again:
+        spin_lock(&fs_info->reada_lock);
+        re = radix_tree_lookup(&fs_info->reada_tree, index);
+        if (re)
+                kref_get(&re->refcnt);
+        spin_unlock(&fs_info->reada_lock);
+
+        if (re || looped)
+                return re;
+
+        re = kzalloc(sizeof(*re), GFP_NOFS);
+        if (!re)
+                return NULL;
+
+        blocksize = btrfs_level_size(root, level);
+        re->logical = logical;
+        re->blocksize = blocksize;
+        re->top = *top;
+        INIT_LIST_HEAD(&re->extctl);
+        spin_lock_init(&re->lock);
+        kref_init(&re->refcnt);
+
+        /*
+         * map block
+         */
+        length = blocksize;
+        ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, &bbio, 0);
+        if (ret || !bbio || length < blocksize)
+                goto error;
+
+        if (bbio->num_stripes > MAX_MIRRORS) {
+                printk(KERN_ERR "btrfs readahead: more than %d copies not "
+                                "supported", MAX_MIRRORS);
+                goto error;
+        }
+
+        for (nzones = 0; nzones < bbio->num_stripes; ++nzones) {
+                struct reada_zone *zone;
+
+                dev = bbio->stripes[nzones].dev;
+                zone = reada_find_zone(fs_info, dev, logical, bbio);
+                if (!zone)
+                        break;
+
+                re->zones[nzones] = zone;
+                spin_lock(&zone->lock);
+                if (!zone->elems)
+                        kref_get(&zone->refcnt);
+                ++zone->elems;
+                spin_unlock(&zone->lock);
+                spin_lock(&fs_info->reada_lock);
+                kref_put(&zone->refcnt, reada_zone_release);
+                spin_unlock(&fs_info->reada_lock);
+        }
+        re->nzones = nzones;
+        if (nzones == 0) {
+                /* not a single zone found, error and out */
+                goto error;
+        }
+
+        /* insert extent in reada_tree + all per-device trees, all or nothing */
+        spin_lock(&fs_info->reada_lock);
+        ret = radix_tree_insert(&fs_info->reada_tree, index, re);
+        if (ret) {
+                spin_unlock(&fs_info->reada_lock);
+                if (ret != -ENOMEM) {
+                        /* someone inserted the extent in the meantime */
+                        looped = 1;
+                }
+                goto error;
+        }
+        for (i = 0; i < nzones; ++i) {
+                dev = bbio->stripes[i].dev;
+                ret = radix_tree_insert(&dev->reada_extents, index, re);
+                if (ret) {
+                        while (--i >= 0) {
+                                dev = bbio->stripes[i].dev;
+                                BUG_ON(dev == NULL);
+                                radix_tree_delete(&dev->reada_extents, index);
+                        }
+                        BUG_ON(fs_info == NULL);
+                        radix_tree_delete(&fs_info->reada_tree, index);
+                        spin_unlock(&fs_info->reada_lock);
+                        goto error;
+                }
+        }
+        spin_unlock(&fs_info->reada_lock);
+
+        kfree(bbio);
+        return re;
+
+error:
+        while (nzones) {
+                struct reada_zone *zone;
+
+                --nzones;
+                zone = re->zones[nzones];
+                kref_get(&zone->refcnt);
+                spin_lock(&zone->lock);
+                --zone->elems;
+                if (zone->elems == 0) {
+                        /*
+                         * no fs_info->reada_lock needed, as this can't be
+                         * the last ref
+                         */
+                        kref_put(&zone->refcnt, reada_zone_release);
+                }
+                spin_unlock(&zone->lock);
+
+                spin_lock(&fs_info->reada_lock);
+                kref_put(&zone->refcnt, reada_zone_release);
+                spin_unlock(&fs_info->reada_lock);
+        }
+        kfree(bbio);
+        kfree(re);
+        if (looped)
+                goto again;
+        return NULL;
+}
+
+static void reada_kref_dummy(struct kref *kr)
+{
+}
+
+static void reada_extent_put(struct btrfs_fs_info *fs_info,
+                             struct reada_extent *re)
+{
+        int i;
+        unsigned long index = re->logical >> PAGE_CACHE_SHIFT;
+
+        spin_lock(&fs_info->reada_lock);
+        if (!kref_put(&re->refcnt, reada_kref_dummy)) {
+                spin_unlock(&fs_info->reada_lock);
+                return;
+        }
+
+        radix_tree_delete(&fs_info->reada_tree, index);
+        for (i = 0; i < re->nzones; ++i) {
+                struct reada_zone *zone = re->zones[i];
+
+                radix_tree_delete(&zone->device->reada_extents, index);
+        }
+
+        spin_unlock(&fs_info->reada_lock);
+
+        for (i = 0; i < re->nzones; ++i) {
+                struct reada_zone *zone = re->zones[i];
+
+                kref_get(&zone->refcnt);
+                spin_lock(&zone->lock);
+                --zone->elems;
+                if (zone->elems == 0) {
+                        /* no fs_info->reada_lock needed, as this can't be
+                         * the last ref */
+                        kref_put(&zone->refcnt, reada_zone_release);
+                }
+                spin_unlock(&zone->lock);
+
+                spin_lock(&fs_info->reada_lock);
+                kref_put(&zone->refcnt, reada_zone_release);
+                spin_unlock(&fs_info->reada_lock);
+        }
+        if (re->scheduled_for)
+                atomic_dec(&re->scheduled_for->reada_in_flight);
+
+        kfree(re);
+}
+
+static void reada_zone_release(struct kref *kref)
+{
+        struct reada_zone *zone = container_of(kref, struct reada_zone, refcnt);
+
+        radix_tree_delete(&zone->device->reada_zones,
+                          zone->end >> PAGE_CACHE_SHIFT);
+
+        kfree(zone);
+}
+
+static void reada_control_release(struct kref *kref)
+{
+        struct reada_control *rc = container_of(kref, struct reada_control,
+                                                refcnt);
+
+        kfree(rc);
+}
+
+static int reada_add_block(struct reada_control *rc, u64 logical,
+                           struct btrfs_key *top, int level, u64 generation)
+{
+        struct btrfs_root *root = rc->root;
+        struct reada_extent *re;
+        struct reada_extctl *rec;
+
+        re = reada_find_extent(root, logical, top, level); /* takes one ref */
+        if (!re)
+                return -1;
+
+        rec = kzalloc(sizeof(*rec), GFP_NOFS);
+        if (!rec) {
+                reada_extent_put(root->fs_info, re);
+                return -1;
+        }
+
+        rec->rc = rc;
+        rec->generation = generation;
+        atomic_inc(&rc->elems);
+
+        spin_lock(&re->lock);
+        list_add_tail(&rec->list, &re->extctl);
+        spin_unlock(&re->lock);
+
+        /* leave the ref on the extent */
+
+        return 0;
+}
+
+/*
+ * called with fs_info->reada_lock held
+ */
+static void reada_peer_zones_set_lock(struct reada_zone *zone, int lock)
+{
+        int i;
+        unsigned long index = zone->end >> PAGE_CACHE_SHIFT;
+
+        for (i = 0; i < zone->ndevs; ++i) {
+                struct reada_zone *peer;
+                peer = radix_tree_lookup(&zone->devs[i]->reada_zones, index);
+                if (peer && peer->device != zone->device)
+                        peer->locked = lock;
+        }
+}
+
+/*
+ * called with fs_info->reada_lock held
+ */
+static int reada_pick_zone(struct btrfs_device *dev)
+{
+        struct reada_zone *top_zone = NULL;
+        struct reada_zone *top_locked_zone = NULL;
+        u64 top_elems = 0;
+        u64 top_locked_elems = 0;
+        unsigned long index = 0;
+        int ret;
+
+        if (dev->reada_curr_zone) {
+                reada_peer_zones_set_lock(dev->reada_curr_zone, 0);
+                kref_put(&dev->reada_curr_zone->refcnt, reada_zone_release);
+                dev->reada_curr_zone = NULL;
+        }
+        /* pick the zone with the most elements */
+        while (1) {
+                struct reada_zone *zone;
+
+                ret = radix_tree_gang_lookup(&dev->reada_zones,
+                                             (void **)&zone, index, 1);
+                if (ret == 0)
+                        break;
+                index = (zone->end >> PAGE_CACHE_SHIFT) + 1;
+                if (zone->locked) {
+                        if (zone->elems > top_locked_elems) {
+                                top_locked_elems = zone->elems;
+                                top_locked_zone = zone;
+                        }
+                } else {
+                        if (zone->elems > top_elems) {
+                                top_elems = zone->elems;
+                                top_zone = zone;
+                        }
+                }
+        }
+        if (top_zone)
+                dev->reada_curr_zone = top_zone;
+        else if (top_locked_zone)
+                dev->reada_curr_zone = top_locked_zone;
+        else
+                return 0;
+
+        dev->reada_next = dev->reada_curr_zone->start;
+        kref_get(&dev->reada_curr_zone->refcnt);
+        reada_peer_zones_set_lock(dev->reada_curr_zone, 1);
+
+        return 1;
+}
+
+static int reada_start_machine_dev(struct btrfs_fs_info *fs_info,
+                                   struct btrfs_device *dev)
+{
+        struct reada_extent *re = NULL;
+        int mirror_num = 0;
+        struct extent_buffer *eb = NULL;
+        u64 logical;
+        u32 blocksize;
+        int ret;
+        int i;
+        int need_kick = 0;
+
+        spin_lock(&fs_info->reada_lock);
+        if (dev->reada_curr_zone == NULL) {
+                ret = reada_pick_zone(dev);
+                if (!ret) {
+                        spin_unlock(&fs_info->reada_lock);
+                        return 0;
+                }
+        }
+        /*
+         * FIXME currently we issue the reads one extent at a time. If we have
+         * a contiguous block of extents, we could also coagulate them or use
+         * plugging to speed things up
+         */
+        ret = radix_tree_gang_lookup(&dev->reada_extents, (void **)&re,
+                                     dev->reada_next >> PAGE_CACHE_SHIFT, 1);
+        if (ret == 0 || re->logical >= dev->reada_curr_zone->end) {
+                ret = reada_pick_zone(dev);
+                if (!ret) {
+                        spin_unlock(&fs_info->reada_lock);
+                        return 0;
+                }
+                re = NULL;
+                ret = radix_tree_gang_lookup(&dev->reada_extents, (void **)&re,
+                                        dev->reada_next >> PAGE_CACHE_SHIFT, 1);
+        }
+        if (ret == 0) {
+                spin_unlock(&fs_info->reada_lock);
+                return 0;
+        }
+        dev->reada_next = re->logical + re->blocksize;
+        kref_get(&re->refcnt);
+
+        spin_unlock(&fs_info->reada_lock);
+
+        /*
+         * find mirror num
+         */
+        for (i = 0; i < re->nzones; ++i) {
+                if (re->zones[i]->device == dev) {
+                        mirror_num = i + 1;
+                        break;
+                }
+        }
+        logical = re->logical;
+        blocksize = re->blocksize;
+
+        spin_lock(&re->lock);
+        if (re->scheduled_for == NULL) {
+                re->scheduled_for = dev;
+                need_kick = 1;
+        }
+        spin_unlock(&re->lock);
+
+        reada_extent_put(fs_info, re);
+
+        if (!need_kick)
+                return 0;
+
+        atomic_inc(&dev->reada_in_flight);
+        ret = reada_tree_block_flagged(fs_info->extent_root, logical, blocksize,
+                         mirror_num, &eb);
+        if (ret)
+                __readahead_hook(fs_info->extent_root, NULL, logical, ret);
+        else if (eb)
+                __readahead_hook(fs_info->extent_root, eb, eb->start, ret);
+
+        if (eb)
+                free_extent_buffer(eb);
+
+        return 1;
+
+}
+
+static void reada_start_machine_worker(struct btrfs_work *work)
+{
+        struct reada_machine_work *rmw;
+        struct btrfs_fs_info *fs_info;
+
+        rmw = container_of(work, struct reada_machine_work, work);
+        fs_info = rmw->fs_info;
+
+        kfree(rmw);
+
+        __reada_start_machine(fs_info);
+}
+
+static void __reada_start_machine(struct btrfs_fs_info *fs_info)
+{
+        struct btrfs_device *device;
+        struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+        u64 enqueued;
+        u64 total = 0;
+        int i;
+
+        do {
+                enqueued = 0;
+                list_for_each_entry(device, &fs_devices->devices, dev_list) {
+                        if (atomic_read(&device->reada_in_flight) <
+                            MAX_IN_FLIGHT)
+                                enqueued += reada_start_machine_dev(fs_info,
+                                                                    device);
+                }
+                total += enqueued;
+        } while (enqueued && total < 10000);
+
+        if (enqueued == 0)
+                return;
+
+        /*
+         * If everything is already in the cache, this is effectively single
+         * threaded. To a) not hold the caller for too long and b) to utilize
+         * more cores, we broke the loop above after 10000 iterations and now
+         * enqueue to workers to finish it. This will distribute the load to
+         * the cores.
+         */
+        for (i = 0; i < 2; ++i)
+                reada_start_machine(fs_info);
+}
+
+static void reada_start_machine(struct btrfs_fs_info *fs_info)
+{
+        struct reada_machine_work *rmw;
+
+        rmw = kzalloc(sizeof(*rmw), GFP_NOFS);
+        if (!rmw) {
+                /* FIXME we cannot handle this properly right now */
+                BUG();
+        }
+        rmw->work.func = reada_start_machine_worker;
+        rmw->fs_info = fs_info;
+
+        btrfs_queue_worker(&fs_info->readahead_workers, &rmw->work);
+}
+
+#ifdef DEBUG
+static void dump_devs(struct btrfs_fs_info *fs_info, int all)
+{
+        struct btrfs_device *device;
+        struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+        unsigned long index;
+        int ret;
+        int i;
+        int j;
+        int cnt;
+
+        spin_lock(&fs_info->reada_lock);
+        list_for_each_entry(device, &fs_devices->devices, dev_list) {
+                printk(KERN_DEBUG "dev %lld has %d in flight\n", device->devid,
+                        atomic_read(&device->reada_in_flight));
+                index = 0;
+                while (1) {
+                        struct reada_zone *zone;
+                        ret = radix_tree_gang_lookup(&device->reada_zones,
+                                                     (void **)&zone, index, 1);
+                        if (ret == 0)
+                                break;
+                        printk(KERN_DEBUG "  zone %llu-%llu elems %llu locked "
+                                "%d devs", zone->start, zone->end, zone->elems,
+                                zone->locked);
+                        for (j = 0; j < zone->ndevs; ++j) {
+                                printk(KERN_CONT " %lld",
+                                        zone->devs[j]->devid);
+                        }
+                        if (device->reada_curr_zone == zone)
+                                printk(KERN_CONT " curr off %llu",
+                                        device->reada_next - zone->start);
+                        printk(KERN_CONT "\n");
+                        index = (zone->end >> PAGE_CACHE_SHIFT) + 1;
+                }
+                cnt = 0;
+                index = 0;
+                while (all) {
+                        struct reada_extent *re = NULL;
+
+                        ret = radix_tree_gang_lookup(&device->reada_extents,
+                                                     (void **)&re, index, 1);
+                        if (ret == 0)
+                                break;
+                        printk(KERN_DEBUG
+                                "  re: logical %llu size %u empty %d for %lld",
+                                re->logical, re->blocksize,
+                                list_empty(&re->extctl), re->scheduled_for ?
+                                re->scheduled_for->devid : -1);
+
+                        for (i = 0; i < re->nzones; ++i) {
+                                printk(KERN_CONT " zone %llu-%llu devs",
+                                        re->zones[i]->start,
+                                        re->zones[i]->end);
+                                for (j = 0; j < re->zones[i]->ndevs; ++j) {
+                                        printk(KERN_CONT " %lld",
+                                                re->zones[i]->devs[j]->devid);
+                                }
+                        }
+                        printk(KERN_CONT "\n");
+                        index = (re->logical >> PAGE_CACHE_SHIFT) + 1;
+                        if (++cnt > 15)
+                                break;
+                }
+        }
+
+        index = 0;
+        cnt = 0;
+        while (all) {
+                struct reada_extent *re = NULL;
+
+                ret = radix_tree_gang_lookup(&fs_info->reada_tree, (void **)&re,
+                                             index, 1);
+                if (ret == 0)
+                        break;
+                if (!re->scheduled_for) {
+                        index = (re->logical >> PAGE_CACHE_SHIFT) + 1;
+                        continue;
+                }
+                printk(KERN_DEBUG
+                        "re: logical %llu size %u list empty %d for %lld",
+                        re->logical, re->blocksize, list_empty(&re->extctl),
+                        re->scheduled_for ? re->scheduled_for->devid : -1);
+                for (i = 0; i < re->nzones; ++i) {
+                        printk(KERN_CONT " zone %llu-%llu devs",
+                                re->zones[i]->start,
+                                re->zones[i]->end);
+                        for (i = 0; i < re->nzones; ++i) {
+                                printk(KERN_CONT " zone %llu-%llu devs",
+                                        re->zones[i]->start,
+                                        re->zones[i]->end);
+                                for (j = 0; j < re->zones[i]->ndevs; ++j) {
+                                        printk(KERN_CONT " %lld",
+                                                re->zones[i]->devs[j]->devid);
+                                }
+                        }
+                }
+                printk(KERN_CONT "\n");
+                index = (re->logical >> PAGE_CACHE_SHIFT) + 1;
+        }
+        spin_unlock(&fs_info->reada_lock);
+}
+#endif
+
+/*
+ * interface
+ */
+struct reada_control *btrfs_reada_add(struct btrfs_root *root,
+                        struct btrfs_key *key_start, struct btrfs_key *key_end)
+{
+        struct reada_control *rc;
+        u64 start;
+        u64 generation;
+        int level;
+        struct extent_buffer *node;
+        static struct btrfs_key max_key = {
+                .objectid = (u64)-1,
+                .type = (u8)-1,
+                .offset = (u64)-1
+        };
+
+        rc = kzalloc(sizeof(*rc), GFP_NOFS);
+        if (!rc)
+                return ERR_PTR(-ENOMEM);
+
+        rc->root = root;
+        rc->key_start = *key_start;
+        rc->key_end = *key_end;
+        atomic_set(&rc->elems, 0);
+        init_waitqueue_head(&rc->wait);
+        kref_init(&rc->refcnt);
+        kref_get(&rc->refcnt); /* one ref for having elements */
+
+        node = btrfs_root_node(root);
+        start = node->start;
+        level = btrfs_header_level(node);
+        generation = btrfs_header_generation(node);
+        free_extent_buffer(node);
+
+        reada_add_block(rc, start, &max_key, level, generation);
+
+        reada_start_machine(root->fs_info);
+
+        return rc;
+}
+
+#ifdef DEBUG
+int btrfs_reada_wait(void *handle)
+{
+        struct reada_control *rc = handle;
+
+        while (atomic_read(&rc->elems)) {
+                wait_event_timeout(rc->wait, atomic_read(&rc->elems) == 0,
+                                   5 * HZ);
+                dump_devs(rc->root->fs_info, rc->elems < 10 ? 1 : 0);
+        }
+
+        dump_devs(rc->root->fs_info, rc->elems < 10 ? 1 : 0);
+
+        kref_put(&rc->refcnt, reada_control_release);
+
+        return 0;
+}
+#else
+int btrfs_reada_wait(void *handle)
+{
+        struct reada_control *rc = handle;
+
+        while (atomic_read(&rc->elems)) {
+                wait_event(rc->wait, atomic_read(&rc->elems) == 0);
+        }
+
+        kref_put(&rc->refcnt, reada_control_release);
+
+        return 0;
+}
+#endif
+
+void btrfs_reada_detach(void *handle)
+{
+        struct reada_control *rc = handle;
+
+        kref_put(&rc->refcnt, reada_control_release);
+}
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 59bb1764273d..24d654ce7a06 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2041,8 +2041,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
                 BUG_ON(IS_ERR(trans));
                 trans->block_rsv = rc->block_rsv;
 
-                ret = btrfs_block_rsv_check(trans, root, rc->block_rsv,
-                                            min_reserved, 0);
+                ret = btrfs_block_rsv_refill(root, rc->block_rsv, min_reserved);
                 if (ret) {
                         BUG_ON(ret != -EAGAIN);
                         ret = btrfs_commit_transaction(trans, root);
@@ -2152,8 +2151,7 @@ int prepare_to_merge(struct reloc_control *rc, int err)
 again:
         if (!err) {
                 num_bytes = rc->merging_rsv_size;
-                ret = btrfs_block_rsv_add(NULL, root, rc->block_rsv,
-                                          num_bytes);
+                ret = btrfs_block_rsv_add(root, rc->block_rsv, num_bytes);
                 if (ret)
                         err = ret;
         }
@@ -2427,7 +2425,7 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans,
         num_bytes = calcu_metadata_size(rc, node, 1) * 2;
 
         trans->block_rsv = rc->block_rsv;
-        ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes);
+        ret = btrfs_block_rsv_add(root, rc->block_rsv, num_bytes);
         if (ret) {
                 if (ret == -EAGAIN)
                         rc->commit_transaction = 1;
@@ -2922,6 +2920,7 @@ static int relocate_file_extent_cluster(struct inode *inode,
         unsigned long last_index;
         struct page *page;
         struct file_ra_state *ra;
+        gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
         int nr = 0;
         int ret = 0;
 
@@ -2956,7 +2955,7 @@ static int relocate_file_extent_cluster(struct inode *inode,
                                                   ra, NULL, index,
                                                   last_index + 1 - index);
                         page = find_or_create_page(inode->i_mapping, index,
-                                                   GFP_NOFS);
+                                                   mask);
                         if (!page) {
                                 btrfs_delalloc_release_metadata(inode,
                                                         PAGE_CACHE_SIZE);
@@ -3323,8 +3322,11 @@ static int find_data_references(struct reloc_control *rc,
         }
 
         key.objectid = ref_objectid;
-        key.offset = ref_offset;
         key.type = BTRFS_EXTENT_DATA_KEY;
+        if (ref_offset > ((u64)-1 << 32))
+                key.offset = 0;
+        else
+                key.offset = ref_offset;
 
         path->search_commit_root = 1;
         path->skip_locking = 1;
@@ -3645,14 +3647,11 @@ int prepare_to_relocate(struct reloc_control *rc)
          * btrfs_init_reloc_root will use them when there
          * is no reservation in transaction handle.
          */
-        ret = btrfs_block_rsv_add(NULL, rc->extent_root, rc->block_rsv,
+        ret = btrfs_block_rsv_add(rc->extent_root, rc->block_rsv,
                                   rc->extent_root->nodesize * 256);
         if (ret)
                 return ret;
 
-        rc->block_rsv->refill_used = 1;
-        btrfs_add_durable_block_rsv(rc->extent_root->fs_info, rc->block_rsv);
-
         memset(&rc->cluster, 0, sizeof(rc->cluster));
         rc->search_start = rc->block_group->key.objectid;
         rc->extents_found = 0;
@@ -3777,8 +3776,7 @@ restart:
                         }
                 }
 
-                ret = btrfs_block_rsv_check(trans, rc->extent_root,
-                                            rc->block_rsv, 0, 5);
+                ret = btrfs_block_rsv_check(rc->extent_root, rc->block_rsv, 5);
                 if (ret < 0) {
                         if (ret != -EAGAIN) {
                                 err = ret;
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index a8d03d5efb5d..ed11d3866afd 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -17,10 +17,14 @@
  */
 
 #include <linux/blkdev.h>
+#include <linux/ratelimit.h>
 #include "ctree.h"
 #include "volumes.h"
 #include "disk-io.h"
 #include "ordered-data.h"
+#include "transaction.h"
+#include "backref.h"
+#include "extent_io.h"
 
 /*
  * This is only the first step towards a full-features scrub. It reads all
@@ -29,15 +33,12 @@
  * any can be found.
  *
  * Future enhancements:
- *  - To enhance the performance, better read-ahead strategies for the
- *    extent-tree can be employed.
  *  - In case an unrepairable extent is encountered, track which files are
  *    affected and report them
  *  - In case of a read error on files with nodatasum, map the file and read
  *    the extent to trigger a writeback of the good copy
  *  - track and record media errors, throw out bad devices
  *  - add a mode to also read unallocated space
- *  - make the prefetch cancellable
  */
 
 struct scrub_bio;
@@ -63,7 +64,7 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix);
 struct scrub_page {
         u64                     flags;  /* extent flags */
         u64                     generation;
-        u64                     mirror_num;
+        int                     mirror_num;
         int                     have_csum;
         u8                      csum[BTRFS_CSUM_SIZE];
 };
@@ -87,6 +88,7 @@ struct scrub_dev {
         int                     first_free;
         int                     curr;
         atomic_t                in_flight;
+        atomic_t                fixup_cnt;
         spinlock_t              list_lock;
         wait_queue_head_t       list_wait;
         u16                     csum_size;
@@ -100,6 +102,27 @@ struct scrub_dev {
         spinlock_t              stat_lock;
 };
 
+struct scrub_fixup_nodatasum {
+        struct scrub_dev        *sdev;
+        u64                     logical;
+        struct btrfs_root       *root;
+        struct btrfs_work       work;
+        int                     mirror_num;
+};
+
+struct scrub_warning {
+        struct btrfs_path       *path;
+        u64                     extent_item_size;
+        char                    *scratch_buf;
+        char                    *msg_buf;
+        const char              *errstr;
+        sector_t                sector;
+        u64                     logical;
+        struct btrfs_device     *dev;
+        int                     msg_bufsize;
+        int                     scratch_bufsize;
+};
+
 static void scrub_free_csums(struct scrub_dev *sdev)
 {
         while (!list_empty(&sdev->csum_list)) {
@@ -175,14 +198,15 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
 
                 if (i != SCRUB_BIOS_PER_DEV-1)
                         sdev->bios[i]->next_free = i + 1;
-                 else
+                else
                         sdev->bios[i]->next_free = -1;
         }
         sdev->first_free = 0;
         sdev->curr = -1;
         atomic_set(&sdev->in_flight, 0);
+        atomic_set(&sdev->fixup_cnt, 0);
         atomic_set(&sdev->cancel_req, 0);
-        sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy);
+        sdev->csum_size = btrfs_super_csum_size(fs_info->super_copy);
         INIT_LIST_HEAD(&sdev->csum_list);
 
         spin_lock_init(&sdev->list_lock);
@@ -195,24 +219,361 @@ nomem:
         return ERR_PTR(-ENOMEM);
 }
 
+static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, void *ctx)
+{
+        u64 isize;
+        u32 nlink;
+        int ret;
+        int i;
+        struct extent_buffer *eb;
+        struct btrfs_inode_item *inode_item;
+        struct scrub_warning *swarn = ctx;
+        struct btrfs_fs_info *fs_info = swarn->dev->dev_root->fs_info;
+        struct inode_fs_paths *ipath = NULL;
+        struct btrfs_root *local_root;
+        struct btrfs_key root_key;
+
+        root_key.objectid = root;
+        root_key.type = BTRFS_ROOT_ITEM_KEY;
+        root_key.offset = (u64)-1;
+        local_root = btrfs_read_fs_root_no_name(fs_info, &root_key);
+        if (IS_ERR(local_root)) {
+                ret = PTR_ERR(local_root);
+                goto err;
+        }
+
+        ret = inode_item_info(inum, 0, local_root, swarn->path);
+        if (ret) {
+                btrfs_release_path(swarn->path);
+                goto err;
+        }
+
+        eb = swarn->path->nodes[0];
+        inode_item = btrfs_item_ptr(eb, swarn->path->slots[0],
+                                        struct btrfs_inode_item);
+        isize = btrfs_inode_size(eb, inode_item);
+        nlink = btrfs_inode_nlink(eb, inode_item);
+        btrfs_release_path(swarn->path);
+
+        ipath = init_ipath(4096, local_root, swarn->path);
+        ret = paths_from_inode(inum, ipath);
+
+        if (ret < 0)
+                goto err;
+
+        /*
+         * we deliberately ignore the bit ipath might have been too small to
+         * hold all of the paths here
+         */
+        for (i = 0; i < ipath->fspath->elem_cnt; ++i)
+                printk(KERN_WARNING "btrfs: %s at logical %llu on dev "
+                        "%s, sector %llu, root %llu, inode %llu, offset %llu, "
+                        "length %llu, links %u (path: %s)\n", swarn->errstr,
+                        swarn->logical, swarn->dev->name,
+                        (unsigned long long)swarn->sector, root, inum, offset,
+                        min(isize - offset, (u64)PAGE_SIZE), nlink,
+                        (char *)ipath->fspath->val[i]);
+
+        free_ipath(ipath);
+        return 0;
+
+err:
+        printk(KERN_WARNING "btrfs: %s at logical %llu on dev "
+                "%s, sector %llu, root %llu, inode %llu, offset %llu: path "
+                "resolving failed with ret=%d\n", swarn->errstr,
+                swarn->logical, swarn->dev->name,
+                (unsigned long long)swarn->sector, root, inum, offset, ret);
+
+        free_ipath(ipath);
+        return 0;
+}
+
+static void scrub_print_warning(const char *errstr, struct scrub_bio *sbio,
+                                int ix)
+{
+        struct btrfs_device *dev = sbio->sdev->dev;
+        struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
+        struct btrfs_path *path;
+        struct btrfs_key found_key;
+        struct extent_buffer *eb;
+        struct btrfs_extent_item *ei;
+        struct scrub_warning swarn;
+        u32 item_size;
+        int ret;
+        u64 ref_root;
+        u8 ref_level;
+        unsigned long ptr = 0;
+        const int bufsize = 4096;
+        u64 extent_offset;
+
+        path = btrfs_alloc_path();
+
+        swarn.scratch_buf = kmalloc(bufsize, GFP_NOFS);
+        swarn.msg_buf = kmalloc(bufsize, GFP_NOFS);
+        swarn.sector = (sbio->physical + ix * PAGE_SIZE) >> 9;
+        swarn.logical = sbio->logical + ix * PAGE_SIZE;
+        swarn.errstr = errstr;
+        swarn.dev = dev;
+        swarn.msg_bufsize = bufsize;
+        swarn.scratch_bufsize = bufsize;
+
+        if (!path || !swarn.scratch_buf || !swarn.msg_buf)
+                goto out;
+
+        ret = extent_from_logical(fs_info, swarn.logical, path, &found_key);
+        if (ret < 0)
+                goto out;
+
+        extent_offset = swarn.logical - found_key.objectid;
+        swarn.extent_item_size = found_key.offset;
+
+        eb = path->nodes[0];
+        ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+        item_size = btrfs_item_size_nr(eb, path->slots[0]);
+
+        if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+                do {
+                        ret = tree_backref_for_extent(&ptr, eb, ei, item_size,
+                                                        &ref_root, &ref_level);
+                        printk(KERN_WARNING "%s at logical %llu on dev %s, "
+                                "sector %llu: metadata %s (level %d) in tree "
+                                "%llu\n", errstr, swarn.logical, dev->name,
+                                (unsigned long long)swarn.sector,
+                                ref_level ? "node" : "leaf",
+                                ret < 0 ? -1 : ref_level,
+                                ret < 0 ? -1 : ref_root);
+                } while (ret != 1);
+        } else {
+                swarn.path = path;
+                iterate_extent_inodes(fs_info, path, found_key.objectid,
+                                        extent_offset,
+                                        scrub_print_warning_inode, &swarn);
+        }
+
+out:
+        btrfs_free_path(path);
+        kfree(swarn.scratch_buf);
+        kfree(swarn.msg_buf);
+}
+
+static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *ctx)
+{
+        struct page *page = NULL;
+        unsigned long index;
+        struct scrub_fixup_nodatasum *fixup = ctx;
+        int ret;
+        int corrected = 0;
+        struct btrfs_key key;
+        struct inode *inode = NULL;
+        u64 end = offset + PAGE_SIZE - 1;
+        struct btrfs_root *local_root;
+
+        key.objectid = root;
+        key.type = BTRFS_ROOT_ITEM_KEY;
+        key.offset = (u64)-1;
+        local_root = btrfs_read_fs_root_no_name(fixup->root->fs_info, &key);
+        if (IS_ERR(local_root))
+                return PTR_ERR(local_root);
+
+        key.type = BTRFS_INODE_ITEM_KEY;
+        key.objectid = inum;
+        key.offset = 0;
+        inode = btrfs_iget(fixup->root->fs_info->sb, &key, local_root, NULL);
+        if (IS_ERR(inode))
+                return PTR_ERR(inode);
+
+        index = offset >> PAGE_CACHE_SHIFT;
+
+        page = find_or_create_page(inode->i_mapping, index, GFP_NOFS);
+        if (!page) {
+                ret = -ENOMEM;
+                goto out;
+        }
+
+        if (PageUptodate(page)) {
+                struct btrfs_mapping_tree *map_tree;
+                if (PageDirty(page)) {
+                        /*
+                         * we need to write the data to the defect sector. the
+                         * data that was in that sector is not in memory,
+                         * because the page was modified. we must not write the
+                         * modified page to that sector.
+                         *
+                         * TODO: what could be done here: wait for the delalloc
+                         *       runner to write out that page (might involve
+                         *       COW) and see whether the sector is still
+                         *       referenced afterwards.
+                         *
+                         * For the meantime, we'll treat this error
+                         * incorrectable, although there is a chance that a
+                         * later scrub will find the bad sector again and that
+                         * there's no dirty page in memory, then.
+                         */
+                        ret = -EIO;
+                        goto out;
+                }
+                map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree;
+                ret = repair_io_failure(map_tree, offset, PAGE_SIZE,
+                                        fixup->logical, page,
+                                        fixup->mirror_num);
+                unlock_page(page);
+                corrected = !ret;
+        } else {
+                /*
+                 * we need to get good data first. the general readpage path
+                 * will call repair_io_failure for us, we just have to make
+                 * sure we read the bad mirror.
+                 */
+                ret = set_extent_bits(&BTRFS_I(inode)->io_tree, offset, end,
+                                        EXTENT_DAMAGED, GFP_NOFS);
+                if (ret) {
+                        /* set_extent_bits should give proper error */
+                        WARN_ON(ret > 0);
+                        if (ret > 0)
+                                ret = -EFAULT;
+                        goto out;
+                }
+
+                ret = extent_read_full_page(&BTRFS_I(inode)->io_tree, page,
+                                                btrfs_get_extent,
+                                                fixup->mirror_num);
+                wait_on_page_locked(page);
+
+                corrected = !test_range_bit(&BTRFS_I(inode)->io_tree, offset,
+                                                end, EXTENT_DAMAGED, 0, NULL);
+                if (!corrected)
+                        clear_extent_bits(&BTRFS_I(inode)->io_tree, offset, end,
+                                                EXTENT_DAMAGED, GFP_NOFS);
+        }
+
+out:
+        if (page)
+                put_page(page);
+        if (inode)
+                iput(inode);
+
+        if (ret < 0)
+                return ret;
+
+        if (ret == 0 && corrected) {
+                /*
+                 * we only need to call readpage for one of the inodes belonging
+                 * to this extent. so make iterate_extent_inodes stop
+                 */
+                return 1;
+        }
+
+        return -EIO;
+}
+
+static void scrub_fixup_nodatasum(struct btrfs_work *work)
+{
+        int ret;
+        struct scrub_fixup_nodatasum *fixup;
+        struct scrub_dev *sdev;
+        struct btrfs_trans_handle *trans = NULL;
+        struct btrfs_fs_info *fs_info;
+        struct btrfs_path *path;
+        int uncorrectable = 0;
+
+        fixup = container_of(work, struct scrub_fixup_nodatasum, work);
+        sdev = fixup->sdev;
+        fs_info = fixup->root->fs_info;
+
+        path = btrfs_alloc_path();
+        if (!path) {
+                spin_lock(&sdev->stat_lock);
+                ++sdev->stat.malloc_errors;
+                spin_unlock(&sdev->stat_lock);
+                uncorrectable = 1;
+                goto out;
+        }
+
+        trans = btrfs_join_transaction(fixup->root);
+        if (IS_ERR(trans)) {
+                uncorrectable = 1;
+                goto out;
+        }
+
+        /*
+         * the idea is to trigger a regular read through the standard path. we
+         * read a page from the (failed) logical address by specifying the
+         * corresponding copynum of the failed sector. thus, that readpage is
+         * expected to fail.
+         * that is the point where on-the-fly error correction will kick in
+         * (once it's finished) and rewrite the failed sector if a good copy
+         * can be found.
+         */
+        ret = iterate_inodes_from_logical(fixup->logical, fixup->root->fs_info,
+                                                path, scrub_fixup_readpage,
+                                                fixup);
+        if (ret < 0) {
+                uncorrectable = 1;
+                goto out;
+        }
+        WARN_ON(ret != 1);
+
+        spin_lock(&sdev->stat_lock);
+        ++sdev->stat.corrected_errors;
+        spin_unlock(&sdev->stat_lock);
+
+out:
+        if (trans && !IS_ERR(trans))
+                btrfs_end_transaction(trans, fixup->root);
+        if (uncorrectable) {
+                spin_lock(&sdev->stat_lock);
+                ++sdev->stat.uncorrectable_errors;
+                spin_unlock(&sdev->stat_lock);
+                printk_ratelimited(KERN_ERR "btrfs: unable to fixup "
+                                        "(nodatasum) error at logical %llu\n",
+                                        fixup->logical);
+        }
+
+        btrfs_free_path(path);
+        kfree(fixup);
+
+        /* see caller why we're pretending to be paused in the scrub counters */
+        mutex_lock(&fs_info->scrub_lock);
+        atomic_dec(&fs_info->scrubs_running);
+        atomic_dec(&fs_info->scrubs_paused);
+        mutex_unlock(&fs_info->scrub_lock);
+        atomic_dec(&sdev->fixup_cnt);
+        wake_up(&fs_info->scrub_pause_wait);
+        wake_up(&sdev->list_wait);
+}
+
 /*
  * scrub_recheck_error gets called when either verification of the page
  * failed or the bio failed to read, e.g. with EIO. In the latter case,
  * recheck_error gets called for every page in the bio, even though only
  * one may be bad
  */
-static void scrub_recheck_error(struct scrub_bio *sbio, int ix)
+static int scrub_recheck_error(struct scrub_bio *sbio, int ix)
 {
+        struct scrub_dev *sdev = sbio->sdev;
+        u64 sector = (sbio->physical + ix * PAGE_SIZE) >> 9;
+        static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL,
+                                        DEFAULT_RATELIMIT_BURST);
+
         if (sbio->err) {
-                if (scrub_fixup_io(READ, sbio->sdev->dev->bdev,
-                                   (sbio->physical + ix * PAGE_SIZE) >> 9,
+                if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, sector,
                                    sbio->bio->bi_io_vec[ix].bv_page) == 0) {
                         if (scrub_fixup_check(sbio, ix) == 0)
-                                return;
+                                return 0;
                 }
+                if (__ratelimit(&_rs))
+                        scrub_print_warning("i/o error", sbio, ix);
+        } else {
+                if (__ratelimit(&_rs))
+                        scrub_print_warning("checksum error", sbio, ix);
         }
 
+        spin_lock(&sdev->stat_lock);
+        ++sdev->stat.read_errors;
+        spin_unlock(&sdev->stat_lock);
+
         scrub_fixup(sbio, ix);
+        return 1;
 }
 
 static int scrub_fixup_check(struct scrub_bio *sbio, int ix)
@@ -250,7 +611,8 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix)
         struct scrub_dev *sdev = sbio->sdev;
         struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
         struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
-        struct btrfs_multi_bio *multi = NULL;
+        struct btrfs_bio *bbio = NULL;
+        struct scrub_fixup_nodatasum *fixup;
         u64 logical = sbio->logical + ix * PAGE_SIZE;
         u64 length;
         int i;
@@ -259,38 +621,57 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix)
 
         if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) &&
             (sbio->spag[ix].have_csum == 0)) {
+                fixup = kzalloc(sizeof(*fixup), GFP_NOFS);
+                if (!fixup)
+                        goto uncorrectable;
+                fixup->sdev = sdev;
+                fixup->logical = logical;
+                fixup->root = fs_info->extent_root;
+                fixup->mirror_num = sbio->spag[ix].mirror_num;
                 /*
-                 * nodatasum, don't try to fix anything
-                 * FIXME: we can do better, open the inode and trigger a
-                 * writeback
+                 * increment scrubs_running to prevent cancel requests from
+                 * completing as long as a fixup worker is running. we must also
+                 * increment scrubs_paused to prevent deadlocking on pause
+                 * requests used for transactions commits (as the worker uses a
+                 * transaction context). it is safe to regard the fixup worker
+                 * as paused for all matters practical. effectively, we only
+                 * avoid cancellation requests from completing.
                  */
-                goto uncorrectable;
+                mutex_lock(&fs_info->scrub_lock);
+                atomic_inc(&fs_info->scrubs_running);
+                atomic_inc(&fs_info->scrubs_paused);
+                mutex_unlock(&fs_info->scrub_lock);
+                atomic_inc(&sdev->fixup_cnt);
+                fixup->work.func = scrub_fixup_nodatasum;
+                btrfs_queue_worker(&fs_info->scrub_workers, &fixup->work);
+                return;
         }
 
         length = PAGE_SIZE;
         ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length,
-                              &multi, 0);
-        if (ret || !multi || length < PAGE_SIZE) {
+                              &bbio, 0);
+        if (ret || !bbio || length < PAGE_SIZE) {
                 printk(KERN_ERR
                        "scrub_fixup: btrfs_map_block failed us for %llu\n",
                        (unsigned long long)logical);
                 WARN_ON(1);
+                kfree(bbio);
                 return;
         }
 
-        if (multi->num_stripes == 1)
+        if (bbio->num_stripes == 1)
                 /* there aren't any replicas */
                 goto uncorrectable;
 
         /*
          * first find a good copy
          */
-        for (i = 0; i < multi->num_stripes; ++i) {
-                if (i == sbio->spag[ix].mirror_num)
+        for (i = 0; i < bbio->num_stripes; ++i) {
+                if (i + 1 == sbio->spag[ix].mirror_num)
                         continue;
 
-                if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev,
-                                   multi->stripes[i].physical >> 9,
+                if (scrub_fixup_io(READ, bbio->stripes[i].dev->bdev,
+                                   bbio->stripes[i].physical >> 9,
                                    sbio->bio->bi_io_vec[ix].bv_page)) {
                         /* I/O-error, this is not a good copy */
                         continue;
@@ -299,7 +680,7 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix)
                 if (scrub_fixup_check(sbio, ix) == 0)
                         break;
         }
-        if (i == multi->num_stripes)
+        if (i == bbio->num_stripes)
                 goto uncorrectable;
 
         if (!sdev->readonly) {
@@ -314,25 +695,23 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix)
                 }
         }
 
-        kfree(multi);
+        kfree(bbio);
         spin_lock(&sdev->stat_lock);
         ++sdev->stat.corrected_errors;
         spin_unlock(&sdev->stat_lock);
 
-        if (printk_ratelimit())
-                printk(KERN_ERR "btrfs: fixed up at %llu\n",
-                       (unsigned long long)logical);
+        printk_ratelimited(KERN_ERR "btrfs: fixed up error at logical %llu\n",
+                               (unsigned long long)logical);
         return;
 
 uncorrectable:
-        kfree(multi);
+        kfree(bbio);
         spin_lock(&sdev->stat_lock);
         ++sdev->stat.uncorrectable_errors;
         spin_unlock(&sdev->stat_lock);
 
-        if (printk_ratelimit())
-                printk(KERN_ERR "btrfs: unable to fixup at %llu\n",
-                         (unsigned long long)logical);
+        printk_ratelimited(KERN_ERR "btrfs: unable to fixup (regular) error at "
+                                "logical %llu\n", (unsigned long long)logical);
 }
 
 static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
@@ -382,8 +761,14 @@ static void scrub_checksum(struct btrfs_work *work)
         int ret;
 
         if (sbio->err) {
+                ret = 0;
                 for (i = 0; i < sbio->count; ++i)
-                        scrub_recheck_error(sbio, i);
+                        ret |= scrub_recheck_error(sbio, i);
+                if (!ret) {
+                        spin_lock(&sdev->stat_lock);
+                        ++sdev->stat.unverified_errors;
+                        spin_unlock(&sdev->stat_lock);
+                }
 
                 sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1);
                 sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
@@ -396,10 +781,6 @@ static void scrub_checksum(struct btrfs_work *work)
                         bi->bv_offset = 0;
                         bi->bv_len = PAGE_SIZE;
                 }
-
-                spin_lock(&sdev->stat_lock);
-                ++sdev->stat.read_errors;
-                spin_unlock(&sdev->stat_lock);
                 goto out;
         }
         for (i = 0; i < sbio->count; ++i) {
@@ -420,8 +801,14 @@ static void scrub_checksum(struct btrfs_work *work)
                         WARN_ON(1);
                 }
                 kunmap_atomic(buffer, KM_USER0);
-                if (ret)
-                        scrub_recheck_error(sbio, i);
+                if (ret) {
+                        ret = scrub_recheck_error(sbio, i);
+                        if (!ret) {
+                                spin_lock(&sdev->stat_lock);
+                                ++sdev->stat.unverified_errors;
+                                spin_unlock(&sdev->stat_lock);
+                        }
+                }
         }
 
 out:
@@ -604,7 +991,7 @@ nomem:
 }
 
 static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len,
-                      u64 physical, u64 flags, u64 gen, u64 mirror_num,
+                      u64 physical, u64 flags, u64 gen, int mirror_num,
                       u8 *csum, int force)
 {
         struct scrub_bio *sbio;
@@ -701,7 +1088,7 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len,
 
 /* scrub extent tries to collect up to 64 kB for each bio */
 static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len,
-                        u64 physical, u64 flags, u64 gen, u64 mirror_num)
+                        u64 physical, u64 flags, u64 gen, int mirror_num)
 {
         int ret;
         u8 csum[BTRFS_CSUM_SIZE];
@@ -741,13 +1128,16 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
         int slot;
         int i;
         u64 nstripes;
-        int start_stripe;
         struct extent_buffer *l;
         struct btrfs_key key;
         u64 physical;
         u64 logical;
         u64 generation;
-        u64 mirror_num;
+        int mirror_num;
+        struct reada_control *reada1;
+        struct reada_control *reada2;
+        struct btrfs_key key_start;
+        struct btrfs_key key_end;
 
         u64 increment = map->stripe_len;
         u64 offset;
@@ -758,102 +1148,88 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
         if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
                 offset = map->stripe_len * num;
                 increment = map->stripe_len * map->num_stripes;
-                mirror_num = 0;
+                mirror_num = 1;
         } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
                 int factor = map->num_stripes / map->sub_stripes;
                 offset = map->stripe_len * (num / map->sub_stripes);
                 increment = map->stripe_len * factor;
-                mirror_num = num % map->sub_stripes;
+                mirror_num = num % map->sub_stripes + 1;
         } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
                 increment = map->stripe_len;
-                mirror_num = num % map->num_stripes;
+                mirror_num = num % map->num_stripes + 1;
         } else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
                 increment = map->stripe_len;
-                mirror_num = num % map->num_stripes;
+                mirror_num = num % map->num_stripes + 1;
         } else {
                 increment = map->stripe_len;
-                mirror_num = 0;
+                mirror_num = 1;
         }
 
         path = btrfs_alloc_path();
         if (!path)
                 return -ENOMEM;
 
-        path->reada = 2;
         path->search_commit_root = 1;
         path->skip_locking = 1;
 
         /*
-         * find all extents for each stripe and just read them to get
-         * them into the page cache
-         * FIXME: we can do better. build a more intelligent prefetching
+         * trigger the readahead for extent tree csum tree and wait for
+         * completion. During readahead, the scrub is officially paused
+         * to not hold off transaction commits
          */
         logical = base + offset;
-        physical = map->stripes[num].physical;
-        ret = 0;
-        for (i = 0; i < nstripes; ++i) {
-                key.objectid = logical;
-                key.type = BTRFS_EXTENT_ITEM_KEY;
-                key.offset = (u64)0;
-
-                ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-                if (ret < 0)
-                        goto out_noplug;
-
-                /*
-                 * we might miss half an extent here, but that doesn't matter,
-                 * as it's only the prefetch
-                 */
-                while (1) {
-                        l = path->nodes[0];
-                        slot = path->slots[0];
-                        if (slot >= btrfs_header_nritems(l)) {
-                                ret = btrfs_next_leaf(root, path);
-                                if (ret == 0)
-                                        continue;
-                                if (ret < 0)
-                                        goto out_noplug;
 
-                                break;
-                        }
-                        btrfs_item_key_to_cpu(l, &key, slot);
+        wait_event(sdev->list_wait,
+                   atomic_read(&sdev->in_flight) == 0);
+        atomic_inc(&fs_info->scrubs_paused);
+        wake_up(&fs_info->scrub_pause_wait);
 
-                        if (key.objectid >= logical + map->stripe_len)
-                                break;
+        /* FIXME it might be better to start readahead at commit root */
+        key_start.objectid = logical;
+        key_start.type = BTRFS_EXTENT_ITEM_KEY;
+        key_start.offset = (u64)0;
+        key_end.objectid = base + offset + nstripes * increment;
+        key_end.type = BTRFS_EXTENT_ITEM_KEY;
+        key_end.offset = (u64)0;
+        reada1 = btrfs_reada_add(root, &key_start, &key_end);
+
+        key_start.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
+        key_start.type = BTRFS_EXTENT_CSUM_KEY;
+        key_start.offset = logical;
+        key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
+        key_end.type = BTRFS_EXTENT_CSUM_KEY;
+        key_end.offset = base + offset + nstripes * increment;
+        reada2 = btrfs_reada_add(csum_root, &key_start, &key_end);
+
+        if (!IS_ERR(reada1))
+                btrfs_reada_wait(reada1);
+        if (!IS_ERR(reada2))
+                btrfs_reada_wait(reada2);
 
-                        path->slots[0]++;
-                }
-                btrfs_release_path(path);
-                logical += increment;
-                physical += map->stripe_len;
-                cond_resched();
+        mutex_lock(&fs_info->scrub_lock);
+        while (atomic_read(&fs_info->scrub_pause_req)) {
+                mutex_unlock(&fs_info->scrub_lock);
+                wait_event(fs_info->scrub_pause_wait,
+                   atomic_read(&fs_info->scrub_pause_req) == 0);
+                mutex_lock(&fs_info->scrub_lock);
         }
+        atomic_dec(&fs_info->scrubs_paused);
+        mutex_unlock(&fs_info->scrub_lock);
+        wake_up(&fs_info->scrub_pause_wait);
 
         /*
          * collect all data csums for the stripe to avoid seeking during
          * the scrub. This might currently (crc32) end up to be about 1MB
          */
-        start_stripe = 0;
         blk_start_plug(&plug);
-again:
-        logical = base + offset + start_stripe * increment;
-        for (i = start_stripe; i < nstripes; ++i) {
-                ret = btrfs_lookup_csums_range(csum_root, logical,
-                                               logical + map->stripe_len - 1,
-                                               &sdev->csum_list, 1);
-                if (ret)
-                        goto out;
 
-                logical += increment;
-                cond_resched();
-        }
         /*
          * now find all extents for each stripe and scrub them
          */
-        logical = base + offset + start_stripe * increment;
-        physical = map->stripes[num].physical + start_stripe * map->stripe_len;
+        logical = base + offset;
+        physical = map->stripes[num].physical;
         ret = 0;
-        for (i = start_stripe; i < nstripes; ++i) {
+        for (i = 0; i < nstripes; ++i) {
                 /*
                  * canceled?
                  */
@@ -882,11 +1258,14 @@ again:
                         atomic_dec(&fs_info->scrubs_paused);
                         mutex_unlock(&fs_info->scrub_lock);
                         wake_up(&fs_info->scrub_pause_wait);
-                        scrub_free_csums(sdev);
-                        start_stripe = i;
-                        goto again;
                 }
 
+                ret = btrfs_lookup_csums_range(csum_root, logical,
+                                               logical + map->stripe_len - 1,
+                                               &sdev->csum_list, 1);
+                if (ret)
+                        goto out;
+
                 key.objectid = logical;
                 key.type = BTRFS_EXTENT_ITEM_KEY;
                 key.offset = (u64)0;
@@ -982,7 +1361,6 @@ next:
 
 out:
         blk_finish_plug(&plug);
-out_noplug:
         btrfs_free_path(path);
         return ret < 0 ? ret : 0;
 }
@@ -1253,10 +1631,11 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
                 ret = scrub_enumerate_chunks(sdev, start, end);
 
         wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0);
-
         atomic_dec(&fs_info->scrubs_running);
         wake_up(&fs_info->scrub_pause_wait);
 
+        wait_event(sdev->list_wait, atomic_read(&sdev->fixup_cnt) == 0);
+
         if (progress)
                 memcpy(progress, &sdev->stat, sizeof(*progress));
 
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 15634d4648d7..57080dffdfc6 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -40,6 +40,7 @@
 #include <linux/magic.h>
 #include <linux/slab.h>
 #include <linux/cleancache.h>
+#include <linux/mnt_namespace.h>
 #include "compat.h"
 #include "delayed-inode.h"
 #include "ctree.h"
@@ -58,6 +59,7 @@
 #include <trace/events/btrfs.h>
 
 static const struct super_operations btrfs_super_ops;
+static struct file_system_type btrfs_fs_type;
 
 static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno,
                                       char nbuf[16])
@@ -162,7 +164,7 @@ enum {
         Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
         Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
         Opt_enospc_debug, Opt_subvolrootid, Opt_defrag,
-        Opt_inode_cache, Opt_err,
+        Opt_inode_cache, Opt_no_space_cache, Opt_recovery, Opt_err,
 };
 
 static match_table_t tokens = {
@@ -195,6 +197,8 @@ static match_table_t tokens = {
         {Opt_subvolrootid, "subvolrootid=%d"},
         {Opt_defrag, "autodefrag"},
         {Opt_inode_cache, "inode_cache"},
+        {Opt_no_space_cache, "no_space_cache"},
+        {Opt_recovery, "recovery"},
         {Opt_err, NULL},
 };
 
@@ -206,14 +210,19 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 {
         struct btrfs_fs_info *info = root->fs_info;
         substring_t args[MAX_OPT_ARGS];
-        char *p, *num, *orig;
+        char *p, *num, *orig = NULL;
+        u64 cache_gen;
         int intarg;
         int ret = 0;
         char *compress_type;
         bool compress_force = false;
 
+        cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy);
+        if (cache_gen)
+                btrfs_set_opt(info->mount_opt, SPACE_CACHE);
+
         if (!options)
-                return 0;
+                goto out;
 
         /*
          * strsep changes the string, duplicate it because parse_options
@@ -360,9 +369,12 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
                         btrfs_set_opt(info->mount_opt, DISCARD);
                         break;
                 case Opt_space_cache:
-                        printk(KERN_INFO "btrfs: enabling disk space caching\n");
                         btrfs_set_opt(info->mount_opt, SPACE_CACHE);
                         break;
+                case Opt_no_space_cache:
+                        printk(KERN_INFO "btrfs: disabling disk space caching\n");
+                        btrfs_clear_opt(info->mount_opt, SPACE_CACHE);
+                        break;
                 case Opt_inode_cache:
                         printk(KERN_INFO "btrfs: enabling inode map caching\n");
                         btrfs_set_opt(info->mount_opt, INODE_MAP_CACHE);
@@ -381,6 +393,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
                         printk(KERN_INFO "btrfs: enabling auto defrag");
                         btrfs_set_opt(info->mount_opt, AUTO_DEFRAG);
                         break;
+                case Opt_recovery:
+                        printk(KERN_INFO "btrfs: enabling auto recovery");
+                        btrfs_set_opt(info->mount_opt, RECOVERY);
+                        break;
                 case Opt_err:
                         printk(KERN_INFO "btrfs: unrecognized mount option "
                                "'%s'\n", p);
@@ -391,6 +407,8 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
                 }
         }
 out:
+        if (!ret && btrfs_test_opt(root, SPACE_CACHE))
+                printk(KERN_INFO "btrfs: disk space caching is enabled\n");
         kfree(orig);
         return ret;
 }
@@ -406,12 +424,12 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags,
                 u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices)
 {
         substring_t args[MAX_OPT_ARGS];
-        char *opts, *orig, *p;
+        char *device_name, *opts, *orig, *p;
         int error = 0;
         int intarg;
 
         if (!options)
-                goto out;
+                return 0;
 
         /*
          * strsep changes the string, duplicate it because parse_options
@@ -457,29 +475,24 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags,
                         }
                         break;
                 case Opt_device:
-                        error = btrfs_scan_one_device(match_strdup(&args[0]),
+                        device_name = match_strdup(&args[0]);
+                        if (!device_name) {
+                                error = -ENOMEM;
+                                goto out;
+                        }
+                        error = btrfs_scan_one_device(device_name,
                                         flags, holder, fs_devices);
+                        kfree(device_name);
                         if (error)
-                                goto out_free_opts;
+                                goto out;
                         break;
                 default:
                         break;
                 }
         }
 
- out_free_opts:
+out:
         kfree(orig);
- out:
-        /*
-         * If no subvolume name is specified we use the default one.  Allocate
-         * a copy of the string "." here so that code later in the
-         * mount path doesn't care if it's the default volume or another one.
-         */
-        if (!*subvol_name) {
-                *subvol_name = kstrdup(".", GFP_KERNEL);
-                if (!*subvol_name)
-                        return -ENOMEM;
-        }
         return error;
 }
 
@@ -492,7 +505,6 @@ static struct dentry *get_default_root(struct super_block *sb,
         struct btrfs_path *path;
         struct btrfs_key location;
         struct inode *inode;
-        struct dentry *dentry;
         u64 dir_id;
         int new = 0;
 
@@ -517,7 +529,7 @@ static struct dentry *get_default_root(struct super_block *sb,
          * will mount by default if we haven't been given a specific subvolume
          * to mount.
          */
-        dir_id = btrfs_super_root_dir(&root->fs_info->super_copy);
+        dir_id = btrfs_super_root_dir(root->fs_info->super_copy);
         di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0);
         if (IS_ERR(di)) {
                 btrfs_free_path(path);
@@ -566,29 +578,7 @@ setup_root:
                 return dget(sb->s_root);
         }
 
-        if (new) {
-                const struct qstr name = { .name = "/", .len = 1 };
-
-                /*
-                 * New inode, we need to make the dentry a sibling of s_root so
-                 * everything gets cleaned up properly on unmount.
-                 */
-                dentry = d_alloc(sb->s_root, &name);
-                if (!dentry) {
-                        iput(inode);
-                        return ERR_PTR(-ENOMEM);
-                }
-                d_splice_alias(inode, dentry);
-        } else {
-                /*
-                 * We found the inode in cache, just find a dentry for it and
-                 * put the reference to the inode we just got.
-                 */
-                dentry = d_find_alias(inode);
-                iput(inode);
-        }
-
-        return dentry;
+        return d_obtain_alias(inode);
 }
 
 static int btrfs_fill_super(struct super_block *sb,
@@ -719,6 +709,8 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs)
                 seq_puts(seq, ",noacl");
         if (btrfs_test_opt(root, SPACE_CACHE))
                 seq_puts(seq, ",space_cache");
+        else
+                seq_puts(seq, ",no_space_cache");
         if (btrfs_test_opt(root, CLEAR_CACHE))
                 seq_puts(seq, ",clear_cache");
         if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED))
@@ -753,6 +745,137 @@ static int btrfs_set_super(struct super_block *s, void *data)
         return set_anon_super(s, data);
 }
 
+/*
+ * subvolumes are identified by ino 256
+ */
+static inline int is_subvolume_inode(struct inode *inode)
+{
+        if (inode && inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+                return 1;
+        return 0;
+}
+
+/*
+ * This will strip out the subvol=%s argument for an argument string and add
+ * subvolid=0 to make sure we get the actual tree root for path walking to the
+ * subvol we want.
+ */
+static char *setup_root_args(char *args)
+{
+        unsigned copied = 0;
+        unsigned len = strlen(args) + 2;
+        char *pos;
+        char *ret;
+
+        /*
+         * We need the same args as before, but minus
+         *
+         * subvol=a
+         *
+         * and add
+         *
+         * subvolid=0
+         *
+         * which is a difference of 2 characters, so we allocate strlen(args) +
+         * 2 characters.
+         */
+        ret = kzalloc(len * sizeof(char), GFP_NOFS);
+        if (!ret)
+                return NULL;
+        pos = strstr(args, "subvol=");
+
+        /* This shouldn't happen, but just in case.. */
+        if (!pos) {
+                kfree(ret);
+                return NULL;
+        }
+
+        /*
+         * The subvol=<> arg is not at the front of the string, copy everybody
+         * up to that into ret.
+         */
+        if (pos != args) {
+                *pos = '\0';
+                strcpy(ret, args);
+                copied += strlen(args);
+                pos++;
+        }
+
+        strncpy(ret + copied, "subvolid=0", len - copied);
+
+        /* Length of subvolid=0 */
+        copied += 10;
+
+        /*
+         * If there is no , after the subvol= option then we know there's no
+         * other options and we can just return.
+         */
+        pos = strchr(pos, ',');
+        if (!pos)
+                return ret;
+
+        /* Copy the rest of the arguments into our buffer */
+        strncpy(ret + copied, pos, len - copied);
+        copied += strlen(pos);
+
+        return ret;
+}
+
+static struct dentry *mount_subvol(const char *subvol_name, int flags,
+                                   const char *device_name, char *data)
+{
+        struct super_block *s;
+        struct dentry *root;
+        struct vfsmount *mnt;
+        struct mnt_namespace *ns_private;
+        char *newargs;
+        struct path path;
+        int error;
+
+        newargs = setup_root_args(data);
+        if (!newargs)
+                return ERR_PTR(-ENOMEM);
+        mnt = vfs_kern_mount(&btrfs_fs_type, flags, device_name,
+                             newargs);
+        kfree(newargs);
+        if (IS_ERR(mnt))
+                return ERR_CAST(mnt);
+
+        ns_private = create_mnt_ns(mnt);
+        if (IS_ERR(ns_private)) {
+                mntput(mnt);
+                return ERR_CAST(ns_private);
+        }
+
+        /*
+         * This will trigger the automount of the subvol so we can just
+         * drop the mnt we have here and return the dentry that we
+         * found.
+         */
+        error = vfs_path_lookup(mnt->mnt_root, mnt, subvol_name,
+                                LOOKUP_FOLLOW, &path);
+        put_mnt_ns(ns_private);
+        if (error)
+                return ERR_PTR(error);
+
+        if (!is_subvolume_inode(path.dentry->d_inode)) {
+                path_put(&path);
+                mntput(mnt);
+                error = -EINVAL;
+                printk(KERN_ERR "btrfs: '%s' is not a valid subvolume\n",
+                                subvol_name);
+                return ERR_PTR(-EINVAL);
+        }
+
+        /* Get a ref to the sb and the dentry we found and return it */
+        s = path.mnt->mnt_sb;
+        atomic_inc(&s->s_active);
+        root = dget(path.dentry);
+        path_put(&path);
+        down_write(&s->s_umount);
+
+        return root;
+}
 
 /*
  * Find a superblock for the given device / mount point.
@@ -784,13 +907,19 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
         if (error)
                 return ERR_PTR(error);
 
+        if (subvol_name) {
+                root = mount_subvol(subvol_name, flags, device_name, data);
+                kfree(subvol_name);
+                return root;
+        }
+
         error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices);
         if (error)
-                goto error_free_subvol_name;
+                return ERR_PTR(error);
 
         error = btrfs_open_devices(fs_devices, mode, fs_type);
         if (error)
-                goto error_free_subvol_name;
+                return ERR_PTR(error);
 
         if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) {
                 error = -EACCES;
@@ -813,88 +942,57 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
         fs_info->fs_devices = fs_devices;
         tree_root->fs_info = fs_info;
 
+        fs_info->super_copy = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS);
+        fs_info->super_for_commit = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS);
+        if (!fs_info->super_copy || !fs_info->super_for_commit) {
+                error = -ENOMEM;
+                goto error_close_devices;
+        }
+
         bdev = fs_devices->latest_bdev;
         s = sget(fs_type, btrfs_test_super, btrfs_set_super, tree_root);
-        if (IS_ERR(s))
-                goto error_s;
+        if (IS_ERR(s)) {
+                error = PTR_ERR(s);
+                goto error_close_devices;
+        }
 
         if (s->s_root) {
                 if ((flags ^ s->s_flags) & MS_RDONLY) {
                         deactivate_locked_super(s);
-                        error = -EBUSY;
-                        goto error_close_devices;
+                        return ERR_PTR(-EBUSY);
                 }
 
                 btrfs_close_devices(fs_devices);
-                kfree(fs_info);
+                free_fs_info(fs_info);
                 kfree(tree_root);
         } else {
                 char b[BDEVNAME_SIZE];
 
                 s->s_flags = flags | MS_NOSEC;
                 strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
+                btrfs_sb(s)->fs_info->bdev_holder = fs_type;
                 error = btrfs_fill_super(s, fs_devices, data,
                                          flags & MS_SILENT ? 1 : 0);
                 if (error) {
                         deactivate_locked_super(s);
-                        goto error_free_subvol_name;
+                        return ERR_PTR(error);
                 }
 
-                btrfs_sb(s)->fs_info->bdev_holder = fs_type;
                 s->s_flags |= MS_ACTIVE;
         }
 
-        /* if they gave us a subvolume name bind mount into that */
-        if (strcmp(subvol_name, ".")) {
-                struct dentry *new_root;
-
-                root = get_default_root(s, subvol_rootid);
-                if (IS_ERR(root)) {
-                        error = PTR_ERR(root);
-                        deactivate_locked_super(s);
-                        goto error_free_subvol_name;
-                }
-
-                mutex_lock(&root->d_inode->i_mutex);
-                new_root = lookup_one_len(subvol_name, root,
-                                      strlen(subvol_name));
-                mutex_unlock(&root->d_inode->i_mutex);
-
-                if (IS_ERR(new_root)) {
-                        dput(root);
-                        deactivate_locked_super(s);
-                        error = PTR_ERR(new_root);
-                        goto error_free_subvol_name;
-                }
-                if (!new_root->d_inode) {
-                        dput(root);
-                        dput(new_root);
-                        deactivate_locked_super(s);
-                        error = -ENXIO;
-                        goto error_free_subvol_name;
-                }
-                dput(root);
-                root = new_root;
-        } else {
-                root = get_default_root(s, subvol_objectid);
-                if (IS_ERR(root)) {
-                        error = PTR_ERR(root);
-                        deactivate_locked_super(s);
-                        goto error_free_subvol_name;
-                }
+        root = get_default_root(s, subvol_objectid);
+        if (IS_ERR(root)) {
+                deactivate_locked_super(s);
+                return root;
         }
 
-        kfree(subvol_name);
         return root;
 
-error_s:
-        error = PTR_ERR(s);
 error_close_devices:
         btrfs_close_devices(fs_devices);
-        kfree(fs_info);
+        free_fs_info(fs_info);
         kfree(tree_root);
-error_free_subvol_name:
-        kfree(subvol_name);
         return ERR_PTR(error);
 }
 
@@ -919,7 +1017,7 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
                 if (root->fs_info->fs_devices->rw_devices == 0)
                         return -EACCES;
 
-                if (btrfs_super_log_root(&root->fs_info->super_copy) != 0)
+                if (btrfs_super_log_root(root->fs_info->super_copy) != 0)
                         return -EINVAL;
 
                 ret = btrfs_cleanup_fs_roots(root->fs_info);
@@ -1085,7 +1183,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes)
 static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 {
         struct btrfs_root *root = btrfs_sb(dentry->d_sb);
-        struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
+        struct btrfs_super_block *disk_super = root->fs_info->super_copy;
         struct list_head *head = &root->fs_info->space_info;
         struct btrfs_space_info *found;
         u64 total_used = 0;
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index e24b7964a155..960835eaf4da 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -55,6 +55,7 @@ static noinline int join_transaction(struct btrfs_root *root, int nofail)
         struct btrfs_transaction *cur_trans;
 
         spin_lock(&root->fs_info->trans_lock);
+loop:
         if (root->fs_info->trans_no_join) {
                 if (!nofail) {
                         spin_unlock(&root->fs_info->trans_lock);
@@ -75,16 +76,18 @@ static noinline int join_transaction(struct btrfs_root *root, int nofail)
         cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS);
         if (!cur_trans)
                 return -ENOMEM;
+
         spin_lock(&root->fs_info->trans_lock);
         if (root->fs_info->running_transaction) {
+                /*
+                 * someone started a transaction after we unlocked.  Make sure
+                 * to redo the trans_no_join checks above
+                 */
                 kmem_cache_free(btrfs_transaction_cachep, cur_trans);
                 cur_trans = root->fs_info->running_transaction;
-                atomic_inc(&cur_trans->use_count);
-                atomic_inc(&cur_trans->num_writers);
-                cur_trans->num_joined++;
-                spin_unlock(&root->fs_info->trans_lock);
-                return 0;
+                goto loop;
         }
+
         atomic_set(&cur_trans->num_writers, 1);
         cur_trans->num_joined = 0;
         init_waitqueue_head(&cur_trans->writer_wait);
@@ -275,7 +278,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
          */
         if (num_items > 0 && root != root->fs_info->chunk_root) {
                 num_bytes = btrfs_calc_trans_metadata_size(root, num_items);
-                ret = btrfs_block_rsv_add(NULL, root,
+                ret = btrfs_block_rsv_add(root,
                                           &root->fs_info->trans_block_rsv,
                                           num_bytes);
                 if (ret)
@@ -418,8 +421,8 @@ static int should_end_transaction(struct btrfs_trans_handle *trans,
                                   struct btrfs_root *root)
 {
         int ret;
-        ret = btrfs_block_rsv_check(trans, root,
-                                    &root->fs_info->global_block_rsv, 0, 5);
+
+        ret = btrfs_block_rsv_check(root, &root->fs_info->global_block_rsv, 5);
         return ret ? 1 : 0;
 }
 
@@ -427,17 +430,26 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
                                  struct btrfs_root *root)
 {
         struct btrfs_transaction *cur_trans = trans->transaction;
+        struct btrfs_block_rsv *rsv = trans->block_rsv;
         int updates;
 
         smp_mb();
         if (cur_trans->blocked || cur_trans->delayed_refs.flushing)
                 return 1;
 
+        /*
+         * We need to do this in case we're deleting csums so the global block
+         * rsv get's used instead of the csum block rsv.
+         */
+        trans->block_rsv = NULL;
+
         updates = trans->delayed_ref_updates;
         trans->delayed_ref_updates = 0;
         if (updates)
                 btrfs_run_delayed_refs(trans, root, updates);
 
+        trans->block_rsv = rsv;
+
         return should_end_transaction(trans, root);
 }
 
@@ -453,6 +465,8 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
                 return 0;
         }
 
+        btrfs_trans_release_metadata(trans, root);
+        trans->block_rsv = NULL;
         while (count < 4) {
                 unsigned long cur = trans->delayed_ref_updates;
                 trans->delayed_ref_updates = 0;
@@ -473,8 +487,6 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
                 count++;
         }
 
-        btrfs_trans_release_metadata(trans, root);
-
         if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) &&
             should_end_transaction(trans, root)) {
                 trans->transaction->blocked = 1;
@@ -562,50 +574,21 @@ int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans,
 int btrfs_write_marked_extents(struct btrfs_root *root,
                                struct extent_io_tree *dirty_pages, int mark)
 {
-        int ret;
         int err = 0;
         int werr = 0;
-        struct page *page;
-        struct inode *btree_inode = root->fs_info->btree_inode;
+        struct address_space *mapping = root->fs_info->btree_inode->i_mapping;
         u64 start = 0;
         u64 end;
-        unsigned long index;
-
-        while (1) {
-                ret = find_first_extent_bit(dirty_pages, start, &start, &end,
-                                            mark);
-                if (ret)
-                        break;
-                while (start <= end) {
-                        cond_resched();
-
-                        index = start >> PAGE_CACHE_SHIFT;
-                        start = (u64)(index + 1) << PAGE_CACHE_SHIFT;
-                        page = find_get_page(btree_inode->i_mapping, index);
-                        if (!page)
-                                continue;
-
-                        btree_lock_page_hook(page);
-                        if (!page->mapping) {
-                                unlock_page(page);
-                                page_cache_release(page);
-                                continue;
-                        }
 
-                        if (PageWriteback(page)) {
-                                if (PageDirty(page))
-                                        wait_on_page_writeback(page);
-                                else {
-                                        unlock_page(page);
-                                        page_cache_release(page);
-                                        continue;
-                                }
-                        }
-                        err = write_one_page(page, 0);
-                        if (err)
-                                werr = err;
-                        page_cache_release(page);
-                }
+        while (!find_first_extent_bit(dirty_pages, start, &start, &end,
+                                      mark)) {
+                convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT, mark,
+                                   GFP_NOFS);
+                err = filemap_fdatawrite_range(mapping, start, end);
+                if (err)
+                        werr = err;
+                cond_resched();
+                start = end + 1;
         }
         if (err)
                 werr = err;
@@ -621,39 +604,20 @@ int btrfs_write_marked_extents(struct btrfs_root *root,
 int btrfs_wait_marked_extents(struct btrfs_root *root,
                               struct extent_io_tree *dirty_pages, int mark)
 {
-        int ret;
         int err = 0;
         int werr = 0;
-        struct page *page;
-        struct inode *btree_inode = root->fs_info->btree_inode;
+        struct address_space *mapping = root->fs_info->btree_inode->i_mapping;
         u64 start = 0;
         u64 end;
-        unsigned long index;
-
-        while (1) {
-                ret = find_first_extent_bit(dirty_pages, start, &start, &end,
-                                            mark);
-                if (ret)
-                        break;
 
-                clear_extent_bits(dirty_pages, start, end, mark, GFP_NOFS);
-                while (start <= end) {
-                        index = start >> PAGE_CACHE_SHIFT;
-                        start = (u64)(index + 1) << PAGE_CACHE_SHIFT;
-                        page = find_get_page(btree_inode->i_mapping, index);
-                        if (!page)
-                                continue;
-                        if (PageDirty(page)) {
-                                btree_lock_page_hook(page);
-                                wait_on_page_writeback(page);
-                                err = write_one_page(page, 0);
-                                if (err)
-                                        werr = err;
-                        }
-                        wait_on_page_writeback(page);
-                        page_cache_release(page);
-                        cond_resched();
-                }
+        while (!find_first_extent_bit(dirty_pages, start, &start, &end,
+                                      EXTENT_NEED_WAIT)) {
+                clear_extent_bits(dirty_pages, start, end, EXTENT_NEED_WAIT, GFP_NOFS);
+                err = filemap_fdatawait_range(mapping, start, end);
+                if (err)
+                        werr = err;
+                cond_resched();
+                start = end + 1;
         }
         if (err)
                 werr = err;
@@ -673,7 +637,12 @@ int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
 
         ret = btrfs_write_marked_extents(root, dirty_pages, mark);
         ret2 = btrfs_wait_marked_extents(root, dirty_pages, mark);
-        return ret || ret2;
+
+        if (ret)
+                return ret;
+        if (ret2)
+                return ret2;
+        return 0;
 }
 
 int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
@@ -911,10 +880,9 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
         }
 
         btrfs_reloc_pre_snapshot(trans, pending, &to_reserve);
-        btrfs_orphan_pre_snapshot(trans, pending, &to_reserve);
 
         if (to_reserve > 0) {
-                ret = btrfs_block_rsv_add(trans, root, &pending->block_rsv,
+                ret = btrfs_block_rsv_add(root, &pending->block_rsv,
                                           to_reserve);
                 if (ret) {
                         pending->error = ret;
@@ -1002,7 +970,6 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
         BUG_ON(IS_ERR(pending->snap));
 
         btrfs_reloc_post_snapshot(trans, pending);
-        btrfs_orphan_post_snapshot(trans, pending);
 fail:
         kfree(new_root_item);
         trans->block_rsv = rsv;
@@ -1032,7 +999,7 @@ static void update_super_roots(struct btrfs_root *root)
         struct btrfs_root_item *root_item;
         struct btrfs_super_block *super;
 
-        super = &root->fs_info->super_copy;
+        super = root->fs_info->super_copy;
 
         root_item = &root->fs_info->chunk_root->root_item;
         super->chunk_root = root_item->bytenr;
@@ -1043,7 +1010,7 @@ static void update_super_roots(struct btrfs_root *root)
         super->root = root_item->bytenr;
         super->generation = root_item->generation;
         super->root_level = root_item->level;
-        if (super->cache_generation != 0 || btrfs_test_opt(root, SPACE_CACHE))
+        if (btrfs_test_opt(root, SPACE_CACHE))
                 super->cache_generation = root_item->generation;
 }
 
@@ -1168,14 +1135,15 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
         btrfs_run_ordered_operations(root, 0);
 
+        btrfs_trans_release_metadata(trans, root);
+        trans->block_rsv = NULL;
+
         /* make a pass through all the delayed refs we have so far
          * any runnings procs may add more while we are here
          */
         ret = btrfs_run_delayed_refs(trans, root, 0);
         BUG_ON(ret);
 
-        btrfs_trans_release_metadata(trans, root);
-
         cur_trans = trans->transaction;
         /*
          * set the flushing flag so procs in this transaction have to
@@ -1341,12 +1309,12 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
         update_super_roots(root);
 
         if (!root->fs_info->log_root_recovering) {
-                btrfs_set_super_log_root(&root->fs_info->super_copy, 0);
-                btrfs_set_super_log_root_level(&root->fs_info->super_copy, 0);
+                btrfs_set_super_log_root(root->fs_info->super_copy, 0);
+                btrfs_set_super_log_root_level(root->fs_info->super_copy, 0);
         }
 
-        memcpy(&root->fs_info->super_for_commit, &root->fs_info->super_copy,
-               sizeof(root->fs_info->super_copy));
+        memcpy(root->fs_info->super_for_commit, root->fs_info->super_copy,
+               sizeof(*root->fs_info->super_copy));
 
         trans->transaction->blocked = 0;
         spin_lock(&root->fs_info->trans_lock);
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 0618aa39740b..3568374d419d 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -276,8 +276,9 @@ static int process_one_buffer(struct btrfs_root *log,
                               struct walk_control *wc, u64 gen)
 {
         if (wc->pin)
-                btrfs_pin_extent(log->fs_info->extent_root,
-                                 eb->start, eb->len, 0);
+                btrfs_pin_extent_for_log_replay(wc->trans,
+                                                log->fs_info->extent_root,
+                                                eb->start, eb->len);
 
         if (btrfs_buffer_uptodate(eb, gen)) {
                 if (wc->write)
@@ -1760,7 +1761,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
                                 WARN_ON(root_owner !=
                                         BTRFS_TREE_LOG_OBJECTID);
-                                ret = btrfs_free_reserved_extent(root,
+                                ret = btrfs_free_and_pin_reserved_extent(root,
                                                          bytenr, blocksize);
                                 BUG_ON(ret);
                         }
@@ -1828,7 +1829,7 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
                                 btrfs_tree_unlock(next);
 
                                 WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
-                                ret = btrfs_free_reserved_extent(root,
+                                ret = btrfs_free_and_pin_reserved_extent(root,
                                                 path->nodes[*level]->start,
                                                 path->nodes[*level]->len);
                                 BUG_ON(ret);
@@ -1897,7 +1898,7 @@ static int walk_log_tree(struct btrfs_trans_handle *trans,
 
                         WARN_ON(log->root_key.objectid !=
                                 BTRFS_TREE_LOG_OBJECTID);
-                        ret = btrfs_free_reserved_extent(log, next->start,
+                        ret = btrfs_free_and_pin_reserved_extent(log, next->start,
                                                          next->len);
                         BUG_ON(ret);
                 }
@@ -2013,10 +2014,10 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
         /* wait for previous tree log sync to complete */
         if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
                 wait_log_commit(trans, root, root->log_transid - 1);
-
         while (1) {
                 unsigned long batch = root->log_batch;
-                if (root->log_multiple_pids) {
+                /* when we're on an ssd, just kick the log commit out */
+                if (!btrfs_test_opt(root, SSD) && root->log_multiple_pids) {
                         mutex_unlock(&root->log_mutex);
                         schedule_timeout_uninterruptible(1);
                         mutex_lock(&root->log_mutex);
@@ -2117,9 +2118,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
         BUG_ON(ret);
         btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
 
-        btrfs_set_super_log_root(&root->fs_info->super_for_commit,
+        btrfs_set_super_log_root(root->fs_info->super_for_commit,
                                 log_root_tree->node->start);
-        btrfs_set_super_log_root_level(&root->fs_info->super_for_commit,
+        btrfs_set_super_log_root_level(root->fs_info->super_for_commit,
                                 btrfs_header_level(log_root_tree->node));
 
         log_root_tree->log_batch = 0;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index f2a4cc79da61..f8e2943101a1 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -366,6 +366,14 @@ static noinline int device_list_add(const char *path,
                 }
                 INIT_LIST_HEAD(&device->dev_alloc_list);
 
+                /* init readahead state */
+                spin_lock_init(&device->reada_lock);
+                device->reada_curr_zone = NULL;
+                atomic_set(&device->reada_in_flight, 0);
+                device->reada_next = 0;
+                INIT_RADIX_TREE(&device->reada_zones, GFP_NOFS & ~__GFP_WAIT);
+                INIT_RADIX_TREE(&device->reada_extents, GFP_NOFS & ~__GFP_WAIT);
+
                 mutex_lock(&fs_devices->device_list_mutex);
                 list_add_rcu(&device->dev_list, &fs_devices->devices);
                 mutex_unlock(&fs_devices->device_list_mutex);
@@ -597,10 +605,8 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
                 set_blocksize(bdev, 4096);
 
                 bh = btrfs_read_dev_super(bdev);
-                if (!bh) {
-                        ret = -EINVAL;
+                if (!bh)
                         goto error_close;
-                }
 
                 disk_super = (struct btrfs_super_block *)bh->b_data;
                 devid = btrfs_stack_device_id(&disk_super->dev_item);
@@ -655,7 +661,7 @@ error:
                 continue;
         }
         if (fs_devices->open_devices == 0) {
-                ret = -EIO;
+                ret = -EINVAL;
                 goto out;
         }
         fs_devices->seeding = seeding;
@@ -1013,8 +1019,13 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
         }
         BUG_ON(ret);
 
-        if (device->bytes_used > 0)
-                device->bytes_used -= btrfs_dev_extent_length(leaf, extent);
+        if (device->bytes_used > 0) {
+                u64 len = btrfs_dev_extent_length(leaf, extent);
+                device->bytes_used -= len;
+                spin_lock(&root->fs_info->free_chunk_lock);
+                root->fs_info->free_chunk_space += len;
+                spin_unlock(&root->fs_info->free_chunk_lock);
+        }
         ret = btrfs_del_item(trans, root, path);
 
 out:
@@ -1356,6 +1367,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
         if (ret)
                 goto error_undo;
 
+        spin_lock(&root->fs_info->free_chunk_lock);
+        root->fs_info->free_chunk_space = device->total_bytes -
+                device->bytes_used;
+        spin_unlock(&root->fs_info->free_chunk_lock);
+
         device->in_fs_metadata = 0;
         btrfs_scrub_cancel_dev(root, device);
 
@@ -1387,8 +1403,8 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
         call_rcu(&device->rcu, free_device);
         mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
-        num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1;
-        btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices);
+        num_devices = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
+        btrfs_set_super_num_devices(root->fs_info->super_copy, num_devices);
 
         if (cur_devices->open_devices == 0) {
                 struct btrfs_fs_devices *fs_devices;
@@ -1450,7 +1466,7 @@ static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans,
         struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
         struct btrfs_fs_devices *old_devices;
         struct btrfs_fs_devices *seed_devices;
-        struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
+        struct btrfs_super_block *disk_super = root->fs_info->super_copy;
         struct btrfs_device *device;
         u64 super_flags;
 
@@ -1691,15 +1707,19 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
                 root->fs_info->fs_devices->num_can_discard++;
         root->fs_info->fs_devices->total_rw_bytes += device->total_bytes;
 
+        spin_lock(&root->fs_info->free_chunk_lock);
+        root->fs_info->free_chunk_space += device->total_bytes;
+        spin_unlock(&root->fs_info->free_chunk_lock);
+
         if (!blk_queue_nonrot(bdev_get_queue(bdev)))
                 root->fs_info->fs_devices->rotating = 1;
 
-        total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
-        btrfs_set_super_total_bytes(&root->fs_info->super_copy,
+        total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy);
+        btrfs_set_super_total_bytes(root->fs_info->super_copy,
                                     total_bytes + device->total_bytes);
 
-        total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
-        btrfs_set_super_num_devices(&root->fs_info->super_copy,
+        total_bytes = btrfs_super_num_devices(root->fs_info->super_copy);
+        btrfs_set_super_num_devices(root->fs_info->super_copy,
                                     total_bytes + 1);
         mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
@@ -1790,7 +1810,7 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
                       struct btrfs_device *device, u64 new_size)
 {
         struct btrfs_super_block *super_copy =
-                &device->dev_root->fs_info->super_copy;
+                device->dev_root->fs_info->super_copy;
         u64 old_total = btrfs_super_total_bytes(super_copy);
         u64 diff = new_size - device->total_bytes;
 
@@ -1849,7 +1869,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
 static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
                         chunk_offset)
 {
-        struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+        struct btrfs_super_block *super_copy = root->fs_info->super_copy;
         struct btrfs_disk_key *disk_key;
         struct btrfs_chunk *chunk;
         u8 *ptr;
@@ -2175,7 +2195,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
         bool retried = false;
         struct extent_buffer *l;
         struct btrfs_key key;
-        struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+        struct btrfs_super_block *super_copy = root->fs_info->super_copy;
         u64 old_total = btrfs_super_total_bytes(super_copy);
         u64 old_size = device->total_bytes;
         u64 diff = device->total_bytes - new_size;
@@ -2192,8 +2212,12 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
         lock_chunks(root);
 
         device->total_bytes = new_size;
-        if (device->writeable)
+        if (device->writeable) {
                 device->fs_devices->total_rw_bytes -= diff;
+                spin_lock(&root->fs_info->free_chunk_lock);
+                root->fs_info->free_chunk_space -= diff;
+                spin_unlock(&root->fs_info->free_chunk_lock);
+        }
         unlock_chunks(root);
 
 again:
@@ -2257,6 +2281,9 @@ again:
                 device->total_bytes = old_size;
                 if (device->writeable)
                         device->fs_devices->total_rw_bytes += diff;
+                spin_lock(&root->fs_info->free_chunk_lock);
+                root->fs_info->free_chunk_space += diff;
+                spin_unlock(&root->fs_info->free_chunk_lock);
                 unlock_chunks(root);
                 goto done;
         }
@@ -2292,7 +2319,7 @@ static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
                            struct btrfs_key *key,
                            struct btrfs_chunk *chunk, int item_size)
 {
-        struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+        struct btrfs_super_block *super_copy = root->fs_info->super_copy;
         struct btrfs_disk_key disk_key;
         u32 array_size;
         u8 *ptr;
@@ -2615,6 +2642,11 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
                 index++;
         }
 
+        spin_lock(&extent_root->fs_info->free_chunk_lock);
+        extent_root->fs_info->free_chunk_space -= (stripe_size *
+                                                   map->num_stripes);
+        spin_unlock(&extent_root->fs_info->free_chunk_lock);
+
         index = 0;
         stripe = &chunk->stripe;
         while (index < map->num_stripes) {
@@ -2848,7 +2880,7 @@ static int find_live_mirror(struct map_lookup *map, int first, int num,
 
 static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
                              u64 logical, u64 *length,
-                             struct btrfs_multi_bio **multi_ret,
+                             struct btrfs_bio **bbio_ret,
                              int mirror_num)
 {
         struct extent_map *em;
@@ -2866,18 +2898,18 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
         int i;
         int num_stripes;
         int max_errors = 0;
-        struct btrfs_multi_bio *multi = NULL;
+        struct btrfs_bio *bbio = NULL;
 
-        if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD)))
+        if (bbio_ret && !(rw & (REQ_WRITE | REQ_DISCARD)))
                 stripes_allocated = 1;
 again:
-        if (multi_ret) {
-                multi = kzalloc(btrfs_multi_bio_size(stripes_allocated),
+        if (bbio_ret) {
+                bbio = kzalloc(btrfs_bio_size(stripes_allocated),
                                 GFP_NOFS);
-                if (!multi)
+                if (!bbio)
                         return -ENOMEM;
 
-                atomic_set(&multi->error, 0);
+                atomic_set(&bbio->error, 0);
         }
 
         read_lock(&em_tree->lock);
@@ -2898,7 +2930,7 @@ again:
         if (mirror_num > map->num_stripes)
                 mirror_num = 0;
 
-        /* if our multi bio struct is too small, back off and try again */
+        /* if our btrfs_bio struct is too small, back off and try again */
         if (rw & REQ_WRITE) {
                 if (map->type & (BTRFS_BLOCK_GROUP_RAID1 |
                                  BTRFS_BLOCK_GROUP_DUP)) {
@@ -2917,11 +2949,11 @@ again:
                         stripes_required = map->num_stripes;
                 }
         }
-        if (multi_ret && (rw & (REQ_WRITE | REQ_DISCARD)) &&
+        if (bbio_ret && (rw & (REQ_WRITE | REQ_DISCARD)) &&
             stripes_allocated < stripes_required) {
                 stripes_allocated = map->num_stripes;
                 free_extent_map(em);
-                kfree(multi);
+                kfree(bbio);
                 goto again;
         }
         stripe_nr = offset;
@@ -2950,7 +2982,7 @@ again:
                 *length = em->len - offset;
         }
 
-        if (!multi_ret)
+        if (!bbio_ret)
                 goto out;
 
         num_stripes = 1;
@@ -2975,13 +3007,17 @@ again:
                         stripe_index = find_live_mirror(map, 0,
                                             map->num_stripes,
                                             current->pid % map->num_stripes);
+                        mirror_num = stripe_index + 1;
                 }
 
         } else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
-                if (rw & (REQ_WRITE | REQ_DISCARD))
+                if (rw & (REQ_WRITE | REQ_DISCARD)) {
                         num_stripes = map->num_stripes;
-                else if (mirror_num)
+                } else if (mirror_num) {
                         stripe_index = mirror_num - 1;
+                } else {
+                        mirror_num = 1;
+                }
 
         } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
                 int factor = map->num_stripes / map->sub_stripes;
@@ -3001,6 +3037,7 @@ again:
                         stripe_index = find_live_mirror(map, stripe_index,
                                               map->sub_stripes, stripe_index +
                                               current->pid % map->sub_stripes);
+                        mirror_num = stripe_index + 1;
                 }
         } else {
                 /*
@@ -3009,15 +3046,16 @@ again:
                  * stripe_index is the number of our device in the stripe array
                  */
                 stripe_index = do_div(stripe_nr, map->num_stripes);
+                mirror_num = stripe_index + 1;
         }
         BUG_ON(stripe_index >= map->num_stripes);
 
         if (rw & REQ_DISCARD) {
                 for (i = 0; i < num_stripes; i++) {
-                        multi->stripes[i].physical =
+                        bbio->stripes[i].physical =
                                 map->stripes[stripe_index].physical +
                                 stripe_offset + stripe_nr * map->stripe_len;
-                        multi->stripes[i].dev = map->stripes[stripe_index].dev;
+                        bbio->stripes[i].dev = map->stripes[stripe_index].dev;
 
                         if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
                                 u64 stripes;
@@ -3038,16 +3076,16 @@ again:
                                 }
                                 stripes = stripe_nr_end - 1 - j;
                                 do_div(stripes, map->num_stripes);
-                                multi->stripes[i].length = map->stripe_len *
+                                bbio->stripes[i].length = map->stripe_len *
                                         (stripes - stripe_nr + 1);
 
                                 if (i == 0) {
-                                        multi->stripes[i].length -=
+                                        bbio->stripes[i].length -=
                                                 stripe_offset;
                                         stripe_offset = 0;
                                 }
                                 if (stripe_index == last_stripe)
-                                        multi->stripes[i].length -=
+                                        bbio->stripes[i].length -=
                                                 stripe_end_offset;
                         } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
                                 u64 stripes;
@@ -3072,11 +3110,11 @@ again:
                                 }
                                 stripes = stripe_nr_end - 1 - j;
                                 do_div(stripes, factor);
-                                multi->stripes[i].length = map->stripe_len *
+                                bbio->stripes[i].length = map->stripe_len *
                                         (stripes - stripe_nr + 1);
 
                                 if (i < map->sub_stripes) {
-                                        multi->stripes[i].length -=
+                                        bbio->stripes[i].length -=
                                                 stripe_offset;
                                         if (i == map->sub_stripes - 1)
                                                 stripe_offset = 0;
@@ -3084,11 +3122,11 @@ again:
                                 if (stripe_index >= last_stripe &&
                                     stripe_index <= (last_stripe +
                                                      map->sub_stripes - 1)) {
-                                        multi->stripes[i].length -=
+                                        bbio->stripes[i].length -=
                                                 stripe_end_offset;
                                 }
                         } else
-                                multi->stripes[i].length = *length;
+                                bbio->stripes[i].length = *length;
 
                         stripe_index++;
                         if (stripe_index == map->num_stripes) {
@@ -3099,19 +3137,20 @@ again:
                 }
         } else {
                 for (i = 0; i < num_stripes; i++) {
-                        multi->stripes[i].physical =
+                        bbio->stripes[i].physical =
                                 map->stripes[stripe_index].physical +
                                 stripe_offset +
                                 stripe_nr * map->stripe_len;
-                        multi->stripes[i].dev =
+                        bbio->stripes[i].dev =
                                 map->stripes[stripe_index].dev;
                         stripe_index++;
                 }
         }
-        if (multi_ret) {
-                *multi_ret = multi;
-                multi->num_stripes = num_stripes;
-                multi->max_errors = max_errors;
+        if (bbio_ret) {
+                *bbio_ret = bbio;
+                bbio->num_stripes = num_stripes;
+                bbio->max_errors = max_errors;
+                bbio->mirror_num = mirror_num;
         }
 out:
         free_extent_map(em);
@@ -3120,9 +3159,9 @@ out:
 
 int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
                       u64 logical, u64 *length,
-                      struct btrfs_multi_bio **multi_ret, int mirror_num)
+                      struct btrfs_bio **bbio_ret, int mirror_num)
 {
-        return __btrfs_map_block(map_tree, rw, logical, length, multi_ret,
+        return __btrfs_map_block(map_tree, rw, logical, length, bbio_ret,
                                  mirror_num);
 }
 
@@ -3191,28 +3230,30 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
         return 0;
 }
 
-static void end_bio_multi_stripe(struct bio *bio, int err)
+static void btrfs_end_bio(struct bio *bio, int err)
 {
-        struct btrfs_multi_bio *multi = bio->bi_private;
+        struct btrfs_bio *bbio = bio->bi_private;
         int is_orig_bio = 0;
 
         if (err)
-                atomic_inc(&multi->error);
+                atomic_inc(&bbio->error);
 
-        if (bio == multi->orig_bio)
+        if (bio == bbio->orig_bio)
                 is_orig_bio = 1;
 
-        if (atomic_dec_and_test(&multi->stripes_pending)) {
+        if (atomic_dec_and_test(&bbio->stripes_pending)) {
                 if (!is_orig_bio) {
                         bio_put(bio);
-                        bio = multi->orig_bio;
+                        bio = bbio->orig_bio;
                 }
-                bio->bi_private = multi->private;
-                bio->bi_end_io = multi->end_io;
+                bio->bi_private = bbio->private;
+                bio->bi_end_io = bbio->end_io;
+                bio->bi_bdev = (struct block_device *)
+                                        (unsigned long)bbio->mirror_num;
                 /* only send an error to the higher layers if it is
                  * beyond the tolerance of the multi-bio
                  */
-                if (atomic_read(&multi->error) > multi->max_errors) {
+                if (atomic_read(&bbio->error) > bbio->max_errors) {
                         err = -EIO;
                 } else if (err) {
                         /*
@@ -3222,7 +3263,7 @@ static void end_bio_multi_stripe(struct bio *bio, int err)
                         set_bit(BIO_UPTODATE, &bio->bi_flags);
                         err = 0;
                 }
-                kfree(multi);
+                kfree(bbio);
 
                 bio_endio(bio, err);
         } else if (!is_orig_bio) {
@@ -3302,20 +3343,20 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
         u64 logical = (u64)bio->bi_sector << 9;
         u64 length = 0;
         u64 map_length;
-        struct btrfs_multi_bio *multi = NULL;
         int ret;
         int dev_nr = 0;
         int total_devs = 1;
+        struct btrfs_bio *bbio = NULL;
 
         length = bio->bi_size;
         map_tree = &root->fs_info->mapping_tree;
         map_length = length;
 
-        ret = btrfs_map_block(map_tree, rw, logical, &map_length, &multi,
+        ret = btrfs_map_block(map_tree, rw, logical, &map_length, &bbio,
                               mirror_num);
         BUG_ON(ret);
 
-        total_devs = multi->num_stripes;
+        total_devs = bbio->num_stripes;
         if (map_length < length) {
                 printk(KERN_CRIT "mapping failed logical %llu bio len %llu "
                        "len %llu\n", (unsigned long long)logical,
@@ -3323,25 +3364,28 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
                        (unsigned long long)map_length);
                 BUG();
         }
-        multi->end_io = first_bio->bi_end_io;
-        multi->private = first_bio->bi_private;
-        multi->orig_bio = first_bio;
-        atomic_set(&multi->stripes_pending, multi->num_stripes);
+
+        bbio->orig_bio = first_bio;
+        bbio->private = first_bio->bi_private;
+        bbio->end_io = first_bio->bi_end_io;
+        atomic_set(&bbio->stripes_pending, bbio->num_stripes);
 
         while (dev_nr < total_devs) {
-                if (total_devs > 1) {
-                        if (dev_nr < total_devs - 1) {
-                                bio = bio_clone(first_bio, GFP_NOFS);
-                                BUG_ON(!bio);
-                        } else {
-                                bio = first_bio;
-                        }
-                        bio->bi_private = multi;
-                        bio->bi_end_io = end_bio_multi_stripe;
+                if (dev_nr < total_devs - 1) {
+                        bio = bio_clone(first_bio, GFP_NOFS);
+                        BUG_ON(!bio);
+                } else {
+                        bio = first_bio;
                 }
-                bio->bi_sector = multi->stripes[dev_nr].physical >> 9;
-                dev = multi->stripes[dev_nr].dev;
+                bio->bi_private = bbio;
+                bio->bi_end_io = btrfs_end_bio;
+                bio->bi_sector = bbio->stripes[dev_nr].physical >> 9;
+                dev = bbio->stripes[dev_nr].dev;
                 if (dev && dev->bdev && (rw != WRITE || dev->writeable)) {
+                        pr_debug("btrfs_map_bio: rw %d, secor=%llu, dev=%lu "
+                                 "(%s id %llu), size=%u\n", rw,
+                                 (u64)bio->bi_sector, (u_long)dev->bdev->bd_dev,
+                                 dev->name, dev->devid, bio->bi_size);
                         bio->bi_bdev = dev->bdev;
                         if (async_submit)
                                 schedule_bio(root, dev, rw, bio);
@@ -3354,8 +3398,6 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
                 }
                 dev_nr++;
         }
-        if (total_devs == 1)
-                kfree(multi);
         return 0;
 }
 
@@ -3616,15 +3658,20 @@ static int read_one_dev(struct btrfs_root *root,
         fill_device_from_item(leaf, dev_item, device);
         device->dev_root = root->fs_info->dev_root;
         device->in_fs_metadata = 1;
-        if (device->writeable)
+        if (device->writeable) {
                 device->fs_devices->total_rw_bytes += device->total_bytes;
+                spin_lock(&root->fs_info->free_chunk_lock);
+                root->fs_info->free_chunk_space += device->total_bytes -
+                        device->bytes_used;
+                spin_unlock(&root->fs_info->free_chunk_lock);
+        }
         ret = 0;
         return ret;
 }
 
 int btrfs_read_sys_array(struct btrfs_root *root)
 {
-        struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+        struct btrfs_super_block *super_copy = root->fs_info->super_copy;
         struct extent_buffer *sb;
         struct btrfs_disk_key *disk_key;
         struct btrfs_chunk *chunk;
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 6d866db4e177..ab5b1c49f352 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -92,6 +92,14 @@ struct btrfs_device {
         struct btrfs_work work;
         struct rcu_head rcu;
         struct work_struct rcu_work;
+
+        /* readahead state */
+        spinlock_t reada_lock;
+        atomic_t reada_in_flight;
+        u64 reada_next;
+        struct reada_zone *reada_curr_zone;
+        struct radix_tree_root reada_zones;
+        struct radix_tree_root reada_extents;
 };
 
 struct btrfs_fs_devices {
@@ -136,7 +144,10 @@ struct btrfs_bio_stripe {
         u64 length; /* only used for discard mappings */
 };
 
-struct btrfs_multi_bio {
+struct btrfs_bio;
+typedef void (btrfs_bio_end_io_t) (struct btrfs_bio *bio, int err);
+
+struct btrfs_bio {
         atomic_t stripes_pending;
         bio_end_io_t *end_io;
         struct bio *orig_bio;
@@ -144,6 +155,7 @@ struct btrfs_multi_bio {
         atomic_t error;
         int max_errors;
         int num_stripes;
+        int mirror_num;
         struct btrfs_bio_stripe stripes[];
 };
 
@@ -171,7 +183,7 @@ struct map_lookup {
 int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
                                    u64 end, u64 *length);
 
-#define btrfs_multi_bio_size(n) (sizeof(struct btrfs_multi_bio) + \
+#define btrfs_bio_size(n) (sizeof(struct btrfs_bio) + \
                             (sizeof(struct btrfs_bio_stripe) * (n)))
 
 int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
@@ -180,7 +192,7 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
                            u64 chunk_offset, u64 start, u64 num_bytes);
 int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
                     u64 logical, u64 *length,
-                    struct btrfs_multi_bio **multi_ret, int mirror_num);
+                    struct btrfs_bio **bbio_ret, int mirror_num);
 int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
                      u64 chunk_start, u64 physical, u64 devid,
                      u64 **logical, int *naddrs, int *stripe_len);
diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
index 426aa464f1af..3848b04e310e 100644
--- a/fs/btrfs/xattr.c
+++ b/fs/btrfs/xattr.c
@@ -127,6 +127,17 @@ static int do_setxattr(struct btrfs_trans_handle *trans,
 again:
         ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode),
                                       name, name_len, value, size);
+        /*
+         * If we're setting an xattr to a new value but the new value is say
+         * exactly BTRFS_MAX_XATTR_SIZE, we could end up with EOVERFLOW getting
+         * back from split_leaf.  This is because it thinks we'll be extending
+         * the existing item size, but we're asking for enough space to add the
+         * item itself.  So if we get EOVERFLOW just set ret to EEXIST and let
+         * the rest of the function figure it out.
+         */
+        if (ret == -EOVERFLOW)
+                ret = -EEXIST;
+
         if (ret == -EEXIST) {
                 if (flags & XATTR_CREATE)
                         goto out;
diff --git a/fs/buffer.c b/fs/buffer.c
index 70a19745cb61..19d8eb7fdc81 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -288,7 +288,7 @@ static void free_more_memory(void)
         struct zone *zone;
         int nid;
 
-        wakeup_flusher_threads(1024);
+        wakeup_flusher_threads(1024, WB_REASON_FREE_MORE_MEM);
         yield();
 
         for_each_online_node(nid) {
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 15b21e35078a..0f327c6c9679 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -487,17 +487,15 @@ static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap,
                 ci->i_rdcache_gen++;
 
         /*
-         * if we are newly issued FILE_SHARED, clear I_COMPLETE; we
+         * if we are newly issued FILE_SHARED, clear D_COMPLETE; we
          * don't know what happened to this directory while we didn't
          * have the cap.
          */
         if ((issued & CEPH_CAP_FILE_SHARED) &&
             (had & CEPH_CAP_FILE_SHARED) == 0) {
                 ci->i_shared_gen++;
-                if (S_ISDIR(ci->vfs_inode.i_mode)) {
-                        dout(" marking %p NOT complete\n", &ci->vfs_inode);
-                        ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
-                }
+                if (S_ISDIR(ci->vfs_inode.i_mode))
+                        ceph_dir_clear_complete(&ci->vfs_inode);
         }
 }
 
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 382abc9a6a54..2abd0dfad7f8 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -108,7 +108,7 @@ static unsigned fpos_off(loff_t p)
  * falling back to a "normal" sync readdir if any dentries in the dir
  * are dropped.
  *
- * I_COMPLETE tells indicates we have all dentries in the dir.  It is
+ * D_COMPLETE tells indicates we have all dentries in the dir.  It is
  * defined IFF we hold CEPH_CAP_FILE_SHARED (which will be revoked by
  * the MDS if/when the directory is modified).
  */
@@ -199,8 +199,8 @@ more:
         filp->f_pos++;
 
         /* make sure a dentry wasn't dropped while we didn't have parent lock */
-        if (!ceph_i_test(dir, CEPH_I_COMPLETE)) {
-                dout(" lost I_COMPLETE on %p; falling back to mds\n", dir);
+        if (!ceph_dir_test_complete(dir)) {
+                dout(" lost D_COMPLETE on %p; falling back to mds\n", dir);
                 err = -EAGAIN;
                 goto out;
         }
@@ -285,7 +285,7 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
         if ((filp->f_pos == 2 || fi->dentry) &&
             !ceph_test_mount_opt(fsc, NOASYNCREADDIR) &&
             ceph_snap(inode) != CEPH_SNAPDIR &&
-            (ci->i_ceph_flags & CEPH_I_COMPLETE) &&
+            ceph_dir_test_complete(inode) &&
             __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) {
                 spin_unlock(&inode->i_lock);
                 err = __dcache_readdir(filp, dirent, filldir);
@@ -351,7 +351,7 @@ more:
 
                 if (!req->r_did_prepopulate) {
                         dout("readdir !did_prepopulate");
-                        fi->dir_release_count--;    /* preclude I_COMPLETE */
+                        fi->dir_release_count--;    /* preclude D_COMPLETE */
                 }
 
                 /* note next offset and last dentry name */
@@ -430,8 +430,7 @@ more:
          */
         spin_lock(&inode->i_lock);
         if (ci->i_release_count == fi->dir_release_count) {
-                dout(" marking %p complete\n", inode);
-                /* ci->i_ceph_flags |= CEPH_I_COMPLETE; */
+                ceph_dir_set_complete(inode);
                 ci->i_max_offset = filp->f_pos;
         }
         spin_unlock(&inode->i_lock);
@@ -614,7 +613,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
                             fsc->mount_options->snapdir_name,
                             dentry->d_name.len) &&
                     !is_root_ceph_dentry(dir, dentry) &&
-                    (ci->i_ceph_flags & CEPH_I_COMPLETE) &&
+                    ceph_dir_test_complete(dir) &&
                     (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) {
                         spin_unlock(&dir->i_lock);
                         dout(" dir %p complete, -ENOENT\n", dir);
@@ -934,7 +933,7 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry,
                  */
 
                 /* d_move screws up d_subdirs order */
-                ceph_i_clear(new_dir, CEPH_I_COMPLETE);
+                ceph_dir_clear_complete(new_dir);
 
                 d_move(old_dentry, new_dentry);
 
@@ -1092,7 +1091,75 @@ static int ceph_snapdir_d_revalidate(struct dentry *dentry,
         return 1;
 }
 
+/*
+ * Set/clear/test dir complete flag on the dir's dentry.
+ */
+static struct dentry * __d_find_any_alias(struct inode *inode)
+{
+        struct dentry *alias;
+
+        if (list_empty(&inode->i_dentry))
+                return NULL;
+        alias = list_first_entry(&inode->i_dentry, struct dentry, d_alias);
+        return alias;
+}
+
+void ceph_dir_set_complete(struct inode *inode)
+{
+        struct dentry *dentry = __d_find_any_alias(inode);
+        
+        if (dentry && ceph_dentry(dentry)) {
+                dout(" marking %p (%p) complete\n", inode, dentry);
+                set_bit(CEPH_D_COMPLETE, &ceph_dentry(dentry)->flags);
+        }
+}
+
+void ceph_dir_clear_complete(struct inode *inode)
+{
+        struct dentry *dentry = __d_find_any_alias(inode);
+
+        if (dentry && ceph_dentry(dentry)) {
+                dout(" marking %p (%p) NOT complete\n", inode, dentry);
+                clear_bit(CEPH_D_COMPLETE, &ceph_dentry(dentry)->flags);
+        }
+}
+
+bool ceph_dir_test_complete(struct inode *inode)
+{
+        struct dentry *dentry = __d_find_any_alias(inode);
+
+        if (dentry && ceph_dentry(dentry))
+                return test_bit(CEPH_D_COMPLETE, &ceph_dentry(dentry)->flags);
+        return false;
+}
+
+/*
+ * When the VFS prunes a dentry from the cache, we need to clear the
+ * complete flag on the parent directory.
+ *
+ * Called under dentry->d_lock.
+ */
+static void ceph_d_prune(struct dentry *dentry)
+{
+        struct ceph_dentry_info *di;
+
+        dout("d_release %p\n", dentry);
+
+        /* do we have a valid parent? */
+        if (!dentry->d_parent || IS_ROOT(dentry))
+                return;
 
+        /* if we are not hashed, we don't affect D_COMPLETE */
+        if (d_unhashed(dentry))
+                return;
+
+        /*
+         * we hold d_lock, so d_parent is stable, and d_fsdata is never
+         * cleared until d_release
+         */
+        di = ceph_dentry(dentry->d_parent);
+        clear_bit(CEPH_D_COMPLETE, &di->flags);
+}
 
 /*
  * read() on a dir.  This weird interface hack only works if mounted
@@ -1306,6 +1373,7 @@ const struct inode_operations ceph_dir_iops = {
 const struct dentry_operations ceph_dentry_ops = {
         .d_revalidate = ceph_d_revalidate,
         .d_release = ceph_d_release,
+        .d_prune = ceph_d_prune,
 };
 
 const struct dentry_operations ceph_snapdir_dentry_ops = {
@@ -1315,4 +1383,5 @@ const struct dentry_operations ceph_snapdir_dentry_ops = {
 
 const struct dentry_operations ceph_snap_dentry_ops = {
         .d_release = ceph_d_release,
+        .d_prune = ceph_d_prune,
 };
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 1616a0d37cbd..e392bfce84a3 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -771,9 +771,9 @@ no_change:
             ceph_snap(inode) == CEPH_NOSNAP &&
             (le32_to_cpu(info->cap.caps) & CEPH_CAP_FILE_SHARED) &&
             (issued & CEPH_CAP_FILE_EXCL) == 0 &&
-            (ci->i_ceph_flags & CEPH_I_COMPLETE) == 0) {
+            !ceph_dir_test_complete(inode)) {
                 dout(" marking %p complete (empty)\n", inode);
-                /* ci->i_ceph_flags |= CEPH_I_COMPLETE; */
+                ceph_dir_set_complete(inode);
                 ci->i_max_offset = 2;
         }
 
@@ -856,7 +856,7 @@ static void ceph_set_dentry_offset(struct dentry *dn)
         di = ceph_dentry(dn);
 
         spin_lock(&inode->i_lock);
-        if ((ceph_inode(inode)->i_ceph_flags & CEPH_I_COMPLETE) == 0) {
+        if (!ceph_dir_test_complete(inode)) {
                 spin_unlock(&inode->i_lock);
                 return;
         }
@@ -1056,7 +1056,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
                          * d_move() puts the renamed dentry at the end of
                          * d_subdirs.  We need to assign it an appropriate
                          * directory offset so we can behave when holding
-                         * I_COMPLETE.
+                         * D_COMPLETE.
                          */
                         ceph_set_dentry_offset(req->r_old_dentry);
                         dout("dn %p gets new offset %lld\n", req->r_old_dentry, 
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 1d72f15fe9f4..264ab701154f 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -619,7 +619,7 @@ static void __unregister_request(struct ceph_mds_client *mdsc,
  *
  * Called under mdsc->mutex.
  */
-struct dentry *get_nonsnap_parent(struct dentry *dentry)
+static struct dentry *get_nonsnap_parent(struct dentry *dentry)
 {
         /*
          * we don't need to worry about protecting the d_parent access
@@ -2002,7 +2002,7 @@ out:
 }
 
 /*
- * Invalidate dir I_COMPLETE, dentry lease state on an aborted MDS
+ * Invalidate dir D_COMPLETE, dentry lease state on an aborted MDS
  * namespace request.
  */
 void ceph_invalidate_dir_request(struct ceph_mds_request *req)
@@ -2010,9 +2010,9 @@ void ceph_invalidate_dir_request(struct ceph_mds_request *req)
         struct inode *inode = req->r_locked_dir;
         struct ceph_inode_info *ci = ceph_inode(inode);
 
-        dout("invalidate_dir_request %p (I_COMPLETE, lease(s))\n", inode);
+        dout("invalidate_dir_request %p (D_COMPLETE, lease(s))\n", inode);
         spin_lock(&inode->i_lock);
-        ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
+        ceph_dir_clear_complete(inode);
         ci->i_release_count++;
         spin_unlock(&inode->i_lock);
 
@@ -3154,7 +3154,7 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc)
 /*
  * true if all sessions are closed, or we force unmount
  */
-bool done_closing_sessions(struct ceph_mds_client *mdsc)
+static bool done_closing_sessions(struct ceph_mds_client *mdsc)
 {
         int i, n = 0;
 
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 788f5ad8e66d..a90846fac759 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -426,7 +426,7 @@ static int extra_mon_dispatch(struct ceph_client *client, struct ceph_msg *msg)
 /*
  * create a new fs client
  */
-struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt,
+static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt,
                                         struct ceph_options *opt)
 {
         struct ceph_fs_client *fsc;
@@ -502,7 +502,7 @@ fail:
         return ERR_PTR(err);
 }
 
-void destroy_fs_client(struct ceph_fs_client *fsc)
+static void destroy_fs_client(struct ceph_fs_client *fsc)
 {
         dout("destroy_fs_client %p\n", fsc);
 
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index b01442aaf278..01bf189e08a9 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -203,6 +203,7 @@ struct ceph_inode_xattr {
  * Ceph dentry state
  */
 struct ceph_dentry_info {
+        unsigned long flags;
         struct ceph_mds_session *lease_session;
         u32 lease_gen, lease_shared_gen;
         u32 lease_seq;
@@ -213,6 +214,18 @@ struct ceph_dentry_info {
         u64 offset;
 };
 
+/*
+ * dentry flags
+ *
+ * The locking for D_COMPLETE is a bit odd:
+ *  - we can clear it at almost any time (see ceph_d_prune)
+ *  - it is only meaningful if:
+ *    - we hold dir inode i_lock
+ *    - we hold dir FILE_SHARED caps
+ *    - the dentry D_COMPLETE is set
+ */
+#define CEPH_D_COMPLETE 1  /* if set, d_u.d_subdirs is complete directory */
+
 struct ceph_inode_xattrs_info {
         /*
          * (still encoded) xattr blob. we avoid the overhead of parsing
@@ -251,7 +264,7 @@ struct ceph_inode_info {
         struct timespec i_rctime;
         u64 i_rbytes, i_rfiles, i_rsubdirs;
         u64 i_files, i_subdirs;
-        u64 i_max_offset;  /* largest readdir offset, set with I_COMPLETE */
+        u64 i_max_offset;  /* largest readdir offset, set with D_COMPLETE */
 
         struct rb_root i_fragtree;
         struct mutex i_fragtree_mutex;
@@ -416,7 +429,6 @@ static inline struct inode *ceph_find_inode(struct super_block *sb,
 /*
  * Ceph inode.
  */
-#define CEPH_I_COMPLETE  1  /* we have complete directory cached */
 #define CEPH_I_NODELAY   4  /* do not delay cap release */
 #define CEPH_I_FLUSH     8  /* do not delay flush of dirty metadata */
 #define CEPH_I_NOFLUSH  16  /* do not flush dirty caps */
@@ -474,6 +486,13 @@ static inline loff_t ceph_make_fpos(unsigned frag, unsigned off)
 }
 
 /*
+ * set/clear directory D_COMPLETE flag
+ */
+void ceph_dir_set_complete(struct inode *inode);
+void ceph_dir_clear_complete(struct inode *inode);
+bool ceph_dir_test_complete(struct inode *inode);
+
+/*
  * caps helpers
  */
 static inline bool __ceph_is_any_real_caps(struct ceph_inode_info *ci)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index c0458c543f17..d6a972df0338 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -37,6 +37,7 @@
 #include <asm/uaccess.h>
 #include <asm/processor.h>
 #include <linux/inet.h>
+#include <linux/module.h>
 #include <net/ipv6.h>
 #include "cifspdu.h"
 #include "cifsglob.h"
diff --git a/fs/exofs/Kconfig b/fs/exofs/Kconfig
index fa9a286c8771..da42f32c49be 100644
--- a/fs/exofs/Kconfig
+++ b/fs/exofs/Kconfig
@@ -5,7 +5,7 @@
 # selected by any of the users.
 config ORE
         tristate
-        depends on EXOFS_FS
+        depends on EXOFS_FS || PNFS_OBJLAYOUT
         select ASYNC_XOR
         default SCSI_OSD_ULD
 
diff --git a/fs/exofs/ore.c b/fs/exofs/ore.c
index fcfa86ae6faf..d271ad837202 100644
--- a/fs/exofs/ore.c
+++ b/fs/exofs/ore.c
@@ -23,6 +23,7 @@
  */
 
 #include <linux/slab.h>
+#include <linux/module.h>
 #include <asm/div64.h>
 #include <linux/lcm.h>
 
diff --git a/fs/exofs/super.c b/fs/exofs/super.c
index 057b237b8b69..e6085ec192d6 100644
--- a/fs/exofs/super.c
+++ b/fs/exofs/super.c
@@ -35,6 +35,7 @@
 #include <linux/parser.h>
 #include <linux/vfs.h>
 #include <linux/random.h>
+#include <linux/module.h>
 #include <linux/exportfs.h>
 #include <linux/slab.h>
 
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index cc5a6da030a1..240f6e2dc7ee 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -2372,7 +2372,7 @@ static int ext4_nonda_switch(struct super_block *sb)
          * start pushing delalloc when 1/2 of free blocks are dirty.
          */
         if (free_blocks < 2 * dirty_blocks)
-                writeback_inodes_sb_if_idle(sb);
+                writeback_inodes_sb_if_idle(sb, WB_REASON_FS_FREE_SPACE);
 
         return 0;
 }
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 04cf3b91e501..73c3992b2bb4 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -41,11 +41,23 @@ struct wb_writeback_work {
         unsigned int for_kupdate:1;
         unsigned int range_cyclic:1;
         unsigned int for_background:1;
+        enum wb_reason reason;          /* why was writeback initiated? */
 
         struct list_head list;          /* pending work list */
         struct completion *done;        /* set if the caller waits */
 };
 
+const char *wb_reason_name[] = {
+        [WB_REASON_BACKGROUND]          = "background",
+        [WB_REASON_TRY_TO_FREE_PAGES]   = "try_to_free_pages",
+        [WB_REASON_SYNC]                = "sync",
+        [WB_REASON_PERIODIC]            = "periodic",
+        [WB_REASON_LAPTOP_TIMER]        = "laptop_timer",
+        [WB_REASON_FREE_MORE_MEM]       = "free_more_memory",
+        [WB_REASON_FS_FREE_SPACE]       = "fs_free_space",
+        [WB_REASON_FORKER_THREAD]       = "forker_thread"
+};
+
 /*
  * Include the creation of the trace points after defining the
  * wb_writeback_work structure so that the definition remains local to this
@@ -115,7 +127,7 @@ static void bdi_queue_work(struct backing_dev_info *bdi,
 
 static void
 __bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages,
-                      bool range_cyclic)
+                      bool range_cyclic, enum wb_reason reason)
 {
         struct wb_writeback_work *work;
 
@@ -135,6 +147,7 @@ __bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages,
         work->sync_mode = WB_SYNC_NONE;
         work->nr_pages  = nr_pages;
         work->range_cyclic = range_cyclic;
+        work->reason    = reason;
 
         bdi_queue_work(bdi, work);
 }
@@ -150,9 +163,10 @@ __bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages,
  *   completion. Caller need not hold sb s_umount semaphore.
  *
  */
-void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages)
+void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages,
+                        enum wb_reason reason)
 {
-        __bdi_start_writeback(bdi, nr_pages, true);
+        __bdi_start_writeback(bdi, nr_pages, true, reason);
 }
 
 /**
@@ -251,7 +265,7 @@ static bool inode_dirtied_after(struct inode *inode, unsigned long t)
  */
 static int move_expired_inodes(struct list_head *delaying_queue,
                                struct list_head *dispatch_queue,
-                               unsigned long *older_than_this)
+                               struct wb_writeback_work *work)
 {
         LIST_HEAD(tmp);
         struct list_head *pos, *node;
@@ -262,8 +276,8 @@ static int move_expired_inodes(struct list_head *delaying_queue,
 
         while (!list_empty(delaying_queue)) {
                 inode = wb_inode(delaying_queue->prev);
-                if (older_than_this &&
-                    inode_dirtied_after(inode, *older_than_this))
+                if (work->older_than_this &&
+                    inode_dirtied_after(inode, *work->older_than_this))
                         break;
                 if (sb && sb != inode->i_sb)
                         do_sb_sort = 1;
@@ -302,13 +316,13 @@ out:
  *                                           |
  *                                           +--> dequeue for IO
  */
-static void queue_io(struct bdi_writeback *wb, unsigned long *older_than_this)
+static void queue_io(struct bdi_writeback *wb, struct wb_writeback_work *work)
 {
         int moved;
         assert_spin_locked(&wb->list_lock);
         list_splice_init(&wb->b_more_io, &wb->b_io);
-        moved = move_expired_inodes(&wb->b_dirty, &wb->b_io, older_than_this);
-        trace_writeback_queue_io(wb, older_than_this, moved);
+        moved = move_expired_inodes(&wb->b_dirty, &wb->b_io, work);
+        trace_writeback_queue_io(wb, work, moved);
 }
 
 static int write_inode(struct inode *inode, struct writeback_control *wbc)
@@ -641,31 +655,40 @@ static long __writeback_inodes_wb(struct bdi_writeback *wb,
         return wrote;
 }
 
-long writeback_inodes_wb(struct bdi_writeback *wb, long nr_pages)
+long writeback_inodes_wb(struct bdi_writeback *wb, long nr_pages,
+                                enum wb_reason reason)
 {
         struct wb_writeback_work work = {
                 .nr_pages       = nr_pages,
                 .sync_mode      = WB_SYNC_NONE,
                 .range_cyclic   = 1,
+                .reason         = reason,
         };
 
         spin_lock(&wb->list_lock);
         if (list_empty(&wb->b_io))
-                queue_io(wb, NULL);
+                queue_io(wb, &work);
         __writeback_inodes_wb(wb, &work);
         spin_unlock(&wb->list_lock);
 
         return nr_pages - work.nr_pages;
 }
 
-static inline bool over_bground_thresh(void)
+static bool over_bground_thresh(struct backing_dev_info *bdi)
 {
         unsigned long background_thresh, dirty_thresh;
 
         global_dirty_limits(&background_thresh, &dirty_thresh);
 
-        return (global_page_state(NR_FILE_DIRTY) +
-                global_page_state(NR_UNSTABLE_NFS) > background_thresh);
+        if (global_page_state(NR_FILE_DIRTY) +
+            global_page_state(NR_UNSTABLE_NFS) > background_thresh)
+                return true;
+
+        if (bdi_stat(bdi, BDI_RECLAIMABLE) >
+                                bdi_dirty_limit(bdi, background_thresh))
+                return true;
+
+        return false;
 }
 
 /*
@@ -675,7 +698,7 @@ static inline bool over_bground_thresh(void)
 static void wb_update_bandwidth(struct bdi_writeback *wb,
                                 unsigned long start_time)
 {
-        __bdi_update_bandwidth(wb->bdi, 0, 0, 0, 0, start_time);
+        __bdi_update_bandwidth(wb->bdi, 0, 0, 0, 0, 0, start_time);
 }
 
 /*
@@ -727,7 +750,7 @@ static long wb_writeback(struct bdi_writeback *wb,
                  * For background writeout, stop when we are below the
                  * background dirty threshold
                  */
-                if (work->for_background && !over_bground_thresh())
+                if (work->for_background && !over_bground_thresh(wb->bdi))
                         break;
 
                 if (work->for_kupdate) {
@@ -738,7 +761,7 @@ static long wb_writeback(struct bdi_writeback *wb,
 
                 trace_writeback_start(wb->bdi, work);
                 if (list_empty(&wb->b_io))
-                        queue_io(wb, work->older_than_this);
+                        queue_io(wb, work);
                 if (work->sb)
                         progress = writeback_sb_inodes(work->sb, wb, work);
                 else
@@ -811,13 +834,14 @@ static unsigned long get_nr_dirty_pages(void)
 
 static long wb_check_background_flush(struct bdi_writeback *wb)
 {
-        if (over_bground_thresh()) {
+        if (over_bground_thresh(wb->bdi)) {
 
                 struct wb_writeback_work work = {
                         .nr_pages       = LONG_MAX,
                         .sync_mode      = WB_SYNC_NONE,
                         .for_background = 1,
                         .range_cyclic   = 1,
+                        .reason         = WB_REASON_BACKGROUND,
                 };
 
                 return wb_writeback(wb, &work);
@@ -851,6 +875,7 @@ static long wb_check_old_data_flush(struct bdi_writeback *wb)
                         .sync_mode      = WB_SYNC_NONE,
                         .for_kupdate    = 1,
                         .range_cyclic   = 1,
+                        .reason         = WB_REASON_PERIODIC,
                 };
 
                 return wb_writeback(wb, &work);
@@ -969,7 +994,7 @@ int bdi_writeback_thread(void *data)
  * Start writeback of `nr_pages' pages.  If `nr_pages' is zero, write back
  * the whole world.
  */
-void wakeup_flusher_threads(long nr_pages)
+void wakeup_flusher_threads(long nr_pages, enum wb_reason reason)
 {
         struct backing_dev_info *bdi;
 
@@ -982,7 +1007,7 @@ void wakeup_flusher_threads(long nr_pages)
         list_for_each_entry_rcu(bdi, &bdi_list, bdi_list) {
                 if (!bdi_has_dirty_io(bdi))
                         continue;
-                __bdi_start_writeback(bdi, nr_pages, false);
+                __bdi_start_writeback(bdi, nr_pages, false, reason);
         }
         rcu_read_unlock();
 }
@@ -1203,7 +1228,9 @@ static void wait_sb_inodes(struct super_block *sb)
  * on how many (if any) will be written, and this function does not wait
  * for IO completion of submitted IO.
  */
-void writeback_inodes_sb_nr(struct super_block *sb, unsigned long nr)
+void writeback_inodes_sb_nr(struct super_block *sb,
+                            unsigned long nr,
+                            enum wb_reason reason)
 {
         DECLARE_COMPLETION_ONSTACK(done);
         struct wb_writeback_work work = {
@@ -1212,6 +1239,7 @@ void writeback_inodes_sb_nr(struct super_block *sb, unsigned long nr)
                 .tagged_writepages      = 1,
                 .done                   = &done,
                 .nr_pages               = nr,
+                .reason                 = reason,
         };
 
         WARN_ON(!rwsem_is_locked(&sb->s_umount));
@@ -1228,9 +1256,9 @@ EXPORT_SYMBOL(writeback_inodes_sb_nr);
  * on how many (if any) will be written, and this function does not wait
  * for IO completion of submitted IO.
  */
-void writeback_inodes_sb(struct super_block *sb)
+void writeback_inodes_sb(struct super_block *sb, enum wb_reason reason)
 {
-        return writeback_inodes_sb_nr(sb, get_nr_dirty_pages());
+        return writeback_inodes_sb_nr(sb, get_nr_dirty_pages(), reason);
 }
 EXPORT_SYMBOL(writeback_inodes_sb);
 
@@ -1241,11 +1269,11 @@ EXPORT_SYMBOL(writeback_inodes_sb);
  * Invoke writeback_inodes_sb if no writeback is currently underway.
  * Returns 1 if writeback was started, 0 if not.
  */
-int writeback_inodes_sb_if_idle(struct super_block *sb)
+int writeback_inodes_sb_if_idle(struct super_block *sb, enum wb_reason reason)
 {
         if (!writeback_in_progress(sb->s_bdi)) {
                 down_read(&sb->s_umount);
-                writeback_inodes_sb(sb);
+                writeback_inodes_sb(sb, reason);
                 up_read(&sb->s_umount);
                 return 1;
         } else
@@ -1262,11 +1290,12 @@ EXPORT_SYMBOL(writeback_inodes_sb_if_idle);
  * Returns 1 if writeback was started, 0 if not.
  */
 int writeback_inodes_sb_nr_if_idle(struct super_block *sb,
-                                   unsigned long nr)
+                                   unsigned long nr,
+                                   enum wb_reason reason)
 {
         if (!writeback_in_progress(sb->s_bdi)) {
                 down_read(&sb->s_umount);
-                writeback_inodes_sb_nr(sb, nr);
+                writeback_inodes_sb_nr(sb, nr, reason);
                 up_read(&sb->s_umount);
                 return 1;
         } else
@@ -1290,6 +1319,7 @@ void sync_inodes_sb(struct super_block *sb)
                 .nr_pages       = LONG_MAX,
                 .range_cyclic   = 0,
                 .done           = &done,
+                .reason         = WB_REASON_SYNC,
         };
 
         WARN_ON(!rwsem_is_locked(&sb->s_umount));
diff --git a/fs/fuse/cuse.c b/fs/fuse/cuse.c
index b6cca47f7b07..3426521f3205 100644
--- a/fs/fuse/cuse.c
+++ b/fs/fuse/cuse.c
@@ -47,6 +47,7 @@
 #include <linux/slab.h>
 #include <linux/spinlock.h>
 #include <linux/stat.h>
+#include <linux/module.h>
 
 #include "fuse_i.h"
 
diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index 7e823bbd2453..cb23c2be731a 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -14,6 +14,7 @@
 #include <linux/buffer_head.h>
 #include <linux/blkdev.h>
 #include <linux/kthread.h>
+#include <linux/export.h>
 #include <linux/namei.h>
 #include <linux/mount.h>
 #include <linux/gfs2_ondisk.h>
diff --git a/fs/ioprio.c b/fs/ioprio.c
index 7da2a06508e5..f79dab83e17b 100644
--- a/fs/ioprio.c
+++ b/fs/ioprio.c
@@ -21,6 +21,7 @@
  */
 #include <linux/gfp.h>
 #include <linux/kernel.h>
+#include <linux/export.h>
 #include <linux/ioprio.h>
 #include <linux/blkdev.h>
 #include <linux/capability.h>
diff --git a/fs/jffs2/compr.c b/fs/jffs2/compr.c
index de4247021d25..5b6c9d1a2fb9 100644
--- a/fs/jffs2/compr.c
+++ b/fs/jffs2/compr.c
@@ -53,6 +53,78 @@ static int jffs2_is_best_compression(struct jffs2_compressor *this,
         return 0;
 }
 
+/*
+ * jffs2_selected_compress:
+ * @compr: Explicit compression type to use (ie, JFFS2_COMPR_ZLIB).
+ *      If 0, just take the first available compression mode.
+ * @data_in: Pointer to uncompressed data
+ * @cpage_out: Pointer to returned pointer to buffer for compressed data
+ * @datalen: On entry, holds the amount of data available for compression.
+ *      On exit, expected to hold the amount of data actually compressed.
+ * @cdatalen: On entry, holds the amount of space available for compressed
+ *      data. On exit, expected to hold the actual size of the compressed
+ *      data.
+ *
+ * Returns: the compression type used.  Zero is used to show that the data
+ * could not be compressed; probably because we couldn't find the requested
+ * compression mode.
+ */
+static int jffs2_selected_compress(u8 compr, unsigned char *data_in,
+                unsigned char **cpage_out, u32 *datalen, u32 *cdatalen)
+{
+        struct jffs2_compressor *this;
+        int err, ret = JFFS2_COMPR_NONE;
+        uint32_t orig_slen, orig_dlen;
+        char *output_buf;
+
+        output_buf = kmalloc(*cdatalen, GFP_KERNEL);
+        if (!output_buf) {
+                printk(KERN_WARNING "JFFS2: No memory for compressor allocation. Compression failed.\n");
+                return ret;
+        }
+        orig_slen = *datalen;
+        orig_dlen = *cdatalen;
+        spin_lock(&jffs2_compressor_list_lock);
+        list_for_each_entry(this, &jffs2_compressor_list, list) {
+                /* Skip decompress-only and disabled modules */
+                if (!this->compress || this->disabled)
+                        continue;
+
+                /* Skip if not the desired compression type */
+                if (compr && (compr != this->compr))
+                        continue;
+
+                /*
+                 * Either compression type was unspecified, or we found our
+                 * compressor; either way, we're good to go.
+                 */
+                this->usecount++;
+                spin_unlock(&jffs2_compressor_list_lock);
+
+                *datalen  = orig_slen;
+                *cdatalen = orig_dlen;
+                err = this->compress(data_in, output_buf, datalen, cdatalen);
+
+                spin_lock(&jffs2_compressor_list_lock);
+                this->usecount--;
+                if (!err) {
+                        /* Success */
+                        ret = this->compr;
+                        this->stat_compr_blocks++;
+                        this->stat_compr_orig_size += *datalen;
+                        this->stat_compr_new_size += *cdatalen;
+                        break;
+                }
+        }
+        spin_unlock(&jffs2_compressor_list_lock);
+        if (ret == JFFS2_COMPR_NONE)
+                kfree(output_buf);
+        else
+                *cpage_out = output_buf;
+
+        return ret;
+}
+
 /* jffs2_compress:
  * @data_in: Pointer to uncompressed data
  * @cpage_out: Pointer to returned pointer to buffer for compressed data
@@ -76,47 +148,23 @@ uint16_t jffs2_compress(struct jffs2_sb_info *c, struct jffs2_inode_info *f,
                         uint32_t *datalen, uint32_t *cdatalen)
 {
         int ret = JFFS2_COMPR_NONE;
-        int compr_ret;
+        int mode, compr_ret;
         struct jffs2_compressor *this, *best=NULL;
         unsigned char *output_buf = NULL, *tmp_buf;
         uint32_t orig_slen, orig_dlen;
         uint32_t best_slen=0, best_dlen=0;
 
-        switch (jffs2_compression_mode) {
+        if (c->mount_opts.override_compr)
+                mode = c->mount_opts.compr;
+        else
+                mode = jffs2_compression_mode;
+
+        switch (mode) {
         case JFFS2_COMPR_MODE_NONE:
                 break;
         case JFFS2_COMPR_MODE_PRIORITY:
-                output_buf = kmalloc(*cdatalen,GFP_KERNEL);
-                if (!output_buf) {
-                        printk(KERN_WARNING "JFFS2: No memory for compressor allocation. Compression failed.\n");
-                        goto out;
-                }
-                orig_slen = *datalen;
-                orig_dlen = *cdatalen;
-                spin_lock(&jffs2_compressor_list_lock);
-                list_for_each_entry(this, &jffs2_compressor_list, list) {
-                        /* Skip decompress-only backwards-compatibility and disabled modules */
-                        if ((!this->compress)||(this->disabled))
-                                continue;
-
-                        this->usecount++;
-                        spin_unlock(&jffs2_compressor_list_lock);
-                        *datalen  = orig_slen;
-                        *cdatalen = orig_dlen;
-                        compr_ret = this->compress(data_in, output_buf, datalen, cdatalen);
-                        spin_lock(&jffs2_compressor_list_lock);
-                        this->usecount--;
-                        if (!compr_ret) {
-                                ret = this->compr;
-                                this->stat_compr_blocks++;
-                                this->stat_compr_orig_size += *datalen;
-                                this->stat_compr_new_size  += *cdatalen;
-                                break;
-                        }
-                }
-                spin_unlock(&jffs2_compressor_list_lock);
-                if (ret == JFFS2_COMPR_NONE)
-                        kfree(output_buf);
+                ret = jffs2_selected_compress(0, data_in, cpage_out, datalen,
+                                cdatalen);
                 break;
         case JFFS2_COMPR_MODE_SIZE:
         case JFFS2_COMPR_MODE_FAVOURLZO:
@@ -174,22 +222,28 @@ uint16_t jffs2_compress(struct jffs2_sb_info *c, struct jffs2_inode_info *f,
                         best->stat_compr_orig_size += best_slen;
                         best->stat_compr_new_size  += best_dlen;
                         ret = best->compr;
+                        *cpage_out = output_buf;
                 }
                 spin_unlock(&jffs2_compressor_list_lock);
                 break;
+        case JFFS2_COMPR_MODE_FORCELZO:
+                ret = jffs2_selected_compress(JFFS2_COMPR_LZO, data_in,
+                                cpage_out, datalen, cdatalen);
+                break;
+        case JFFS2_COMPR_MODE_FORCEZLIB:
+                ret = jffs2_selected_compress(JFFS2_COMPR_ZLIB, data_in,
+                                cpage_out, datalen, cdatalen);
+                break;
         default:
                 printk(KERN_ERR "JFFS2: unknown compression mode.\n");
         }
- out:
+
         if (ret == JFFS2_COMPR_NONE) {
                 *cpage_out = data_in;
                 *datalen = *cdatalen;
                 none_stat_compr_blocks++;
                 none_stat_compr_size += *datalen;
         }
-        else {
-                *cpage_out = output_buf;
-        }
         return ret;
 }
 
diff --git a/fs/jffs2/compr.h b/fs/jffs2/compr.h
index 13bb7597ab39..5e91d578f4ed 100644
--- a/fs/jffs2/compr.h
+++ b/fs/jffs2/compr.h
@@ -40,6 +40,8 @@
 #define JFFS2_COMPR_MODE_PRIORITY   1
 #define JFFS2_COMPR_MODE_SIZE       2
 #define JFFS2_COMPR_MODE_FAVOURLZO  3
+#define JFFS2_COMPR_MODE_FORCELZO   4
+#define JFFS2_COMPR_MODE_FORCEZLIB  5
 
 #define FAVOUR_LZO_PERCENT 80
 
diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c
index 7286e44ac665..4b8afe39a87f 100644
--- a/fs/jffs2/fs.c
+++ b/fs/jffs2/fs.c
@@ -379,7 +379,7 @@ void jffs2_dirty_inode(struct inode *inode, int flags)
         jffs2_do_setattr(inode, &iattr);
 }
 
-int jffs2_remount_fs (struct super_block *sb, int *flags, char *data)
+int jffs2_do_remount_fs(struct super_block *sb, int *flags, char *data)
 {
         struct jffs2_sb_info *c = JFFS2_SB_INFO(sb);
 
diff --git a/fs/jffs2/jffs2_fs_sb.h b/fs/jffs2/jffs2_fs_sb.h
index 0bc6a6c80a56..55a0c1dceadf 100644
--- a/fs/jffs2/jffs2_fs_sb.h
+++ b/fs/jffs2/jffs2_fs_sb.h
@@ -29,6 +29,11 @@
 
 struct jffs2_inodirty;
 
+struct jffs2_mount_opts {
+        bool override_compr;
+        unsigned int compr;
+};
+
 /* A struct for the overall file system control.  Pointers to
    jffs2_sb_info structs are named `c' in the source code.
    Nee jffs_control
@@ -126,6 +131,7 @@ struct jffs2_sb_info {
 #endif
 
         struct jffs2_summary *summary;          /* Summary information */
+        struct jffs2_mount_opts mount_opts;
 
 #ifdef CONFIG_JFFS2_FS_XATTR
 #define XATTRINDEX_HASHSIZE     (57)
diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h
index 6c1755c59c0f..ab65ee3ec858 100644
--- a/fs/jffs2/os-linux.h
+++ b/fs/jffs2/os-linux.h
@@ -176,7 +176,7 @@ void jffs2_dirty_inode(struct inode *inode, int flags);
 struct inode *jffs2_new_inode (struct inode *dir_i, umode_t mode,
                                struct jffs2_raw_inode *ri);
 int jffs2_statfs (struct dentry *, struct kstatfs *);
-int jffs2_remount_fs (struct super_block *, int *, char *);
+int jffs2_do_remount_fs(struct super_block *, int *, char *);
 int jffs2_do_fill_super(struct super_block *sb, void *data, int silent);
 void jffs2_gc_release_inode(struct jffs2_sb_info *c,
                             struct jffs2_inode_info *f);
diff --git a/fs/jffs2/scan.c b/fs/jffs2/scan.c
index 8d8cd3419d02..28107ca136e4 100644
--- a/fs/jffs2/scan.c
+++ b/fs/jffs2/scan.c
@@ -275,9 +275,7 @@ int jffs2_scan_medium(struct jffs2_sb_info *c)
         else
                 c->mtd->unpoint(c->mtd, 0, c->mtd->size);
 #endif
-        if (s)
-                kfree(s);
-
+        kfree(s);
         return ret;
 }
 
diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c
index 853b8e300084..e7e974454115 100644
--- a/fs/jffs2/super.c
+++ b/fs/jffs2/super.c
@@ -17,11 +17,13 @@
 #include <linux/fs.h>
 #include <linux/err.h>
 #include <linux/mount.h>
+#include <linux/parser.h>
 #include <linux/jffs2.h>
 #include <linux/pagemap.h>
 #include <linux/mtd/super.h>
 #include <linux/ctype.h>
 #include <linux/namei.h>
+#include <linux/seq_file.h>
 #include <linux/exportfs.h>
 #include "compr.h"
 #include "nodelist.h"
@@ -75,6 +77,37 @@ static void jffs2_write_super(struct super_block *sb)
         unlock_super(sb);
 }
 
+static const char *jffs2_compr_name(unsigned int compr)
+{
+        switch (compr) {
+        case JFFS2_COMPR_MODE_NONE:
+                return "none";
+#ifdef CONFIG_JFFS2_LZO
+        case JFFS2_COMPR_MODE_FORCELZO:
+                return "lzo";
+#endif
+#ifdef CONFIG_JFFS2_ZLIB
+        case JFFS2_COMPR_MODE_FORCEZLIB:
+                return "zlib";
+#endif
+        default:
+                /* should never happen; programmer error */
+                WARN_ON(1);
+                return "";
+        }
+}
+
+static int jffs2_show_options(struct seq_file *s, struct vfsmount *mnt)
+{
+        struct jffs2_sb_info *c = JFFS2_SB_INFO(mnt->mnt_sb);
+        struct jffs2_mount_opts *opts = &c->mount_opts;
+
+        if (opts->override_compr)
+                seq_printf(s, ",compr=%s", jffs2_compr_name(opts->compr));
+
+        return 0;
+}
+
 static int jffs2_sync_fs(struct super_block *sb, int wait)
 {
         struct jffs2_sb_info *c = JFFS2_SB_INFO(sb);
@@ -133,6 +166,85 @@ static const struct export_operations jffs2_export_ops = {
         .fh_to_parent = jffs2_fh_to_parent,
 };
 
+/*
+ * JFFS2 mount options.
+ *
+ * Opt_override_compr: override default compressor
+ * Opt_err: just end of array marker
+ */
+enum {
+        Opt_override_compr,
+        Opt_err,
+};
+
+static const match_table_t tokens = {
+        {Opt_override_compr, "compr=%s"},
+        {Opt_err, NULL},
+};
+
+static int jffs2_parse_options(struct jffs2_sb_info *c, char *data)
+{
+        substring_t args[MAX_OPT_ARGS];
+        char *p, *name;
+
+        if (!data)
+                return 0;
+
+        while ((p = strsep(&data, ","))) {
+                int token;
+
+                if (!*p)
+                        continue;
+
+                token = match_token(p, tokens, args);
+                switch (token) {
+                case Opt_override_compr:
+                        name = match_strdup(&args[0]);
+
+                        if (!name)
+                                return -ENOMEM;
+                        if (!strcmp(name, "none"))
+                                c->mount_opts.compr = JFFS2_COMPR_MODE_NONE;
+#ifdef CONFIG_JFFS2_LZO
+                        else if (!strcmp(name, "lzo"))
+                                c->mount_opts.compr = JFFS2_COMPR_MODE_FORCELZO;
+#endif
+#ifdef CONFIG_JFFS2_ZLIB
+                        else if (!strcmp(name, "zlib"))
+                                c->mount_opts.compr =
+                                                JFFS2_COMPR_MODE_FORCEZLIB;
+#endif
+                        else {
+                                printk(KERN_ERR "JFFS2 Error: unknown compressor \"%s\"",
+                                                name);
+                                kfree(name);
+                                return -EINVAL;
+                        }
+                        kfree(name);
+                        c->mount_opts.override_compr = true;
+                        break;
+                default:
+                        printk(KERN_ERR "JFFS2 Error: unrecognized mount option '%s' or missing value\n",
+                                        p);
+                        return -EINVAL;
+                }
+        }
+
+        return 0;
+}
+
+static int jffs2_remount_fs(struct super_block *sb, int *flags, char *data)
+{
+        struct jffs2_sb_info *c = JFFS2_SB_INFO(sb);
+        int err;
+
+        err = jffs2_parse_options(c, data);
+        if (err)
+                return -EINVAL;
+
+        return jffs2_do_remount_fs(sb, flags, data);
+}
+
 static const struct super_operations jffs2_super_operations =
 {
         .alloc_inode =  jffs2_alloc_inode,
@@ -143,6 +255,7 @@ static const struct super_operations jffs2_super_operations =
         .remount_fs =   jffs2_remount_fs,
         .evict_inode =  jffs2_evict_inode,
         .dirty_inode =  jffs2_dirty_inode,
+        .show_options = jffs2_show_options,
         .sync_fs =      jffs2_sync_fs,
 };
 
@@ -166,6 +279,12 @@ static int jffs2_fill_super(struct super_block *sb, void *data, int silent)
         c->os_priv = sb;
         sb->s_fs_info = c;
 
+        ret = jffs2_parse_options(c, data);
+        if (ret) {
+                kfree(c);
+                return -EINVAL;
+        }
+
         /* Initialize JFFS2 superblock locks, the further initialization will
          * be done later */
         mutex_init(&c->alloc_sem);
diff --git a/fs/jffs2/wbuf.c b/fs/jffs2/wbuf.c
index 4515bea0268f..b09e51d2f81f 100644
--- a/fs/jffs2/wbuf.c
+++ b/fs/jffs2/wbuf.c
@@ -578,8 +578,7 @@ static int __jffs2_flush_wbuf(struct jffs2_sb_info *c, int pad)
         if (!jffs2_is_writebuffered(c))
                 return 0;
 
-        if (mutex_trylock(&c->alloc_sem)) {
-                mutex_unlock(&c->alloc_sem);
+        if (!mutex_is_locked(&c->alloc_sem)) {
                 printk(KERN_CRIT "jffs2_flush_wbuf() called with alloc_sem not locked!\n");
                 BUG();
         }
@@ -1026,7 +1025,7 @@ int jffs2_check_oob_empty(struct jffs2_sb_info *c,
         int cmlen = min_t(int, c->oobavail, OOB_CM_SIZE);
         struct mtd_oob_ops ops;
 
-        ops.mode = MTD_OOB_AUTO;
+        ops.mode = MTD_OPS_AUTO_OOB;
         ops.ooblen = NR_OOB_SCAN_PAGES * c->oobavail;
         ops.oobbuf = c->oobbuf;
         ops.len = ops.ooboffs = ops.retlen = ops.oobretlen = 0;
@@ -1069,7 +1068,7 @@ int jffs2_check_nand_cleanmarker(struct jffs2_sb_info *c,
         struct mtd_oob_ops ops;
         int ret, cmlen = min_t(int, c->oobavail, OOB_CM_SIZE);
 
-        ops.mode = MTD_OOB_AUTO;
+        ops.mode = MTD_OPS_AUTO_OOB;
         ops.ooblen = cmlen;
         ops.oobbuf = c->oobbuf;
         ops.len = ops.ooboffs = ops.retlen = ops.oobretlen = 0;
@@ -1095,7 +1094,7 @@ int jffs2_write_nand_cleanmarker(struct jffs2_sb_info *c,
         struct mtd_oob_ops ops;
         int cmlen = min_t(int, c->oobavail, OOB_CM_SIZE);
 
-        ops.mode = MTD_OOB_AUTO;
+        ops.mode = MTD_OPS_AUTO_OOB;
         ops.ooblen = cmlen;
         ops.oobbuf = (uint8_t *)&oob_cleanmarker;
         ops.len = ops.ooboffs = ops.retlen = ops.oobretlen = 0;
diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c
index 583636f745e5..cc5f811ed383 100644
--- a/fs/jfs/jfs_logmgr.c
+++ b/fs/jfs/jfs_logmgr.c
@@ -67,6 +67,7 @@
 #include <linux/buffer_head.h>          /* for sync_blockdev() */
 #include <linux/bio.h>
 #include <linux/freezer.h>
+#include <linux/export.h>
 #include <linux/delay.h>
 #include <linux/mutex.h>
 #include <linux/seq_file.h>
diff --git a/fs/logfs/super.c b/fs/logfs/super.c
index f2697e4df109..e795c234ea33 100644
--- a/fs/logfs/super.c
+++ b/fs/logfs/super.c
@@ -13,6 +13,7 @@
 #include <linux/bio.h>
 #include <linux/slab.h>
 #include <linux/blkdev.h>
+#include <linux/module.h>
 #include <linux/mtd/mtd.h>
 #include <linux/statfs.h>
 #include <linux/buffer_head.h>
diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c
index 918ad647afea..726e59a9e50f 100644
--- a/fs/nfs/callback_xdr.c
+++ b/fs/nfs/callback_xdr.c
@@ -488,17 +488,18 @@ static __be32 decode_recallany_args(struct svc_rqst *rqstp,
                                       struct xdr_stream *xdr,
                                       struct cb_recallanyargs *args)
 {
-        __be32 *p;
+        uint32_t bitmap[2];
+        __be32 *p, status;
 
         args->craa_addr = svc_addr(rqstp);
         p = read_buf(xdr, 4);
         if (unlikely(p == NULL))
                 return htonl(NFS4ERR_BADXDR);
         args->craa_objs_to_keep = ntohl(*p++);
-        p = read_buf(xdr, 4);
-        if (unlikely(p == NULL))
-                return htonl(NFS4ERR_BADXDR);
-        args->craa_type_mask = ntohl(*p);
+        status = decode_bitmap(xdr, bitmap);
+        if (unlikely(status))
+                return status;
+        args->craa_type_mask = bitmap[0];
 
         return 0;
 }
@@ -986,4 +987,5 @@ struct svc_version nfs4_callback_version4 = {
         .vs_proc = nfs4_callback_procedures1,
         .vs_xdrsize = NFS4_CALLBACK_XDRSIZE,
         .vs_dispatch = NULL,
+        .vs_hidden = 1,
 };
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 91c01f0a4c3b..0a1f8312b4dc 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -137,11 +137,9 @@ nfs_file_open(struct inode *inode, struct file *filp)
 static int
 nfs_file_release(struct inode *inode, struct file *filp)
 {
-        struct dentry *dentry = filp->f_path.dentry;
-
         dprintk("NFS: release(%s/%s)\n",
-                        dentry->d_parent->d_name.name,
-                        dentry->d_name.name);
+                        filp->f_path.dentry->d_parent->d_name.name,
+                        filp->f_path.dentry->d_name.name);
 
         nfs_inc_stats(inode, NFSIOS_VFSRELEASE);
         return nfs_release(inode, filp);
@@ -228,14 +226,13 @@ nfs_file_read(struct kiocb *iocb, const struct iovec *iov,
         struct dentry * dentry = iocb->ki_filp->f_path.dentry;
         struct inode * inode = dentry->d_inode;
         ssize_t result;
-        size_t count = iov_length(iov, nr_segs);
 
         if (iocb->ki_filp->f_flags & O_DIRECT)
                 return nfs_file_direct_read(iocb, iov, nr_segs, pos);
 
         dprintk("NFS: read(%s/%s, %lu@%lu)\n",
                 dentry->d_parent->d_name.name, dentry->d_name.name,
-                (unsigned long) count, (unsigned long) pos);
+                (unsigned long) iov_length(iov, nr_segs), (unsigned long) pos);
 
         result = nfs_revalidate_mapping(inode, iocb->ki_filp->f_mapping);
         if (!result) {
diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index 09119418402f..a62d36b9a99e 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -31,6 +31,7 @@
 
 #include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
+#include <linux/module.h>
 
 #include "internal.h"
 #include "nfs4filelayout.h"
@@ -449,9 +450,8 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 
         fl->dsaddr = dsaddr;
 
-        if (fl->first_stripe_index < 0 ||
-            fl->first_stripe_index >= dsaddr->stripe_count) {
-                dprintk("%s Bad first_stripe_index %d\n",
+        if (fl->first_stripe_index >= dsaddr->stripe_count) {
+                dprintk("%s Bad first_stripe_index %u\n",
                                 __func__, fl->first_stripe_index);
                 goto out_put;
         }
@@ -552,7 +552,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo,
 
         /* Note that a zero value for num_fh is legal for STRIPE_SPARSE.
          * Futher checking is done in filelayout_check_layout */
-        if (fl->num_fh < 0 || fl->num_fh >
+        if (fl->num_fh >
             max(NFS4_PNFS_MAX_STRIPE_CNT, NFS4_PNFS_MAX_MULTI_CNT))
                 goto out_err;
 
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index d2ae413c986a..b60fddf606f7 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -5950,6 +5950,7 @@ static void nfs4_layoutcommit_release(void *calldata)
 {
         struct nfs4_layoutcommit_data *data = calldata;
         struct pnfs_layout_segment *lseg, *tmp;
+        unsigned long *bitlock = &NFS_I(data->args.inode)->flags;
 
         pnfs_cleanup_layoutcommit(data);
         /* Matched by references in pnfs_set_layoutcommit */
@@ -5959,6 +5960,11 @@ static void nfs4_layoutcommit_release(void *calldata)
                                        &lseg->pls_flags))
                         put_lseg(lseg);
         }
+
+        clear_bit_unlock(NFS_INO_LAYOUTCOMMITTING, bitlock);
+        smp_mb__after_clear_bit();
+        wake_up_bit(bitlock, NFS_INO_LAYOUTCOMMITTING);
+
         put_rpccred(data->cred);
         kfree(data);
 }
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 1dce12f41a4f..e6161b213ed1 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -6602,8 +6602,6 @@ static int nfs4_xdr_dec_secinfo(struct rpc_rqst *rqstp,
         if (status)
                 goto out;
         status = decode_secinfo(xdr, res);
-        if (status)
-                goto out;
 out:
         return status;
 }
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index d0cda12fddc3..c807ab93140e 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -38,21 +38,15 @@
  */
 
 #include <linux/module.h>
-#include <scsi/osd_initiator.h>
+#include <scsi/osd_ore.h>
 
 #include "objlayout.h"
 
 #define NFSDBG_FACILITY         NFSDBG_PNFS_LD
 
-#define _LLU(x) ((unsigned long long)x)
-
-enum { BIO_MAX_PAGES_KMALLOC =
-                (PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),
-};
-
 struct objio_dev_ent {
         struct nfs4_deviceid_node id_node;
-        struct osd_dev *od;
+        struct ore_dev od;
 };
 
 static void
@@ -60,8 +54,8 @@ objio_free_deviceid_node(struct nfs4_deviceid_node *d)
 {
         struct objio_dev_ent *de = container_of(d, struct objio_dev_ent, id_node);
 
-        dprintk("%s: free od=%p\n", __func__, de->od);
-        osduld_put_device(de->od);
+        dprintk("%s: free od=%p\n", __func__, de->od.od);
+        osduld_put_device(de->od.od);
         kfree(de);
 }
 
@@ -98,12 +92,12 @@ _dev_list_add(const struct nfs_server *nfss,
                                 nfss->pnfs_curr_ld,
                                 nfss->nfs_client,
                                 d_id);
-        de->od = od;
+        de->od.od = od;
 
         d = nfs4_insert_deviceid_node(&de->id_node);
         n = container_of(d, struct objio_dev_ent, id_node);
         if (n != de) {
-                dprintk("%s: Race with other n->od=%p\n", __func__, n->od);
+                dprintk("%s: Race with other n->od=%p\n", __func__, n->od.od);
                 objio_free_deviceid_node(&de->id_node);
                 de = n;
         }
@@ -111,28 +105,11 @@ _dev_list_add(const struct nfs_server *nfss,
         return de;
 }
 
-struct caps_buffers {
-        u8 caps_key[OSD_CRYPTO_KEYID_SIZE];
-        u8 creds[OSD_CAP_LEN];
-};
-
 struct objio_segment {
         struct pnfs_layout_segment lseg;
 
-        struct pnfs_osd_object_cred *comps;
-
-        unsigned mirrors_p1;
-        unsigned stripe_unit;
-        unsigned group_width;   /* Data stripe_units without integrity comps */
-        u64 group_depth;
-        unsigned group_count;
-
-        unsigned max_io_size;
-
-        unsigned comps_index;
-        unsigned num_comps;
-        /* variable length */
-        struct objio_dev_ent *ods[];
+        struct ore_layout layout;
+        struct ore_components oc;
 };
 
 static inline struct objio_segment *
@@ -141,59 +118,44 @@ OBJIO_LSEG(struct pnfs_layout_segment *lseg)
         return container_of(lseg, struct objio_segment, lseg);
 }
 
-struct objio_state;
-typedef ssize_t (*objio_done_fn)(struct objio_state *ios);
-
 struct objio_state {
         /* Generic layer */
-        struct objlayout_io_state ol_state;
-
-        struct objio_segment *layout;
-
-        struct kref kref;
-        objio_done_fn done;
-        void *private;
-
-        unsigned long length;
-        unsigned numdevs; /* Actually used devs in this IO */
-        /* A per-device variable array of size numdevs */
-        struct _objio_per_comp {
-                struct bio *bio;
-                struct osd_request *or;
-                unsigned long length;
-                u64 offset;
-                unsigned dev;
-        } per_dev[];
+        struct objlayout_io_res oir;
+
+        bool sync;
+        /*FIXME: Support for extra_bytes at ore_get_rw_state() */
+        struct ore_io_state *ios;
 };
 
 /* Send and wait for a get_device_info of devices in the layout,
    then look them up with the osd_initiator library */
-static struct objio_dev_ent *_device_lookup(struct pnfs_layout_hdr *pnfslay,
-                                struct objio_segment *objio_seg, unsigned comp,
-                                gfp_t gfp_flags)
+static int objio_devices_lookup(struct pnfs_layout_hdr *pnfslay,
+        struct objio_segment *objio_seg, unsigned c, struct nfs4_deviceid *d_id,
+        gfp_t gfp_flags)
 {
         struct pnfs_osd_deviceaddr *deviceaddr;
-        struct nfs4_deviceid *d_id;
         struct objio_dev_ent *ode;
         struct osd_dev *od;
         struct osd_dev_info odi;
         int err;
 
-        d_id = &objio_seg->comps[comp].oc_object_id.oid_device_id;
-
         ode = _dev_list_find(NFS_SERVER(pnfslay->plh_inode), d_id);
-        if (ode)
-                return ode;
+        if (ode) {
+                objio_seg->oc.ods[c] = &ode->od; /* must use container_of */
+                return 0;
+        }
 
         err = objlayout_get_deviceinfo(pnfslay, d_id, &deviceaddr, gfp_flags);
         if (unlikely(err)) {
                 dprintk("%s: objlayout_get_deviceinfo dev(%llx:%llx) =>%d\n",
                         __func__, _DEVID_LO(d_id), _DEVID_HI(d_id), err);
-                return ERR_PTR(err);
+                return err;
         }
 
         odi.systemid_len = deviceaddr->oda_systemid.len;
         if (odi.systemid_len > sizeof(odi.systemid)) {
+                dprintk("%s: odi.systemid_len > sizeof(systemid=%zd)\n",
+                        __func__, sizeof(odi.systemid));
                 err = -EINVAL;
                 goto out;
         } else if (odi.systemid_len)
@@ -218,96 +180,53 @@ static struct objio_dev_ent *_device_lookup(struct pnfs_layout_hdr *pnfslay,
 
         ode = _dev_list_add(NFS_SERVER(pnfslay->plh_inode), d_id, od,
                             gfp_flags);
-
+        objio_seg->oc.ods[c] = &ode->od; /* must use container_of */
+        dprintk("Adding new dev_id(%llx:%llx)\n",
+                _DEVID_LO(d_id), _DEVID_HI(d_id));
 out:
-        dprintk("%s: return=%d\n", __func__, err);
         objlayout_put_deviceinfo(deviceaddr);
-        return err ? ERR_PTR(err) : ode;
+        return err;
 }
 
-static int objio_devices_lookup(struct pnfs_layout_hdr *pnfslay,
-        struct objio_segment *objio_seg,
-        gfp_t gfp_flags)
+static void copy_single_comp(struct ore_components *oc, unsigned c,
+                             struct pnfs_osd_object_cred *src_comp)
 {
-        unsigned i;
-        int err;
+        struct ore_comp *ocomp = &oc->comps[c];
 
-        /* lookup all devices */
-        for (i = 0; i < objio_seg->num_comps; i++) {
-                struct objio_dev_ent *ode;
+        WARN_ON(src_comp->oc_cap_key.cred_len > 0); /* libosd is NO_SEC only */
+        WARN_ON(src_comp->oc_cap.cred_len > sizeof(ocomp->cred));
 
-                ode = _device_lookup(pnfslay, objio_seg, i, gfp_flags);
-                if (unlikely(IS_ERR(ode))) {
-                        err = PTR_ERR(ode);
-                        goto out;
-                }
-                objio_seg->ods[i] = ode;
-        }
-        err = 0;
+        ocomp->obj.partition = src_comp->oc_object_id.oid_partition_id;
+        ocomp->obj.id = src_comp->oc_object_id.oid_object_id;
 
-out:
-        dprintk("%s: return=%d\n", __func__, err);
-        return err;
+        memcpy(ocomp->cred, src_comp->oc_cap.cred, sizeof(ocomp->cred));
 }
 
-static int _verify_data_map(struct pnfs_osd_layout *layout)
+int __alloc_objio_seg(unsigned numdevs, gfp_t gfp_flags,
+                       struct objio_segment **pseg)
 {
-        struct pnfs_osd_data_map *data_map = &layout->olo_map;
-        u64 stripe_length;
-        u32 group_width;
-
-/* FIXME: Only raid0 for now. if not go through MDS */
-        if (data_map->odm_raid_algorithm != PNFS_OSD_RAID_0) {
-                printk(KERN_ERR "Only RAID_0 for now\n");
-                return -ENOTSUPP;
-        }
-        if (0 != (data_map->odm_num_comps % (data_map->odm_mirror_cnt + 1))) {
-                printk(KERN_ERR "Data Map wrong, num_comps=%u mirrors=%u\n",
-                          data_map->odm_num_comps, data_map->odm_mirror_cnt);
-                return -EINVAL;
-        }
+        struct __alloc_objio_segment {
+                struct objio_segment olseg;
+                struct ore_dev *ods[numdevs];
+                struct ore_comp comps[numdevs];
+        } *aolseg;
 
-        if (data_map->odm_group_width)
-                group_width = data_map->odm_group_width;
-        else
-                group_width = data_map->odm_num_comps /
-                                                (data_map->odm_mirror_cnt + 1);
-
-        stripe_length = (u64)data_map->odm_stripe_unit * group_width;
-        if (stripe_length >= (1ULL << 32)) {
-                printk(KERN_ERR "Total Stripe length(0x%llx)"
-                          " >= 32bit is not supported\n", _LLU(stripe_length));
-                return -ENOTSUPP;
+        aolseg = kzalloc(sizeof(*aolseg), gfp_flags);
+        if (unlikely(!aolseg)) {
+                dprintk("%s: Faild allocation numdevs=%d size=%zd\n", __func__,
+                        numdevs, sizeof(*aolseg));
+                return -ENOMEM;
         }
 
-        if (0 != (data_map->odm_stripe_unit & ~PAGE_MASK)) {
-                printk(KERN_ERR "Stripe Unit(0x%llx)"
-                          " must be Multples of PAGE_SIZE(0x%lx)\n",
-                          _LLU(data_map->odm_stripe_unit), PAGE_SIZE);
-                return -ENOTSUPP;
-        }
+        aolseg->olseg.oc.numdevs = numdevs;
+        aolseg->olseg.oc.single_comp = EC_MULTPLE_COMPS;
+        aolseg->olseg.oc.comps = aolseg->comps;
+        aolseg->olseg.oc.ods = aolseg->ods;
 
+        *pseg = &aolseg->olseg;
         return 0;
 }
 
-static void copy_single_comp(struct pnfs_osd_object_cred *cur_comp,
-                             struct pnfs_osd_object_cred *src_comp,
-                             struct caps_buffers *caps_p)
-{
-        WARN_ON(src_comp->oc_cap_key.cred_len > sizeof(caps_p->caps_key));
-        WARN_ON(src_comp->oc_cap.cred_len > sizeof(caps_p->creds));
-
-        *cur_comp = *src_comp;
-
-        memcpy(caps_p->caps_key, src_comp->oc_cap_key.cred,
-               sizeof(caps_p->caps_key));
-        cur_comp->oc_cap_key.cred = caps_p->caps_key;
-
-        memcpy(caps_p->creds, src_comp->oc_cap.cred,
-               sizeof(caps_p->creds));
-        cur_comp->oc_cap.cred = caps_p->creds;
-}
-
 int objio_alloc_lseg(struct pnfs_layout_segment **outp,
         struct pnfs_layout_hdr *pnfslay,
         struct pnfs_layout_range *range,
@@ -317,59 +236,43 @@ int objio_alloc_lseg(struct pnfs_layout_segment **outp,
         struct objio_segment *objio_seg;
         struct pnfs_osd_xdr_decode_layout_iter iter;
         struct pnfs_osd_layout layout;
-        struct pnfs_osd_object_cred *cur_comp, src_comp;
-        struct caps_buffers *caps_p;
+        struct pnfs_osd_object_cred src_comp;
+        unsigned cur_comp;
         int err;
 
         err = pnfs_osd_xdr_decode_layout_map(&layout, &iter, xdr);
         if (unlikely(err))
                 return err;
 
-        err = _verify_data_map(&layout);
+        err = __alloc_objio_seg(layout.olo_num_comps, gfp_flags, &objio_seg);
         if (unlikely(err))
                 return err;
 
-        objio_seg = kzalloc(sizeof(*objio_seg) +
-                            sizeof(objio_seg->ods[0]) * layout.olo_num_comps +
-                            sizeof(*objio_seg->comps) * layout.olo_num_comps +
-                            sizeof(struct caps_buffers) * layout.olo_num_comps,
-                            gfp_flags);
-        if (!objio_seg)
-                return -ENOMEM;
+        objio_seg->layout.stripe_unit = layout.olo_map.odm_stripe_unit;
+        objio_seg->layout.group_width = layout.olo_map.odm_group_width;
+        objio_seg->layout.group_depth = layout.olo_map.odm_group_depth;
+        objio_seg->layout.mirrors_p1 = layout.olo_map.odm_mirror_cnt + 1;
+        objio_seg->layout.raid_algorithm = layout.olo_map.odm_raid_algorithm;
 
-        objio_seg->comps = (void *)(objio_seg->ods + layout.olo_num_comps);
-        cur_comp = objio_seg->comps;
-        caps_p = (void *)(cur_comp + layout.olo_num_comps);
-        while (pnfs_osd_xdr_decode_layout_comp(&src_comp, &iter, xdr, &err))
-                copy_single_comp(cur_comp++, &src_comp, caps_p++);
+        err = ore_verify_layout(layout.olo_map.odm_num_comps,
+                                          &objio_seg->layout);
         if (unlikely(err))
                 goto err;
 
-        objio_seg->num_comps = layout.olo_num_comps;
-        objio_seg->comps_index = layout.olo_comps_index;
-        err = objio_devices_lookup(pnfslay, objio_seg, gfp_flags);
-        if (err)
-                goto err;
-
-        objio_seg->mirrors_p1 = layout.olo_map.odm_mirror_cnt + 1;
-        objio_seg->stripe_unit = layout.olo_map.odm_stripe_unit;
-        if (layout.olo_map.odm_group_width) {
-                objio_seg->group_width = layout.olo_map.odm_group_width;
-                objio_seg->group_depth = layout.olo_map.odm_group_depth;
-                objio_seg->group_count = layout.olo_map.odm_num_comps /
-                                                objio_seg->mirrors_p1 /
-                                                objio_seg->group_width;
-        } else {
-                objio_seg->group_width = layout.olo_map.odm_num_comps /
-                                                objio_seg->mirrors_p1;
-                objio_seg->group_depth = -1;
-                objio_seg->group_count = 1;
+        objio_seg->oc.first_dev = layout.olo_comps_index;
+        cur_comp = 0;
+        while (pnfs_osd_xdr_decode_layout_comp(&src_comp, &iter, xdr, &err)) {
+                copy_single_comp(&objio_seg->oc, cur_comp, &src_comp);
+                err = objio_devices_lookup(pnfslay, objio_seg, cur_comp,
+                                           &src_comp.oc_object_id.oid_device_id,
+                                           gfp_flags);
+                if (err)
+                        goto err;
+                ++cur_comp;
         }
-
-        /* Cache this calculation it will hit for every page */
-        objio_seg->max_io_size = (BIO_MAX_PAGES_KMALLOC * PAGE_SIZE -
-                                  objio_seg->stripe_unit) *
-                                 objio_seg->group_width;
+        /* pnfs_osd_xdr_decode_layout_comp returns false on error */
+        if (unlikely(err))
+                goto err;
 
         *outp = &objio_seg->lseg;
         return 0;
@@ -386,43 +289,63 @@ void objio_free_lseg(struct pnfs_layout_segment *lseg)
         int i;
         struct objio_segment *objio_seg = OBJIO_LSEG(lseg);
 
-        for (i = 0; i < objio_seg->num_comps; i++) {
-                if (!objio_seg->ods[i])
+        for (i = 0; i < objio_seg->oc.numdevs; i++) {
+                struct ore_dev *od = objio_seg->oc.ods[i];
+                struct objio_dev_ent *ode;
+
+                if (!od)
                         break;
-                nfs4_put_deviceid_node(&objio_seg->ods[i]->id_node);
+                ode = container_of(od, typeof(*ode), od);
+                nfs4_put_deviceid_node(&ode->id_node);
         }
         kfree(objio_seg);
 }
 
-int objio_alloc_io_state(struct pnfs_layout_segment *lseg,
-                         struct objlayout_io_state **outp,
-                         gfp_t gfp_flags)
+static int
+objio_alloc_io_state(struct pnfs_layout_hdr *pnfs_layout_type, bool is_reading,
+        struct pnfs_layout_segment *lseg, struct page **pages, unsigned pgbase,
+        loff_t offset, size_t count, void *rpcdata, gfp_t gfp_flags,
+        struct objio_state **outp)
 {
         struct objio_segment *objio_seg = OBJIO_LSEG(lseg);
-        struct objio_state *ios;
-        const unsigned first_size = sizeof(*ios) +
-                                objio_seg->num_comps * sizeof(ios->per_dev[0]);
-        const unsigned sec_size = objio_seg->num_comps *
-                                                sizeof(ios->ol_state.ioerrs[0]);
-
-        ios = kzalloc(first_size + sec_size, gfp_flags);
-        if (unlikely(!ios))
+        struct ore_io_state *ios;
+        int ret;
+        struct __alloc_objio_state {
+                struct objio_state objios;
+                struct pnfs_osd_ioerr ioerrs[objio_seg->oc.numdevs];
+        } *aos;
+
+        aos = kzalloc(sizeof(*aos), gfp_flags);
+        if (unlikely(!aos))
                 return -ENOMEM;
 
-        ios->layout = objio_seg;
-        ios->ol_state.ioerrs = ((void *)ios) + first_size;
-        ios->ol_state.num_comps = objio_seg->num_comps;
+        objlayout_init_ioerrs(&aos->objios.oir, objio_seg->oc.numdevs,
+                        aos->ioerrs, rpcdata, pnfs_layout_type);
 
-        *outp = &ios->ol_state;
+        ret = ore_get_rw_state(&objio_seg->layout, &objio_seg->oc, is_reading,
+                               offset, count, &ios);
+        if (unlikely(ret)) {
+                kfree(aos);
+                return ret;
+        }
+
+        ios->pages = pages;
+        ios->pgbase = pgbase;
+        ios->private = aos;
+        BUG_ON(ios->nr_pages > (pgbase + count + PAGE_SIZE - 1) >> PAGE_SHIFT);
+
+        aos->objios.sync = 0;
+        aos->objios.ios = ios;
+        *outp = &aos->objios;
         return 0;
 }
 
-void objio_free_io_state(struct objlayout_io_state *ol_state)
+void objio_free_result(struct objlayout_io_res *oir)
 {
-        struct objio_state *ios = container_of(ol_state, struct objio_state,
-                                               ol_state);
+        struct objio_state *objios = container_of(oir, struct objio_state, oir);
 
-        kfree(ios);
+        ore_put_io_state(objios->ios);
+        kfree(objios);
 }
 
 enum pnfs_osd_errno osd_pri_2_pnfs_err(enum osd_err_priority oep)
@@ -455,539 +378,152 @@ enum pnfs_osd_errno osd_pri_2_pnfs_err(enum osd_err_priority oep)
         }
 }
 
-static void _clear_bio(struct bio *bio)
+static void __on_dev_error(struct ore_io_state *ios,
+        struct ore_dev *od, unsigned dev_index, enum osd_err_priority oep,
+        u64 dev_offset, u64  dev_len)
 {
-        struct bio_vec *bv;
-        unsigned i;
-
-        __bio_for_each_segment(bv, bio, i, 0) {
-                unsigned this_count = bv->bv_len;
-
-                if (likely(PAGE_SIZE == this_count))
-                        clear_highpage(bv->bv_page);
-                else
-                        zero_user(bv->bv_page, bv->bv_offset, this_count);
-        }
-}
-
-static int _io_check(struct objio_state *ios, bool is_write)
-{
-        enum osd_err_priority oep = OSD_ERR_PRI_NO_ERROR;
-        int lin_ret = 0;
-        int i;
-
-        for (i = 0; i <  ios->numdevs; i++) {
-                struct osd_sense_info osi;
-                struct osd_request *or = ios->per_dev[i].or;
-                int ret;
-
-                if (!or)
-                        continue;
+        struct objio_state *objios = ios->private;
+        struct pnfs_osd_objid pooid;
+        struct objio_dev_ent *ode = container_of(od, typeof(*ode), od);
+        /* FIXME: what to do with more-then-one-group layouts. We need to
+         * translate from ore_io_state index to oc->comps index
+         */
+        unsigned comp = dev_index;
 
-                ret = osd_req_decode_sense(or, &osi);
-                if (likely(!ret))
-                        continue;
+        pooid.oid_device_id = ode->id_node.deviceid;
+        pooid.oid_partition_id = ios->oc->comps[comp].obj.partition;
+        pooid.oid_object_id = ios->oc->comps[comp].obj.id;
 
-                if (OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) {
-                        /* start read offset passed endof file */
-                        BUG_ON(is_write);
-                        _clear_bio(ios->per_dev[i].bio);
-                        dprintk("%s: start read offset passed end of file "
-                                "offset=0x%llx, length=0x%lx\n", __func__,
-                                _LLU(ios->per_dev[i].offset),
-                                ios->per_dev[i].length);
-
-                        continue; /* we recovered */
-                }
-                objlayout_io_set_result(&ios->ol_state, i,
-                                        &ios->layout->comps[i].oc_object_id,
-                                        osd_pri_2_pnfs_err(osi.osd_err_pri),
-                                        ios->per_dev[i].offset,
-                                        ios->per_dev[i].length,
-                                        is_write);
-
-                if (osi.osd_err_pri >= oep) {
-                        oep = osi.osd_err_pri;
-                        lin_ret = ret;
-                }
-        }
-
-        return lin_ret;
-}
-
-/*
- * Common IO state helpers.
- */
-static void _io_free(struct objio_state *ios)
-{
-        unsigned i;
-
-        for (i = 0; i < ios->numdevs; i++) {
-                struct _objio_per_comp *per_dev = &ios->per_dev[i];
-
-                if (per_dev->or) {
-                        osd_end_request(per_dev->or);
-                        per_dev->or = NULL;
-                }
-
-                if (per_dev->bio) {
-                        bio_put(per_dev->bio);
-                        per_dev->bio = NULL;
-                }
-        }
-}
-
-struct osd_dev *_io_od(struct objio_state *ios, unsigned dev)
-{
-        unsigned min_dev = ios->layout->comps_index;
-        unsigned max_dev = min_dev + ios->layout->num_comps;
-
-        BUG_ON(dev < min_dev || max_dev <= dev);
-        return ios->layout->ods[dev - min_dev]->od;
-}
-
-struct _striping_info {
-        u64 obj_offset;
-        u64 group_length;
-        unsigned dev;
-        unsigned unit_off;
-};
-
-static void _calc_stripe_info(struct objio_state *ios, u64 file_offset,
-                              struct _striping_info *si)
-{
-        u32     stripe_unit = ios->layout->stripe_unit;
-        u32     group_width = ios->layout->group_width;
-        u64     group_depth = ios->layout->group_depth;
-        u32     U = stripe_unit * group_width;
-
-        u64     T = U * group_depth;
-        u64     S = T * ios->layout->group_count;
-        u64     M = div64_u64(file_offset, S);
-
-        /*
-        G = (L - (M * S)) / T
-        H = (L - (M * S)) % T
-        */
-        u64     LmodU = file_offset - M * S;
-        u32     G = div64_u64(LmodU, T);
-        u64     H = LmodU - G * T;
-
-        u32     N = div_u64(H, U);
-
-        div_u64_rem(file_offset, stripe_unit, &si->unit_off);
-        si->obj_offset = si->unit_off + (N * stripe_unit) +
-                                  (M * group_depth * stripe_unit);
-
-        /* "H - (N * U)" is just "H % U" so it's bound to u32 */
-        si->dev = (u32)(H - (N * U)) / stripe_unit + G * group_width;
-        si->dev *= ios->layout->mirrors_p1;
-
-        si->group_length = T - H;
-}
-
-static int _add_stripe_unit(struct objio_state *ios,  unsigned *cur_pg,
-                unsigned pgbase, struct _objio_per_comp *per_dev, int len,
-                gfp_t gfp_flags)
-{
-        unsigned pg = *cur_pg;
-        int cur_len = len;
-        struct request_queue *q =
-                        osd_request_queue(_io_od(ios, per_dev->dev));
-
-        if (per_dev->bio == NULL) {
-                unsigned pages_in_stripe = ios->layout->group_width *
-                                      (ios->layout->stripe_unit / PAGE_SIZE);
-                unsigned bio_size = (ios->ol_state.nr_pages + pages_in_stripe) /
-                                    ios->layout->group_width;
-
-                if (BIO_MAX_PAGES_KMALLOC < bio_size)
-                        bio_size = BIO_MAX_PAGES_KMALLOC;
-
-                per_dev->bio = bio_kmalloc(gfp_flags, bio_size);
-                if (unlikely(!per_dev->bio)) {
-                        dprintk("Faild to allocate BIO size=%u\n", bio_size);
-                        return -ENOMEM;
-                }
-        }
-
-        while (cur_len > 0) {
-                unsigned pglen = min_t(unsigned, PAGE_SIZE - pgbase, cur_len);
-                unsigned added_len;
-
-                BUG_ON(ios->ol_state.nr_pages <= pg);
-                cur_len -= pglen;
-
-                added_len = bio_add_pc_page(q, per_dev->bio,
-                                        ios->ol_state.pages[pg], pglen, pgbase);
-                if (unlikely(pglen != added_len))
-                        return -ENOMEM;
-                pgbase = 0;
-                ++pg;
-        }
-        BUG_ON(cur_len);
-
-        per_dev->length += len;
-        *cur_pg = pg;
-        return 0;
-}
-
-static int _prepare_one_group(struct objio_state *ios, u64 length,
-                              struct _striping_info *si, unsigned *last_pg,
-                              gfp_t gfp_flags)
-{
-        unsigned stripe_unit = ios->layout->stripe_unit;
-        unsigned mirrors_p1 = ios->layout->mirrors_p1;
-        unsigned devs_in_group = ios->layout->group_width * mirrors_p1;
-        unsigned dev = si->dev;
-        unsigned first_dev = dev - (dev % devs_in_group);
-        unsigned max_comp = ios->numdevs ? ios->numdevs - mirrors_p1 : 0;
-        unsigned cur_pg = *last_pg;
-        int ret = 0;
-
-        while (length) {
-                struct _objio_per_comp *per_dev = &ios->per_dev[dev - first_dev];
-                unsigned cur_len, page_off = 0;
-
-                if (!per_dev->length) {
-                        per_dev->dev = dev;
-                        if (dev < si->dev) {
-                                per_dev->offset = si->obj_offset + stripe_unit -
-                                                                   si->unit_off;
-                                cur_len = stripe_unit;
-                        } else if (dev == si->dev) {
-                                per_dev->offset = si->obj_offset;
-                                cur_len = stripe_unit - si->unit_off;
-                                page_off = si->unit_off & ~PAGE_MASK;
-                                BUG_ON(page_off &&
-                                      (page_off != ios->ol_state.pgbase));
-                        } else { /* dev > si->dev */
-                                per_dev->offset = si->obj_offset - si->unit_off;
-                                cur_len = stripe_unit;
-                        }
-
-                        if (max_comp < dev - first_dev)
-                                max_comp = dev - first_dev;
-                } else {
-                        cur_len = stripe_unit;
-                }
-                if (cur_len >= length)
-                        cur_len = length;
-
-                ret = _add_stripe_unit(ios, &cur_pg, page_off , per_dev,
-                                       cur_len, gfp_flags);
-                if (unlikely(ret))
-                        goto out;
-
-                dev += mirrors_p1;
-                dev = (dev % devs_in_group) + first_dev;
-
-                length -= cur_len;
-                ios->length += cur_len;
-        }
-out:
-        ios->numdevs = max_comp + mirrors_p1;
-        *last_pg = cur_pg;
-        return ret;
-}
-
-static int _io_rw_pagelist(struct objio_state *ios, gfp_t gfp_flags)
-{
-        u64 length = ios->ol_state.count;
-        u64 offset = ios->ol_state.offset;
-        struct _striping_info si;
-        unsigned last_pg = 0;
-        int ret = 0;
-
-        while (length) {
-                _calc_stripe_info(ios, offset, &si);
-
-                if (length < si.group_length)
-                        si.group_length = length;
-
-                ret = _prepare_one_group(ios, si.group_length, &si, &last_pg, gfp_flags);
-                if (unlikely(ret))
-                        goto out;
-
-                offset += si.group_length;
-                length -= si.group_length;
-        }
-
-out:
-        if (!ios->length)
-                return ret;
-
-        return 0;
-}
-
-static ssize_t _sync_done(struct objio_state *ios)
-{
-        struct completion *waiting = ios->private;
-
-        complete(waiting);
-        return 0;
-}
-
-static void _last_io(struct kref *kref)
-{
-        struct objio_state *ios = container_of(kref, struct objio_state, kref);
-
-        ios->done(ios);
-}
-
-static void _done_io(struct osd_request *or, void *p)
-{
-        struct objio_state *ios = p;
-
-        kref_put(&ios->kref, _last_io);
-}
-
-static ssize_t _io_exec(struct objio_state *ios)
-{
-        DECLARE_COMPLETION_ONSTACK(wait);
-        ssize_t status = 0; /* sync status */
-        unsigned i;
-        objio_done_fn saved_done_fn = ios->done;
-        bool sync = ios->ol_state.sync;
-
-        if (sync) {
-                ios->done = _sync_done;
-                ios->private = &wait;
-        }
-
-        kref_init(&ios->kref);
-
-        for (i = 0; i < ios->numdevs; i++) {
-                struct osd_request *or = ios->per_dev[i].or;
-
-                if (!or)
-                        continue;
-
-                kref_get(&ios->kref);
-                osd_execute_request_async(or, _done_io, ios);
-        }
-
-        kref_put(&ios->kref, _last_io);
-
-        if (sync) {
-                wait_for_completion(&wait);
-                status = saved_done_fn(ios);
-        }
-
-        return status;
+        objlayout_io_set_result(&objios->oir, comp,
+                                &pooid, osd_pri_2_pnfs_err(oep),
+                                dev_offset, dev_len, !ios->reading);
 }
 
 /*
  * read
  */
-static ssize_t _read_done(struct objio_state *ios)
+static void _read_done(struct ore_io_state *ios, void *private)
 {
+        struct objio_state *objios = private;
         ssize_t status;
-        int ret = _io_check(ios, false);
+        int ret = ore_check_io(ios, &__on_dev_error);
 
-        _io_free(ios);
+        /* FIXME: _io_free(ios) can we dealocate the libosd resources; */
 
         if (likely(!ret))
                 status = ios->length;
         else
                 status = ret;
 
-        objlayout_read_done(&ios->ol_state, status, ios->ol_state.sync);
-        return status;
+        objlayout_read_done(&objios->oir, status, objios->sync);
 }
 
-static int _read_mirrors(struct objio_state *ios, unsigned cur_comp)
+int objio_read_pagelist(struct nfs_read_data *rdata)
 {
-        struct osd_request *or = NULL;
-        struct _objio_per_comp *per_dev = &ios->per_dev[cur_comp];
-        unsigned dev = per_dev->dev;
-        struct pnfs_osd_object_cred *cred =
-                        &ios->layout->comps[cur_comp];
-        struct osd_obj_id obj = {
-                .partition = cred->oc_object_id.oid_partition_id,
-                .id = cred->oc_object_id.oid_object_id,
-        };
+        struct objio_state *objios;
         int ret;
 
-        or = osd_start_request(_io_od(ios, dev), GFP_KERNEL);
-        if (unlikely(!or)) {
-                ret = -ENOMEM;
-                goto err;
-        }
-        per_dev->or = or;
-
-        osd_req_read(or, &obj, per_dev->offset, per_dev->bio, per_dev->length);
-
-        ret = osd_finalize_request(or, 0, cred->oc_cap.cred, NULL);
-        if (ret) {
-                dprintk("%s: Faild to osd_finalize_request() => %d\n",
-                        __func__, ret);
-                goto err;
-        }
-
-        dprintk("%s:[%d] dev=%d obj=0x%llx start=0x%llx length=0x%lx\n",
-                __func__, cur_comp, dev, obj.id, _LLU(per_dev->offset),
-                per_dev->length);
-
-err:
-        return ret;
-}
-
-static ssize_t _read_exec(struct objio_state *ios)
-{
-        unsigned i;
-        int ret;
-
-        for (i = 0; i < ios->numdevs; i += ios->layout->mirrors_p1) {
-                if (!ios->per_dev[i].length)
-                        continue;
-                ret = _read_mirrors(ios, i);
-                if (unlikely(ret))
-                        goto err;
-        }
-
-        ios->done = _read_done;
-        return _io_exec(ios); /* In sync mode exec returns the io status */
-
-err:
-        _io_free(ios);
-        return ret;
-}
-
-ssize_t objio_read_pagelist(struct objlayout_io_state *ol_state)
-{
-        struct objio_state *ios = container_of(ol_state, struct objio_state,
-                                               ol_state);
-        int ret;
-
-        ret = _io_rw_pagelist(ios, GFP_KERNEL);
+        ret = objio_alloc_io_state(NFS_I(rdata->inode)->layout, true,
+                        rdata->lseg, rdata->args.pages, rdata->args.pgbase,
+                        rdata->args.offset, rdata->args.count, rdata,
+                        GFP_KERNEL, &objios);
         if (unlikely(ret))
                 return ret;
 
-        return _read_exec(ios);
+        objios->ios->done = _read_done;
+        dprintk("%s: offset=0x%llx length=0x%x\n", __func__,
+                rdata->args.offset, rdata->args.count);
+        return ore_read(objios->ios);
 }
 
 /*
  * write
  */
-static ssize_t _write_done(struct objio_state *ios)
+static void _write_done(struct ore_io_state *ios, void *private)
 {
+        struct objio_state *objios = private;
         ssize_t status;
-        int ret = _io_check(ios, true);
+        int ret = ore_check_io(ios, &__on_dev_error);
 
-        _io_free(ios);
+        /* FIXME: _io_free(ios) can we dealocate the libosd resources; */
 
         if (likely(!ret)) {
                 /* FIXME: should be based on the OSD's persistence model
                  * See OSD2r05 Section 4.13 Data persistence model */
-                ios->ol_state.committed = NFS_FILE_SYNC;
+                objios->oir.committed = NFS_FILE_SYNC;
                 status = ios->length;
         } else {
                 status = ret;
         }
 
-        objlayout_write_done(&ios->ol_state, status, ios->ol_state.sync);
-        return status;
+        objlayout_write_done(&objios->oir, status, objios->sync);
 }
 
-static int _write_mirrors(struct objio_state *ios, unsigned cur_comp)
+static struct page *__r4w_get_page(void *priv, u64 offset, bool *uptodate)
 {
-        struct _objio_per_comp *master_dev = &ios->per_dev[cur_comp];
-        unsigned dev = ios->per_dev[cur_comp].dev;
-        unsigned last_comp = cur_comp + ios->layout->mirrors_p1;
-        int ret;
-
-        for (; cur_comp < last_comp; ++cur_comp, ++dev) {
-                struct osd_request *or = NULL;
-                struct pnfs_osd_object_cred *cred =
-                                        &ios->layout->comps[cur_comp];
-                struct osd_obj_id obj = {
-                        .partition = cred->oc_object_id.oid_partition_id,
-                        .id = cred->oc_object_id.oid_object_id,
-                };
-                struct _objio_per_comp *per_dev = &ios->per_dev[cur_comp];
-                struct bio *bio;
-
-                or = osd_start_request(_io_od(ios, dev), GFP_NOFS);
-                if (unlikely(!or)) {
-                        ret = -ENOMEM;
-                        goto err;
-                }
-                per_dev->or = or;
-
-                if (per_dev != master_dev) {
-                        bio = bio_kmalloc(GFP_NOFS,
-                                          master_dev->bio->bi_max_vecs);
-                        if (unlikely(!bio)) {
-                                dprintk("Faild to allocate BIO size=%u\n",
-                                        master_dev->bio->bi_max_vecs);
-                                ret = -ENOMEM;
-                                goto err;
-                        }
-
-                        __bio_clone(bio, master_dev->bio);
-                        bio->bi_bdev = NULL;
-                        bio->bi_next = NULL;
-                        per_dev->bio = bio;
-                        per_dev->dev = dev;
-                        per_dev->length = master_dev->length;
-                        per_dev->offset =  master_dev->offset;
-                } else {
-                        bio = master_dev->bio;
-                        bio->bi_rw |= REQ_WRITE;
-                }
-
-                osd_req_write(or, &obj, per_dev->offset, bio, per_dev->length);
+        struct objio_state *objios = priv;
+        struct nfs_write_data *wdata = objios->oir.rpcdata;
+        pgoff_t index = offset / PAGE_SIZE;
+        struct page *page = find_get_page(wdata->inode->i_mapping, index);
 
-                ret = osd_finalize_request(or, 0, cred->oc_cap.cred, NULL);
-                if (ret) {
-                        dprintk("%s: Faild to osd_finalize_request() => %d\n",
-                                __func__, ret);
-                        goto err;
+        if (!page) {
+                page = find_or_create_page(wdata->inode->i_mapping,
+                                                index, GFP_NOFS);
+                if (unlikely(!page)) {
+                        dprintk("%s: grab_cache_page Failed index=0x%lx\n",
+                                __func__, index);
+                        return NULL;
                 }
-
-                dprintk("%s:[%d] dev=%d obj=0x%llx start=0x%llx length=0x%lx\n",
-                        __func__, cur_comp, dev, obj.id, _LLU(per_dev->offset),
-                        per_dev->length);
+                unlock_page(page);
         }
+        if (PageDirty(page) || PageWriteback(page))
+                *uptodate = true;
+        else
+                *uptodate = PageUptodate(page);
+        dprintk("%s: index=0x%lx uptodate=%d\n", __func__, index, *uptodate);
+        return page;
+}
 
-err:
-        return ret;
+static void __r4w_put_page(void *priv, struct page *page)
+{
+        dprintk("%s: index=0x%lx\n", __func__, page->index);
+        page_cache_release(page);
+        return;
 }
 
-static ssize_t _write_exec(struct objio_state *ios)
+static const struct _ore_r4w_op _r4w_op = {
+        .get_page = &__r4w_get_page,
+        .put_page = &__r4w_put_page,
+};
+
+int objio_write_pagelist(struct nfs_write_data *wdata, int how)
 {
-        unsigned i;
+        struct objio_state *objios;
         int ret;
 
-        for (i = 0; i < ios->numdevs; i += ios->layout->mirrors_p1) {
-                if (!ios->per_dev[i].length)
-                        continue;
-                ret = _write_mirrors(ios, i);
-                if (unlikely(ret))
-                        goto err;
-        }
-
-        ios->done = _write_done;
-        return _io_exec(ios); /* In sync mode exec returns the io->status */
+        ret = objio_alloc_io_state(NFS_I(wdata->inode)->layout, false,
+                        wdata->lseg, wdata->args.pages, wdata->args.pgbase,
+                        wdata->args.offset, wdata->args.count, wdata, GFP_NOFS,
+                        &objios);
+        if (unlikely(ret))
+                return ret;
 
-err:
-        _io_free(ios);
-        return ret;
-}
+        objios->sync = 0 != (how & FLUSH_SYNC);
+        objios->ios->r4w = &_r4w_op;
 
-ssize_t objio_write_pagelist(struct objlayout_io_state *ol_state, bool stable)
-{
-        struct objio_state *ios = container_of(ol_state, struct objio_state,
-                                               ol_state);
-        int ret;
+        if (!objios->sync)
+                objios->ios->done = _write_done;
 
-        /* TODO: ios->stable = stable; */
-        ret = _io_rw_pagelist(ios, GFP_NOFS);
+        dprintk("%s: offset=0x%llx length=0x%x\n", __func__,
+                wdata->args.offset, wdata->args.count);
+        ret = ore_write(objios->ios);
         if (unlikely(ret))
                 return ret;
 
-        return _write_exec(ios);
+        if (objios->sync)
+                _write_done(objios->ios, objios);
+
+        return 0;
 }
 
 static bool objio_pg_test(struct nfs_pageio_descriptor *pgio,
@@ -997,7 +533,7 @@ static bool objio_pg_test(struct nfs_pageio_descriptor *pgio,
                 return false;
 
         return pgio->pg_count + req->wb_bytes <=
-                        OBJIO_LSEG(pgio->pg_lseg)->max_io_size;
+                        OBJIO_LSEG(pgio->pg_lseg)->layout.max_io_length;
 }
 
 static const struct nfs_pageio_ops objio_pg_read_ops = {
diff --git a/fs/nfs/objlayout/objlayout.c b/fs/nfs/objlayout/objlayout.c
index 1d06f8e2adea..72074e3a04f9 100644
--- a/fs/nfs/objlayout/objlayout.c
+++ b/fs/nfs/objlayout/objlayout.c
@@ -156,77 +156,39 @@ last_byte_offset(u64 start, u64 len)
         return end > start ? end - 1 : NFS4_MAX_UINT64;
 }
 
-static struct objlayout_io_state *
-objlayout_alloc_io_state(struct pnfs_layout_hdr *pnfs_layout_type,
-                        struct page **pages,
-                        unsigned pgbase,
-                        loff_t offset,
-                        size_t count,
-                        struct pnfs_layout_segment *lseg,
-                        void *rpcdata,
-                        gfp_t gfp_flags)
+void _fix_verify_io_params(struct pnfs_layout_segment *lseg,
+                           struct page ***p_pages, unsigned *p_pgbase,
+                           u64 offset, unsigned long count)
 {
-        struct objlayout_io_state *state;
         u64 lseg_end_offset;
 
-        dprintk("%s: allocating io_state\n", __func__);
-        if (objio_alloc_io_state(lseg, &state, gfp_flags))
-                return NULL;
-
         BUG_ON(offset < lseg->pls_range.offset);
         lseg_end_offset = end_offset(lseg->pls_range.offset,
                                      lseg->pls_range.length);
         BUG_ON(offset >= lseg_end_offset);
-        if (offset + count > lseg_end_offset) {
-                count = lseg->pls_range.length -
-                                (offset - lseg->pls_range.offset);
-                dprintk("%s: truncated count %Zd\n", __func__, count);
-        }
+        WARN_ON(offset + count > lseg_end_offset);
 
-        if (pgbase > PAGE_SIZE) {
-                pages += pgbase >> PAGE_SHIFT;
-                pgbase &= ~PAGE_MASK;
+        if (*p_pgbase > PAGE_SIZE) {
+                dprintk("%s: pgbase(0x%x) > PAGE_SIZE\n", __func__, *p_pgbase);
+                *p_pages += *p_pgbase >> PAGE_SHIFT;
+                *p_pgbase &= ~PAGE_MASK;
         }
-
-        INIT_LIST_HEAD(&state->err_list);
-        state->lseg = lseg;
-        state->rpcdata = rpcdata;
-        state->pages = pages;
-        state->pgbase = pgbase;
-        state->nr_pages = (pgbase + count + PAGE_SIZE - 1) >> PAGE_SHIFT;
-        state->offset = offset;
-        state->count = count;
-        state->sync = 0;
-
-        return state;
-}
-
-static void
-objlayout_free_io_state(struct objlayout_io_state *state)
-{
-        dprintk("%s: freeing io_state\n", __func__);
-        if (unlikely(!state))
-                return;
-
-        objio_free_io_state(state);
 }
 
 /*
  * I/O done common code
  */
 static void
-objlayout_iodone(struct objlayout_io_state *state)
+objlayout_iodone(struct objlayout_io_res *oir)
 {
-        dprintk("%s: state %p status\n", __func__, state);
-
-        if (likely(state->status >= 0)) {
-                objlayout_free_io_state(state);
+        if (likely(oir->status >= 0)) {
+                objio_free_result(oir);
         } else {
-                struct objlayout *objlay = OBJLAYOUT(state->lseg->pls_layout);
+                struct objlayout *objlay = oir->objlay;
 
                 spin_lock(&objlay->lock);
                 objlay->delta_space_valid = OBJ_DSU_INVALID;
-                list_add(&objlay->err_list, &state->err_list);
+                list_add(&objlay->err_list, &oir->err_list);
                 spin_unlock(&objlay->lock);
         }
 }
@@ -238,13 +200,13 @@ objlayout_iodone(struct objlayout_io_state *state)
  * the error for later reporting at layout-return.
  */
 void
-objlayout_io_set_result(struct objlayout_io_state *state, unsigned index,
+objlayout_io_set_result(struct objlayout_io_res *oir, unsigned index,
                         struct pnfs_osd_objid *pooid, int osd_error,
                         u64 offset, u64 length, bool is_write)
 {
-        struct pnfs_osd_ioerr *ioerr = &state->ioerrs[index];
+        struct pnfs_osd_ioerr *ioerr = &oir->ioerrs[index];
 
-        BUG_ON(index >= state->num_comps);
+        BUG_ON(index >= oir->num_comps);
         if (osd_error) {
                 ioerr->oer_component = *pooid;
                 ioerr->oer_comp_offset = offset;
@@ -285,21 +247,18 @@ static void _rpc_read_complete(struct work_struct *work)
 }
 
 void
-objlayout_read_done(struct objlayout_io_state *state, ssize_t status, bool sync)
+objlayout_read_done(struct objlayout_io_res *oir, ssize_t status, bool sync)
 {
-        int eof = state->eof;
-        struct nfs_read_data *rdata;
+        struct nfs_read_data *rdata = oir->rpcdata;
 
-        state->status = status;
-        dprintk("%s: Begin status=%zd eof=%d\n", __func__, status, eof);
-        rdata = state->rpcdata;
-        rdata->task.tk_status = status;
-        if (status >= 0) {
+        oir->status = rdata->task.tk_status = status;
+        if (status >= 0)
                 rdata->res.count = status;
-                rdata->res.eof = eof;
-        }
-        objlayout_iodone(state);
-        /* must not use state after this point */
+        objlayout_iodone(oir);
+        /* must not use oir after this point */
+
+        dprintk("%s: Return status=%zd eof=%d sync=%d\n", __func__,
+                status, rdata->res.eof, sync);
 
         if (sync)
                 pnfs_ld_read_done(rdata);
@@ -317,40 +276,36 @@ objlayout_read_pagelist(struct nfs_read_data *rdata)
 {
         loff_t offset = rdata->args.offset;
         size_t count = rdata->args.count;
-        struct objlayout_io_state *state;
-        ssize_t status = 0;
+        int err;
         loff_t eof;
 
-        dprintk("%s: Begin inode %p offset %llu count %d\n",
-                __func__, rdata->inode, offset, (int)count);
-
         eof = i_size_read(rdata->inode);
         if (unlikely(offset + count > eof)) {
                 if (offset >= eof) {
-                        status = 0;
+                        err = 0;
                         rdata->res.count = 0;
                         rdata->res.eof = 1;
+                        /*FIXME: do we need to call pnfs_ld_read_done() */
                         goto out;
                 }
                 count = eof - offset;
         }
 
-        state = objlayout_alloc_io_state(NFS_I(rdata->inode)->layout,
-                                         rdata->args.pages, rdata->args.pgbase,
-                                         offset, count,
-                                         rdata->lseg, rdata,
-                                         GFP_KERNEL);
-        if (unlikely(!state)) {
-                status = -ENOMEM;
-                goto out;
-        }
+        rdata->res.eof = (offset + count) >= eof;
+        _fix_verify_io_params(rdata->lseg, &rdata->args.pages,
+                              &rdata->args.pgbase,
+                              rdata->args.offset, rdata->args.count);
 
-        state->eof = state->offset + state->count >= eof;
+        dprintk("%s: inode(%lx) offset 0x%llx count 0x%Zx eof=%d\n",
+                __func__, rdata->inode->i_ino, offset, count, rdata->res.eof);
 
-        status = objio_read_pagelist(state);
+        err = objio_read_pagelist(rdata);
  out:
-        dprintk("%s: Return status %Zd\n", __func__, status);
-        rdata->pnfs_error = status;
+        if (unlikely(err)) {
+                rdata->pnfs_error = err;
+                dprintk("%s: Returned Error %d\n", __func__, err);
+                return PNFS_NOT_ATTEMPTED;
+        }
         return PNFS_ATTEMPTED;
 }
 
@@ -371,26 +326,20 @@ static void _rpc_write_complete(struct work_struct *work)
 }
 
 void
-objlayout_write_done(struct objlayout_io_state *state, ssize_t status,
-                     bool sync)
+objlayout_write_done(struct objlayout_io_res *oir, ssize_t status, bool sync)
 {
-        struct nfs_write_data *wdata;
+        struct nfs_write_data *wdata = oir->rpcdata;
 
-        dprintk("%s: Begin\n", __func__);
-        wdata = state->rpcdata;
-        state->status = status;
-        wdata->task.tk_status = status;
+        oir->status = wdata->task.tk_status = status;
         if (status >= 0) {
                 wdata->res.count = status;
-                wdata->verf.committed = state->committed;
-                dprintk("%s: Return status %d committed %d\n",
-                        __func__, wdata->task.tk_status,
-                        wdata->verf.committed);
-        } else
-                dprintk("%s: Return status %d\n",
-                        __func__, wdata->task.tk_status);
-        objlayout_iodone(state);
-        /* must not use state after this point */
+                wdata->verf.committed = oir->committed;
+        }
+        objlayout_iodone(oir);
+        /* must not use oir after this point */
+
+        dprintk("%s: Return status %zd committed %d sync=%d\n", __func__,
+                status, wdata->verf.committed, sync);
 
         if (sync)
                 pnfs_ld_write_done(wdata);
@@ -407,30 +356,18 @@ enum pnfs_try_status
 objlayout_write_pagelist(struct nfs_write_data *wdata,
                          int how)
 {
-        struct objlayout_io_state *state;
-        ssize_t status;
-
-        dprintk("%s: Begin inode %p offset %llu count %u\n",
-                __func__, wdata->inode, wdata->args.offset, wdata->args.count);
-
-        state = objlayout_alloc_io_state(NFS_I(wdata->inode)->layout,
-                                         wdata->args.pages,
-                                         wdata->args.pgbase,
-                                         wdata->args.offset,
-                                         wdata->args.count,
-                                         wdata->lseg, wdata,
-                                         GFP_NOFS);
-        if (unlikely(!state)) {
-                status = -ENOMEM;
-                goto out;
-        }
+        int err;
 
-        state->sync = how & FLUSH_SYNC;
+        _fix_verify_io_params(wdata->lseg, &wdata->args.pages,
+                              &wdata->args.pgbase,
+                              wdata->args.offset, wdata->args.count);
 
-        status = objio_write_pagelist(state, how & FLUSH_STABLE);
- out:
-        dprintk("%s: Return status %Zd\n", __func__, status);
-        wdata->pnfs_error = status;
+        err = objio_write_pagelist(wdata, how);
+        if (unlikely(err)) {
+                wdata->pnfs_error = err;
+                dprintk("%s: Returned Error %d\n", __func__, err);
+                return PNFS_NOT_ATTEMPTED;
+        }
         return PNFS_ATTEMPTED;
 }
 
@@ -537,14 +474,14 @@ merge_ioerr(struct pnfs_osd_ioerr *dest_err,
 static void
 encode_accumulated_error(struct objlayout *objlay, __be32 *p)
 {
-        struct objlayout_io_state *state, *tmp;
+        struct objlayout_io_res *oir, *tmp;
         struct pnfs_osd_ioerr accumulated_err = {.oer_errno = 0};
 
-        list_for_each_entry_safe(state, tmp, &objlay->err_list, err_list) {
+        list_for_each_entry_safe(oir, tmp, &objlay->err_list, err_list) {
                 unsigned i;
 
-                for (i = 0; i < state->num_comps; i++) {
-                        struct pnfs_osd_ioerr *ioerr = &state->ioerrs[i];
+                for (i = 0; i < oir->num_comps; i++) {
+                        struct pnfs_osd_ioerr *ioerr = &oir->ioerrs[i];
 
                         if (!ioerr->oer_errno)
                                 continue;
@@ -563,8 +500,8 @@ encode_accumulated_error(struct objlayout *objlay, __be32 *p)
 
                         merge_ioerr(&accumulated_err, ioerr);
                 }
-                list_del(&state->err_list);
-                objlayout_free_io_state(state);
+                list_del(&oir->err_list);
+                objio_free_result(oir);
         }
 
         pnfs_osd_xdr_encode_ioerr(p, &accumulated_err);
@@ -576,7 +513,7 @@ objlayout_encode_layoutreturn(struct pnfs_layout_hdr *pnfslay,
                               const struct nfs4_layoutreturn_args *args)
 {
         struct objlayout *objlay = OBJLAYOUT(pnfslay);
-        struct objlayout_io_state *state, *tmp;
+        struct objlayout_io_res *oir, *tmp;
         __be32 *start;
 
         dprintk("%s: Begin\n", __func__);
@@ -585,13 +522,13 @@ objlayout_encode_layoutreturn(struct pnfs_layout_hdr *pnfslay,
 
         spin_lock(&objlay->lock);
 
-        list_for_each_entry_safe(state, tmp, &objlay->err_list, err_list) {
+        list_for_each_entry_safe(oir, tmp, &objlay->err_list, err_list) {
                 __be32 *last_xdr = NULL, *p;
                 unsigned i;
                 int res = 0;
 
-                for (i = 0; i < state->num_comps; i++) {
-                        struct pnfs_osd_ioerr *ioerr = &state->ioerrs[i];
+                for (i = 0; i < oir->num_comps; i++) {
+                        struct pnfs_osd_ioerr *ioerr = &oir->ioerrs[i];
 
                         if (!ioerr->oer_errno)
                                 continue;
@@ -615,7 +552,7 @@ objlayout_encode_layoutreturn(struct pnfs_layout_hdr *pnfslay,
                         }
 
                         last_xdr = p;
-                        pnfs_osd_xdr_encode_ioerr(p, &state->ioerrs[i]);
+                        pnfs_osd_xdr_encode_ioerr(p, &oir->ioerrs[i]);
                 }
 
                 /* TODO: use xdr_write_pages */
@@ -631,8 +568,8 @@ objlayout_encode_layoutreturn(struct pnfs_layout_hdr *pnfslay,
                         encode_accumulated_error(objlay, last_xdr);
                         goto loop_done;
                 }
-                list_del(&state->err_list);
-                objlayout_free_io_state(state);
+                list_del(&oir->err_list);
+                objio_free_result(oir);
         }
 loop_done:
         spin_unlock(&objlay->lock);
diff --git a/fs/nfs/objlayout/objlayout.h b/fs/nfs/objlayout/objlayout.h
index a8244c8e042d..8ec34727ed21 100644
--- a/fs/nfs/objlayout/objlayout.h
+++ b/fs/nfs/objlayout/objlayout.h
@@ -74,19 +74,11 @@ OBJLAYOUT(struct pnfs_layout_hdr *lo)
  * per-I/O operation state
  * embedded in objects provider io_state data structure
  */
-struct objlayout_io_state {
-        struct pnfs_layout_segment *lseg;
-
-        struct page **pages;
-        unsigned pgbase;
-        unsigned nr_pages;
-        unsigned long count;
-        loff_t offset;
-        bool sync;
+struct objlayout_io_res {
+        struct objlayout *objlay;
 
         void *rpcdata;
         int status;             /* res */
-        int eof;                /* res */
         int committed;          /* res */
 
         /* Error reporting (layout_return) */
@@ -100,6 +92,18 @@ struct objlayout_io_state {
         struct pnfs_osd_ioerr *ioerrs;
 };
 
+static inline
+void objlayout_init_ioerrs(struct objlayout_io_res *oir, unsigned num_comps,
+                        struct pnfs_osd_ioerr *ioerrs, void *rpcdata,
+                        struct pnfs_layout_hdr *pnfs_layout_type)
+{
+        oir->objlay = OBJLAYOUT(pnfs_layout_type);
+        oir->rpcdata = rpcdata;
+        INIT_LIST_HEAD(&oir->err_list);
+        oir->num_comps = num_comps;
+        oir->ioerrs = ioerrs;
+}
+
 /*
  * Raid engine I/O API
  */
@@ -110,28 +114,24 @@ extern int objio_alloc_lseg(struct pnfs_layout_segment **outp,
         gfp_t gfp_flags);
 extern void objio_free_lseg(struct pnfs_layout_segment *lseg);
 
-extern int objio_alloc_io_state(
-        struct pnfs_layout_segment *lseg,
-        struct objlayout_io_state **outp,
-        gfp_t gfp_flags);
-extern void objio_free_io_state(struct objlayout_io_state *state);
+/* objio_free_result will free these @oir structs recieved from
+ * objlayout_{read,write}_done
+ */
+extern void objio_free_result(struct objlayout_io_res *oir);
 
-extern ssize_t objio_read_pagelist(struct objlayout_io_state *ol_state);
-extern ssize_t objio_write_pagelist(struct objlayout_io_state *ol_state,
-                                    bool stable);
+extern int objio_read_pagelist(struct nfs_read_data *rdata);
+extern int objio_write_pagelist(struct nfs_write_data *wdata, int how);
 
 /*
  * callback API
  */
-extern void objlayout_io_set_result(struct objlayout_io_state *state,
+extern void objlayout_io_set_result(struct objlayout_io_res *oir,
                         unsigned index, struct pnfs_osd_objid *pooid,
                         int osd_error, u64 offset, u64 length, bool is_write);
 
 static inline void
-objlayout_add_delta_space_used(struct objlayout_io_state *state, s64 space_used)
+objlayout_add_delta_space_used(struct objlayout *objlay, s64 space_used)
 {
-        struct objlayout *objlay = OBJLAYOUT(state->lseg->pls_layout);
-
         /* If one of the I/Os errored out and the delta_space_used was
          * invalid we render the complete report as invalid. Protocol mandate
          * the DSU be accurate or not reported.
@@ -144,9 +144,9 @@ objlayout_add_delta_space_used(struct objlayout_io_state *state, s64 space_used)
         spin_unlock(&objlay->lock);
 }
 
-extern void objlayout_read_done(struct objlayout_io_state *state,
+extern void objlayout_read_done(struct objlayout_io_res *oir,
                                 ssize_t status, bool sync);
-extern void objlayout_write_done(struct objlayout_io_state *state,
+extern void objlayout_write_done(struct objlayout_io_res *oir,
                                  ssize_t status, bool sync);
 
 extern int objlayout_get_deviceinfo(struct pnfs_layout_hdr *pnfslay,
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index b60970cc7f1f..5668f7c54c41 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -18,6 +18,7 @@
 #include <linux/nfs_page.h>
 #include <linux/nfs_fs.h>
 #include <linux/nfs_mount.h>
+#include <linux/export.h>
 
 #include "internal.h"
 #include "pnfs.h"
@@ -41,7 +42,7 @@ nfs_page_free(struct nfs_page *p)
 
 /**
  * nfs_create_request - Create an NFS read/write request.
- * @file: file descriptor to use
+ * @ctx: open context to use
  * @inode: inode to which the request is attached
  * @page: page to write
  * @offset: starting offset within the page for the write
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index ee73d9a4f700..baf73536bc04 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -29,6 +29,7 @@
 
 #include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
+#include <linux/module.h>
 #include "internal.h"
 #include "pnfs.h"
 #include "iostat.h"
@@ -1443,17 +1444,31 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
         /* Note kzalloc ensures data->res.seq_res.sr_slot == NULL */
         data = kzalloc(sizeof(*data), GFP_NOFS);
         if (!data) {
-                mark_inode_dirty_sync(inode);
                 status = -ENOMEM;
                 goto out;
         }
 
+        if (!test_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags))
+                goto out_free;
+
+        if (test_and_set_bit(NFS_INO_LAYOUTCOMMITTING, &nfsi->flags)) {
+                if (!sync) {
+                        status = -EAGAIN;
+                        goto out_free;
+                }
+                status = wait_on_bit_lock(&nfsi->flags, NFS_INO_LAYOUTCOMMITTING,
+                                        nfs_wait_bit_killable, TASK_KILLABLE);
+                if (status)
+                        goto out_free;
+        }
+
         INIT_LIST_HEAD(&data->lseg_list);
         spin_lock(&inode->i_lock);
         if (!test_and_clear_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags)) {
+                clear_bit(NFS_INO_LAYOUTCOMMITTING, &nfsi->flags);
                 spin_unlock(&inode->i_lock);
-                kfree(data);
-                goto out;
+                wake_up_bit(&nfsi->flags, NFS_INO_LAYOUTCOMMITTING);
+                goto out_free;
         }
 
         pnfs_list_write_lseg(inode, &data->lseg_list);
@@ -1475,6 +1490,11 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
 
         status = nfs4_proc_layoutcommit(data, sync);
 out:
+        if (status)
+                mark_inode_dirty_sync(inode);
         dprintk("<-- %s status %d\n", __func__, status);
         return status;
+out_free:
+        kfree(data);
+        goto out;
 }
diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c
index 6fda5228ef56..4f359d2a26eb 100644
--- a/fs/nfs/pnfs_dev.c
+++ b/fs/nfs/pnfs_dev.c
@@ -28,6 +28,7 @@
  *  such damages.
  */
 
+#include <linux/export.h>
 #include "pnfs.h"
 
 #define NFSDBG_FACILITY         NFSDBG_PNFS
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 2219c88d96b2..1dda78db6a73 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -20,6 +20,7 @@
 #include <linux/nfs_mount.h>
 #include <linux/nfs_page.h>
 #include <linux/backing-dev.h>
+#include <linux/export.h>
 
 #include <asm/uaccess.h>
 
@@ -1243,7 +1244,6 @@ void nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 {
         struct nfs_writeargs    *argp = &data->args;
         struct nfs_writeres     *resp = &data->res;
-        struct nfs_server       *server = NFS_SERVER(data->inode);
         int status;
 
         dprintk("NFS: %5u nfs_writeback_done (status %d)\n",
@@ -1277,7 +1277,7 @@ void nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
                 if (time_before(complain, jiffies)) {
                         dprintk("NFS:       faulty NFS server %s:"
                                 " (committed = %d) != (stable = %d)\n",
-                                server->nfs_client->cl_hostname,
+                                NFS_SERVER(data->inode)->nfs_client->cl_hostname,
                                 resp->verf->committed, argp->stable);
                         complain = jiffies + 300 * HZ;
                 }
diff --git a/fs/nfsd/nfs4acl.c b/fs/nfsd/nfs4acl.c
index ad88f1c0a4c3..9c51aff02ae2 100644
--- a/fs/nfsd/nfs4acl.c
+++ b/fs/nfsd/nfs4acl.c
@@ -36,6 +36,7 @@
 
 #include <linux/slab.h>
 #include <linux/nfs_fs.h>
+#include <linux/export.h>
 #include "acl.h"
 
 
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index db34a585e112..c45a2ea4a090 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -13,6 +13,7 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/gss_api.h>
 #include <linux/sunrpc/gss_krb5_enctypes.h>
+#include <linux/module.h>
 
 #include "idmap.h"
 #include "nfsd.h"
diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index dc5a1bf476b1..eda7d7e55e05 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -8,6 +8,7 @@
 
 #include <linux/sched.h>
 #include <linux/freezer.h>
+#include <linux/module.h>
 #include <linux/fs_struct.h>
 #include <linux/swap.h>
 
@@ -256,6 +257,8 @@ static void nfsd_last_thread(struct svc_serv *serv)
         nfsd_serv = NULL;
         nfsd_shutdown();
 
+        svc_rpcb_cleanup(serv);
+
         printk(KERN_WARNING "nfsd: last server has exited, flushing export "
                             "cache\n");
         nfsd_export_flush();
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index db5ee4b4f47a..ad7d0c155de4 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -59,6 +59,7 @@
 #include <linux/idr.h>
 #include <linux/kref.h>
 #include <linux/net.h>
+#include <linux/export.h>
 #include <net/tcp.h>
 
 #include <asm/uaccess.h>
diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c
index 56f82cb912e3..0e28e242226d 100644
--- a/fs/ocfs2/dlm/dlmdebug.c
+++ b/fs/ocfs2/dlm/dlmdebug.c
@@ -30,6 +30,7 @@
 #include <linux/sysctl.h>
 #include <linux/spinlock.h>
 #include <linux/debugfs.h>
+#include <linux/export.h>
 
 #include "cluster/heartbeat.h"
 #include "cluster/nodemanager.h"
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index cd99bf557650..b0f450a2bb7c 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -12,6 +12,7 @@
 #include <linux/user.h>
 #include <linux/elf.h>
 #include <linux/elfcore.h>
+#include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/bootmem.h>
diff --git a/fs/quota/quota.c b/fs/quota/quota.c
index aae0edb95c6c..35f4b0ecdeb3 100644
--- a/fs/quota/quota.c
+++ b/fs/quota/quota.c
@@ -286,7 +286,7 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id,
                 /* caller already holds s_umount */
                 if (sb->s_flags & MS_RDONLY)
                         return -EROFS;
-                writeback_inodes_sb(sb);
+                writeback_inodes_sb(sb, WB_REASON_SYNC);
                 return 0;
         default:
                 return -EINVAL;
diff --git a/fs/squashfs/Kconfig b/fs/squashfs/Kconfig
index 048b59d5b2f0..c70111ebefd4 100644
--- a/fs/squashfs/Kconfig
+++ b/fs/squashfs/Kconfig
@@ -78,6 +78,28 @@ config SQUASHFS_XZ
 
           If unsure, say N.
 
+config SQUASHFS_4K_DEVBLK_SIZE
+        bool "Use 4K device block size?"
+        depends on SQUASHFS
+        help
+          By default Squashfs sets the dev block size (sb_min_blocksize)
+          to 1K or the smallest block size supported by the block device
+          (if larger).  This, because blocks are packed together and
+          unaligned in Squashfs, should reduce latency.
+
+          This, however, gives poor performance on MTD NAND devices where
+          the optimal I/O size is 4K (even though the devices can support
+          smaller block sizes).
+
+          Using a 4K device block size may also improve overall I/O
+          performance for some file access patterns (e.g. sequential
+          accesses of files in filesystem order) on all media.
+
+          Setting this option will force Squashfs to use a 4K device block
+          size by default.
+
+          If unsure, say N.
+
 config SQUASHFS_EMBEDDED
         bool "Additional option for memory-constrained systems"
         depends on SQUASHFS
diff --git a/fs/squashfs/squashfs_fs.h b/fs/squashfs/squashfs_fs.h
index b4a4e539a08c..e8e14645de9a 100644
--- a/fs/squashfs/squashfs_fs.h
+++ b/fs/squashfs/squashfs_fs.h
@@ -36,6 +36,13 @@
 #define SQUASHFS_FILE_SIZE              131072
 #define SQUASHFS_FILE_LOG               17
 
+/* default size of block device I/O */
+#ifdef CONFIG_SQUASHFS_4K_DEVBLK_SIZE
+#define SQUASHFS_DEVBLK_SIZE 4096
+#else
+#define SQUASHFS_DEVBLK_SIZE 1024
+#endif
+
 #define SQUASHFS_FILE_MAX_SIZE          1048576
 #define SQUASHFS_FILE_MAX_LOG           20
 
diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c
index 7438850c62d0..2da1715452ac 100644
--- a/fs/squashfs/super.c
+++ b/fs/squashfs/super.c
@@ -95,7 +95,7 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
         }
         msblk = sb->s_fs_info;
 
-        msblk->devblksize = sb_min_blocksize(sb, BLOCK_SIZE);
+        msblk->devblksize = sb_min_blocksize(sb, SQUASHFS_DEVBLK_SIZE);
         msblk->devblksize_log2 = ffz(~msblk->devblksize);
 
         mutex_init(&msblk->read_data_mutex);
diff --git a/fs/statfs.c b/fs/statfs.c
index 8244924dec55..9cf04a118965 100644
--- a/fs/statfs.c
+++ b/fs/statfs.c
@@ -76,7 +76,7 @@ EXPORT_SYMBOL(vfs_statfs);
 int user_statfs(const char __user *pathname, struct kstatfs *st)
 {
         struct path path;
-        int error = user_path(pathname, &path);
+        int error = user_path_at(AT_FDCWD, pathname, LOOKUP_FOLLOW|LOOKUP_AUTOMOUNT, &path);
         if (!error) {
                 error = vfs_statfs(&path, st);
                 path_put(&path);
diff --git a/fs/sync.c b/fs/sync.c
index c98a7477edfd..101b8ef901d7 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -43,7 +43,7 @@ static int __sync_filesystem(struct super_block *sb, int wait)
         if (wait)
                 sync_inodes_sb(sb);
         else
-                writeback_inodes_sb(sb);
+                writeback_inodes_sb(sb, WB_REASON_SYNC);
 
         if (sb->s_op->sync_fs)
                 sb->s_op->sync_fs(sb, wait);
@@ -98,7 +98,7 @@ static void sync_filesystems(int wait)
  */
 SYSCALL_DEFINE0(sync)
 {
-        wakeup_flusher_threads(0);
+        wakeup_flusher_threads(0, WB_REASON_SYNC);
         sync_filesystems(0);
         sync_filesystems(1);
         if (unlikely(laptop_mode))
diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c
index 315de66e52b2..bc4f94b28706 100644
--- a/fs/ubifs/budget.c
+++ b/fs/ubifs/budget.c
@@ -63,7 +63,7 @@
 static void shrink_liability(struct ubifs_info *c, int nr_to_write)
 {
         down_read(&c->vfs_sb->s_umount);
-        writeback_inodes_sb(c->vfs_sb);
+        writeback_inodes_sb(c->vfs_sb, WB_REASON_FS_FREE_SPACE);
         up_read(&c->vfs_sb->s_umount);
 }
 
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index eef109a1a927..b09ba2dd8b62 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -870,6 +870,22 @@ void dbg_dump_lpt_info(struct ubifs_info *c)
         spin_unlock(&dbg_lock);
 }
 
+void dbg_dump_sleb(const struct ubifs_info *c,
+                   const struct ubifs_scan_leb *sleb, int offs)
+{
+        struct ubifs_scan_node *snod;
+
+        printk(KERN_DEBUG "(pid %d) start dumping scanned data from LEB %d:%d\n",
+               current->pid, sleb->lnum, offs);
+
+        list_for_each_entry(snod, &sleb->nodes, list) {
+                cond_resched();
+                printk(KERN_DEBUG "Dumping node at LEB %d:%d len %d\n", sleb->lnum,
+                       snod->offs, snod->len);
+                dbg_dump_node(c, snod->node);
+        }
+}
+
 void dbg_dump_leb(const struct ubifs_info *c, int lnum)
 {
         struct ubifs_scan_leb *sleb;
diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h
index feb361e252ac..8d9c46810189 100644
--- a/fs/ubifs/debug.h
+++ b/fs/ubifs/debug.h
@@ -269,6 +269,8 @@ void dbg_dump_lprop(const struct ubifs_info *c, const struct ubifs_lprops *lp);
 void dbg_dump_lprops(struct ubifs_info *c);
 void dbg_dump_lpt_info(struct ubifs_info *c);
 void dbg_dump_leb(const struct ubifs_info *c, int lnum);
+void dbg_dump_sleb(const struct ubifs_info *c,
+                   const struct ubifs_scan_leb *sleb, int offs);
 void dbg_dump_znode(const struct ubifs_info *c,
                     const struct ubifs_znode *znode);
 void dbg_dump_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat);
@@ -387,6 +389,9 @@ static inline void dbg_dump_lpt_info(struct ubifs_info *c)        { return; }
 static inline void dbg_dump_leb(const struct ubifs_info *c,
                                 int lnum)                         { return; }
 static inline void
+dbg_dump_sleb(const struct ubifs_info *c,
+              const struct ubifs_scan_leb *sleb, int offs)        { return; }
+static inline void
 dbg_dump_znode(const struct ubifs_info *c,
                const struct ubifs_znode *znode)                   { return; }
 static inline void dbg_dump_heap(struct ubifs_info *c,
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
index af02790d9328..ee4f43f4bb99 100644
--- a/fs/ubifs/recovery.c
+++ b/fs/ubifs/recovery.c
@@ -983,7 +983,7 @@ int ubifs_recover_inl_heads(struct ubifs_info *c, void *sbuf)
 }
 
 /**
- *  clean_an_unclean_leb - read and write a LEB to remove corruption.
+ * clean_an_unclean_leb - read and write a LEB to remove corruption.
  * @c: UBIFS file-system description object
  * @ucleb: unclean LEB information
  * @sbuf: LEB-sized buffer to use
diff --git a/fs/ubifs/sb.c b/fs/ubifs/sb.c
index 93d938ad3d2a..6094c5a5d7a8 100644
--- a/fs/ubifs/sb.c
+++ b/fs/ubifs/sb.c
@@ -247,7 +247,7 @@ static int create_default_filesystem(struct ubifs_info *c)
         mst->total_dirty = cpu_to_le64(tmp64);
 
         /*  The indexing LEB does not contribute to dark space */
-        tmp64 = (c->main_lebs - 1) * c->dark_wm;
+        tmp64 = ((long long)(c->main_lebs - 1) * c->dark_wm);
         mst->total_dark = cpu_to_le64(tmp64);
 
         mst->total_used = cpu_to_le64(UBIFS_INO_NODE_SZ);