Non-existent device

[arpruss]

I followed the direction as best I could by running make EMULATOR=simh, and then ./start, on my Raspberry PI 3. Here's what I got:

PDP-10 simulator V4.0-0 Current        git commit id: e1db7dec
out/simh/boot-6> at dz0 10004
Listening on port 10004
out/simh/boot-8> at dz0 line=7,10018
Line 7 Listening on port 10018
out/simh/boot-10> at tty line=6,10019
Non-existent device

 DSKDMP

I then typed its and got DIRNF.

Is there a step I missed?

[larsbrinkhoff]

The line that says at tty line=6,10019 has a mistake. The tty should be dz0. But that's not fatal.

Getting DIRNF is a problem. I will try to reproduce it myself.

Some things you can try:

• Rebuild using make clean all EMULATOR=simh.
• Get disk images directly from here: http://hactrn.org/images/
• Build using another emulator. Do make clean first.

[larsbrinkhoff]

I have confirmed I get the same result myself. Thanks for reporting! Will look into it soon.

[larsbrinkhoff]

I have located the bad commit: 36be2f753073fbf918ab25177647553b1518b101

[jeffhhk]

The title of this issue and the repro have two different root causes. I'm going to comment based on the title of the issue. I will follow up with comments on the higher-impact repro on a separate case.

When I build with:

make clean all EMULATOR=simh

The following code removes the (low impact) "Non-existent device" message:

diff --git a/build/mchn/DB/boot b/build/mchn/DB/boot
index e9d1d845..dd22deb1 100644
--- a/build/mchn/DB/boot
+++ b/build/mchn/DB/boot
@@ -7,7 +7,7 @@ at dz0 10004
 # VT52
 at dz0 line=7,10018
 # GT40
-at tty line=6,10019
+#at tty line=6,10019
 set rp0 rp06
 at rp0 out/simh/rp0.dsk
 b rp0
diff --git a/build/simh/boot b/build/simh/boot
index 08d2e656..207aa7ed 100644
--- a/build/simh/boot
+++ b/build/simh/boot
@@ -7,7 +7,7 @@ at dz0 10004
 # VT52
 at dz0 line=7,10018
 # GT40
-at dz0 line=6,10019
+#at dz0 line=6,10019
 # Dial-up
 at dz0 line=5,10015
 set rp0 rp06

[larsbrinkhoff]

Thanks for checking, @jeffhhk. The mistake is that the simh configuration refers to tty instead of dz0.

[rcornwell]

Solution to this can be: Boot using tape dskdmp. At DSKDMP run:

dskdmp bin

It will respond with DSKDMP. You can from then on boot from disk. I am not sure why the older version of DSKDMP is still in the DSKDMP swap space.

[larsbrinkhoff]

I haven't checked this, but I'm pretty sure what happens is that the tape DSKDMP writes itself to disk. So when DSKDMP BIN is run after that, it will write the new version that is supposed to be on disk. This became a problem because one more round of rebooting was introduced and the last reboot didn't invoke the new DSKDMP BIN.

[eswenson1]

We should create a new dskdmp tape as soon as we’ve assemble DSKDMP from current sources. And use that from then on. Using the one with DSKDMP RPA6 anytime thereafter would presumably break DSKDMP disk boot again.

[rcornwell]

There is an issue with dskdmp rp06 when dealing with emulators. The disk responds too quick and it ends up overwriting itself before it can finish loading.

[larsbrinkhoff]

There is another issue #2140 for updating the bootstrap DSKDMP.

Note that a new "ndskdmp.tape" is generated in the KS10 build.