xfs: proper replay of deferred ops queued during log recovery

When we replay unfinished intent items that have been recovered from the
log, it's possible that the replay will cause the creation of more
deferred work items.  As outlined in commit 509955823cc9c ("xfs: log
recovery should replay deferred ops in order"), later work items have an
implicit ordering dependency on earlier work items.  Therefore, recovery
must replay the items (both recovered and created) in the same order
that they would have been during normal operation.

For log recovery, we enforce this ordering by using an empty transaction
to collect deferred ops that get created in the process of recovering a
log intent item to prevent them from being committed before the rest of
the recovered intent items.  After we finish committing all the
recovered log items, we allocate a transaction with an enormous block
reservation, splice our huge list of created deferred ops into that
transaction, and commit it, thereby finishing all those ops.

This is /really/ hokey -- it's the one place in XFS where we allow
nested transactions; the splicing of the defer ops list is is inelegant
and has to be done twice per recovery function; and the broken way we
handle inode pointers and block reservations cause subtle use-after-free
and allocator problems that will be fixed by this patch and the two
patches after it.

Therefore, replace the hokey empty transaction with a structure designed
to capture each chain of deferred ops that are created as part of
recovering a single unfinished log intent.  Finally, refactor the loop
that replays those chains to do so using one transaction per chain.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>

2 files changed, 100 insertions, 8 deletions

diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c
index 36c103c14bc9..85c371d29e8d 100644
--- a/fs/xfs/libxfs/xfs_defer.c
+++ b/fs/xfs/libxfs/xfs_defer.c
@@ -549,14 +549,89 @@ xfs_defer_move(
  *
  * Create and log intent items for all the work that we're capturing so that we
  * can be assured that the items will get replayed if the system goes down
- * before log recovery gets a chance to finish the work it put off.  Then we
- * move the chain from stp to dtp.
+ * before log recovery gets a chance to finish the work it put off.  The entire
+ * deferred ops state is transferred to the capture structure and the
+ * transaction is then ready for the caller to commit it.  If there are no
+ * intent items to capture, this function returns NULL.
  */
+static struct xfs_defer_capture *
+xfs_defer_ops_capture(
+	struct xfs_trans		*tp)
+{
+	struct xfs_defer_capture	*dfc;
+
+	if (list_empty(&tp->t_dfops))
+		return NULL;
+
+	/* Create an object to capture the defer ops. */
+	dfc = kmem_zalloc(sizeof(*dfc), KM_NOFS);
+	INIT_LIST_HEAD(&dfc->dfc_list);
+	INIT_LIST_HEAD(&dfc->dfc_dfops);
+
+	xfs_defer_create_intents(tp);
+
+	/* Move the dfops chain and transaction state to the capture struct. */
+	list_splice_init(&tp->t_dfops, &dfc->dfc_dfops);
+	dfc->dfc_tpflags = tp->t_flags & XFS_TRANS_LOWMODE;
+	tp->t_flags &= ~XFS_TRANS_LOWMODE;
+
+	return dfc;
+}
+
+/* Release all resources that we used to capture deferred ops. */
 void
-xfs_defer_capture(
-	struct xfs_trans	*dtp,
-	struct xfs_trans	*stp)
+xfs_defer_ops_release(
+	struct xfs_mount		*mp,
+	struct xfs_defer_capture	*dfc)
+{
+	xfs_defer_cancel_list(mp, &dfc->dfc_dfops);
+	kmem_free(dfc);
+}
+
+/*
+ * Capture any deferred ops and commit the transaction.  This is the last step
+ * needed to finish a log intent item that we recovered from the log.
+ */
+int
+xfs_defer_ops_capture_and_commit(
+	struct xfs_trans		*tp,
+	struct list_head		*capture_list)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_defer_capture	*dfc;
+	int				error;
+
+	/* If we don't capture anything, commit transaction and exit. */
+	dfc = xfs_defer_ops_capture(tp);
+	if (!dfc)
+		return xfs_trans_commit(tp);
+
+	/* Commit the transaction and add the capture structure to the list. */
+	error = xfs_trans_commit(tp);
+	if (error) {
+		xfs_defer_ops_release(mp, dfc);
+		return error;
+	}
+
+	list_add_tail(&dfc->dfc_list, capture_list);
+	return 0;
+}
+
+/*
+ * Attach a chain of captured deferred ops to a new transaction and free the
+ * capture structure.
+ */
+void
+xfs_defer_ops_continue(
+	struct xfs_defer_capture	*dfc,
+	struct xfs_trans		*tp)
 {
-	xfs_defer_create_intents(stp);
-	xfs_defer_move(dtp, stp);
+	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
+	ASSERT(!(tp->t_flags & XFS_TRANS_DIRTY));
+
+	/* Move captured dfops chain and state to the transaction. */
+	list_splice_init(&dfc->dfc_dfops, &tp->t_dfops);
+	tp->t_flags |= dfc->dfc_tpflags;
+
+	kmem_free(dfc);
 }
diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
index 3164199162b6..3af82ebc1249 100644
--- a/fs/xfs/libxfs/xfs_defer.h
+++ b/fs/xfs/libxfs/xfs_defer.h
@@ -8,6 +8,7 @@
 
 struct xfs_btree_cur;
 struct xfs_defer_op_type;
+struct xfs_defer_capture;
 
 /*
  * Header for deferred operation list.
@@ -64,9 +65,25 @@ extern const struct xfs_defer_op_type xfs_extent_free_defer_type;
 extern const struct xfs_defer_op_type xfs_agfl_free_defer_type;
 
 /*
+ * This structure enables a dfops user to detach the chain of deferred
+ * operations from a transaction so that they can be continued later.
+ */
+struct xfs_defer_capture {
+	/* List of other capture structures. */
+	struct list_head	dfc_list;
+
+	/* Deferred ops state saved from the transaction. */
+	struct list_head	dfc_dfops;
+	unsigned int		dfc_tpflags;
+};
+
+/*
  * Functions to capture a chain of deferred operations and continue them later.
  * This doesn't normally happen except log recovery.
  */
-void xfs_defer_capture(struct xfs_trans *dtp, struct xfs_trans *stp);
+int xfs_defer_ops_capture_and_commit(struct xfs_trans *tp,
+		struct list_head *capture_list);
+void xfs_defer_ops_continue(struct xfs_defer_capture *d, struct xfs_trans *tp);
+void xfs_defer_ops_release(struct xfs_mount *mp, struct xfs_defer_capture *d);
 
 #endif /* __XFS_DEFER_H__ */