27 Jul
2015
27 Jul
'15
2:39 a.m.
On 2015-07-27 08:16, Vincent Slyngstad wrote:
Correct. Pre-Omnibus machines had a rather different set of instructions
in in those opcodes.
From: Brian L. Stuart: Sunday, July 26, 2015 8:51 PM
To Stuart - are you sure your code does it right. Could you post it?
Running the GTF instruction on my 8/M
when the inhibit FF is set results in a 0 in AC3.
Curiouser. Furthermore,
both simh and Doug Jones' emulator omit the interrupt inhibit
bit in the GTF instruction. In fact Dr Jones' emulator has a /*?*/
comment in the code. Does anyone know the origin of this
discrepancy? Was the behavior the same on pre-Omnibus
versions of the 8, or did earlier address extensions include
that bit in the GTF result?
I don't believe that the pre-Omnibus memory extensions implement GTF. The IM6100 datasheet says bits 3 and 5-11 are under
control of the
memory extension. Reading the description for the IM6102, it says GTF
will place INT INH in AC3.
The HD6210 datasheet says bit 3 is the PWRON flag for GTF, and bit 3 is
unchanged by RIB.
The design of the KT8A is sufficiently opaque that I haven't figured out
yet how GTF works there.
he KT8A is very different. If someone is curious, the manuals do explain
it, or ask and I can try and explain it as well.
But essentially, AC3 in GTF in the KT8A is targeted for use to detect if
an IOT was executed while a delayed CIF was in effect in user mode.
Essentially saying to the OS that the user mode program is barfed and
cannot be restarted.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol