> On Jun 9, 2016, at 12:56 AM, Don North <ak...@mindspring.com> wrote:
>
> On 6/8/2016 5:56 PM, Paul Koning wrote:
>>> On Jun 8, 2016, at 6:46 PM, Jay West <jw...@classiccmp.org> wrote:
>>>
>>> I wrote...
>>> On Thu, Jun 9, 2016 at 6:34 AM, Jay West <jw...@classiccmp.org> wrote:
>>>> 23-E39A9 is still lost to time, afaik. TMSCP - TU81
>>> To which mike replied...
>>> ----
>>> Could that not be reverse-engineered from the boot code in e.g. Emulex
>>> UC17 ROMs? They could do TMSCP...
>>>
>>> BTW what PROM blower would folks recommend for creating/imaging M9312 ROMs?
>>> ----
>>> Possibly, but some (me) are sticklers for original code. Plus, I am not
>>> sure, but I think someone said this rom did something really bizarre to fit
>>> in the available rom space - self modifying code or something...
>> MSCP isn't all that hard. The RSTS secondary loader fits in one block, and
>> it contains not just code to speak MSCP but also enough room for a map of
>> pointers to where the code to load lives (as opposed to the primary boot
>> which just has to load one block from address zero).
>>
>> paul
>
> In an M9312 boot PROM there are 49. words of bootstrap space that are
> available for code and data. Getting an MSCP boot to fit in that space
> required playing some tricks like treating some specially crafted
> instructions as data values (and vice versa).
Oh... I didn't realize it was that small. Yes, that does make it harder.
paul
