It should be the same as hard_reset_dma_out_of_bounds from a kernel/recovery perspective, but it's useful for testing hang debugging tools.