xfs: allow queued AG intents to drain before scrubbingscrub-drain-intents_2021-09-17

Currently, online scrub isn't sufficiently careful about quiescing
allocation groups before checking them.  While scrub does take the AG
header locks, it doesn't serialize against chains of AG update intents
that are being processed concurrently.  If there's a collision,
cross-referencing between data structures (e.g. rmapbt and refcountbt)
can yield false corruption events; if repair is running, this results in
incorrect repairs.

Fix this by adding to the perag structure the count of active intents
and make scrub wait until there aren't any to continue.  This is a
little stupid since transactions can queue intents without taking buffer
locks, but we'll also wait for those transactions.

XXX: should have instead a per-ag rwsem that gets taken as soon as the
AG[IF] are locked and stays held until the transaction commits or moves
on to the next AG?  would we rather have a six lock so that intents can
take an ix lock, and not have to upgrade to x until we actually want to
make changes to that ag?  is that how those even work??

Signed-off-by: Darrick J. Wong <djwong@kernel.org>

1 files changed, 26 insertions, 0 deletions

diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index 006bd1086e50..f9d67cdf9c5d 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -364,9 +364,12 @@ xfs_refcount_update_finish_item(
 	struct xfs_btree_cur		**state)
 {
 	struct xfs_refcount_intent	*ri;
+	struct xfs_mount		*mp = tp->t_mountp;
+	xfs_fsblock_t			orig_startblock;
 	int				error;
 
 	ri = container_of(item, struct xfs_refcount_intent, ri_list);
+	orig_startblock = ri->ri_startblock;
 	error = xfs_trans_log_finish_refcount_update(tp, CUD_ITEM(done), ri,
 			state);
 
@@ -376,6 +379,13 @@ xfs_refcount_update_finish_item(
 		       ri->ri_type == XFS_REFCOUNT_DECREASE);
 		return -EAGAIN;
 	}
+
+	/*
+	 * Drop our intent counter reference now that we've finished all the
+	 * work or failed.  Be careful to use the original startblock because
+	 * the finishing functions can update the intent state.
+	 */
+	xfs_fs_drop_intents(mp, ri->ri_realtime, orig_startblock);
 	kmem_free(ri);
 	return error;
 }
@@ -391,14 +401,29 @@ xfs_refcount_update_abort_intent(
 /* Cancel a deferred refcount update. */
 STATIC void
 xfs_refcount_update_cancel_item(
+	struct xfs_mount		*mp,
 	struct list_head		*item)
 {
 	struct xfs_refcount_intent	*ri;
 
 	ri = container_of(item, struct xfs_refcount_intent, ri_list);
+	xfs_fs_drop_intents(mp, ri->ri_realtime, ri->ri_startblock);
 	kmem_free(ri);
 }
 
+/* Add a deferred refcount update. */
+STATIC void
+xfs_refcount_update_add_item(
+	struct xfs_mount		*mp,
+	const struct list_head		*item)
+{
+	const struct xfs_refcount_intent *ri;
+
+	/* Grab an intent counter reference for this intent item. */
+	ri = container_of(item, struct xfs_refcount_intent, ri_list);
+	xfs_fs_bump_intents(mp, ri->ri_realtime, ri->ri_startblock);
+}
+
 const struct xfs_defer_op_type xfs_refcount_update_defer_type = {
 	.max_items	= XFS_CUI_MAX_FAST_EXTENTS,
 	.create_intent	= xfs_refcount_update_create_intent,
@@ -407,6 +432,7 @@ const struct xfs_defer_op_type xfs_refcount_update_defer_type = {
 	.finish_item	= xfs_refcount_update_finish_item,
 	.finish_cleanup = xfs_refcount_finish_one_cleanup,
 	.cancel_item	= xfs_refcount_update_cancel_item,
+	.add_item	= xfs_refcount_update_add_item,
 };
 
 /* Is this recovered CUI ok? */