Merge tag 'xfs-5.2-merge-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs updates from Darrick Wong:
 "Here's a big pile of new stuff for XFS for 5.2. XFS has grown the
  ability to report metadata health status to userspace after online
  fsck checks the filesystem. The online metadata checking code is (I
  really hope) feature complete with the addition of checks for the
  global fs counters, though it'll remain EXPERIMENTAL for now.

  There are also fixes for thundering herds of writeback completions and
  some other deadlocks, fixes for theoretical integer overflow attacks
  on space accounting, and removal of the long-defunct 'mntpt' option
  which was deprecated in the mid-2000s and (it turns out) totally
  broken since 2011 (and nobody complained...).

  Summary:

   - Fix some more buffer deadlocks when performing an unmount after a
     hard shutdown.

   - Fix some minor space accounting issues.

   - Fix some use after free problems.

   - Make the (undocumented) FITRIM behavior consistent with other
     filesystems.

   - Embiggen the xfs geometry ioctl's data structure.

   - Introduce a new AG geometry ioctl.

   - Introduce a new online health reporting infrastructure and ioctl
     for userspace to query a filesystem's health status.

   - Enhance online scrub and repair to update the health reports.

   - Reduce thundering herd problems when writeback io completes.

   - Fix some transaction reservation type errors.

   - Fix integer overflow problems with delayed alloc reservation
     counters.

   - Fix some problems where we would exit to userspace without
     unlocking.

   - Fix inconsistent behavior when finishing deferred ops fails.

   - Strengthen scrub to check incore data against ondisk metadata.

   - Remove long-broken mntpt mount option.

   - Add an online scrub function for the filesystem summary counters,
     which should make online metadata scrub more or less feature
     complete for now.

   - Various cleanups"

* tag 'xfs-5.2-merge-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (38 commits)
  xfs: change some error-less functions to void types
  xfs: add online scrub for superblock counters
  xfs: don't parse the mtpt mount option
  xfs: always rejoin held resources during defer roll
  xfs: add missing error check in xfs_prepare_shift()
  xfs: scrub should check incore counters against ondisk headers
  xfs: allow scrubbers to pause background reclaim
  xfs: rename the speculative block allocation reclaim toggle functions
  xfs: track delayed allocation reservations across the filesystem
  xfs: fix broken bhold behavior in xrep_roll_ag_trans
  xfs: unlock inode when xfs_ioctl_setattr_get_trans can't get transaction
  xfs: kill the xfs_dqtrx_t typedef
  xfs: widen inode delalloc block counter to 64-bits
  xfs: widen quota block counters to 64-bit integers
  xfs: abort unaligned nowait directio early
  xfs: assert that we don't enter agfl freeing with a non-permanent transaction
  xfs: make tr_growdata a permanent transaction
  xfs: merge adjacent io completions of the same type
  xfs: remove unused m_data_workqueue
  xfs: implement per-inode writeback completion queues
  ...

1 files changed, 124 insertions, 11 deletions

diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 3619e9e8d359..09ac1bb4c2b7 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -234,11 +234,10 @@ xfs_setfilesize_ioend(
  * IO write completion.
  */
 STATIC void
-xfs_end_io(
-	struct work_struct *work)
+xfs_end_ioend(
+	struct xfs_ioend	*ioend)
 {
-	struct xfs_ioend	*ioend =
-		container_of(work, struct xfs_ioend, io_work);
+	struct list_head	ioend_list;
 	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
 	xfs_off_t		offset = ioend->io_offset;
 	size_t			size = ioend->io_size;
@@ -275,7 +274,116 @@ xfs_end_io(
 done:
 	if (ioend->io_append_trans)
 		error = xfs_setfilesize_ioend(ioend, error);
+	list_replace_init(&ioend->io_list, &ioend_list);
 	xfs_destroy_ioend(ioend, error);
+
+	while (!list_empty(&ioend_list)) {
+		ioend = list_first_entry(&ioend_list, struct xfs_ioend,
+				io_list);
+		list_del_init(&ioend->io_list);
+		xfs_destroy_ioend(ioend, error);
+	}
+}
+
+/*
+ * We can merge two adjacent ioends if they have the same set of work to do.
+ */
+static bool
+xfs_ioend_can_merge(
+	struct xfs_ioend	*ioend,
+	int			ioend_error,
+	struct xfs_ioend	*next)
+{
+	int			next_error;
+
+	next_error = blk_status_to_errno(next->io_bio->bi_status);
+	if (ioend_error != next_error)
+		return false;
+	if ((ioend->io_fork == XFS_COW_FORK) ^ (next->io_fork == XFS_COW_FORK))
+		return false;
+	if ((ioend->io_state == XFS_EXT_UNWRITTEN) ^
+	    (next->io_state == XFS_EXT_UNWRITTEN))
+		return false;
+	if (ioend->io_offset + ioend->io_size != next->io_offset)
+		return false;
+	if (xfs_ioend_is_append(ioend) != xfs_ioend_is_append(next))
+		return false;
+	return true;
+}
+
+/* Try to merge adjacent completions. */
+STATIC void
+xfs_ioend_try_merge(
+	struct xfs_ioend	*ioend,
+	struct list_head	*more_ioends)
+{
+	struct xfs_ioend	*next_ioend;
+	int			ioend_error;
+	int			error;
+
+	if (list_empty(more_ioends))
+		return;
+
+	ioend_error = blk_status_to_errno(ioend->io_bio->bi_status);
+
+	while (!list_empty(more_ioends)) {
+		next_ioend = list_first_entry(more_ioends, struct xfs_ioend,
+				io_list);
+		if (!xfs_ioend_can_merge(ioend, ioend_error, next_ioend))
+			break;
+		list_move_tail(&next_ioend->io_list, &ioend->io_list);
+		ioend->io_size += next_ioend->io_size;
+		if (ioend->io_append_trans) {
+			error = xfs_setfilesize_ioend(next_ioend, 1);
+			ASSERT(error == 1);
+		}
+	}
+}
+
+/* list_sort compare function for ioends */
+static int
+xfs_ioend_compare(
+	void			*priv,
+	struct list_head	*a,
+	struct list_head	*b)
+{
+	struct xfs_ioend	*ia;
+	struct xfs_ioend	*ib;
+
+	ia = container_of(a, struct xfs_ioend, io_list);
+	ib = container_of(b, struct xfs_ioend, io_list);
+	if (ia->io_offset < ib->io_offset)
+		return -1;
+	else if (ia->io_offset > ib->io_offset)
+		return 1;
+	return 0;
+}
+
+/* Finish all pending io completions. */
+void
+xfs_end_io(
+	struct work_struct	*work)
+{
+	struct xfs_inode	*ip;
+	struct xfs_ioend	*ioend;
+	struct list_head	completion_list;
+	unsigned long		flags;
+
+	ip = container_of(work, struct xfs_inode, i_ioend_work);
+
+	spin_lock_irqsave(&ip->i_ioend_lock, flags);
+	list_replace_init(&ip->i_ioend_list, &completion_list);
+	spin_unlock_irqrestore(&ip->i_ioend_lock, flags);
+
+	list_sort(NULL, &completion_list, xfs_ioend_compare);
+
+	while (!list_empty(&completion_list)) {
+		ioend = list_first_entry(&completion_list, struct xfs_ioend,
+				io_list);
+		list_del_init(&ioend->io_list);
+		xfs_ioend_try_merge(ioend, &completion_list);
+		xfs_end_ioend(ioend);
+	}
 }
 
 STATIC void
@@ -283,14 +391,20 @@ xfs_end_bio(
 	struct bio		*bio)
 {
 	struct xfs_ioend	*ioend = bio->bi_private;
-	struct xfs_mount	*mp = XFS_I(ioend->io_inode)->i_mount;
+	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	unsigned long		flags;
 
 	if (ioend->io_fork == XFS_COW_FORK ||
-	    ioend->io_state == XFS_EXT_UNWRITTEN)
-		queue_work(mp->m_unwritten_workqueue, &ioend->io_work);
-	else if (ioend->io_append_trans)
-		queue_work(mp->m_data_workqueue, &ioend->io_work);
-	else
+	    ioend->io_state == XFS_EXT_UNWRITTEN ||
+	    ioend->io_append_trans != NULL) {
+		spin_lock_irqsave(&ip->i_ioend_lock, flags);
+		if (list_empty(&ip->i_ioend_list))
+			WARN_ON_ONCE(!queue_work(mp->m_unwritten_workqueue,
+						 &ip->i_ioend_work));
+		list_add_tail(&ioend->io_list, &ip->i_ioend_list);
+		spin_unlock_irqrestore(&ip->i_ioend_lock, flags);
+	} else
 		xfs_destroy_ioend(ioend, blk_status_to_errno(bio->bi_status));
 }
 
@@ -594,7 +708,6 @@ xfs_alloc_ioend(
 	ioend->io_inode = inode;
 	ioend->io_size = 0;
 	ioend->io_offset = offset;
-	INIT_WORK(&ioend->io_work, xfs_end_io);
 	ioend->io_append_trans = NULL;
 	ioend->io_bio = bio;
 	return ioend;