Add support for USB Panda Display

[DavidGriffith]

Since most people have machines with internal USB ports now, I created a USB-based Panda Display (see https://github.com/DavidGriffith/panda-display) in imitation of the old parallel-port Panda Display from Spare Time Gizmos (see http://sparetimegizmos.com/Hardware/Panda.htm). I made three boards, built two of them, and sold the third. It works as far as directly manipulating it with a Linux test program. I cannot get Mark Crispin's pdp10 program for exercising the Panda Display to compile. The machine-monitoring functionality I added to my fork of klh10 doesn't seem to use the display correctly.

I will start adding to this repo my changes and then delete my klh10 repo.

[DavidGriffith]

I got this test program from Jan de Rie who got it from Mark Crispin. This program is controlling the parallel port Panda Display in this video: https://www.youtube.com/watch?v=g_LcQ5apODg

        TITLE HACK
        SEARCH MONSYM

HACK:   SKIPA 4,[^D4]
HACK0:   SOJL 4,TRAILL
        SETZB 1,3
HACK1:  TLNE 1,777777
         JRST HACK0
        MOVE 2,1
        CIRC 2,-^D18
        SKIPN 4
         SKIPA 1,3
          IOR 1,3
        LITES%
         HALT
        HRRZ 2,1
        MOVEI 1,30
        DISMS%
        MOVE 1,2
        AOJA 1,HACK1

TRAILL: SETZ 1,
TRALL0: LITES%
         HALT
        MOVE 2,1
        MOVEI 1,30
        DISMS%
        JUMPL 2,TRAILR
        LSHC 1,^D37
        AOJA 1,TRALL0

TRAILR: MOVSI 1,400000
TRALR0: LITES%
         HALT
        MOVE 2,1
        MOVEI 1,30
        DISMS%
        MOVE 1,2
        TRNE 1,1
         JRST HACK
        ASH 1,-1
        JRST TRALR0

        END HACK

[Rhialto]

Such an addition sounds useful, especially if people can get that display for their own use. And it is configurable to co-exist with the current lights (at least compile-time configurable; even cleaner would probably be if it is run-time configurable).

[DavidGriffith]

I think the most practical would be to have the type of Panda Display configurable at compile time.

[DavidGriffith]

Is anyone interested in investigating this problem? I have one completely built USB Panda Display available to whoever will do it.

[larsbrinkhoff]

I'm really too busy, but maybe if no one else has the spare cycles, I can chip in. I'd be testing it from ITS.

[frink0]

I'd be willing to have a look. Like Lars, it would be from ITS, though I suppose I could try TOPS-20 afterward too.

[larsbrinkhoff]

Hello @frink0 and @DavidGriffith,

If you want a more interactive conversation about ITS and/or PDP-10s, there is an IRC cannel called # PDP-10 at freenode.net.

[DavidGriffith]

I think I'll make up some more Panda Display boards. Besides @larsbrinkhoff and @frink0, who else wants one (or more)?

[larsbrinkhoff]

This may not be the best place to reach the intended audience. I think a post to the its-hackers mailing list would reach more people.

[larsbrinkhoff]

I took the offer to help with testing.

[larsbrinkhoff]

@DavidGriffith, I think I have all parts now except the Maxim display drivers. Maybe that's enough to do some testing?

[DavidGriffith]

No. You'll need the Maxim chip otherwise you won't know if any of your testing is doing anything useful.

[frink0]

@DavidGriffith, I forgot to say back in May: I much appreciate your offer to make some, as I lack the skills needed to do it myself presently. But I wouldn't want to put you out of pocket, so I'm willing to pay for the parts required and shipping afterward, and if you would like some sort of remuneration for your time besides, I'm sure we could work that out too.

[DavidGriffith]

@frink0 I can work with that. However, I'm not going to do a new run of boards until I finalize the design for a 3D-printed carrier to allow the board to easily be mounted in a 5.25-inch drive bay (see https://github.com/PDP-10/its/issues/1127)

[frink0]

@DavidGriffith No worries, this is hardly pressing. Thanks again!

[larsbrinkhoff]

I will have to remind the Hong Kong dealer they were supposed to deliver the Maxim part today.

[larsbrinkhoff]

The last part arrived just now! I think I'll be able to test it tonight.

[larsbrinkhoff]

Got the firmware programmed. Setting LEDs with ptest works.

[larsbrinkhoff]

I'm not sure how this should be accessed from KLH10. The KL10 doesn't have any front panel, just a PDP-11/40 front end. The KS10 similarly has an 8080 front end and no lights.

The MIT MC KL10 had a custom KL-UDGE board attached as an I/O bus device. But we don't have an emulator that can run KL10 ITS.

The most straigtforward option is @rcornwell's KA10 simulator.

I wonder how the Panda TOPS-20 could get access to the Panda Display?

[Rhialto]

src/dvlites.c has code to access it.

[larsbrinkhoff]

Thanks @Rhialto, I'll examine that file.

I see the old display could have four units attached! It had 36 main lights, and 3 auxiliary. The USB display has 40 + 4.

CONO do the "host" device sets the aux lights. DATAO PI, sets the main lights. As far as I can see this should work on both KL and KS!

[larsbrinkhoff]

I have the HACK sample program running in ITS now, using the KA10 simulator though.

Something in ITS is interfering.

[larsbrinkhoff]

These.

SEAREP+13/   DATAO PI,B
SCHDN2+5/   DATAO PI,116057
QSETUP+56/   DATAO PI,
CKML1A+1/   DATAO PI,

[larsbrinkhoff]

To make ITS leave lights alone: MIPGDS/ -1

[larsbrinkhoff]

The KS10 of course has a completely different set of I/O instructions, but fortunately the opcode occupied by DATAO PI, is free!

[larsbrinkhoff]

I made this branch:
https://github.com/PDP-10/klh10/commits/lars/lights

As you can see, it's at the hacking stage still.

[larsbrinkhoff]

I suppose we should retain support for the old parallel port display?

I think adding to the existing dvlites.c is the way to go, rather than adding a new file. But I'm open to suggestions.

Should this be a runtime option? Or a configure "--enable" option?

[larsbrinkhoff]

This isn't quite ready for merging yet, but almost.

I made it a runtime option. If you give the command lights usb, it'll try to use the USB Panda Display.

It doesn't work on a KS10, and I still don't know why.

[larsbrinkhoff]

Now checked with KS10 and ITS.

[larsbrinkhoff]

Ok, merging this then.