On Dec 7, 19:26, Tony Duell wrote:
> OK, many machines used Western Digital hard disk controller chipsets
and
> had relatively standard formats, but others didn't. Some
manufacturers
> used homebrew controllers -- 2910-based state machines, 8X300-based
> microcontrollers, ASICs, etc.
RQDXn controllers come to mind :-(
> > the interface is set to write and the relevant drive is selected,
and
> > stop when it isn't. Similarly it can replay the data stream when
> > requested. There's nothing magical about a sector, it's just a
stream
>
> A minor correction : There is no 'request' for outputting data, other
> than selecting the drive, head, and cylinder. When that's done, you
have
> to keep on squirting data to the controller.
I meant it can stop as soon as the drive is no longer selected. Most
systms, if they want to read a sector, assert the select, wait for the
correct header to come around, and then de-select as soon as they've
got it. A lot don't of course; they wait until they've verified the
checksum, and if it fails, they expect to see the same sector come
around again. Or they may keep the drive selected for a while in case
whatever software is asking for sectors, asks for another one.
> > So Tony's idea could handle everything including an LLF, for any
format
> > and any encoding scheme you like.
>
> Which, IMHO, is important.
No, it's essential if emulation is going to work for anything other
than the simplest case of a PC with an ST506 controller (which is
trivial to replace with IDE anyway). :-)
BTW, it occured to me that you can alleviate the transfer rate problem
on the real drive by splitting the data stream, like RAID systems do.
--
Pete Peter Turnbull
Network Manager
University of York
Received on Sun Dec 07 2003 - 17:15:48 GMT