17 Nov
2014
17 Nov
'14
3:35 p.m.
From: Ben Sinclair
I wasn't clear though on the discussion about the
line clock... If my
machine is having problems with the line clock, would I need to do
anything to the aux switch, other than leave it in a certain position?
Well, I wouldn't call it a 'problem'. The thing is that the 11/03 and 11/23
were, unlike all other PDP-11s, designed with a line clock which _the software
could not enable/disable_. [*] The only way to turn the LTC on/off on those
machines is with that front-panel switch. When running software which _does
not_ handle line clock interrupts, turning the LTC on will blow the software
away - it will get an un-handled interrupt. (As happened to you.)
So if you're running code which does not use/handle the LTC (like XXDP,
apparently), turn it off. If you're running something that does want/need it,
turn it on. (And some software may require that it be _off_ while booting, and
turned _on_ once the system has started - Unix V6 falls into this category.)
* Two caveats. First, the 11/23-PLUS _does_ have software control over the
LTC; there's an LTC register on that board; it's only the dual-height 11/23
which does not. Second, the BDV11 card has this really elegant kludge that
basically adds an LTC register to a 11/03 or an 11/23, so on an 11/03 or 11/23
with a BDV11, you can leave the LTC switch on all the time, and the software
can enable line clock interrupts if it wants them.
Noel
17 Nov
17 Nov
4:14 p.m.
Since I do have a BDV11, am I correct in that I should just leave the aux
switch alone and continue debugging elsewhere?
Thanks!
On Mon, Nov 17, 2014 at 9:35 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
wrote:
--
Ben Sinclair
ben at bensinclair.com
From: Ben
Sinclair
I wasn't clear though on the discussion about
the line clock... If my
machine is having problems with the line clock, would I need to do
anything to the aux switch, other than leave it in a certain
position?
Well, I wouldn't call it a 'problem'. The thing is that the 11/03 and 11/23
were, unlike all other PDP-11s, designed with a line clock which _the
software
could not enable/disable_. [*] The only way to turn the LTC on/off on those
machines is with that front-panel switch. When running software which _does
not_ handle line clock interrupts, turning the LTC on will blow the
software
away - it will get an un-handled interrupt. (As happened to you.)
So if you're running code which does not use/handle the LTC (like XXDP,
apparently), turn it off. If you're running something that does want/need
it,
turn it on. (And some software may require that it be _off_ while booting,
and
turned _on_ once the system has started - Unix V6 falls into this
category.)
* Two caveats. First, the 11/23-PLUS _does_ have software control over the
LTC; there's an LTC register on that board; it's only the dual-height 11/23
which does not. Second, the BDV11 card has this really elegant kludge that
basically adds an LTC register to a 11/03 or an 11/23, so on an 11/03 or
11/23
with a BDV11, you can leave the LTC switch on all the time, and the
software
can enable line clock interrupts if it wants them.
Noel
4:56 p.m.
On Mon, Nov 17, 2014 at 8:14 AM, Ben Sinclair <bensinc at gmail.com> wrote:
Since I do have a BDV11, am I correct in that I should
just leave the aux
switch alone and continue debugging elsewhere?
Do you have jumper W4 installed or removed on your M8186? If you are
using a BDV11 with the M8186 system then I believe you would normally
have jumper W4 installed on the M8186.
Then on the BDV11 you would set the E21 Switch 5 to the On/Closed
position so that the LTC function is software controlled by Bit 6 of
the resister at 177546.
EK-KDF11-UG-PR2_Mar79.pdf
Page 2-2, 2.2.2 Event Line - W4
EK-BDV11-TM-001_Mar78.pdf
Page 3-3,Table 3-1, BEVNT L Switch
Page 4-22, Figure 4-11, BEVNT Logic
-Glen
7:08 p.m.
Thanks! I did not have a jumper on W4, so I added one, then set E21 switch
5 to "on."
I've been messing with a few things today, and it seems like the system
kind of works. I can boot the TU58 images I have from AK6DN's archive. He
has a "home grown" diagnostics image, and they all seem to run fine, though
I'm not sure exactly what they're testing. I ran the plain XXDP image and
was able to run xteko. From the 11/23 XXDP images I can run the MMU test,
which does pass after pass with no errors.
The test I can't run, and think I should be able to run before I try to do
anything else, is the JKDBDO CPU test. When I try to run this one it
doesn't print anything and drops me to ODT at 037526.
I believe it should print a "JKDBDO DCFlI AA CPU DIAGNOSTIC" banner when it
starts, but it doesn't seem to get that far.
Also, and I should have mentioned this before, from the time I got this
system the LEDs on the BDV11 are always all on. I know green and one of the
reds should be on to indicate proper power, but I believe the rest should
be off if the system is running normally. However, I'm not sure I
understand how those lights are controlled. The BDV11 manual says they are
controlled through a register via software, so maybe I'm just not running
anything that sets them. The only boot rom I have on the BDV11 is a TU58
boot loader, so if it's up to some other boot rom to set that LED register,
maybe mine just doesn't do it.
Thanks!
On Mon, Nov 17, 2014 at 10:56 AM, Glen Slick <glen.slick at gmail.com> wrote:
On Mon, Nov 17, 2014 at 8:14 AM, Ben Sinclair
<bensinc at gmail.com> wrote:
--
Ben Sinclair
ben at bensinclair.com
Since I do have a BDV11, am I correct in that I
should just leave the aux
switch alone and continue debugging elsewhere?
Do you have jumper W4 installed or removed on your M8186? If you are
using a BDV11 with the M8186 system then I believe you would normally
have jumper W4 installed on the M8186.
Then on the BDV11 you would set the E21 Switch 5 to the On/Closed
position so that the LTC function is software controlled by Bit 6 of
the resister at 177546.
EK-KDF11-UG-PR2_Mar79.pdf
Page 2-2, 2.2.2 Event Line - W4
EK-BDV11-TM-001_Mar78.pdf
Page 3-3,Table 3-1, BEVNT L Switch
Page 4-22, Figure 4-11, BEVNT Logic
-Glen
3694
days inactive
3694
days old
3 comments
3 participants
participants (3)
-
bensinc＠gmail.com -
glen.slick＠gmail.com -
jnc＠mercury.lcs.mit.edu