fs/btrfs/delayed-inode.c

44 	atomic_set(&delayed_root->items, 0);
411 	atomic_inc(&delayed_node->root->fs_info->delayed_root->items);
420 	if ((atomic_dec_return(&delayed_root->items) <
647  * Insert a single delayed item or a batch of delayed items, as many as possible
648  * that fit in a leaf. The delayed items (dir index keys) are sorted by their key
649  * in the rbtree, and if there's a gap between two consecutive dir index items,
654  * fits within a gap between delayed dir index items.
679 	 * increasing, so we can batch insert all items as there will not be any
684 	 * order to ensure we do not end up with out of order items in our leaf.
690 	 * For delayed items to insert, we track reserved metadata bytes based
774 	 * Now release our path before releasing the delayed items and their
784 	 * items, so if there are more items to process we can decrement
785 	 * index_item_leaves by 1 as we inserted 1 leaf's worth of items.
788 	 * worth of items, we may have not had continuous items, so decrementing
790 	 * only clean up the accounting when there are no items left.
794 		 * We inserted one batch of items into a leaf a there are more
795 		 * items to flush in a future batch, now release one unit of
803 		 * There are no more items to insert. We can have a number of
805 		 * items are added and then removed before they are flushed (file
877 	 * leaf - if so, we can add it to the batch of items to delete from the
903 	/* In case of BTRFS_FS_LOG_RECOVERING items won't have reserved space */
950 			 * delayed items. We ignore errors when running delayed
951 			 * items from an async context, through a work queue job
953 			 * release delayed items that failed to complete. This
955 			 * commit time we always run delayed items and will
958 			 * So just release delayed items for which we can't find
974 		 * need to add delayed items to this delayed node.
1147  * outstanding delayed items cleaned up.
1184 		 * node. If the commit of delayed items was successful the path
1195 	 * mutex. If another task starts running delayed items before we take
1346 		if (atomic_read(&delayed_root->items) <
1416 	if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
1426 	if ((atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) ||
1430 	if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) {
1518 	 * node's delayed items can be run.
1580 	 * For delayed items to insert, we track reserved metadata bytes based
1591 	 * it belongs to. With the current limit on delayed items, we rarely
1592 	 * accumulate enough dir index items to fill more than one leaf (even
1604 	/* If we now have no more dir index items, we can release all leaves. */
1700 	 * We can only do one readdir with delayed items at a time because of
1727 	 * insert/delete delayed items in this period. So we also needn't
1951 	atomic_inc(&root->fs_info->delayed_root->items);
1995 	atomic_inc(&fs_info->delayed_root->items);
2122 		 * Task A collected the delayed items into a log list while
2124 		 * only releases the items after logging the inodes they point
2130 		 * delayed items. This can happen for example when logging some
2135 		 * If this happens, just ignore delayed items already in a log
2165 	 * delete delayed items.
2208 	 * delete delayed items.