Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: need a bit of early boot help



Den 2025-11-11 kl. 11:52, skrev Johnny Billquist:
On 2025-11-11 11:39, Anders Magnusson wrote:
Den 2025-11-11 kl. 11:14, skrev Johnny Billquist:
On 2025-11-11 06:00, Mouse wrote:
That may not be the relevant thing, though.  I don't know 1.6.1 in
enough detail, but perhaps kernel output is polled but userland is
interrupt-driven?  Maybe something is wrong with TXCS-driven
interrupts?

That would be the case, yes. Unless I remember wrong.
Since output from the kernel works, we can at least say that there is nothing wrong with the serial port as such. And the booting process went fine.

And VMS also hangs after printing a banner... Which would suggest that there is something funny either with the printing on the console, but which is interrupt related, or something else is getting the machine stuck at an elevated privilege mode, preventing interrupts from being processed.

I would actually probably suspect something else isn't completing, and the processor is stuck at some elevated prio. The console is probably working just fine, including when interrupt driven. But something else is not.

I assume we're talking about a real machine here. So it has actual Unibus. Simple suggestions would be to remove everything not absolutely required to get the machine running, and really make sure you have bus grants in all Unibus slots not having a card in it. Also, make sure the NPR jumper is in place. This one can be a bit tricky, since it's actually a wire on the back side of the backplane, which might have been cut. There a double height bus grant cards that also covers the NPR signal, but the small square ones do not.

And if the Unibus isn't properly configured, you can probably get things hung/stuck for some time.
It's a 780, Unibus is not involved here at all :-)
The PDP11 is connected directly to the backplane (SBI).

Well. Yes an no.
The PDP11, which is the FE, don't have any Unibus at all. And the machine boots, and there is terminal I/O to the console, and the system starts running.

But something happens somewhere along the way. And the *VAX* have a Unibus, and it will be probed somewhere along the way, and that is where I wonder what might be going on.
Well, since it is on a separate adapter (nexus) it cannot do much harm (actually, more than one Unibus is quite common). When something on a Unibus behaves badly the adapter just tells the OS that something is broken.  Not uncommon.

It might be something on the SBI that locks up any IPL above 14, but that should have been noted earlier in the boot process.  The kernel autoconfig runs at IPL0.  Feels more like console transmit interrupt do not reach the CPU.

I don't think there are any problems at all with the PDP11 FE.
No, not likely, since then it would not have booted anything :-)

Therefore I recommend to run the diagnostics (simpler than ripping out hardware).  The 780 diagnostics are actually quite good; sometimes it even can get down to "circuit E12 at card M1234 seems broken, replace that" which is extremely helpful :-)

.. R


Home | Main Index | Thread Index | Old Index