25 Feb
2022
25 Feb
'22
2 a.m.
On Feb 24, 2022, at 8:24 PM, Noel Chiappa via cctalk
<cctalk at classiccmp.org> wrote:
? at first I thought that maybe you were thinking of the M7850 Parity Controller (which
is
actually a memory option, not KD11-E/EA specific; more below), but that's a
dual card.
Yes, I was thinking of the M7850, but I had forgotten this was just a dual card!
I think I had come to think of it as a CPU option because at some point during the /34
restoration that I did last year I got to the place of ?oh, I needn?t bother trying to
running the parity MAINDECs because I don?t have an M7850?, which felt sort of like ?I
needn?t bother trying to run floating point MAINDECs because I don?t have floating
point...? So alas, nothing exotic for you here, Noel -- just some sloppy thinking on my
part :-)
The complication was that when I powered the machine
on, it turned out that
something was asserting SACK when the machine was halted; if I put it into a
'BR .' loop, that goes away.
That is quite interesting, and not what I would have expected!
Looking at the KY11-LB manual, it does in fact assert
SACK (after it has sent
the KD11 a 'halt request, and receives a 'halt acknowledge'), to recognize
the CPU's acknowledgement of the halt request.
I ran into this when I started using one of Jeorg?s ?Unibone? cards with my /34. When the
Unibone is in place and jumped to take grants, those grant chains are open at power up
until the Unibone?s own control software can boot. So after power on my machine (w/
KY11-LB) would be jammed and inoperable from the front panel until the Unibone was
booted.
The thing that's puzzling me is that the M8264
seems to exactly replicate the
functionality of the M9302, with an 'unused' bus grant being turned into a
SACK. So I don't understand the point of the M8264.
I think the only difference would be that since the M8264 is timer based, it doesn?t need
the intact end-to-end path required for turnaround. So your bus won?t lock even if you
have a broken grant chain or a poorly behaved or hung device eating grants. I think my
/45 has this built-in (will check later). I would *guess* that the M9302 was just a ?poor
man?s SACK timeout?, and the 8264 was offered later to shore that up to a real timeout for
applications where it mattered?
cheers!
?FritzM.