1 files changed, 462 insertions, 360 deletions

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index b3ad168a0bfc..2bfdc641d4e3 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -36,6 +36,7 @@
 #include <linux/xattr.h>
 #include <linux/posix_acl.h>
 #include <linux/falloc.h>
+#include <linux/slab.h>
 #include "compat.h"
 #include "ctree.h"
 #include "disk-io.h"
@@ -88,13 +89,14 @@ static noinline int cow_file_range(struct inode *inode,
                                    u64 start, u64 end, int *page_started,
                                    unsigned long *nr_written, int unlock);
 
-static int btrfs_init_inode_security(struct inode *inode,  struct inode *dir)
+static int btrfs_init_inode_security(struct btrfs_trans_handle *trans,
+                                     struct inode *inode,  struct inode *dir)
 {
         int err;
 
-        err = btrfs_init_acl(inode, dir);
+        err = btrfs_init_acl(trans, inode, dir);
         if (!err)
-                err = btrfs_xattr_security_init(inode, dir);
+                err = btrfs_xattr_security_init(trans, inode, dir);
         return err;
 }
 
@@ -188,8 +190,18 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
         btrfs_mark_buffer_dirty(leaf);
         btrfs_free_path(path);
 
+        /*
+         * we're an inline extent, so nobody can
+         * extend the file past i_size without locking
+         * a page we already have locked.
+         *
+         * We must do any isize and inode updates
+         * before we unlock the pages.  Otherwise we
+         * could end up racing with unlink.
+         */
         BTRFS_I(inode)->disk_i_size = inode->i_size;
         btrfs_update_inode(trans, root, inode);
+
         return 0;
 fail:
         btrfs_free_path(path);
@@ -230,8 +242,7 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
                 return 1;
         }
 
-        ret = btrfs_drop_extents(trans, root, inode, start,
-                                 aligned_end, aligned_end, start,
+        ret = btrfs_drop_extents(trans, inode, start, aligned_end,
                                  &hint_byte, 1);
         BUG_ON(ret);
 
@@ -369,7 +380,8 @@ again:
          * change at any time if we discover bad compression ratios.
          */
         if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) &&
-            btrfs_test_opt(root, COMPRESS)) {
+            (btrfs_test_opt(root, COMPRESS) ||
+             (BTRFS_I(inode)->force_compress))) {
                 WARN_ON(pages);
                 pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS);
 
@@ -416,7 +428,6 @@ again:
                                                     start, end,
                                                     total_compressed, pages);
                 }
-                btrfs_end_transaction(trans, root);
                 if (ret == 0) {
                         /*
                          * inline extent creation worked, we don't need
@@ -430,9 +441,11 @@ again:
                              EXTENT_CLEAR_DELALLOC |
                              EXTENT_CLEAR_ACCOUNTING |
                              EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK);
-                        ret = 0;
+
+                        btrfs_end_transaction(trans, root);
                         goto free_pages_out;
                 }
+                btrfs_end_transaction(trans, root);
         }
 
         if (will_compress) {
@@ -472,7 +485,10 @@ again:
                 nr_pages_ret = 0;
 
                 /* flag the file so we don't compress in the future */
-                BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
+                if (!btrfs_test_opt(root, FORCE_COMPRESS) &&
+                    !(BTRFS_I(inode)->force_compress)) {
+                        BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
+                }
         }
         if (will_compress) {
                 *num_added += 1;
@@ -543,7 +559,6 @@ static noinline int submit_compressed_extents(struct inode *inode,
         if (list_empty(&async_cow->extents))
                 return 0;
 
-        trans = btrfs_join_transaction(root, 1);
 
         while (!list_empty(&async_cow->extents)) {
                 async_extent = list_entry(async_cow->extents.next,
@@ -559,8 +574,8 @@ retry:
                         unsigned long nr_written = 0;
 
                         lock_extent(io_tree, async_extent->start,
-                                    async_extent->start +
-                                    async_extent->ram_size - 1, GFP_NOFS);
+                                         async_extent->start +
+                                         async_extent->ram_size - 1, GFP_NOFS);
 
                         /* allocate blocks */
                         ret = cow_file_range(inode, async_cow->locked_page,
@@ -590,19 +605,15 @@ retry:
                 lock_extent(io_tree, async_extent->start,
                             async_extent->start + async_extent->ram_size - 1,
                             GFP_NOFS);
-                /*
-                 * here we're doing allocation and writeback of the
-                 * compressed pages
-                 */
-                btrfs_drop_extent_cache(inode, async_extent->start,
-                                        async_extent->start +
-                                        async_extent->ram_size - 1, 0);
 
+                trans = btrfs_join_transaction(root, 1);
                 ret = btrfs_reserve_extent(trans, root,
                                            async_extent->compressed_size,
                                            async_extent->compressed_size,
                                            0, alloc_hint,
                                            (u64)-1, &ins, 1);
+                btrfs_end_transaction(trans, root);
+
                 if (ret) {
                         int i;
                         for (i = 0; i < async_extent->nr_pages; i++) {
@@ -618,6 +629,14 @@ retry:
                         goto retry;
                 }
 
+                /*
+                 * here we're doing allocation and writeback of the
+                 * compressed pages
+                 */
+                btrfs_drop_extent_cache(inode, async_extent->start,
+                                        async_extent->start +
+                                        async_extent->ram_size - 1, 0);
+
                 em = alloc_extent_map(GFP_NOFS);
                 em->start = async_extent->start;
                 em->len = async_extent->ram_size;
@@ -649,8 +668,6 @@ retry:
                                                BTRFS_ORDERED_COMPRESSED);
                 BUG_ON(ret);
 
-                btrfs_end_transaction(trans, root);
-
                 /*
                  * clear dirty, set writeback and unlock the pages.
                  */
@@ -672,13 +689,11 @@ retry:
                                     async_extent->nr_pages);
 
                 BUG_ON(ret);
-                trans = btrfs_join_transaction(root, 1);
                 alloc_hint = ins.objectid + ins.offset;
                 kfree(async_extent);
                 cond_resched();
         }
 
-        btrfs_end_transaction(trans, root);
         return 0;
 }
 
@@ -742,6 +757,7 @@ static noinline int cow_file_range(struct inode *inode,
                                      EXTENT_CLEAR_DIRTY |
                                      EXTENT_SET_WRITEBACK |
                                      EXTENT_END_WRITEBACK);
+
                         *nr_written = *nr_written +
                              (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE;
                         *page_started = 1;
@@ -781,7 +797,7 @@ static noinline int cow_file_range(struct inode *inode,
         while (disk_num_bytes > 0) {
                 unsigned long op;
 
-                cur_alloc_size = min(disk_num_bytes, root->fs_info->max_extent);
+                cur_alloc_size = disk_num_bytes;
                 ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
                                            root->sectorsize, 0, alloc_hint,
                                            (u64)-1, &ins, 1);
@@ -1199,7 +1215,8 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page,
         else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC)
                 ret = run_delalloc_nocow(inode, locked_page, start, end,
                                          page_started, 0, nr_written);
-        else if (!btrfs_test_opt(root, COMPRESS))
+        else if (!btrfs_test_opt(root, COMPRESS) &&
+                 !(BTRFS_I(inode)->force_compress))
                 ret = cow_file_range(inode, locked_page, start, end,
                                       page_started, nr_written, 1);
         else
@@ -1211,30 +1228,9 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page,
 static int btrfs_split_extent_hook(struct inode *inode,
                                     struct extent_state *orig, u64 split)
 {
-        struct btrfs_root *root = BTRFS_I(inode)->root;
-        u64 size;
-
         if (!(orig->state & EXTENT_DELALLOC))
                 return 0;
 
-        size = orig->end - orig->start + 1;
-        if (size > root->fs_info->max_extent) {
-                u64 num_extents;
-                u64 new_size;
-
-                new_size = orig->end - split + 1;
-                num_extents = div64_u64(size + root->fs_info->max_extent - 1,
-                                        root->fs_info->max_extent);
-
-                /*
-                 * if we break a large extent up then leave oustanding_extents
-                 * be, since we've already accounted for the large extent.
-                 */
-                if (div64_u64(new_size + root->fs_info->max_extent - 1,
-                              root->fs_info->max_extent) < num_extents)
-                        return 0;
-        }
-
         spin_lock(&BTRFS_I(inode)->accounting_lock);
         BTRFS_I(inode)->outstanding_extents++;
         spin_unlock(&BTRFS_I(inode)->accounting_lock);
@@ -1252,38 +1248,10 @@ static int btrfs_merge_extent_hook(struct inode *inode,
                                    struct extent_state *new,
                                    struct extent_state *other)
 {
-        struct btrfs_root *root = BTRFS_I(inode)->root;
-        u64 new_size, old_size;
-        u64 num_extents;
-
         /* not delalloc, ignore it */
         if (!(other->state & EXTENT_DELALLOC))
                 return 0;
 
-        old_size = other->end - other->start + 1;
-        if (new->start < other->start)
-                new_size = other->end - new->start + 1;
-        else
-                new_size = new->end - other->start + 1;
-
-        /* we're not bigger than the max, unreserve the space and go */
-        if (new_size <= root->fs_info->max_extent) {
-                spin_lock(&BTRFS_I(inode)->accounting_lock);
-                BTRFS_I(inode)->outstanding_extents--;
-                spin_unlock(&BTRFS_I(inode)->accounting_lock);
-                return 0;
-        }
-
-        /*
-         * If we grew by another max_extent, just return, we want to keep that
-         * reserved amount.
-         */
-        num_extents = div64_u64(old_size + root->fs_info->max_extent - 1,
-                                root->fs_info->max_extent);
-        if (div64_u64(new_size + root->fs_info->max_extent - 1,
-                      root->fs_info->max_extent) > num_extents)
-                return 0;
-
         spin_lock(&BTRFS_I(inode)->accounting_lock);
         BTRFS_I(inode)->outstanding_extents--;
         spin_unlock(&BTRFS_I(inode)->accounting_lock);
@@ -1312,6 +1280,7 @@ static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
                 BTRFS_I(inode)->outstanding_extents++;
                 spin_unlock(&BTRFS_I(inode)->accounting_lock);
                 btrfs_delalloc_reserve_space(root, inode, end - start + 1);
+
                 spin_lock(&root->fs_info->delalloc_lock);
                 BTRFS_I(inode)->delalloc_bytes += end - start + 1;
                 root->fs_info->delalloc_bytes += end - start + 1;
@@ -1340,6 +1309,7 @@ static int btrfs_clear_bit_hook(struct inode *inode,
 
                 if (bits & EXTENT_DO_ACCOUNTING) {
                         spin_lock(&BTRFS_I(inode)->accounting_lock);
+                        WARN_ON(!BTRFS_I(inode)->outstanding_extents);
                         BTRFS_I(inode)->outstanding_extents--;
                         spin_unlock(&BTRFS_I(inode)->accounting_lock);
                         btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
@@ -1496,12 +1466,13 @@ static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
         return 0;
 }
 
-int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end)
+int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
+                              struct extent_state **cached_state)
 {
         if ((end & (PAGE_CACHE_SIZE - 1)) == 0)
                 WARN_ON(1);
         return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
-                                   GFP_NOFS);
+                                   cached_state, GFP_NOFS);
 }
 
 /* see btrfs_writepage_start_hook for details on why this is required */
@@ -1514,6 +1485,7 @@ static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
 {
         struct btrfs_writepage_fixup *fixup;
         struct btrfs_ordered_extent *ordered;
+        struct extent_state *cached_state = NULL;
         struct page *page;
         struct inode *inode;
         u64 page_start;
@@ -1532,7 +1504,8 @@ again:
         page_start = page_offset(page);
         page_end = page_offset(page) + PAGE_CACHE_SIZE - 1;
 
-        lock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+        lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, 0,
+                         &cached_state, GFP_NOFS);
 
         /* already ordered? We're done */
         if (PagePrivate2(page))
@@ -1540,17 +1513,18 @@ again:
 
         ordered = btrfs_lookup_ordered_extent(inode, page_start);
         if (ordered) {
-                unlock_extent(&BTRFS_I(inode)->io_tree, page_start,
-                              page_end, GFP_NOFS);
+                unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start,
+                                     page_end, &cached_state, GFP_NOFS);
                 unlock_page(page);
                 btrfs_start_ordered_extent(inode, ordered, 1);
                 goto again;
         }
 
-        btrfs_set_extent_delalloc(inode, page_start, page_end);
+        btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state);
         ClearPageChecked(page);
 out:
-        unlock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+        unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end,
+                             &cached_state, GFP_NOFS);
 out_page:
         unlock_page(page);
         page_cache_release(page);
@@ -1596,7 +1570,6 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
                                        struct inode *inode, u64 file_pos,
                                        u64 disk_bytenr, u64 disk_num_bytes,
                                        u64 num_bytes, u64 ram_bytes,
-                                       u64 locked_end,
                                        u8 compression, u8 encryption,
                                        u16 other_encoding, int extent_type)
 {
@@ -1622,9 +1595,8 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
          * the caller is expected to unpin it and allow it to be merged
          * with the others.
          */
-        ret = btrfs_drop_extents(trans, root, inode, file_pos,
-                                 file_pos + num_bytes, locked_end,
-                                 file_pos, &hint, 0);
+        ret = btrfs_drop_extents(trans, inode, file_pos, file_pos + num_bytes,
+                                 &hint, 0);
         BUG_ON(ret);
 
         ins.objectid = inode->i_ino;
@@ -1671,24 +1643,6 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
  * before we start the transaction.  It limits the amount of btree
  * reads required while inside the transaction.
  */
-static noinline void reada_csum(struct btrfs_root *root,
-                                struct btrfs_path *path,
-                                struct btrfs_ordered_extent *ordered_extent)
-{
-        struct btrfs_ordered_sum *sum;
-        u64 bytenr;
-
-        sum = list_entry(ordered_extent->list.next, struct btrfs_ordered_sum,
-                         list);
-        bytenr = sum->sums[0].bytenr;
-
-        /*
-         * we don't care about the results, the point of this search is
-         * just to get the btree leaves into ram
-         */
-        btrfs_lookup_csum(NULL, root->fs_info->csum_root, path, bytenr, 0);
-}
-
 /* as ordered data IO finishes, this gets called so we can finish
  * an ordered extent if the range of bytes in the file it covers are
  * fully written.
@@ -1699,54 +1653,39 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
         struct btrfs_trans_handle *trans;
         struct btrfs_ordered_extent *ordered_extent = NULL;
         struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
-        struct btrfs_path *path;
+        struct extent_state *cached_state = NULL;
         int compressed = 0;
         int ret;
 
-        ret = btrfs_dec_test_ordered_pending(inode, start, end - start + 1);
+        ret = btrfs_dec_test_ordered_pending(inode, &ordered_extent, start,
+                                             end - start + 1);
         if (!ret)
                 return 0;
+        BUG_ON(!ordered_extent);
 
-        /*
-         * before we join the transaction, try to do some of our IO.
-         * This will limit the amount of IO that we have to do with
-         * the transaction running.  We're unlikely to need to do any
-         * IO if the file extents are new, the disk_i_size checks
-         * covers the most common case.
-         */
-        if (start < BTRFS_I(inode)->disk_i_size) {
-                path = btrfs_alloc_path();
-                if (path) {
-                        ret = btrfs_lookup_file_extent(NULL, root, path,
-                                                       inode->i_ino,
-                                                       start, 0);
-                        ordered_extent = btrfs_lookup_ordered_extent(inode,
-                                                                     start);
-                        if (!list_empty(&ordered_extent->list)) {
-                                btrfs_release_path(root, path);
-                                reada_csum(root, path, ordered_extent);
-                        }
-                        btrfs_free_path(path);
+        if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
+                BUG_ON(!list_empty(&ordered_extent->list));
+                ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
+                if (!ret) {
+                        trans = btrfs_join_transaction(root, 1);
+                        ret = btrfs_update_inode(trans, root, inode);
+                        BUG_ON(ret);
+                        btrfs_end_transaction(trans, root);
                 }
+                goto out;
         }
 
-        trans = btrfs_join_transaction(root, 1);
-
-        if (!ordered_extent)
-                ordered_extent = btrfs_lookup_ordered_extent(inode, start);
-        BUG_ON(!ordered_extent);
-        if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags))
-                goto nocow;
+        lock_extent_bits(io_tree, ordered_extent->file_offset,
+                         ordered_extent->file_offset + ordered_extent->len - 1,
+                         0, &cached_state, GFP_NOFS);
 
-        lock_extent(io_tree, ordered_extent->file_offset,
-                    ordered_extent->file_offset + ordered_extent->len - 1,
-                    GFP_NOFS);
+        trans = btrfs_join_transaction(root, 1);
 
         if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
                 compressed = 1;
         if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) {
                 BUG_ON(compressed);
-                ret = btrfs_mark_extent_written(trans, root, inode,
+                ret = btrfs_mark_extent_written(trans, inode,
                                                 ordered_extent->file_offset,
                                                 ordered_extent->file_offset +
                                                 ordered_extent->len);
@@ -1758,8 +1697,6 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
                                                 ordered_extent->disk_len,
                                                 ordered_extent->len,
                                                 ordered_extent->len,
-                                                ordered_extent->file_offset +
-                                                ordered_extent->len,
                                                 compressed, 0, 0,
                                                 BTRFS_FILE_EXTENT_REG);
                 unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
@@ -1767,25 +1704,24 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
                                    ordered_extent->len);
                 BUG_ON(ret);
         }
-        unlock_extent(io_tree, ordered_extent->file_offset,
-                    ordered_extent->file_offset + ordered_extent->len - 1,
-                    GFP_NOFS);
-nocow:
+        unlock_extent_cached(io_tree, ordered_extent->file_offset,
+                             ordered_extent->file_offset +
+                             ordered_extent->len - 1, &cached_state, GFP_NOFS);
+
         add_pending_csums(trans, inode, ordered_extent->file_offset,
                           &ordered_extent->list);
 
-        mutex_lock(&BTRFS_I(inode)->extent_mutex);
-        btrfs_ordered_update_i_size(inode, ordered_extent);
-        btrfs_update_inode(trans, root, inode);
-        btrfs_remove_ordered_extent(inode, ordered_extent);
-        mutex_unlock(&BTRFS_I(inode)->extent_mutex);
-
+        /* this also removes the ordered extent from the tree */
+        btrfs_ordered_update_i_size(inode, 0, ordered_extent);
+        ret = btrfs_update_inode(trans, root, inode);
+        BUG_ON(ret);
+        btrfs_end_transaction(trans, root);
+out:
         /* once for us */
         btrfs_put_ordered_extent(ordered_extent);
         /* once for the tree */
         btrfs_put_ordered_extent(ordered_extent);
 
-        btrfs_end_transaction(trans, root);
         return 0;
 }
 
@@ -2008,6 +1944,54 @@ zeroit:
         return -EIO;
 }
 
+struct delayed_iput {
+        struct list_head list;
+        struct inode *inode;
+};
+
+void btrfs_add_delayed_iput(struct inode *inode)
+{
+        struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
+        struct delayed_iput *delayed;
+
+        if (atomic_add_unless(&inode->i_count, -1, 1))
+                return;
+
+        delayed = kmalloc(sizeof(*delayed), GFP_NOFS | __GFP_NOFAIL);
+        delayed->inode = inode;
+
+        spin_lock(&fs_info->delayed_iput_lock);
+        list_add_tail(&delayed->list, &fs_info->delayed_iputs);
+        spin_unlock(&fs_info->delayed_iput_lock);
+}
+
+void btrfs_run_delayed_iputs(struct btrfs_root *root)
+{
+        LIST_HEAD(list);
+        struct btrfs_fs_info *fs_info = root->fs_info;
+        struct delayed_iput *delayed;
+        int empty;
+
+        spin_lock(&fs_info->delayed_iput_lock);
+        empty = list_empty(&fs_info->delayed_iputs);
+        spin_unlock(&fs_info->delayed_iput_lock);
+        if (empty)
+                return;
+
+        down_read(&root->fs_info->cleanup_work_sem);
+        spin_lock(&fs_info->delayed_iput_lock);
+        list_splice_init(&fs_info->delayed_iputs, &list);
+        spin_unlock(&fs_info->delayed_iput_lock);
+
+        while (!list_empty(&list)) {
+                delayed = list_entry(list.next, struct delayed_iput, list);
+                list_del(&delayed->list);
+                iput(delayed->inode);
+                kfree(delayed);
+        }
+        up_read(&root->fs_info->cleanup_work_sem);
+}
+
 /*
  * This creates an orphan entry for the given inode in case something goes
  * wrong in the middle of an unlink/truncate.
@@ -2080,16 +2064,17 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
         struct inode *inode;
         int ret = 0, nr_unlink = 0, nr_truncate = 0;
 
-        path = btrfs_alloc_path();
-        if (!path)
+        if (!xchg(&root->clean_orphans, 0))
                 return;
+
+        path = btrfs_alloc_path();
+        BUG_ON(!path);
         path->reada = -1;
 
         key.objectid = BTRFS_ORPHAN_OBJECTID;
         btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
         key.offset = (u64)-1;
 
-
         while (1) {
                 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
                 if (ret < 0) {
@@ -2131,7 +2116,7 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
                 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);
+                inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL);
                 if (IS_ERR(inode))
                         break;
 
@@ -2834,37 +2819,40 @@ out:
  * min_type is the minimum key type to truncate down to.  If set to 0, this
  * will kill all the items on this inode, including the INODE_ITEM_KEY.
  */
-noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
-                                        struct btrfs_root *root,
-                                        struct inode *inode,
-                                        u64 new_size, u32 min_type)
+int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
+                               struct btrfs_root *root,
+                               struct inode *inode,
+                               u64 new_size, u32 min_type)
 {
-        int ret;
         struct btrfs_path *path;
-        struct btrfs_key key;
-        struct btrfs_key found_key;
-        u32 found_type = (u8)-1;
         struct extent_buffer *leaf;
         struct btrfs_file_extent_item *fi;
+        struct btrfs_key key;
+        struct btrfs_key found_key;
         u64 extent_start = 0;
         u64 extent_num_bytes = 0;
         u64 extent_offset = 0;
         u64 item_end = 0;
+        u64 mask = root->sectorsize - 1;
+        u32 found_type = (u8)-1;
         int found_extent;
         int del_item;
         int pending_del_nr = 0;
         int pending_del_slot = 0;
         int extent_type = -1;
         int encoding;
-        u64 mask = root->sectorsize - 1;
+        int ret;
+        int err = 0;
+
+        BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY);
 
         if (root->ref_cows)
                 btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0);
+
         path = btrfs_alloc_path();
         BUG_ON(!path);
         path->reada = -1;
 
-        /* FIXME, add redo link to tree so we don't leak on crash */
         key.objectid = inode->i_ino;
         key.offset = (u64)-1;
         key.type = (u8)-1;
@@ -2872,17 +2860,17 @@ noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 search_again:
         path->leave_spinning = 1;
         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
-        if (ret < 0)
-                goto error;
+        if (ret < 0) {
+                err = ret;
+                goto out;
+        }
 
         if (ret > 0) {
                 /* there are no items in the tree for us to truncate, we're
                  * done
                  */
-                if (path->slots[0] == 0) {
-                        ret = 0;
-                        goto error;
-                }
+                if (path->slots[0] == 0)
+                        goto out;
                 path->slots[0]--;
         }
 
@@ -2917,28 +2905,17 @@ search_again:
                         }
                         item_end--;
                 }
-                if (item_end < new_size) {
-                        if (found_type == BTRFS_DIR_ITEM_KEY)
-                                found_type = BTRFS_INODE_ITEM_KEY;
-                        else if (found_type == BTRFS_EXTENT_ITEM_KEY)
-                                found_type = BTRFS_EXTENT_DATA_KEY;
-                        else if (found_type == BTRFS_EXTENT_DATA_KEY)
-                                found_type = BTRFS_XATTR_ITEM_KEY;
-                        else if (found_type == BTRFS_XATTR_ITEM_KEY)
-                                found_type = BTRFS_INODE_REF_KEY;
-                        else if (found_type)
-                                found_type--;
-                        else
+                if (found_type > min_type) {
+                        del_item = 1;
+                } else {
+                        if (item_end < new_size)
                                 break;
-                        btrfs_set_key_type(&key, found_type);
-                        goto next;
+                        if (found_key.offset >= new_size)
+                                del_item = 1;
+                        else
+                                del_item = 0;
                 }
-                if (found_key.offset >= new_size)
-                        del_item = 1;
-                else
-                        del_item = 0;
                 found_extent = 0;
-
                 /* FIXME, shrink the extent if the ref count is only 1 */
                 if (found_type != BTRFS_EXTENT_DATA_KEY)
                         goto delete;
@@ -3025,42 +3002,36 @@ delete:
                                                 inode->i_ino, extent_offset);
                         BUG_ON(ret);
                 }
-next:
-                if (path->slots[0] == 0) {
-                        if (pending_del_nr)
-                                goto del_pending;
-                        btrfs_release_path(root, path);
-                        if (found_type == BTRFS_INODE_ITEM_KEY)
-                                break;
-                        goto search_again;
-                }
 
-                path->slots[0]--;
-                if (pending_del_nr &&
-                    path->slots[0] + 1 != pending_del_slot) {
-                        struct btrfs_key debug;
-del_pending:
-                        btrfs_item_key_to_cpu(path->nodes[0], &debug,
-                                              pending_del_slot);
-                        ret = btrfs_del_items(trans, root, path,
-                                              pending_del_slot,
-                                              pending_del_nr);
-                        BUG_ON(ret);
-                        pending_del_nr = 0;
+                if (found_type == BTRFS_INODE_ITEM_KEY)
+                        break;
+
+                if (path->slots[0] == 0 ||
+                    path->slots[0] != pending_del_slot) {
+                        if (root->ref_cows) {
+                                err = -EAGAIN;
+                                goto out;
+                        }
+                        if (pending_del_nr) {
+                                ret = btrfs_del_items(trans, root, path,
+                                                pending_del_slot,
+                                                pending_del_nr);
+                                BUG_ON(ret);
+                                pending_del_nr = 0;
+                        }
                         btrfs_release_path(root, path);
-                        if (found_type == BTRFS_INODE_ITEM_KEY)
-                                break;
                         goto search_again;
+                } else {
+                        path->slots[0]--;
                 }
         }
-        ret = 0;
-error:
+out:
         if (pending_del_nr) {
                 ret = btrfs_del_items(trans, root, path, pending_del_slot,
                                       pending_del_nr);
         }
         btrfs_free_path(path);
-        return ret;
+        return err;
 }
 
 /*
@@ -3073,6 +3044,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
         struct btrfs_root *root = BTRFS_I(inode)->root;
         struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
         struct btrfs_ordered_extent *ordered;
+        struct extent_state *cached_state = NULL;
         char *kaddr;
         u32 blocksize = root->sectorsize;
         pgoff_t index = from >> PAGE_CACHE_SHIFT;
@@ -3119,12 +3091,14 @@ again:
         }
         wait_on_page_writeback(page);
 
-        lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+        lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state,
+                         GFP_NOFS);
         set_page_extent_mapped(page);
 
         ordered = btrfs_lookup_ordered_extent(inode, page_start);
         if (ordered) {
-                unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+                unlock_extent_cached(io_tree, page_start, page_end,
+                                     &cached_state, GFP_NOFS);
                 unlock_page(page);
                 page_cache_release(page);
                 btrfs_start_ordered_extent(inode, ordered, 1);
@@ -3132,13 +3106,15 @@ again:
                 goto again;
         }
 
-        clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
+        clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
                           EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING,
-                          GFP_NOFS);
+                          0, 0, &cached_state, GFP_NOFS);
 
-        ret = btrfs_set_extent_delalloc(inode, page_start, page_end);
+        ret = btrfs_set_extent_delalloc(inode, page_start, page_end,
+                                        &cached_state);
         if (ret) {
-                unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+                unlock_extent_cached(io_tree, page_start, page_end,
+                                     &cached_state, GFP_NOFS);
                 goto out_unlock;
         }
 
@@ -3151,7 +3127,8 @@ again:
         }
         ClearPageChecked(page);
         set_page_dirty(page);
-        unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+        unlock_extent_cached(io_tree, page_start, page_end, &cached_state,
+                             GFP_NOFS);
 
 out_unlock:
         if (ret)
@@ -3169,6 +3146,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
         struct btrfs_root *root = BTRFS_I(inode)->root;
         struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
         struct extent_map *em;
+        struct extent_state *cached_state = NULL;
         u64 mask = root->sectorsize - 1;
         u64 hole_start = (inode->i_size + mask) & ~mask;
         u64 block_end = (size + mask) & ~mask;
@@ -3180,25 +3158,20 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
         if (size <= hole_start)
                 return 0;
 
-        err = btrfs_truncate_page(inode->i_mapping, inode->i_size);
-        if (err)
-                return err;
-
         while (1) {
                 struct btrfs_ordered_extent *ordered;
                 btrfs_wait_ordered_range(inode, hole_start,
                                          block_end - hole_start);
-                lock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+                lock_extent_bits(io_tree, hole_start, block_end - 1, 0,
+                                 &cached_state, GFP_NOFS);
                 ordered = btrfs_lookup_ordered_extent(inode, hole_start);
                 if (!ordered)
                         break;
-                unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+                unlock_extent_cached(io_tree, hole_start, block_end - 1,
+                                     &cached_state, GFP_NOFS);
                 btrfs_put_ordered_extent(ordered);
         }
 
-        trans = btrfs_start_transaction(root, 1);
-        btrfs_set_trans_block_group(trans, inode);
-
         cur_offset = hole_start;
         while (1) {
                 em = btrfs_get_extent(inode, NULL, 0, cur_offset,
@@ -3206,40 +3179,121 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
                 BUG_ON(IS_ERR(em) || !em);
                 last_byte = min(extent_map_end(em), block_end);
                 last_byte = (last_byte + mask) & ~mask;
-                if (test_bit(EXTENT_FLAG_VACANCY, &em->flags)) {
+                if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) {
                         u64 hint_byte = 0;
                         hole_size = last_byte - cur_offset;
-                        err = btrfs_drop_extents(trans, root, inode,
-                                                 cur_offset,
-                                                 cur_offset + hole_size,
-                                                 block_end,
-                                                 cur_offset, &hint_byte, 1);
-                        if (err)
-                                break;
 
-                        err = btrfs_reserve_metadata_space(root, 1);
+                        err = btrfs_reserve_metadata_space(root, 2);
                         if (err)
                                 break;
 
+                        trans = btrfs_start_transaction(root, 1);
+                        btrfs_set_trans_block_group(trans, inode);
+
+                        err = btrfs_drop_extents(trans, inode, cur_offset,
+                                                 cur_offset + hole_size,
+                                                 &hint_byte, 1);
+                        BUG_ON(err);
+
                         err = btrfs_insert_file_extent(trans, root,
                                         inode->i_ino, cur_offset, 0,
                                         0, hole_size, 0, hole_size,
                                         0, 0, 0);
+                        BUG_ON(err);
+
                         btrfs_drop_extent_cache(inode, hole_start,
                                         last_byte - 1, 0);
-                        btrfs_unreserve_metadata_space(root, 1);
+
+                        btrfs_end_transaction(trans, root);
+                        btrfs_unreserve_metadata_space(root, 2);
                 }
                 free_extent_map(em);
                 cur_offset = last_byte;
-                if (err || cur_offset >= block_end)
+                if (cur_offset >= block_end)
                         break;
         }
 
-        btrfs_end_transaction(trans, root);
-        unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+        unlock_extent_cached(io_tree, hole_start, block_end - 1, &cached_state,
+                             GFP_NOFS);
         return err;
 }
 
+static int btrfs_setattr_size(struct inode *inode, struct iattr *attr)
+{
+        struct btrfs_root *root = BTRFS_I(inode)->root;
+        struct btrfs_trans_handle *trans;
+        unsigned long nr;
+        int ret;
+
+        if (attr->ia_size == inode->i_size)
+                return 0;
+
+        if (attr->ia_size > inode->i_size) {
+                unsigned long limit;
+                limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;
+                if (attr->ia_size > inode->i_sb->s_maxbytes)
+                        return -EFBIG;
+                if (limit != RLIM_INFINITY && attr->ia_size > limit) {
+                        send_sig(SIGXFSZ, current, 0);
+                        return -EFBIG;
+                }
+        }
+
+        ret = btrfs_reserve_metadata_space(root, 1);
+        if (ret)
+                return ret;
+
+        trans = btrfs_start_transaction(root, 1);
+        btrfs_set_trans_block_group(trans, inode);
+
+        ret = btrfs_orphan_add(trans, inode);
+        BUG_ON(ret);
+
+        nr = trans->blocks_used;
+        btrfs_end_transaction(trans, root);
+        btrfs_unreserve_metadata_space(root, 1);
+        btrfs_btree_balance_dirty(root, nr);
+
+        if (attr->ia_size > inode->i_size) {
+                ret = btrfs_cont_expand(inode, attr->ia_size);
+                if (ret) {
+                        btrfs_truncate(inode);
+                        return ret;
+                }
+
+                i_size_write(inode, attr->ia_size);
+                btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
+
+                trans = btrfs_start_transaction(root, 1);
+                btrfs_set_trans_block_group(trans, inode);
+
+                ret = btrfs_update_inode(trans, root, inode);
+                BUG_ON(ret);
+                if (inode->i_nlink > 0) {
+                        ret = btrfs_orphan_del(trans, inode);
+                        BUG_ON(ret);
+                }
+                nr = trans->blocks_used;
+                btrfs_end_transaction(trans, root);
+                btrfs_btree_balance_dirty(root, nr);
+                return 0;
+        }
+
+        /*
+         * We're truncating a file that used to have good data down to
+         * zero. Make sure it gets into the ordered flush list so that
+         * any new writes get down to disk quickly.
+         */
+        if (attr->ia_size == 0)
+                BTRFS_I(inode)->ordered_data_close = 1;
+
+        /* we don't support swapfiles, so vmtruncate shouldn't fail */
+        ret = vmtruncate(inode, attr->ia_size);
+        BUG_ON(ret);
+
+        return 0;
+}
+
 static int btrfs_setattr(struct dentry *dentry, struct iattr *attr)
 {
         struct inode *inode = dentry->d_inode;
@@ -3250,23 +3304,14 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr)
                 return err;
 
         if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
-                if (attr->ia_size > inode->i_size) {
-                        err = btrfs_cont_expand(inode, attr->ia_size);
-                        if (err)
-                                return err;
-                } else if (inode->i_size > 0 &&
-                           attr->ia_size == 0) {
-
-                        /* we're truncating a file that used to have good
-                         * data down to zero.  Make sure it gets into
-                         * the ordered flush list so that any new writes
-                         * get down to disk quickly.
-                         */
-                        BTRFS_I(inode)->ordered_data_close = 1;
-                }
+                err = btrfs_setattr_size(inode, attr);
+                if (err)
+                        return err;
         }
+        attr->ia_valid &= ~ATTR_SIZE;
 
-        err = inode_setattr(inode, attr);
+        if (attr->ia_valid)
+                err = inode_setattr(inode, attr);
 
         if (!err && ((attr->ia_valid & ATTR_MODE)))
                 err = btrfs_acl_chmod(inode);
@@ -3287,36 +3332,43 @@ void btrfs_delete_inode(struct inode *inode)
         }
         btrfs_wait_ordered_range(inode, 0, (u64)-1);
 
+        if (root->fs_info->log_root_recovering) {
+                BUG_ON(!list_empty(&BTRFS_I(inode)->i_orphan));
+                goto no_delete;
+        }
+
         if (inode->i_nlink > 0) {
                 BUG_ON(btrfs_root_refs(&root->root_item) != 0);
                 goto no_delete;
         }
 
         btrfs_i_size_write(inode, 0);
-        trans = btrfs_join_transaction(root, 1);
 
-        btrfs_set_trans_block_group(trans, inode);
-        ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, 0);
-        if (ret) {
-                btrfs_orphan_del(NULL, inode);
-                goto no_delete_lock;
-        }
+        while (1) {
+                trans = btrfs_start_transaction(root, 1);
+                btrfs_set_trans_block_group(trans, inode);
+                ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
 
-        btrfs_orphan_del(trans, inode);
+                if (ret != -EAGAIN)
+                        break;
 
-        nr = trans->blocks_used;
-        clear_inode(inode);
+                nr = trans->blocks_used;
+                btrfs_end_transaction(trans, root);
+                trans = NULL;
+                btrfs_btree_balance_dirty(root, nr);
+        }
 
-        btrfs_end_transaction(trans, root);
-        btrfs_btree_balance_dirty(root, nr);
-        return;
+        if (ret == 0) {
+                ret = btrfs_orphan_del(trans, inode);
+                BUG_ON(ret);
+        }
 
-no_delete_lock:
         nr = trans->blocks_used;
         btrfs_end_transaction(trans, root);
         btrfs_btree_balance_dirty(root, nr);
 no_delete:
         clear_inode(inode);
+        return;
 }
 
 /*
@@ -3560,6 +3612,7 @@ static noinline void init_btrfs_i(struct inode *inode)
         bi->index_cnt = (u64)-1;
         bi->last_unlink_trans = 0;
         bi->ordered_data_close = 0;
+        bi->force_compress = 0;
         extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
         extent_io_tree_init(&BTRFS_I(inode)->io_tree,
                              inode->i_mapping, GFP_NOFS);
@@ -3569,7 +3622,6 @@ static noinline void init_btrfs_i(struct inode *inode)
         INIT_LIST_HEAD(&BTRFS_I(inode)->ordered_operations);
         RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
         btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
-        mutex_init(&BTRFS_I(inode)->extent_mutex);
         mutex_init(&BTRFS_I(inode)->log_mutex);
 }
 
@@ -3609,7 +3661,7 @@ static struct inode *btrfs_iget_locked(struct super_block *s,
  * Returns in *is_new if the inode was read from disk
  */
 struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
-                         struct btrfs_root *root)
+                         struct btrfs_root *root, int *new)
 {
         struct inode *inode;
 
@@ -3624,6 +3676,8 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
 
                 inode_tree_add(inode);
                 unlock_new_inode(inode);
+                if (new)
+                        *new = 1;
         }
 
         return inode;
@@ -3676,7 +3730,7 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
                 return NULL;
 
         if (location.type == BTRFS_INODE_ITEM_KEY) {
-                inode = btrfs_iget(dir->i_sb, &location, root);
+                inode = btrfs_iget(dir->i_sb, &location, root, NULL);
                 return inode;
         }
 
@@ -3691,10 +3745,17 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
                 else
                         inode = new_simple_dir(dir->i_sb, &location, sub_root);
         } else {
-                inode = btrfs_iget(dir->i_sb, &location, sub_root);
+                inode = btrfs_iget(dir->i_sb, &location, sub_root, NULL);
         }
         srcu_read_unlock(&root->fs_info->subvol_srcu, index);
 
+        if (root != sub_root) {
+                down_read(&root->fs_info->cleanup_work_sem);
+                if (!(inode->i_sb->s_flags & MS_RDONLY))
+                        btrfs_orphan_cleanup(sub_root);
+                up_read(&root->fs_info->cleanup_work_sem);
+        }
+
         return inode;
 }
 
@@ -3869,7 +3930,11 @@ skip:
 
         /* Reached end of directory/root. Bump pos past the last item. */
         if (key_type == BTRFS_DIR_INDEX_KEY)
-                filp->f_pos = INT_LIMIT(off_t);
+                /*
+                 * 32-bit glibc will use getdents64, but then strtol -
+                 * so the last number we can serve is this.
+                 */
+                filp->f_pos = 0x7fffffff;
         else
                 filp->f_pos++;
 nopos:
@@ -3879,7 +3944,7 @@ err:
         return ret;
 }
 
-int btrfs_write_inode(struct inode *inode, int wait)
+int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc)
 {
         struct btrfs_root *root = BTRFS_I(inode)->root;
         struct btrfs_trans_handle *trans;
@@ -3888,7 +3953,7 @@ int btrfs_write_inode(struct inode *inode, int wait)
         if (root->fs_info->btree_inode == inode)
                 return 0;
 
-        if (wait) {
+        if (wbc->sync_mode == WB_SYNC_ALL) {
                 trans = btrfs_join_transaction(root, 1);
                 btrfs_set_trans_block_group(trans, inode);
                 ret = btrfs_commit_transaction(trans, root);
@@ -4219,7 +4284,7 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
         if (IS_ERR(inode))
                 goto out_unlock;
 
-        err = btrfs_init_inode_security(inode, dir);
+        err = btrfs_init_inode_security(trans, inode, dir);
         if (err) {
                 drop_inode = 1;
                 goto out_unlock;
@@ -4290,7 +4355,7 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
         if (IS_ERR(inode))
                 goto out_unlock;
 
-        err = btrfs_init_inode_security(inode, dir);
+        err = btrfs_init_inode_security(trans, inode, dir);
         if (err) {
                 drop_inode = 1;
                 goto out_unlock;
@@ -4336,6 +4401,10 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
         if (inode->i_nlink == 0)
                 return -ENOENT;
 
+        /* do not allow sys_link's with other subvols of the same device */
+        if (root->objectid != BTRFS_I(inode)->root->objectid)
+                return -EPERM;
+
         /*
          * 1 item for inode ref
          * 2 items for dir items
@@ -4408,7 +4477,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
         err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
         if (err) {
                 err = -ENOSPC;
-                goto out_unlock;
+                goto out_fail;
         }
 
         inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
@@ -4423,7 +4492,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 
         drop_on_err = 1;
 
-        err = btrfs_init_inode_security(inode, dir);
+        err = btrfs_init_inode_security(trans, inode, dir);
         if (err)
                 goto out_fail;
 
@@ -4886,6 +4955,7 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
 {
         struct extent_io_tree *tree;
         struct btrfs_ordered_extent *ordered;
+        struct extent_state *cached_state = NULL;
         u64 page_start = page_offset(page);
         u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
 
@@ -4904,7 +4974,8 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
                 btrfs_releasepage(page, GFP_NOFS);
                 return;
         }
-        lock_extent(tree, page_start, page_end, GFP_NOFS);
+        lock_extent_bits(tree, page_start, page_end, 0, &cached_state,
+                         GFP_NOFS);
         ordered = btrfs_lookup_ordered_extent(page->mapping->host,
                                            page_offset(page));
         if (ordered) {
@@ -4915,7 +4986,7 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
                 clear_extent_bit(tree, page_start, page_end,
                                  EXTENT_DIRTY | EXTENT_DELALLOC |
                                  EXTENT_LOCKED | EXTENT_DO_ACCOUNTING, 1, 0,
-                                 NULL, GFP_NOFS);
+                                 &cached_state, GFP_NOFS);
                 /*
                  * whoever cleared the private bit is responsible
                  * for the finish_ordered_io
@@ -4925,11 +4996,13 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
                                                 page_start, page_end);
                 }
                 btrfs_put_ordered_extent(ordered);
-                lock_extent(tree, page_start, page_end, GFP_NOFS);
+                cached_state = NULL;
+                lock_extent_bits(tree, page_start, page_end, 0, &cached_state,
+                                 GFP_NOFS);
         }
         clear_extent_bit(tree, page_start, page_end,
                  EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
-                 EXTENT_DO_ACCOUNTING, 1, 1, NULL, GFP_NOFS);
+                 EXTENT_DO_ACCOUNTING, 1, 1, &cached_state, GFP_NOFS);
         __btrfs_releasepage(page, GFP_NOFS);
 
         ClearPageChecked(page);
@@ -4962,6 +5035,7 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
         struct btrfs_root *root = BTRFS_I(inode)->root;
         struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
         struct btrfs_ordered_extent *ordered;
+        struct extent_state *cached_state = NULL;
         char *kaddr;
         unsigned long zero_start;
         loff_t size;
@@ -5000,7 +5074,8 @@ again:
         }
         wait_on_page_writeback(page);
 
-        lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+        lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state,
+                         GFP_NOFS);
         set_page_extent_mapped(page);
 
         /*
@@ -5009,7 +5084,8 @@ again:
          */
         ordered = btrfs_lookup_ordered_extent(inode, page_start);
         if (ordered) {
-                unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+                unlock_extent_cached(io_tree, page_start, page_end,
+                                     &cached_state, GFP_NOFS);
                 unlock_page(page);
                 btrfs_start_ordered_extent(inode, ordered, 1);
                 btrfs_put_ordered_extent(ordered);
@@ -5023,13 +5099,15 @@ again:
          * is probably a better way to do this, but for now keep consistent with
          * prepare_pages in the normal write path.
          */
-        clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
+        clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
                           EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING,
-                          GFP_NOFS);
+                          0, 0, &cached_state, GFP_NOFS);
 
-        ret = btrfs_set_extent_delalloc(inode, page_start, page_end);
+        ret = btrfs_set_extent_delalloc(inode, page_start, page_end,
+                                        &cached_state);
         if (ret) {
-                unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+                unlock_extent_cached(io_tree, page_start, page_end,
+                                     &cached_state, GFP_NOFS);
                 ret = VM_FAULT_SIGBUS;
                 btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
                 goto out_unlock;
@@ -5055,7 +5133,7 @@ again:
         BTRFS_I(inode)->last_trans = root->fs_info->generation;
         BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid;
 
-        unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+        unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS);
 
 out_unlock:
         btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
@@ -5074,17 +5152,20 @@ static void btrfs_truncate(struct inode *inode)
         unsigned long nr;
         u64 mask = root->sectorsize - 1;
 
-        if (!S_ISREG(inode->i_mode))
-                return;
-        if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+        if (!S_ISREG(inode->i_mode)) {
+                WARN_ON(1);
                 return;
+        }
 
         ret = btrfs_truncate_page(inode->i_mapping, inode->i_size);
         if (ret)
                 return;
+
         btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
+        btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
 
         trans = btrfs_start_transaction(root, 1);
+        btrfs_set_trans_block_group(trans, inode);
 
         /*
          * setattr is responsible for setting the ordered_data_close flag,
@@ -5106,21 +5187,32 @@ static void btrfs_truncate(struct inode *inode)
         if (inode->i_size == 0 && BTRFS_I(inode)->ordered_data_close)
                 btrfs_add_ordered_operation(trans, root, inode);
 
-        btrfs_set_trans_block_group(trans, inode);
-        btrfs_i_size_write(inode, inode->i_size);
+        while (1) {
+                ret = btrfs_truncate_inode_items(trans, root, inode,
+                                                 inode->i_size,
+                                                 BTRFS_EXTENT_DATA_KEY);
+                if (ret != -EAGAIN)
+                        break;
 
-        ret = btrfs_orphan_add(trans, inode);
-        if (ret)
-                goto out;
-        /* FIXME, add redo link to tree so we don't leak on crash */
-        ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size,
-                                      BTRFS_EXTENT_DATA_KEY);
-        btrfs_update_inode(trans, root, inode);
+                ret = btrfs_update_inode(trans, root, inode);
+                BUG_ON(ret);
+
+                nr = trans->blocks_used;
+                btrfs_end_transaction(trans, root);
+                btrfs_btree_balance_dirty(root, nr);
+
+                trans = btrfs_start_transaction(root, 1);
+                btrfs_set_trans_block_group(trans, inode);
+        }
+
+        if (ret == 0 && inode->i_nlink > 0) {
+                ret = btrfs_orphan_del(trans, inode);
+                BUG_ON(ret);
+        }
 
-        ret = btrfs_orphan_del(trans, inode);
+        ret = btrfs_update_inode(trans, root, inode);
         BUG_ON(ret);
 
-out:
         nr = trans->blocks_used;
         ret = btrfs_end_transaction_throttle(trans, root);
         BUG_ON(ret);
@@ -5217,9 +5309,9 @@ void btrfs_destroy_inode(struct inode *inode)
 
         spin_lock(&root->list_lock);
         if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
-                printk(KERN_ERR "BTRFS: inode %lu: inode still on the orphan"
-                       " list\n", inode->i_ino);
-                dump_stack();
+                printk(KERN_INFO "BTRFS: inode %lu still on the orphan list\n",
+                       inode->i_ino);
+                list_del_init(&BTRFS_I(inode)->i_orphan);
         }
         spin_unlock(&root->list_lock);
 
@@ -5246,7 +5338,6 @@ free:
 void btrfs_drop_inode(struct inode *inode)
 {
         struct btrfs_root *root = BTRFS_I(inode)->root;
-
         if (inode->i_nlink > 0 && btrfs_root_refs(&root->root_item) == 0)
                 generic_delete_inode(inode);
         else
@@ -5476,7 +5567,7 @@ out_fail:
  * some fairly slow code that needs optimization. This walks the list
  * of all the inodes with pending delalloc and forces them to disk.
  */
-int btrfs_start_delalloc_inodes(struct btrfs_root *root)
+int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
 {
         struct list_head *head = &root->fs_info->delalloc_inodes;
         struct btrfs_inode *binode;
@@ -5495,7 +5586,10 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root)
                 spin_unlock(&root->fs_info->delalloc_lock);
                 if (inode) {
                         filemap_flush(inode->i_mapping);
-                        iput(inode);
+                        if (delay_iput)
+                                btrfs_add_delayed_iput(inode);
+                        else
+                                iput(inode);
                 }
                 cond_resched();
                 spin_lock(&root->fs_info->delalloc_lock);
@@ -5569,7 +5663,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
         if (IS_ERR(inode))
                 goto out_unlock;
 
-        err = btrfs_init_inode_security(inode, dir);
+        err = btrfs_init_inode_security(trans, inode, dir);
         if (err) {
                 drop_inode = 1;
                 goto out_unlock;
@@ -5641,62 +5735,80 @@ out_fail:
         return err;
 }
 
-static int prealloc_file_range(struct btrfs_trans_handle *trans,
-                               struct inode *inode, u64 start, u64 end,
-                               u64 locked_end, u64 alloc_hint, int mode)
+static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
+                        u64 alloc_hint, int mode, loff_t actual_len)
 {
+        struct btrfs_trans_handle *trans;
         struct btrfs_root *root = BTRFS_I(inode)->root;
         struct btrfs_key ins;
-        u64 alloc_size;
         u64 cur_offset = start;
         u64 num_bytes = end - start;
         int ret = 0;
+        u64 i_size;
 
         while (num_bytes > 0) {
-                alloc_size = min(num_bytes, root->fs_info->max_extent);
+                trans = btrfs_start_transaction(root, 1);
 
-                ret = btrfs_reserve_metadata_space(root, 1);
-                if (ret)
-                        goto out;
-
-                ret = btrfs_reserve_extent(trans, root, alloc_size,
+                ret = btrfs_reserve_extent(trans, root, num_bytes,
                                            root->sectorsize, 0, alloc_hint,
                                            (u64)-1, &ins, 1);
                 if (ret) {
                         WARN_ON(1);
-                        goto out;
+                        goto stop_trans;
                 }
+
+                ret = btrfs_reserve_metadata_space(root, 3);
+                if (ret) {
+                        btrfs_free_reserved_extent(root, ins.objectid,
+                                                   ins.offset);
+                        goto stop_trans;
+                }
+
                 ret = insert_reserved_file_extent(trans, inode,
                                                   cur_offset, ins.objectid,
                                                   ins.offset, ins.offset,
-                                                  ins.offset, locked_end,
-                                                  0, 0, 0,
+                                                  ins.offset, 0, 0, 0,
                                                   BTRFS_FILE_EXTENT_PREALLOC);
                 BUG_ON(ret);
                 btrfs_drop_extent_cache(inode, cur_offset,
                                         cur_offset + ins.offset -1, 0);
+
                 num_bytes -= ins.offset;
                 cur_offset += ins.offset;
                 alloc_hint = ins.objectid + ins.offset;
-                btrfs_unreserve_metadata_space(root, 1);
-        }
-out:
-        if (cur_offset > start) {
+
                 inode->i_ctime = CURRENT_TIME;
                 BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC;
                 if (!(mode & FALLOC_FL_KEEP_SIZE) &&
-                    cur_offset > i_size_read(inode))
-                        btrfs_i_size_write(inode, cur_offset);
+                        (actual_len > inode->i_size) &&
+                        (cur_offset > inode->i_size)) {
+
+                        if (cur_offset > actual_len)
+                                i_size  = actual_len;
+                        else
+                                i_size = cur_offset;
+                        i_size_write(inode, i_size);
+                        btrfs_ordered_update_i_size(inode, i_size, NULL);
+                }
+
                 ret = btrfs_update_inode(trans, root, inode);
                 BUG_ON(ret);
+
+                btrfs_end_transaction(trans, root);
+                btrfs_unreserve_metadata_space(root, 3);
         }
+        return ret;
 
+stop_trans:
+        btrfs_end_transaction(trans, root);
         return ret;
+
 }
 
 static long btrfs_fallocate(struct inode *inode, int mode,
                             loff_t offset, loff_t len)
 {
+        struct extent_state *cached_state = NULL;
         u64 cur_offset;
         u64 last_byte;
         u64 alloc_start;
@@ -5705,8 +5817,6 @@ static long btrfs_fallocate(struct inode *inode, int mode,
         u64 locked_end;
         u64 mask = BTRFS_I(inode)->root->sectorsize - 1;
         struct extent_map *em;
-        struct btrfs_trans_handle *trans;
-        struct btrfs_root *root;
         int ret;
 
         alloc_start = offset & ~mask;
@@ -5725,9 +5835,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
                         goto out;
         }
 
-        root = BTRFS_I(inode)->root;
-
-        ret = btrfs_check_data_free_space(root, inode,
+        ret = btrfs_check_data_free_space(BTRFS_I(inode)->root, inode,
                                           alloc_end - alloc_start);
         if (ret)
                 goto out;
@@ -5736,27 +5844,20 @@ static long btrfs_fallocate(struct inode *inode, int mode,
         while (1) {
                 struct btrfs_ordered_extent *ordered;
 
-                trans = btrfs_start_transaction(BTRFS_I(inode)->root, 1);
-                if (!trans) {
-                        ret = -EIO;
-                        goto out_free;
-                }
-
                 /* the extent lock is ordered inside the running
                  * transaction
                  */
-                lock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
-                            GFP_NOFS);
+                lock_extent_bits(&BTRFS_I(inode)->io_tree, alloc_start,
+                                 locked_end, 0, &cached_state, GFP_NOFS);
                 ordered = btrfs_lookup_first_ordered_extent(inode,
                                                             alloc_end - 1);
                 if (ordered &&
                     ordered->file_offset + ordered->len > alloc_start &&
                     ordered->file_offset < alloc_end) {
                         btrfs_put_ordered_extent(ordered);
-                        unlock_extent(&BTRFS_I(inode)->io_tree,
-                                      alloc_start, locked_end, GFP_NOFS);
-                        btrfs_end_transaction(trans, BTRFS_I(inode)->root);
-
+                        unlock_extent_cached(&BTRFS_I(inode)->io_tree,
+                                             alloc_start, locked_end,
+                                             &cached_state, GFP_NOFS);
                         /*
                          * we can't wait on the range with the transaction
                          * running or with the extent lock held
@@ -5777,10 +5878,12 @@ static long btrfs_fallocate(struct inode *inode, int mode,
                 BUG_ON(IS_ERR(em) || !em);
                 last_byte = min(extent_map_end(em), alloc_end);
                 last_byte = (last_byte + mask) & ~mask;
-                if (em->block_start == EXTENT_MAP_HOLE) {
-                        ret = prealloc_file_range(trans, inode, cur_offset,
-                                        last_byte, locked_end + 1,
-                                        alloc_hint, mode);
+                if (em->block_start == EXTENT_MAP_HOLE ||
+                    (cur_offset >= inode->i_size &&
+                     !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
+                        ret = prealloc_file_range(inode,
+                                                  cur_offset, last_byte,
+                                                alloc_hint, mode, offset+len);
                         if (ret < 0) {
                                 free_extent_map(em);
                                 break;
@@ -5796,12 +5899,11 @@ static long btrfs_fallocate(struct inode *inode, int mode,
                         break;
                 }
         }
-        unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
-                      GFP_NOFS);
+        unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
+                             &cached_state, GFP_NOFS);
 
-        btrfs_end_transaction(trans, BTRFS_I(inode)->root);
-out_free:
-        btrfs_free_reserved_data_space(root, inode, alloc_end - alloc_start);
+        btrfs_free_reserved_data_space(BTRFS_I(inode)->root, inode,
+                                       alloc_end - alloc_start);
 out:
         mutex_unlock(&inode->i_mutex);
         return ret;