[HECnet] two dumb RSX questions

[Jerome H. Fine]

>Johnny Billquist wrote:

>On 2013-12-31 02:53, Dave McGuire wrote:

>On 12/27/2013 04:16 AM, goran ahling wrote:

As I remember the SYSGEN/NETGEN task, the controller intended for DECnet
should NOT be taken into SYSGEN at all, but later, at the totally
separate task of performing NETGEN once Your OS. is up and running, it
should be included there. So, no need to redo your SYSGEN when playing
with different Ethernet cards.


    Ahh, excellent.   Thank you.

    I've run into a snag, though.   I'm trying to allow sysgen to
autoconfigure the hardware via ACF, and it hangs on my 11/24 system.
I've tried with both 4.3 and 4.6.   Thinking it might have something to
do with my recently-installed and untested DELUA, I removed it, but I
get the same result.

    So I completed sysgens via manual hardware configuration entry, both
with and without the DELUA.   Both sysgen runs completed successfully,
but the system hangs when I try to boot the new executive.

    The hardware is, as far as I can tell, otherwise fine.   I built the
system up from parts a few weeks ago, and have exercised it with RT-11.
  Everything seems ok.   The baseline RSX system boots just fine.

    I've set up an identically-configured instance of simh, and run
through the same process on that, and it works.

    My next step is to pull it out of the rack and triple check things
like the memory boards' CSRs and such, etc., but I thought I'd ask here
first.

    I last ran an RSX sysgen almost thirty years ago, and I barely
remember it at all...but I'm certain nothing like this happened.   That
was on an 11/34 with RL01s, RL02s, and RK07s.

    Any thoughts?


Others have come with some suggestions. I'll stick my neck out and say that some device isn't working properly. RSX probes and kicks devices at startup (check that they respond to their CSRs, and for some check that they generate interrupts). I don't think RT-11 do that as much. Also, it might be that the device in question isn't even configured in RT-11.

When both autoconfigure fails and a generated system fails, it's one of the controllers that you have genned in that is causing you problems.

Anyway. My first suspicion would be a DMA controller with the NPR wire still in place. But there might be some other controller broken as well.

I do not think you have power problems, although I wouldn't totally rule it out. But I would expect you to see problems with RT-11 as well, if that was the case. Or with the Baseline RSX system.
What you do see is typical of one specific controller blocking you. Unfortunately, with 11M it isn't that easy to spot which one, as the probing of the controller are done automatically right at the start of the booting.
In M+ it is done in a different way, which makes it easier to troubleshoot.

Happy New Year to everyone who uses the Common Era calendar,
since by the time this reply is read, New Year's day of 2014 will have
already started for some individuals.

In addition, if anyone who is reading this still uses RT-11 to change
RT-11 programs and uses the RT-11 Symbolic Debugger, SD:, I
have a question at the very end.

I can offer a few comments on the basic operations which occur when
RT-11 boots, but only at the operating system level.   I am not aware of
exactly what occurs within each device driver, let alone the exact details
of the code in the installation section of any device driver.

During an RT-11 boot, other than the device driver which is used as the
system device, RT-11 only attempts to INSTALL the other device drivers.
If I remember correctly, RT-11 checks all the files on the system device
to determine if they are valid device drivers for the operating system
which is being booted.   Both Mapped and UnMapped operating
systems (specifically the file for the operating system) can coexist on
the same system device along with separate device drivers for both
types of operating systems.

The primary boot code for RT-11 (at least after V04.00 of RT-11
in 1980) was a single block in block zero of the system device and was
obtained from the device driver and written into block zero along
with the secondary boot code in blocks 2, 3, 4 and 5 via:
COPY/BOOT   DL0:RT11FB.SYS   DL0:
as just one example for the RL01 / RL02 devices with an UnMapped
variant of the RT-11 operating system.   The primary boot code is from
the device driver, DL0:DL.SYS, while the secondary boot code is from
the operating system file, DL0:RT11FB.SYS

During the RT-11 boot, RT-11 attempts to INSTALL all the device
drivers on the system device which are compatible with the operating
system.   In general, RT-11 attempts to execute the installation code
from each compatible device driver which, except for the resident
system device, consists of an access to the IOPAGE registers that
are used by the device driver.   Sometimes the device driver file
contains additional code, but if I remember correctly, no data device
driver (all devices are data devices if they are not the system device)
in RT-11 requires any interrupts.

So if booting RSX-11 requires device interrupts, that might be the reason.
You could attempt to isolate the device responsible in RT-11 by an
attempt to use each device driver after RT-11 boots.

As for using SD: (the RT-11 Symbolic Debugger), does anyone find
that SDHX.SYS uses too much low memory (1125 words)?   Would
having an option to retain the Program Counter Locations (at least
the last 1000 addresses executed - in a circular buffer, of course) be
of help in debugging RT-11 code?   Another option which I would
appreciate would be the ability to support the temporary execution
of other jobs while the current job being execution is placed into
.Wait state.   Does anyone else agree?   Under TSX-Plus and RSX-11,
that happens automatically, in addition to using VDT under RT-11.
BUT, SDHX.SYS freezes the complete RT-11 system when the
code being debugged is stopped at a break point.

Jerome Fine