Go back to resetting the refresh counter after a DMA. Probably not quite correct as it is probably reset on VDP triggered refresh, but this is close enough for now given the general limitations with my refresh code. VDP FIFO Testing seems to be passing 100% reliably again (was occassionally failing still with the last commit)

#ifndef REALTEC_H_
#define REALTEC_H_

uint8_t realtec_detect(uint8_t *rom, uint32_t rom_size);
rom_info realtec_configure_rom(uint8_t *rom, uint32_t rom_size, memmap_chunk const *base_map, uint32_t base_chunks);

#endif //REALTEC_H_