diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index b94c4abcda3a16..73a6b88c6511fb 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -1957,6 +1957,54 @@ static int can_nocow_file_extent(struct btrfs_path *path,
 	return ret < 0 ? ret : can_nocow;
 }
 
+/*
+ * To cleanup the dirty folios which will never be submitted due to
+ * error.
+ *
+ * When running a delalloc range, we may need to split the ranges (due to
+ * fragmentation or NOCOW limit). If we hit an error in the later part,
+ * we will error out and previously successfully executed range will never
+ * be submitted, thus we have to cleanup those folios by clear their
+ * dirty flag, start and finish the writeback.
+ */
+static void cleanup_dirty_folios(struct btrfs_inode *inode,
+				 struct folio *locked_folio,
+				 u64 start, u64 end, int error)
+{
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
+	struct address_space *mapping = inode->vfs_inode.i_mapping;
+	pgoff_t start_index = start >> PAGE_SHIFT;
+	pgoff_t end_index = end >> PAGE_SHIFT;
+	u32 len;
+
+	ASSERT(end + 1 - start < U32_MAX);
+	ASSERT(IS_ALIGNED(start, fs_info->sectorsize) &&
+	       IS_ALIGNED(end + 1, fs_info->sectorsize));
+	len = end + 1 - start;
+
+	/*
+	 * Handle the locked folio first.
+	 * btrfs_folio_clamp_*() helpers can handle range out of the folio case.
+	 */
+	btrfs_folio_clamp_finish_io(fs_info, locked_folio, start, len);
+
+	for (pgoff_t index = start_index; index <= end_index; index++) {
+		struct folio *folio;
+
+		/* Already handled at the beginning. */
+		if (index == locked_folio->index)
+			continue;
+		folio = __filemap_get_folio(mapping, index, FGP_LOCK, GFP_NOFS);
+		/* Cache already dropped, no need to do any cleanup. */
+		if (IS_ERR(folio))
+			continue;
+		btrfs_folio_clamp_finish_io(fs_info, locked_folio, start, len);
+		folio_unlock(folio);
+		folio_put(folio);
+	}
+	mapping_set_error(mapping, error);
+}
+
 /*
  * when nowcow writeback call back.  This checks for snapshots or COW copies
  * of the extents that exist in the file, and COWs the file as required.
@@ -1972,6 +2020,11 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 	struct btrfs_root *root = inode->root;
 	struct btrfs_path *path;
 	u64 cow_start = (u64)-1;
+	/*
+	 * If not 0, represents the inclusive end of the last fallback_to_cow()
+	 * range. Only for error handling.
+	 */
+	u64 cow_end = 0;
 	u64 cur_offset = start;
 	int ret;
 	bool check_prev = true;
@@ -2132,6 +2185,7 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 					      found_key.offset - 1);
 			cow_start = (u64)-1;
 			if (ret) {
+				cow_end = found_key.offset - 1;
 				btrfs_dec_nocow_writers(nocow_bg);
 				goto error;
 			}
@@ -2205,24 +2259,54 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 		cow_start = cur_offset;
 
 	if (cow_start != (u64)-1) {
-		cur_offset = end;
 		ret = fallback_to_cow(inode, locked_folio, cow_start, end);
 		cow_start = (u64)-1;
-		if (ret)
+		if (ret) {
+			cow_end = end;
 			goto error;
+		}
 	}
 
 	btrfs_free_path(path);
 	return 0;
 
 error:
+	/*
+	 * There are several error cases:
+	 *
+	 * 1) Failed without falling back to COW
+	 *    start         cur_offset             end
+	 *    |/////////////|                      |
+	 *
+	 *    For range [start, cur_offset) the folios are already unlocked (except
+	 *    @locked_folio), EXTENT_DELALLOC already removed.
+	 *    Only need to clear the dirty flag as they will never be submitted.
+	 *    Ordered extent and extent maps are handled by
+	 *    btrfs_mark_ordered_io_finished() inside run_delalloc_range().
+	 *
+	 * 2) Failed with error from fallback_to_cow()
+	 *    start         cur_offset  cow_end    end
+	 *    |/////////////|-----------|          |
+	 *
+	 *    For range [start, cur_offset) it's the same as case 1).
+	 *    But for range [cur_offset, cow_end), the folios have dirty flag
+	 *    cleared and unlocked, EXTENT_DEALLLOC cleared by cow_file_range().
+	 *
+	 *    Thus we should not call extent_clear_unlock_delalloc() on range
+	 *    [cur_offset, cow_end), as the folios are already unlocked.
+	 *
+	 * So clear the folio dirty flags for [start, cur_offset) first.
+	 */
+	if (cur_offset > start)
+		cleanup_dirty_folios(inode, locked_folio, start, cur_offset - 1, ret);
+
 	/*
 	 * If an error happened while a COW region is outstanding, cur_offset
-	 * needs to be reset to cow_start to ensure the COW region is unlocked
-	 * as well.
+	 * needs to be reset to @cow_end + 1 to skip the COW range, as
+	 * cow_file_range() will do the proper cleanup at error.
 	 */
-	if (cow_start != (u64)-1)
-		cur_offset = cow_start;
+	if (cow_end)
+		cur_offset = cow_end + 1;
 
 	/*
 	 * We need to lock the extent here because we're clearing DELALLOC and
diff --git a/fs/btrfs/subpage.h b/fs/btrfs/subpage.h
index 428fa9389fd49e..44fff1f4eac482 100644
--- a/fs/btrfs/subpage.h
+++ b/fs/btrfs/subpage.h
@@ -137,6 +137,19 @@ DECLARE_BTRFS_SUBPAGE_OPS(writeback);
 DECLARE_BTRFS_SUBPAGE_OPS(ordered);
 DECLARE_BTRFS_SUBPAGE_OPS(checked);
 
+/*
+ * Helper for error cleanup, where a folio will have its dirty flag cleared,
+ * with writeback started and finished.
+ */
+static inline void btrfs_folio_clamp_finish_io(struct btrfs_fs_info *fs_info,
+					       struct folio *locked_folio,
+					       u64 start, u32 len)
+{
+	btrfs_folio_clamp_clear_dirty(fs_info, locked_folio, start, len);
+	btrfs_folio_clamp_set_writeback(fs_info, locked_folio, start, len);
+	btrfs_folio_clamp_clear_writeback(fs_info, locked_folio, start, len);
+}
+
 bool btrfs_subpage_clear_and_test_dirty(const struct btrfs_fs_info *fs_info,
 					struct folio *folio, u64 start, u32 len);