14 files changed, 348 insertions, 134 deletions

diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 8855aad3929c..22c64fff1bd5 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -683,7 +683,7 @@ static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref,
                 return PTR_ERR(fspath);
 
         if (fspath > fspath_min) {
-                ipath->fspath->val[i] = (u64)fspath;
+                ipath->fspath->val[i] = (u64)(unsigned long)fspath;
                 ++ipath->fspath->elem_cnt;
                 ipath->fspath->bytes_left = fspath - fspath_min;
         } else {
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 0fe615e4ea38..dede441bdeee 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -514,10 +514,25 @@ static inline int should_cow_block(struct btrfs_trans_handle *trans,
                                    struct btrfs_root *root,
                                    struct extent_buffer *buf)
 {
+        /* ensure we can see the force_cow */
+        smp_rmb();
+
+        /*
+         * We do not need to cow a block if
+         * 1) this block is not created or changed in this transaction;
+         * 2) this block does not belong to TREE_RELOC tree;
+         * 3) the root is not forced COW.
+         *
+         * What is forced COW:
+         *    when we create snapshot during commiting the transaction,
+         *    after we've finished coping src root, we must COW the shared
+         *    block to ensure the metadata consistency.
+         */
         if (btrfs_header_generation(buf) == trans->transid &&
             !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) &&
             !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
-              btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)))
+              btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)) &&
+            !root->force_cow)
                 return 0;
         return 1;
 }
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index b9ba59ff9292..50634abef9b4 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -848,7 +848,8 @@ struct btrfs_free_cluster {
 enum btrfs_caching_type {
         BTRFS_CACHE_NO          = 0,
         BTRFS_CACHE_STARTED     = 1,
-        BTRFS_CACHE_FINISHED    = 2,
+        BTRFS_CACHE_FAST        = 2,
+        BTRFS_CACHE_FINISHED    = 3,
 };
 
 enum btrfs_disk_cache_state {
@@ -1271,6 +1272,8 @@ struct btrfs_root {
          * for stat.  It may be used for more later
          */
         dev_t anon_dev;
+
+        int force_cow;
 };
 
 struct btrfs_ioctl_defrag_range_args {
@@ -2366,6 +2369,9 @@ int btrfs_block_rsv_check(struct btrfs_root *root,
 int btrfs_block_rsv_refill(struct btrfs_root *root,
                           struct btrfs_block_rsv *block_rsv,
                           u64 min_reserved);
+int btrfs_block_rsv_refill_noflush(struct btrfs_root *root,
+                                   struct btrfs_block_rsv *block_rsv,
+                                   u64 min_reserved);
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
                             struct btrfs_block_rsv *dst_rsv,
                             u64 num_bytes);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 62afe5c5694e..632f8f3cc9db 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -620,7 +620,7 @@ out:
 
 static int btree_io_failed_hook(struct bio *failed_bio,
                          struct page *page, u64 start, u64 end,
-                         u64 mirror_num, struct extent_state *state)
+                         int mirror_num, struct extent_state *state)
 {
         struct extent_io_tree *tree;
         unsigned long len;
@@ -2573,22 +2573,10 @@ static int write_dev_supers(struct btrfs_device *device,
         int errors = 0;
         u32 crc;
         u64 bytenr;
-        int last_barrier = 0;
 
         if (max_mirrors == 0)
                 max_mirrors = BTRFS_SUPER_MIRROR_MAX;
 
-        /* make sure only the last submit_bh does a barrier */
-        if (do_barriers) {
-                for (i = 0; i < max_mirrors; i++) {
-                        bytenr = btrfs_sb_offset(i);
-                        if (bytenr + BTRFS_SUPER_INFO_SIZE >=
-                            device->total_bytes)
-                                break;
-                        last_barrier = i;
-                }
-        }
-
         for (i = 0; i < max_mirrors; i++) {
                 bytenr = btrfs_sb_offset(i);
                 if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes)
@@ -2634,17 +2622,136 @@ static int write_dev_supers(struct btrfs_device *device,
                         bh->b_end_io = btrfs_end_buffer_write_sync;
                 }
 
-                if (i == last_barrier && do_barriers)
-                        ret = submit_bh(WRITE_FLUSH_FUA, bh);
-                else
-                        ret = submit_bh(WRITE_SYNC, bh);
-
+                /*
+                 * we fua the first super.  The others we allow
+                 * to go down lazy.
+                 */
+                ret = submit_bh(WRITE_FUA, bh);
                 if (ret)
                         errors++;
         }
         return errors < i ? 0 : -1;
 }
 
+/*
+ * endio for the write_dev_flush, this will wake anyone waiting
+ * for the barrier when it is done
+ */
+static void btrfs_end_empty_barrier(struct bio *bio, int err)
+{
+        if (err) {
+                if (err == -EOPNOTSUPP)
+                        set_bit(BIO_EOPNOTSUPP, &bio->bi_flags);
+                clear_bit(BIO_UPTODATE, &bio->bi_flags);
+        }
+        if (bio->bi_private)
+                complete(bio->bi_private);
+        bio_put(bio);
+}
+
+/*
+ * trigger flushes for one the devices.  If you pass wait == 0, the flushes are
+ * sent down.  With wait == 1, it waits for the previous flush.
+ *
+ * any device where the flush fails with eopnotsupp are flagged as not-barrier
+ * capable
+ */
+static int write_dev_flush(struct btrfs_device *device, int wait)
+{
+        struct bio *bio;
+        int ret = 0;
+
+        if (device->nobarriers)
+                return 0;
+
+        if (wait) {
+                bio = device->flush_bio;
+                if (!bio)
+                        return 0;
+
+                wait_for_completion(&device->flush_wait);
+
+                if (bio_flagged(bio, BIO_EOPNOTSUPP)) {
+                        printk("btrfs: disabling barriers on dev %s\n",
+                               device->name);
+                        device->nobarriers = 1;
+                }
+                if (!bio_flagged(bio, BIO_UPTODATE)) {
+                        ret = -EIO;
+                }
+
+                /* drop the reference from the wait == 0 run */
+                bio_put(bio);
+                device->flush_bio = NULL;
+
+                return ret;
+        }
+
+        /*
+         * one reference for us, and we leave it for the
+         * caller
+         */
+        device->flush_bio = NULL;;
+        bio = bio_alloc(GFP_NOFS, 0);
+        if (!bio)
+                return -ENOMEM;
+
+        bio->bi_end_io = btrfs_end_empty_barrier;
+        bio->bi_bdev = device->bdev;
+        init_completion(&device->flush_wait);
+        bio->bi_private = &device->flush_wait;
+        device->flush_bio = bio;
+
+        bio_get(bio);
+        submit_bio(WRITE_FLUSH, bio);
+
+        return 0;
+}
+
+/*
+ * send an empty flush down to each device in parallel,
+ * then wait for them
+ */
+static int barrier_all_devices(struct btrfs_fs_info *info)
+{
+        struct list_head *head;
+        struct btrfs_device *dev;
+        int errors = 0;
+        int ret;
+
+        /* send down all the barriers */
+        head = &info->fs_devices->devices;
+        list_for_each_entry_rcu(dev, head, dev_list) {
+                if (!dev->bdev) {
+                        errors++;
+                        continue;
+                }
+                if (!dev->in_fs_metadata || !dev->writeable)
+                        continue;
+
+                ret = write_dev_flush(dev, 0);
+                if (ret)
+                        errors++;
+        }
+
+        /* wait for all the barriers */
+        list_for_each_entry_rcu(dev, head, dev_list) {
+                if (!dev->bdev) {
+                        errors++;
+                        continue;
+                }
+                if (!dev->in_fs_metadata || !dev->writeable)
+                        continue;
+
+                ret = write_dev_flush(dev, 1);
+                if (ret)
+                        errors++;
+        }
+        if (errors)
+                return -EIO;
+        return 0;
+}
+
 int write_all_supers(struct btrfs_root *root, int max_mirrors)
 {
         struct list_head *head;
@@ -2666,6 +2773,10 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 
         mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
         head = &root->fs_info->fs_devices->devices;
+
+        if (do_barriers)
+                barrier_all_devices(root->fs_info);
+
         list_for_each_entry_rcu(dev, head, dev_list) {
                 if (!dev->bdev) {
                         total_errors++;
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index b232150b5b6b..f0d5718d2587 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -467,13 +467,59 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
                              struct btrfs_root *root,
                              int load_cache_only)
 {
+        DEFINE_WAIT(wait);
         struct btrfs_fs_info *fs_info = cache->fs_info;
         struct btrfs_caching_control *caching_ctl;
         int ret = 0;
 
-        smp_mb();
-        if (cache->cached != BTRFS_CACHE_NO)
+        caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS);
+        BUG_ON(!caching_ctl);
+
+        INIT_LIST_HEAD(&caching_ctl->list);
+        mutex_init(&caching_ctl->mutex);
+        init_waitqueue_head(&caching_ctl->wait);
+        caching_ctl->block_group = cache;
+        caching_ctl->progress = cache->key.objectid;
+        atomic_set(&caching_ctl->count, 1);
+        caching_ctl->work.func = caching_thread;
+
+        spin_lock(&cache->lock);
+        /*
+         * This should be a rare occasion, but this could happen I think in the
+         * case where one thread starts to load the space cache info, and then
+         * some other thread starts a transaction commit which tries to do an
+         * allocation while the other thread is still loading the space cache
+         * info.  The previous loop should have kept us from choosing this block
+         * group, but if we've moved to the state where we will wait on caching
+         * block groups we need to first check if we're doing a fast load here,
+         * so we can wait for it to finish, otherwise we could end up allocating
+         * from a block group who's cache gets evicted for one reason or
+         * another.
+         */
+        while (cache->cached == BTRFS_CACHE_FAST) {
+                struct btrfs_caching_control *ctl;
+
+                ctl = cache->caching_ctl;
+                atomic_inc(&ctl->count);
+                prepare_to_wait(&ctl->wait, &wait, TASK_UNINTERRUPTIBLE);
+                spin_unlock(&cache->lock);
+
+                schedule();
+
+                finish_wait(&ctl->wait, &wait);
+                put_caching_control(ctl);
+                spin_lock(&cache->lock);
+        }
+
+        if (cache->cached != BTRFS_CACHE_NO) {
+                spin_unlock(&cache->lock);
+                kfree(caching_ctl);
                 return 0;
+        }
+        WARN_ON(cache->caching_ctl);
+        cache->caching_ctl = caching_ctl;
+        cache->cached = BTRFS_CACHE_FAST;
+        spin_unlock(&cache->lock);
 
         /*
          * We can't do the read from on-disk cache during a commit since we need
@@ -484,56 +530,51 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
         if (trans && (!trans->transaction->in_commit) &&
             (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);
-                        return 0;
-                }
-                cache->cached = BTRFS_CACHE_STARTED;
-                spin_unlock(&cache->lock);
-
                 ret = load_free_space_cache(fs_info, cache);
 
                 spin_lock(&cache->lock);
                 if (ret == 1) {
+                        cache->caching_ctl = NULL;
                         cache->cached = BTRFS_CACHE_FINISHED;
                         cache->last_byte_to_unpin = (u64)-1;
                 } else {
-                        cache->cached = BTRFS_CACHE_NO;
+                        if (load_cache_only) {
+                                cache->caching_ctl = NULL;
+                                cache->cached = BTRFS_CACHE_NO;
+                        } else {
+                                cache->cached = BTRFS_CACHE_STARTED;
+                        }
                 }
                 spin_unlock(&cache->lock);
+                wake_up(&caching_ctl->wait);
                 if (ret == 1) {
+                        put_caching_control(caching_ctl);
                         free_excluded_extents(fs_info->extent_root, cache);
                         return 0;
                 }
+        } else {
+                /*
+                 * We are not going to do the fast caching, set cached to the
+                 * appropriate value and wakeup any waiters.
+                 */
+                spin_lock(&cache->lock);
+                if (load_cache_only) {
+                        cache->caching_ctl = NULL;
+                        cache->cached = BTRFS_CACHE_NO;
+                } else {
+                        cache->cached = BTRFS_CACHE_STARTED;
+                }
+                spin_unlock(&cache->lock);
+                wake_up(&caching_ctl->wait);
         }
 
-        if (load_cache_only)
-                return 0;
-
-        caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS);
-        BUG_ON(!caching_ctl);
-
-        INIT_LIST_HEAD(&caching_ctl->list);
-        mutex_init(&caching_ctl->mutex);
-        init_waitqueue_head(&caching_ctl->wait);
-        caching_ctl->block_group = cache;
-        caching_ctl->progress = cache->key.objectid;
-        /* one for caching kthread, one for caching block group list */
-        atomic_set(&caching_ctl->count, 2);
-        caching_ctl->work.func = caching_thread;
-
-        spin_lock(&cache->lock);
-        if (cache->cached != BTRFS_CACHE_NO) {
-                spin_unlock(&cache->lock);
-                kfree(caching_ctl);
+        if (load_cache_only) {
+                put_caching_control(caching_ctl);
                 return 0;
         }
-        cache->caching_ctl = caching_ctl;
-        cache->cached = BTRFS_CACHE_STARTED;
-        spin_unlock(&cache->lock);
 
         down_write(&fs_info->extent_commit_sem);
+        atomic_inc(&caching_ctl->count);
         list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups);
         up_write(&fs_info->extent_commit_sem);
 
@@ -3847,9 +3888,9 @@ int btrfs_block_rsv_check(struct btrfs_root *root,
         return ret;
 }
 
-int btrfs_block_rsv_refill(struct btrfs_root *root,
-                          struct btrfs_block_rsv *block_rsv,
-                          u64 min_reserved)
+static inline int __btrfs_block_rsv_refill(struct btrfs_root *root,
+                                           struct btrfs_block_rsv *block_rsv,
+                                           u64 min_reserved, int flush)
 {
         u64 num_bytes = 0;
         int ret = -ENOSPC;
@@ -3868,7 +3909,7 @@ int btrfs_block_rsv_refill(struct btrfs_root *root,
         if (!ret)
                 return 0;
 
-        ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 1);
+        ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush);
         if (!ret) {
                 block_rsv_add_bytes(block_rsv, num_bytes, 0);
                 return 0;
@@ -3877,6 +3918,20 @@ int btrfs_block_rsv_refill(struct btrfs_root *root,
         return ret;
 }
 
+int btrfs_block_rsv_refill(struct btrfs_root *root,
+                           struct btrfs_block_rsv *block_rsv,
+                           u64 min_reserved)
+{
+        return __btrfs_block_rsv_refill(root, block_rsv, min_reserved, 1);
+}
+
+int btrfs_block_rsv_refill_noflush(struct btrfs_root *root,
+                                   struct btrfs_block_rsv *block_rsv,
+                                   u64 min_reserved)
+{
+        return __btrfs_block_rsv_refill(root, block_rsv, min_reserved, 0);
+}
+
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
                             struct btrfs_block_rsv *dst_rsv,
                             u64 num_bytes)
@@ -5178,13 +5233,15 @@ search:
                 }
 
 have_block_group:
-                if (unlikely(block_group->cached == BTRFS_CACHE_NO)) {
+                cached = block_group_cache_done(block_group);
+                if (unlikely(!cached)) {
                         u64 free_percent;
 
+                        found_uncached_bg = true;
                         ret = cache_block_group(block_group, trans,
                                                 orig_root, 1);
                         if (block_group->cached == BTRFS_CACHE_FINISHED)
-                                goto have_block_group;
+                                goto alloc;
 
                         free_percent = btrfs_block_group_used(&block_group->item);
                         free_percent *= 100;
@@ -5206,7 +5263,6 @@ have_block_group:
                                                         orig_root, 0);
                                 BUG_ON(ret);
                         }
-                        found_uncached_bg = true;
 
                         /*
                          * If loop is set for cached only, try the next block
@@ -5216,17 +5272,14 @@ have_block_group:
                                 goto loop;
                 }
 
-                cached = block_group_cache_done(block_group);
-                if (unlikely(!cached))
-                        found_uncached_bg = true;
-
+alloc:
                 if (unlikely(block_group->ro))
                         goto loop;
 
                 spin_lock(&block_group->free_space_ctl->tree_lock);
                 if (cached &&
                     block_group->free_space_ctl->free_space <
-                    num_bytes + empty_size) {
+                    num_bytes + empty_cluster + empty_size) {
                         spin_unlock(&block_group->free_space_ctl->tree_lock);
                         goto loop;
                 }
@@ -5247,12 +5300,10 @@ have_block_group:
                          * people trying to start a new cluster
                          */
                         spin_lock(&last_ptr->refill_lock);
-                        if (last_ptr->block_group &&
-                            (last_ptr->block_group->ro ||
-                            !block_group_bits(last_ptr->block_group, data))) {
-                                offset = 0;
+                        if (!last_ptr->block_group ||
+                            last_ptr->block_group->ro ||
+                            !block_group_bits(last_ptr->block_group, data))
                                 goto refill_cluster;
-                        }
 
                         offset = btrfs_alloc_from_cluster(block_group, last_ptr,
                                                  num_bytes, search_start);
@@ -5303,7 +5354,7 @@ refill_cluster:
                         /* allocate a cluster in this block group */
                         ret = btrfs_find_space_cluster(trans, root,
                                                block_group, last_ptr,
-                                               offset, num_bytes,
+                                               search_start, num_bytes,
                                                empty_cluster + empty_size);
                         if (ret == 0) {
                                 /*
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 1f87c4d0e7a0..be1bf627a14b 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -2285,16 +2285,22 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
                                 clean_io_failure(start, page);
                 }
                 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);
+                        int failed_mirror;
+                        failed_mirror = (int)(unsigned long)bio->bi_bdev;
+                        /*
+                         * The generic bio_readpage_error handles errors the
+                         * following way: If possible, new read requests are
+                         * created and submitted and will end up in
+                         * end_bio_extent_readpage as well (if we're lucky, not
+                         * in the !uptodate case). In that case it returns 0 and
+                         * we just go on with the next page in our bio. If it
+                         * can't handle the error it will return -EIO and we
+                         * remain responsible for that page.
+                         */
+                        ret = bio_readpage_error(bio, page, start, end,
+                                                        failed_mirror, NULL);
                         if (ret == 0) {
+error_handled:
                                 uptodate =
                                         test_bit(BIO_UPTODATE, &bio->bi_flags);
                                 if (err)
@@ -2302,6 +2308,13 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
                                 uncache_state(&cached);
                                 continue;
                         }
+                        if (tree->ops && tree->ops->readpage_io_failed_hook) {
+                                ret = tree->ops->readpage_io_failed_hook(
+                                                        bio, page, start, end,
+                                                        failed_mirror, state);
+                                if (ret == 0)
+                                        goto error_handled;
+                        }
                 }
 
                 if (uptodate) {
@@ -3366,6 +3379,9 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
                 return -ENOMEM;
         path->leave_spinning = 1;
 
+        start = ALIGN(start, BTRFS_I(inode)->root->sectorsize);
+        len = ALIGN(len, BTRFS_I(inode)->root->sectorsize);
+
         /*
          * lookup the last file extent.  We're not using i_size here
          * because there might be preallocation past i_size
@@ -3413,7 +3429,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
         lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0,
                          &cached_state, GFP_NOFS);
 
-        em = get_extent_skip_holes(inode, off, last_for_get_extent,
+        em = get_extent_skip_holes(inode, start, last_for_get_extent,
                                    get_extent);
         if (!em)
                 goto out;
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index feb9be0e23bc..7604c3001322 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -70,7 +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 failed_mirror,
+                                       u64 start, u64 end, int failed_mirror,
                                        struct extent_state *state);
         int (*writepage_io_failed_hook)(struct bio *bio, struct page *page,
                                         u64 start, u64 end,
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 181760f9d2ab..ec23d43d0c35 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -351,6 +351,11 @@ static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode,
                 }
         }
 
+        for (i = 0; i < io_ctl->num_pages; i++) {
+                clear_page_dirty_for_io(io_ctl->pages[i]);
+                set_page_extent_mapped(io_ctl->pages[i]);
+        }
+
         return 0;
 }
 
@@ -1465,6 +1470,7 @@ static void add_new_bitmap(struct btrfs_free_space_ctl *ctl,
 {
         info->offset = offset_to_bitmap(ctl, offset);
         info->bytes = 0;
+        INIT_LIST_HEAD(&info->list);
         link_free_space(ctl, info);
         ctl->total_bitmaps++;
 
@@ -1844,7 +1850,13 @@ again:
                 info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
                                           1, 0);
                 if (!info) {
-                        WARN_ON(1);
+                        /* the tree logging code might be calling us before we
+                         * have fully loaded the free space rbtree for this
+                         * block group.  So it is possible the entry won't
+                         * be in the rbtree yet at all.  The caching code
+                         * will make sure not to put it in the rbtree if
+                         * the logging code has pinned it.
+                         */
                         goto out_lock;
                 }
         }
@@ -2308,6 +2320,7 @@ again:
 
         if (!found) {
                 start = i;
+                cluster->max_size = 0;
                 found = true;
         }
 
@@ -2451,16 +2464,23 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
 {
         struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
         struct btrfs_free_space *entry;
-        struct rb_node *node;
         int ret = -ENOSPC;
+        u64 bitmap_offset = offset_to_bitmap(ctl, offset);
 
         if (ctl->total_bitmaps == 0)
                 return -ENOSPC;
 
         /*
-         * First check our cached list of bitmaps and see if there is an entry
-         * here that will work.
+         * The bitmap that covers offset won't be in the list unless offset
+         * is just its start offset.
          */
+        entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
+        if (entry->offset != bitmap_offset) {
+                entry = tree_search_offset(ctl, bitmap_offset, 1, 0);
+                if (entry && list_empty(&entry->list))
+                        list_add(&entry->list, bitmaps);
+        }
+
         list_for_each_entry(entry, bitmaps, list) {
                 if (entry->bytes < min_bytes)
                         continue;
@@ -2471,38 +2491,10 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
         }
 
         /*
-         * If we do have entries on our list and we are here then we didn't find
-         * anything, so go ahead and get the next entry after the last entry in
-         * this list and start the search from there.
+         * The bitmaps list has all the bitmaps that record free space
+         * starting after offset, so no more search is required.
          */
-        if (!list_empty(bitmaps)) {
-                entry = list_entry(bitmaps->prev, struct btrfs_free_space,
-                                   list);
-                node = rb_next(&entry->offset_index);
-                if (!node)
-                        return -ENOSPC;
-                entry = rb_entry(node, struct btrfs_free_space, offset_index);
-                goto search;
-        }
-
-        entry = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 0, 1);
-        if (!entry)
-                return -ENOSPC;
-
-search:
-        node = &entry->offset_index;
-        do {
-                entry = rb_entry(node, struct btrfs_free_space, offset_index);
-                node = rb_next(&entry->offset_index);
-                if (!entry->bitmap)
-                        continue;
-                if (entry->bytes < min_bytes)
-                        continue;
-                ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset,
-                                           bytes, min_bytes);
-        } while (ret && node);
-
-        return ret;
+        return -ENOSPC;
 }
 
 /*
@@ -2520,8 +2512,8 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
                              u64 offset, u64 bytes, u64 empty_size)
 {
         struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
-        struct list_head bitmaps;
         struct btrfs_free_space *entry, *tmp;
+        LIST_HEAD(bitmaps);
         u64 min_bytes;
         int ret;
 
@@ -2560,7 +2552,6 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
                 goto out;
         }
 
-        INIT_LIST_HEAD(&bitmaps);
         ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset,
                                       bytes, min_bytes);
         if (ret)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 116ab67a06df..2c984f7d4c2a 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -3490,7 +3490,7 @@ void btrfs_evict_inode(struct inode *inode)
          * doing the truncate.
          */
         while (1) {
-                ret = btrfs_block_rsv_refill(root, rsv, min_size);
+                ret = btrfs_block_rsv_refill_noflush(root, rsv, min_size);
 
                 /*
                  * Try and steal from the global reserve since we will
@@ -6794,11 +6794,13 @@ static int btrfs_getattr(struct vfsmount *mnt,
                          struct dentry *dentry, struct kstat *stat)
 {
         struct inode *inode = dentry->d_inode;
+        u32 blocksize = inode->i_sb->s_blocksize;
+
         generic_fillattr(inode, stat);
         stat->dev = BTRFS_I(inode)->root->anon_dev;
         stat->blksize = PAGE_CACHE_SIZE;
-        stat->blocks = (inode_get_bytes(inode) +
-                        BTRFS_I(inode)->delalloc_bytes) >> 9;
+        stat->blocks = (ALIGN(inode_get_bytes(inode), blocksize) +
+                ALIGN(BTRFS_I(inode)->delalloc_bytes, blocksize)) >> 9;
         return 0;
 }
 
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 4a34c472f126..72d461656f60 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -1216,12 +1216,12 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root,
                 *devstr = '\0';
                 devstr = vol_args->name;
                 devid = simple_strtoull(devstr, &end, 10);
-                printk(KERN_INFO "resizing devid %llu\n",
+                printk(KERN_INFO "btrfs: resizing devid %llu\n",
                        (unsigned long long)devid);
         }
         device = btrfs_find_device(root, devid, NULL, NULL);
         if (!device) {
-                printk(KERN_INFO "resizer unable to find device %llu\n",
+                printk(KERN_INFO "btrfs: resizer unable to find device %llu\n",
                        (unsigned long long)devid);
                 ret = -EINVAL;
                 goto out_unlock;
@@ -1267,7 +1267,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root,
         do_div(new_size, root->sectorsize);
         new_size *= root->sectorsize;
 
-        printk(KERN_INFO "new size for %s is %llu\n",
+        printk(KERN_INFO "btrfs: new size for %s is %llu\n",
                 device->name, (unsigned long long)new_size);
 
         if (new_size > old_size) {
@@ -1278,7 +1278,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root,
                 }
                 ret = btrfs_grow_device(trans, device, new_size);
                 btrfs_commit_transaction(trans, root);
-        } else {
+        } else if (new_size < old_size) {
                 ret = btrfs_shrink_device(device, new_size);
         }
 
@@ -2930,11 +2930,13 @@ static long btrfs_ioctl_ino_to_path(struct btrfs_root *root, void __user *arg)
                 goto out;
 
         for (i = 0; i < ipath->fspath->elem_cnt; ++i) {
-                rel_ptr = ipath->fspath->val[i] - (u64)ipath->fspath->val;
+                rel_ptr = ipath->fspath->val[i] -
+                          (u64)(unsigned long)ipath->fspath->val;
                 ipath->fspath->val[i] = rel_ptr;
         }
 
-        ret = copy_to_user((void *)ipa->fspath, (void *)ipath->fspath, size);
+        ret = copy_to_user((void *)(unsigned long)ipa->fspath,
+                           (void *)(unsigned long)ipath->fspath, size);
         if (ret) {
                 ret = -EFAULT;
                 goto out;
@@ -3017,7 +3019,8 @@ static long btrfs_ioctl_logical_to_ino(struct btrfs_root *root,
         if (ret < 0)
                 goto out;
 
-        ret = copy_to_user((void *)loi->inodes, (void *)inodes, size);
+        ret = copy_to_user((void *)(unsigned long)loi->inodes,
+                           (void *)(unsigned long)inodes, size);
         if (ret)
                 ret = -EFAULT;
 
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index f4190f22edfb..c27bcb67f330 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -256,6 +256,11 @@ static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, void *ctx)
         btrfs_release_path(swarn->path);
 
         ipath = init_ipath(4096, local_root, swarn->path);
+        if (IS_ERR(ipath)) {
+                ret = PTR_ERR(ipath);
+                ipath = NULL;
+                goto err;
+        }
         ret = paths_from_inode(inum, ipath);
 
         if (ret < 0)
@@ -272,7 +277,7 @@ static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, void *ctx)
                         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]);
+                        (char *)(unsigned long)ipath->fspath->val[i]);
 
         free_ipath(ipath);
         return 0;
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 17ee7fc5e64e..e28ad4baf483 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -1057,7 +1057,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes)
         int i = 0, nr_devices;
         int ret;
 
-        nr_devices = fs_info->fs_devices->rw_devices;
+        nr_devices = fs_info->fs_devices->open_devices;
         BUG_ON(!nr_devices);
 
         devices_info = kmalloc(sizeof(*devices_info) * nr_devices,
@@ -1079,8 +1079,8 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes)
         else
                 min_stripe_size = BTRFS_STRIPE_LEN;
 
-        list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) {
-                if (!device->in_fs_metadata)
+        list_for_each_entry(device, &fs_devices->devices, dev_list) {
+                if (!device->in_fs_metadata || !device->bdev)
                         continue;
 
                 avail_space = device->total_bytes - device->bytes_used;
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 6a0574e923bc..81376d94cd3c 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -785,6 +785,10 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 
                         btrfs_save_ino_cache(root, trans);
 
+                        /* see comments in should_cow_block() */
+                        root->force_cow = 0;
+                        smp_wmb();
+
                         if (root->commit_root != root->node) {
                                 mutex_lock(&root->fs_commit_mutex);
                                 switch_commit_root(root);
@@ -947,6 +951,10 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
         btrfs_tree_unlock(old);
         free_extent_buffer(old);
 
+        /* see comments in should_cow_block() */
+        root->force_cow = 1;
+        smp_wmb();
+
         btrfs_set_root_node(new_root_item, tmp);
         /* record when the snapshot was created in key.offset */
         key.offset = trans->transid;
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index ab5b1c49f352..78f2d4d4f37f 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -100,6 +100,12 @@ struct btrfs_device {
         struct reada_zone *reada_curr_zone;
         struct radix_tree_root reada_zones;
         struct radix_tree_root reada_extents;
+
+        /* for sending down flush barriers */
+        struct bio *flush_bio;
+        struct completion flush_wait;
+        int nobarriers;
+
 };
 
 struct btrfs_fs_devices {