I'm not quite sure yet how we want to do this. A good first step
would be to resurrect the media error code Andries was working on
but I assume was never completed.
Well, I have been running my changes to sd.c for a few weeks
to full satisfaction. Not doing any bus resets on a media error
is a big improvement.
Yup, not forcing resets was the easy part.
There's still some work to be done so that requests can partially
succeed when the medium error is not for the first block.
I think the code I once sent you for scsi.c / sd.c does this already.
(But similar code is required for sr.c etc.)
I've been looking over the code you sent, and I'm afraid it doesn't really work
right. In the case of ISA bounce buffers being used, for example, the data for
the partial transfer will not have been transferred to the correct memory
locations. Allowing a command to partially complete is harder than it looks at
first glance.
Leonard