Just curious, why not add a new block_device_operations member 'alternative_gpt_sector', then move all the code calculating it to the mmc layer?
Just curious, why not add a new block_device_operations member 'alternative_gpt_sector', then move all the code calculating it to the mmc layer?