asm: fix misplaced padding breaking offset calculation in ivy_assembler_finish

asm/assembler/assembler.c

@@ -109,13 +109,13 @@ enum ivy_status ivy_assembler_finish(struct ivy_assembler *as)
 
 	unsigned short nr_table_entries = 0;
 
+	pad(as, 16);
+
 	struct ivy_bin_header header = {0};
 	header.h_table_offset = b_i64_htob(as->as_data_offset);
 	header.h_table_len = b_i16_htob(0);
 	header.h_magic = b_i32_htob(IVY_BIN_MAGIC);
 
-	pad(as, 16);
-
 	b_queue_iterator it = {0};
 	b_queue_foreach (&it, &as->as_table) {
 		struct asm_table_entry *e