Merge tag 'for-5.6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull more btrfs updates from David Sterba:
"Fixes that arrived after the merge window freeze, mostly stable
material.

- fix race in tree-mod-log element tracking

- fix bio flushing inside extent writepages

- fix assertion when in-memory tracking of discarded extents finds an
empty tree (eg. after adding a new device)

- update logic of temporary read-only block groups to take into
account overcommit

- fix some fixup worker corner cases:
- page could not go through proper COW cycle and the dirty status
is lost due to page migration
- deadlock if delayed allocation is performed under page lock

- fix send emitting invalid clones within the same file

- fix statfs reporting 0 free space when global block reserve size is
larger than remaining free space but there is still space for new
chunks"

* tag 'for-5.6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: do not zero f_bavail if we have available space
Btrfs: send, fix emission of invalid clone operations within the same file
btrfs: do not do delalloc reservation under page lock
btrfs: drop the -EBUSY case in __extent_writepage_io
Btrfs: keep pages dirty when using btrfs_writepage_fixup_worker
btrfs: take overcommit into account in inc_block_group_ro
btrfs: fix force usage in inc_block_group_ro
btrfs: Correctly handle empty trees in find_first_clear_extent_bit
btrfs: flush write bio if we loop in extent_write_cache_pages
Btrfs: fix race between adding and putting tree mod seq elements and nodes

fs/btrfs/block-group.c

···
 {
 	struct btrfs_space_info *sinfo = cache->space_info;
 	u64 num_bytes;
-	u64 sinfo_used;
 	int ret = -ENOSPC;
 
 	spin_lock(&sinfo->lock);
···
 
 	num_bytes = cache->length - cache->reserved - cache->pinned -
 		    cache->bytes_super - cache->used;
-	sinfo_used = btrfs_space_info_used(sinfo, true);
 
 	/*
-	 * sinfo_used + num_bytes should always <= sinfo->total_bytes.
-	 *
-	 * Here we make sure if we mark this bg RO, we still have enough
-	 * free space as buffer.
+	 * Data never overcommits, even in mixed mode, so do just the straight
+	 * check of left over space in how much we have allocated.
 	 */
-	if (sinfo_used + num_bytes <= sinfo->total_bytes) {
+	if (force) {
+		ret = 0;
+	} else if (sinfo->flags & BTRFS_BLOCK_GROUP_DATA) {
+		u64 sinfo_used = btrfs_space_info_used(sinfo, true);
+
+		/*
+		 * Here we make sure if we mark this bg RO, we still have enough
+		 * free space as buffer.
+		 */
+		if (sinfo_used + num_bytes <= sinfo->total_bytes)
+			ret = 0;
+	} else {
+		/*
+		 * We overcommit metadata, so we need to do the
+		 * btrfs_can_overcommit check here, and we need to pass in
+		 * BTRFS_RESERVE_NO_FLUSH to give ourselves the most amount of
+		 * leeway to allow us to mark this block group as read only.
+		 */
+		if (btrfs_can_overcommit(cache->fs_info, sinfo, num_bytes,
+					 BTRFS_RESERVE_NO_FLUSH))
+			ret = 0;
+	}
+
+	if (!ret) {
 		sinfo->bytes_readonly += num_bytes;
 		cache->ro++;
 		list_add_tail(&cache->ro_list, &sinfo->ro_bgs);
-		ret = 0;
 	}
 out:
 	spin_unlock(&cache->lock);
···
 	if (ret == -ENOSPC && btrfs_test_opt(cache->fs_info, ENOSPC_DEBUG)) {
 		btrfs_info(cache->fs_info,
 			"unable to make block group %llu ro", cache->start);
-		btrfs_info(cache->fs_info,
-			"sinfo_used=%llu bg_num_bytes=%llu",
-			sinfo_used, num_bytes);
 		btrfs_dump_space_info(cache->fs_info, cache->space_info, 0, 0);
 	}
 	return ret;
···
 		}
 	}
 
-	ret = inc_block_group_ro(cache, !do_chunk_alloc);
+	ret = inc_block_group_ro(cache, 0);
 	if (!do_chunk_alloc)
 		goto unlock_out;
 	if (!ret)

fs/btrfs/ctree.c

···
 			   struct seq_list *elem)
 {
 	write_lock(&fs_info->tree_mod_log_lock);
-	spin_lock(&fs_info->tree_mod_seq_lock);
 	if (!elem->seq) {
 		elem->seq = btrfs_inc_tree_mod_seq(fs_info);
 		list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
 	}
-	spin_unlock(&fs_info->tree_mod_seq_lock);
 	write_unlock(&fs_info->tree_mod_log_lock);
 
 	return elem->seq;
···
 	if (!seq_putting)
 		return;
 
-	spin_lock(&fs_info->tree_mod_seq_lock);
+	write_lock(&fs_info->tree_mod_log_lock);
 	list_del(&elem->list);
 	elem->seq = 0;
 
···
 				 * blocker with lower sequence number exists, we
 				 * cannot remove anything from the log
 				 */
-				spin_unlock(&fs_info->tree_mod_seq_lock);
+				write_unlock(&fs_info->tree_mod_log_lock);
 				return;
 			}
 			min_seq = cur_elem->seq;
 		}
 	}
-	spin_unlock(&fs_info->tree_mod_seq_lock);
 
 	/*
 	 * anything that's lower than the lowest existing (read: blocked)
 	 * sequence number can be removed from the tree.
 	 */
-	write_lock(&fs_info->tree_mod_log_lock);
 	tm_root = &fs_info->tree_mod_log;
 	for (node = rb_first(tm_root); node; node = next) {
 		next = rb_next(node);

fs/btrfs/ctree.h

···
 	atomic_t nr_delayed_iputs;
 	wait_queue_head_t delayed_iputs_wait;
 
-	/* this protects tree_mod_seq_list */
-	spinlock_t tree_mod_seq_lock;
 	atomic64_t tree_mod_seq;
-	struct list_head tree_mod_seq_list;
 
-	/* this protects tree_mod_log */
+	/* this protects tree_mod_log and tree_mod_seq_list */
 	rwlock_t tree_mod_log_lock;
 	struct rb_root tree_mod_log;
+	struct list_head tree_mod_seq_list;
 
 	atomic_t async_delalloc_pages;
 

fs/btrfs/delayed-ref.c

···
 	if (head->is_data)
 		return;
 
-	spin_lock(&fs_info->tree_mod_seq_lock);
+	read_lock(&fs_info->tree_mod_log_lock);
 	if (!list_empty(&fs_info->tree_mod_seq_list)) {
 		struct seq_list *elem;
 
···
 					struct seq_list, list);
 		seq = elem->seq;
 	}
-	spin_unlock(&fs_info->tree_mod_seq_lock);
+	read_unlock(&fs_info->tree_mod_log_lock);
 
 again:
 	for (node = rb_first_cached(&head->ref_tree); node;
···
 	struct seq_list *elem;
 	int ret = 0;
 
-	spin_lock(&fs_info->tree_mod_seq_lock);
+	read_lock(&fs_info->tree_mod_log_lock);
 	if (!list_empty(&fs_info->tree_mod_seq_list)) {
 		elem = list_first_entry(&fs_info->tree_mod_seq_list,
 					struct seq_list, list);
···
 		}
 	}
 
-	spin_unlock(&fs_info->tree_mod_seq_lock);
+	read_unlock(&fs_info->tree_mod_log_lock);
 	return ret;
 }
 

fs/btrfs/disk-io.c

···
 	spin_lock_init(&fs_info->fs_roots_radix_lock);
 	spin_lock_init(&fs_info->delayed_iput_lock);
 	spin_lock_init(&fs_info->defrag_inodes_lock);
-	spin_lock_init(&fs_info->tree_mod_seq_lock);
 	spin_lock_init(&fs_info->super_lock);
 	spin_lock_init(&fs_info->buffer_lock);
 	spin_lock_init(&fs_info->unused_bgs_lock);

fs/btrfs/extent_io.c

···
 	/* Find first extent with bits cleared */
 	while (1) {
 		node = __etree_search(tree, start, &next, &prev, NULL, NULL);
-		if (!node) {
+		if (!node && !next && !prev) {
+			/*
+			 * Tree is completely empty, send full range and let
+			 * caller deal with it
+			 */
+			*start_ret = 0;
+			*end_ret = -1;
+			goto out;
+		} else if (!node && !next) {
+			/*
+			 * We are past the last allocated chunk, set start at
+			 * the end of the last extent.
+			 */
+			state = rb_entry(prev, struct extent_state, rb_node);
+			*start_ret = state->end + 1;
+			*end_ret = -1;
+			goto out;
+		} else if (!node) {
 			node = next;
-			if (!node) {
-				/*
-				 * We are past the last allocated chunk,
-				 * set start at the end of the last extent. The
-				 * device alloc tree should never be empty so
-				 * prev is always set.
-				 */
-				ASSERT(prev);
-				state = rb_entry(prev, struct extent_state, rb_node);
-				*start_ret = state->end + 1;
-				*end_ret = -1;
-				goto out;
-			}
 		}
 		/*
 		 * At this point 'node' either contains 'start' or start is
···
 	ret = btrfs_writepage_cow_fixup(page, start, page_end);
 	if (ret) {
 		/* Fixup worker will requeue */
-		if (ret == -EBUSY)
-			wbc->pages_skipped++;
-		else
-			redirty_page_for_writepage(wbc, page);
-
+		redirty_page_for_writepage(wbc, page);
 		update_nr_written(wbc, nr_written);
 		unlock_page(page);
 		return 1;
···
 		 */
 		scanned = 1;
 		index = 0;
-		goto retry;
+
+		/*
+		 * If we're looping we could run into a page that is locked by a
+		 * writer and that writer could be waiting on writeback for a
+		 * page in our current bio, and thus deadlock, so flush the
+		 * write bio here.
+		 */
+		ret = flush_write_bio(epd);
+		if (!ret)
+			goto retry;
 	}
 
 	if (wbc->range_cyclic || (wbc->nr_to_write > 0 && range_whole))

fs/btrfs/inode.c

···
 /* see btrfs_writepage_start_hook for details on why this is required */
 struct btrfs_writepage_fixup {
 	struct page *page;
+	struct inode *inode;
 	struct btrfs_work work;
 };
 
···
 	struct inode *inode;
 	u64 page_start;
 	u64 page_end;
-	int ret;
+	int ret = 0;
+	bool free_delalloc_space = true;
 
 	fixup = container_of(work, struct btrfs_writepage_fixup, work);
 	page = fixup->page;
+	inode = fixup->inode;
+	page_start = page_offset(page);
+	page_end = page_offset(page) + PAGE_SIZE - 1;
+
+	/*
+	 * This is similar to page_mkwrite, we need to reserve the space before
+	 * we take the page lock.
+	 */
+	ret = btrfs_delalloc_reserve_space(inode, &data_reserved, page_start,
+					   PAGE_SIZE);
 again:
 	lock_page(page);
+
+	/*
+	 * Before we queued this fixup, we took a reference on the page.
+	 * page->mapping may go NULL, but it shouldn't be moved to a different
+	 * address space.
+	 */
 	if (!page->mapping || !PageDirty(page) || !PageChecked(page)) {
-		ClearPageChecked(page);
+		/*
+		 * Unfortunately this is a little tricky, either
+		 *
+		 * 1) We got here and our page had already been dealt with and
+		 *    we reserved our space, thus ret == 0, so we need to just
+		 *    drop our space reservation and bail.  This can happen the
+		 *    first time we come into the fixup worker, or could happen
+		 *    while waiting for the ordered extent.
+		 * 2) Our page was already dealt with, but we happened to get an
+		 *    ENOSPC above from the btrfs_delalloc_reserve_space.  In
+		 *    this case we obviously don't have anything to release, but
+		 *    because the page was already dealt with we don't want to
+		 *    mark the page with an error, so make sure we're resetting
+		 *    ret to 0.  This is why we have this check _before_ the ret
+		 *    check, because we do not want to have a surprise ENOSPC
+		 *    when the page was already properly dealt with.
+		 */
+		if (!ret) {
+			btrfs_delalloc_release_extents(BTRFS_I(inode),
+						       PAGE_SIZE);
+			btrfs_delalloc_release_space(inode, data_reserved,
+						     page_start, PAGE_SIZE,
+						     true);
+		}
+		ret = 0;
 		goto out_page;
 	}
 
-	inode = page->mapping->host;
-	page_start = page_offset(page);
-	page_end = page_offset(page) + PAGE_SIZE - 1;
+	/*
+	 * We can't mess with the page state unless it is locked, so now that
+	 * it is locked bail if we failed to make our space reservation.
+	 */
+	if (ret)
+		goto out_page;
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
 			 &cached_state);
 
 	/* already ordered? We're done */
 	if (PagePrivate2(page))
-		goto out;
+		goto out_reserved;
 
 	ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start,
 					PAGE_SIZE);
···
 		goto again;
 	}
 
-	ret = btrfs_delalloc_reserve_space(inode, &data_reserved, page_start,
-					   PAGE_SIZE);
-	if (ret) {
-		mapping_set_error(page->mapping, ret);
-		end_extent_writepage(page, ret, page_start, page_end);
-		ClearPageChecked(page);
-		goto out;
-	 }
-
 	ret = btrfs_set_extent_delalloc(inode, page_start, page_end, 0,
 					&cached_state);
-	if (ret) {
-		mapping_set_error(page->mapping, ret);
-		end_extent_writepage(page, ret, page_start, page_end);
-		ClearPageChecked(page);
+	if (ret)
 		goto out_reserved;
-	}
 
-	ClearPageChecked(page);
-	set_page_dirty(page);
+	/*
+	 * Everything went as planned, we're now the owner of a dirty page with
+	 * delayed allocation bits set and space reserved for our COW
+	 * destination.
+	 *
+	 * The page was dirty when we started, nothing should have cleaned it.
+	 */
+	BUG_ON(!PageDirty(page));
+	free_delalloc_space = false;
 out_reserved:
 	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
-	if (ret)
+	if (free_delalloc_space)
 		btrfs_delalloc_release_space(inode, data_reserved, page_start,
 					     PAGE_SIZE, true);
-out:
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end,
 			     &cached_state);
 out_page:
+	if (ret) {
+		/*
+		 * We hit ENOSPC or other errors.  Update the mapping and page
+		 * to reflect the errors and clean the page.
+		 */
+		mapping_set_error(page->mapping, ret);
+		end_extent_writepage(page, ret, page_start, page_end);
+		clear_page_dirty_for_io(page);
+		SetPageError(page);
+	}
+	ClearPageChecked(page);
 	unlock_page(page);
 	put_page(page);
 	kfree(fixup);
 	extent_changeset_free(data_reserved);
+	/*
+	 * As a precaution, do a delayed iput in case it would be the last iput
+	 * that could need flushing space. Recursing back to fixup worker would
+	 * deadlock.
+	 */
+	btrfs_add_delayed_iput(inode);
 }
 
 /*
···
 	if (TestClearPagePrivate2(page))
 		return 0;
 
+	/*
+	 * PageChecked is set below when we create a fixup worker for this page,
+	 * don't try to create another one if we're already PageChecked()
+	 *
+	 * The extent_io writepage code will redirty the page if we send back
+	 * EAGAIN.
+	 */
 	if (PageChecked(page))
 		return -EAGAIN;
 
···
 	if (!fixup)
 		return -EAGAIN;
 
+	/*
+	 * We are already holding a reference to this inode from
+	 * write_cache_pages.  We need to hold it because the space reservation
+	 * takes place outside of the page lock, and we can't trust
+	 * page->mapping outside of the page lock.
+	 */
+	ihold(inode);
 	SetPageChecked(page);
 	get_page(page);
 	btrfs_init_work(&fixup->work, btrfs_writepage_fixup_worker, NULL, NULL);
 	fixup->page = page;
+	fixup->inode = inode;
 	btrfs_queue_work(fs_info->fixup_workers, &fixup->work);
-	return -EBUSY;
+
+	return -EAGAIN;
 }
 
 static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,

fs/btrfs/send.c

···
 		 * destination of the stream.
 		 */
 		if (ino == bctx->cur_objectid &&
-		    offset >= bctx->sctx->cur_inode_next_write_offset)
+		    offset + bctx->extent_len >
+		    bctx->sctx->cur_inode_next_write_offset)
 			return 0;
 	}
 

fs/btrfs/space-info.c

···
 	return (global->size << 1);
 }
 
-static int can_overcommit(struct btrfs_fs_info *fs_info,
-			  struct btrfs_space_info *space_info, u64 bytes,
-			  enum btrfs_reserve_flush_enum flush)
+int btrfs_can_overcommit(struct btrfs_fs_info *fs_info,
+			 struct btrfs_space_info *space_info, u64 bytes,
+			 enum btrfs_reserve_flush_enum flush)
 {
 	u64 profile;
 	u64 avail;
···
 
 		/* Check and see if our ticket can be satisified now. */
 		if ((used + ticket->bytes <= space_info->total_bytes) ||
-		    can_overcommit(fs_info, space_info, ticket->bytes, flush)) {
+		    btrfs_can_overcommit(fs_info, space_info, ticket->bytes,
+					 flush)) {
 			btrfs_space_info_update_bytes_may_use(fs_info,
 							      space_info,
 							      ticket->bytes);
···
 		return to_reclaim;
 
 	to_reclaim = min_t(u64, num_online_cpus() * SZ_1M, SZ_16M);
-	if (can_overcommit(fs_info, space_info, to_reclaim,
-			   BTRFS_RESERVE_FLUSH_ALL))
+	if (btrfs_can_overcommit(fs_info, space_info, to_reclaim,
+				 BTRFS_RESERVE_FLUSH_ALL))
 		return 0;
 
 	used = btrfs_space_info_used(space_info, true);
 
-	if (can_overcommit(fs_info, space_info, SZ_1M, BTRFS_RESERVE_FLUSH_ALL))
+	if (btrfs_can_overcommit(fs_info, space_info, SZ_1M,
+				 BTRFS_RESERVE_FLUSH_ALL))
 		expected = div_factor_fine(space_info->total_bytes, 95);
 	else
 		expected = div_factor_fine(space_info->total_bytes, 90);
···
 	 */
 	if (!pending_tickets &&
 	    ((used + orig_bytes <= space_info->total_bytes) ||
-	     can_overcommit(fs_info, space_info, orig_bytes, flush))) {
+	     btrfs_can_overcommit(fs_info, space_info, orig_bytes, flush))) {
 		btrfs_space_info_update_bytes_may_use(fs_info, space_info,
 						      orig_bytes);
 		ret = 0;

fs/btrfs/space-info.h

···
 				 enum btrfs_reserve_flush_enum flush);
 void btrfs_try_granting_tickets(struct btrfs_fs_info *fs_info,
 				struct btrfs_space_info *space_info);
+int btrfs_can_overcommit(struct btrfs_fs_info *fs_info,
+			 struct btrfs_space_info *space_info, u64 bytes,
+			 enum btrfs_reserve_flush_enum flush);
 
 static inline void btrfs_space_info_free_bytes_may_use(
 				struct btrfs_fs_info *fs_info,

fs/btrfs/super.c

···
 	 */
 	thresh = SZ_4M;
 
-	if (!mixed && total_free_meta - thresh < block_rsv->size)
+	/*
+	 * We only want to claim there's no available space if we can no longer
+	 * allocate chunks for our metadata profile and our global reserve will
+	 * not fit in the free metadata space.  If we aren't ->full then we
+	 * still can allocate chunks and thus are fine using the currently
+	 * calculated f_bavail.
+	 */
+	if (!mixed && block_rsv->space_info->full &&
+	    total_free_meta - thresh < block_rsv->size)
 		buf->f_bavail = 0;
 
 	buf->f_type = BTRFS_SUPER_MAGIC;

fs/btrfs/tests/btrfs-tests.c

···
 	spin_lock_init(&fs_info->qgroup_lock);
 	spin_lock_init(&fs_info->super_lock);
 	spin_lock_init(&fs_info->fs_roots_radix_lock);
-	spin_lock_init(&fs_info->tree_mod_seq_lock);
 	mutex_init(&fs_info->qgroup_ioctl_lock);
 	mutex_init(&fs_info->qgroup_rescan_lock);
 	rwlock_init(&fs_info->tree_mod_log_lock);

fs/btrfs/tests/extent-io-tests.c

···
 	int ret = -EINVAL;
 
 	test_msg("running find_first_clear_extent_bit test");
+
 	extent_io_tree_init(NULL, &tree, IO_TREE_SELFTEST, NULL);
 
+	/* Test correct handling of empty tree */
+	find_first_clear_extent_bit(&tree, 0, &start, &end, CHUNK_TRIMMED);
+	if (start != 0 || end != -1) {
+		test_err(
+	"error getting a range from completely empty tree: start %llu end %llu",
+			 start, end);
+		goto out;
+	}
 	/*
 	 * Set 1M-4M alloc/discard and 32M-64M thus leaving a hole between
 	 * 4M-32M