Merge branch 'for-linus-4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull btrfs fixes and cleanups from Chris Mason:
 "Some of this got cherry-picked from a github repo this week, but I
  verified the patches.

  We have three small scrub cleanups and a collection of fixes"

* 'for-linus-4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  btrfs: Use fs_info directly in btrfs_delete_unused_bgs
  btrfs: Fix lost-data-profile caused by balance bg
  btrfs: Fix lost-data-profile caused by auto removing bg
  btrfs: Remove len argument from scrub_find_csum
  btrfs: Reduce unnecessary arguments in scrub_recheck_block
  btrfs: Use scrub_checksum_data and scrub_checksum_tree_block for scrub_recheck_block_checksum
  btrfs: Reset sblock->xxx_error stats before calling scrub_recheck_block_checksum
  btrfs: scrub: setup all fields for sblock_to_check
  btrfs: scrub: set error stats when tree block spanning stripes
  Btrfs: fix race when listing an inode's xattrs
  Btrfs: fix race leading to BUG_ON when running delalloc for nodatacow
  Btrfs: fix race leading to incorrect item deletion when dropping extents
  Btrfs: fix sleeping inside atomic context in qgroup rescan worker
  Btrfs: fix race waiting for qgroup rescan worker
  btrfs: qgroup: exit the rescan worker during umount
  Btrfs: fix extent accounting for partial direct IO writes

1 files changed, 45 insertions, 17 deletions

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 0e4f2bfcc37d..994490d5fa64 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -1304,8 +1304,14 @@ next_slot:
                 num_bytes = 0;
                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 
-                if (found_key.objectid > ino ||
-                    found_key.type > BTRFS_EXTENT_DATA_KEY ||
+                if (found_key.objectid > ino)
+                        break;
+                if (WARN_ON_ONCE(found_key.objectid < ino) ||
+                    found_key.type < BTRFS_EXTENT_DATA_KEY) {
+                        path->slots[0]++;
+                        goto next_slot;
+                }
+                if (found_key.type > BTRFS_EXTENT_DATA_KEY ||
                     found_key.offset > end)
                         break;
 
@@ -7503,6 +7509,28 @@ struct btrfs_dio_data {
         u64 reserve;
 };
 
+static void adjust_dio_outstanding_extents(struct inode *inode,
+                                           struct btrfs_dio_data *dio_data,
+                                           const u64 len)
+{
+        unsigned num_extents;
+
+        num_extents = (unsigned) div64_u64(len + BTRFS_MAX_EXTENT_SIZE - 1,
+                                           BTRFS_MAX_EXTENT_SIZE);
+        /*
+         * If we have an outstanding_extents count still set then we're
+         * within our reservation, otherwise we need to adjust our inode
+         * counter appropriately.
+         */
+        if (dio_data->outstanding_extents) {
+                dio_data->outstanding_extents -= num_extents;
+        } else {
+                spin_lock(&BTRFS_I(inode)->lock);
+                BTRFS_I(inode)->outstanding_extents += num_extents;
+                spin_unlock(&BTRFS_I(inode)->lock);
+        }
+}
+
 static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
                                    struct buffer_head *bh_result, int create)
 {
@@ -7538,8 +7566,11 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
          * If this errors out it's because we couldn't invalidate pagecache for
          * this range and we need to fallback to buffered.
          */
-        if (lock_extent_direct(inode, lockstart, lockend, &cached_state, create))
-                return -ENOTBLK;
+        if (lock_extent_direct(inode, lockstart, lockend, &cached_state,
+                               create)) {
+                ret = -ENOTBLK;
+                goto err;
+        }
 
         em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
         if (IS_ERR(em)) {
@@ -7657,19 +7688,7 @@ unlock:
                 if (start + len > i_size_read(inode))
                         i_size_write(inode, start + len);
 
-                /*
-                 * If we have an outstanding_extents count still set then we're
-                 * within our reservation, otherwise we need to adjust our inode
-                 * counter appropriately.
-                 */
-                if (dio_data->outstanding_extents) {
-                        (dio_data->outstanding_extents)--;
-                } else {
-                        spin_lock(&BTRFS_I(inode)->lock);
-                        BTRFS_I(inode)->outstanding_extents++;
-                        spin_unlock(&BTRFS_I(inode)->lock);
-                }
-
+                adjust_dio_outstanding_extents(inode, dio_data, len);
                 btrfs_free_reserved_data_space(inode, start, len);
                 WARN_ON(dio_data->reserve < len);
                 dio_data->reserve -= len;
@@ -7696,8 +7715,17 @@ unlock:
 unlock_err:
         clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
                          unlock_bits, 1, 0, &cached_state, GFP_NOFS);
+err:
         if (dio_data)
                 current->journal_info = dio_data;
+        /*
+         * Compensate the delalloc release we do in btrfs_direct_IO() when we
+         * write less data then expected, so that we don't underflow our inode's
+         * outstanding extents counter.
+         */
+        if (create && dio_data)
+                adjust_dio_outstanding_extents(inode, dio_data, len);
+
         return ret;
 }