Home Page Forums MAME MAME Disk image oddity.

I added a delay so that the current byte would finish before the write is ended and then the crc was written correctly. I still got an error though because the verify read was interrupted before the crc was read. Seems to me that if a read or write are interrupted it should complete the current sector like the 765 does. I have no machine with a wd1790 to try it on though.

It does seem likely that it would wait for the end of the current sector, but a test on hardware would certainly be nice.

would it be possible to see the code you added for that?

Well it would seem that the Nimbus relies on this behavior, and it does seem to work this way on the actual hardware, as I've not seen the same problem at all.

Cheers.

Phill.

Well, I reverted it but it sets a flag in interrupt_start if main_state is WRITE_SECTOR and cur_live.state is WRITE_BYTE. Then in live_run when cur_live.state is WRITE_BYTE_DONE it calls interrupt_start and returns.

Cheers,

Implementing that deffo seems to fix the problems I was seeing, and it makes sense that it would work this way as the datasheet says that the only way of terminating a write multiple is by a foced interrupt.

Cheers.

Phill.