Fix long_free_dirty accounting for small files (#16264)

For files smaller than recordsize, it's most likely that they don't have
L1 blocks. However, current calculation will always return at least 1 L1
block.

In this change, we check dnode level to figure out if it has L1 blocks
or not, and return 0 if it doesn't. This will reduce the chance of
unnecessary throttling when deleting a large number of small files.

Signed-off-by: Chunwei Chen <[email protected]>
Co-authored-by: Chunwei Chen <[email protected]>
Reviewed-by: Alexander Motin <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>

module/zfs/dmu.c

@@ -815,6 +815,13 @@ get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t minimum, uint64_t *l1blks)
 
 	ASSERT3U(minimum, <=, *start);
 
+	/* dn_nlevels == 1 means we don't have any L1 blocks */
+	if (dn->dn_nlevels <= 1) {
+		*l1blks = 0;
+		*start = minimum;
+		return (0);
+	}
+
 	/*
 	 * Check if we can free the entire range assuming that all of the
 	 * L1 blocks in this range have data. If we can, we use this