Merge tag 'xfs-for-linus-4.10-rc1' of git://git.kernel.org/pub/scm/li…

…nux/kernel/git/dgc/linux-xfs

Pull xfs updates from Dave Chinner:
 "There is quite a varied bunch of stuff in this update, and some of it
  you will have already merged through the ext4 tree which imported the
  dax-4.10-iomap-pmd topic branch from the XFS tree.

  There is also a new direct IO implementation that uses the iomap
  infrastructure. It's much simpler, faster, and has lower IO latency
  than the existing direct IO infrastructure.

  Summary:
   - DAX PMD faults via iomap infrastructure
   - Direct-io support in iomap infrastructure
   - removal of now-redundant XFS inode iolock, replaced with VFS
     i_rwsem
   - synchronisation with fixes and changes in userspace libxfs code
   - extent tree lookup helpers
   - lots of little corruption detection improvements to verifiers
   - optimised CRC calculations
   - faster buffer cache lookups
   - deprecation of barrier/nobarrier mount options - we always use
     REQ_FUA/REQ_FLUSH where appropriate for data integrity now
   - cleanups to speculative preallocation
   - miscellaneous minor bug fixes and cleanups"

* tag 'xfs-for-linus-4.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs: (63 commits)
  xfs: nuke unused tracepoint definitions
  xfs: use GPF_NOFS when allocating btree cursors
  xfs: use xfs_vn_setattr_size to check on new size
  xfs: deprecate barrier/nobarrier mount option
  xfs: Always flush caches when integrity is required
  xfs: ignore leaf attr ichdr.count in verifier during log replay
  xfs: use rhashtable to track buffer cache
  xfs: optimise CRC updates
  xfs: make xfs btree stats less huge
  xfs: don't cap maximum dedupe request length
  xfs: don't allow di_size with high bit set
  xfs: error out if trying to add attrs and anextents > 0
  xfs: don't crash if reading a directory results in an unexpected hole
  xfs: complain if we don't get nextents bmap records
  xfs: check for bogus values in btree block headers
  xfs: forbid AG btrees with level == 0
  xfs: several xattr functions can be void
  xfs: handle cow fork in xfs_bmap_trace_exlist
  xfs: pass state not whichfork to trace_xfs_extlist
  xfs: Move AGI buffer type setting to xfs_read_agi
  ...

Documentation/filesystems/xfs.txt

@@ -51,13 +51,6 @@ default behaviour.
 	CRC enabled filesystems always use the attr2 format, and so
 	will reject the noattr2 mount option if it is set.
 
-  barrier (*)
-  nobarrier
-	Enables/disables the use of block layer write barriers for
-	writes into the journal and for data integrity operations.
-	This allows for drive level write caching to be enabled, for
-	devices that support write barriers.
-
   discard
   nodiscard (*)
 	Enable/disable the issuing of commands to let the block
@@ -228,7 +221,10 @@ default behaviour.
 Deprecated Mount Options
 ========================
 
-None at present.
+  Name				Removal Schedule
+  ----				----------------
+  barrier			no earlier than v4.15
+  nobarrier			no earlier than v4.15
 
 
 Removed Mount Options

fs/direct-io.c

@@ -554,7 +554,7 @@ static inline int dio_bio_reap(struct dio *dio, struct dio_submit *sdio)
  * filesystems that don't need it and also allows us to create the workqueue
  * late enough so the we can include s_id in the name of the workqueue.
  */
-static int sb_init_dio_done_wq(struct super_block *sb)
+int sb_init_dio_done_wq(struct super_block *sb)
 {
 	struct workqueue_struct *old;
 	struct workqueue_struct *wq = alloc_workqueue("dio/%s",

fs/internal.h

@@ -184,3 +184,6 @@ typedef loff_t (*iomap_actor_t)(struct inode *inode, loff_t pos, loff_t len,
 loff_t iomap_apply(struct inode *inode, loff_t pos, loff_t length,
 		unsigned flags, struct iomap_ops *ops, void *data,
 		iomap_actor_t actor);
+
+/* direct-io.c: */
+int sb_init_dio_done_wq(struct super_block *sb);