Hey Rob,

I've rebuilt my SBBS docker image and on buster it core dumps. On stretch it seems to run fine.

I did re-build it without RELEASE=1, however, it doesnt seem to have created debugging systems (from the gdb output) - but here is a backtrace:

Reading symbols from exec/sbbs...(no debugging symbols found)...done.
[New LWP 21]
[New LWP 14]
[New LWP 19]
[New LWP 20]
[New LWP 18]
[New LWP 15]
[New LWP 16]
[New LWP 17]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `sbbs'.
Program terminated with signal SIGILL, Illegal instruction.
#0 0x00007f058492a4a0 in aclConsistent.isra.1.part.2 () from /opt/sbbs/exec/libsbbs.so
[Current thread is 1 (Thread 0x7f057a7fc700 (LWP 21))]
(gdb) bt
#0 0x00007f058492a4a0 in aclConsistent.isra.1.part.2 () from /opt/sbbs/exec/libsbbs.so
#1 0x00007f058492b1e1 in initAttributeACL () from /opt/sbbs/exec/libsbbs.so
#2 0x00007f05848f0629 in krnlBeginInit () from /opt/sbbs/exec/libsbbs.so
#3 0x00007f058485df14 in initCryptlib () from /opt/sbbs/exec/libsbbs.so
#4 0x00007f0584858109 in cryptInit () from /opt/sbbs/exec/libsbbs.so
#5 0x00007f058484ab99 in internal_do_cryptInit () from /opt/sbbs/exec/libsbbs.so
#6 0x00007f0583be4759 in __pthread_once_slow (once_control=0x7f0584c72728 <crypt_init_once>, init_routine=0x7f058484ab90 <internal_do_cryptInit>)
at pthread_once.c:116
#7 0x00007f058484b177 in do_cryptInit () from /opt/sbbs/exec/libsbbs.so
#8 0x00007f058484b1f3 in get_ssl_cert () from /opt/sbbs/exec/libsbbs.so
#9 0x00007f0584253249 in web_server () from /opt/sbbs/exec/libwebsrvr.so
#10 0x00007f0583bdd4a4 in start_thread (arg=0x7f057a7fc700) at pthread_create.c:456
#11 0x00007f0583082d0f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:97

My build line is:
make USE_DOSEMU=1 NO_X=1 JSINCLUDE=/usr/include/js JSLIB=mozjs185 SBBSDIR=/opt/sbbs install

Any ideas?
...ëîå*

... There is an order of things in this universe. Apollo, stardate 3468.1.

---
þ Synchronet þ Alterant | an SBBS in Docker on Pi!

Re: Coredump on Buster
By: Alterego to Digital Man on Wed Sep 18 2019 08:32 pm

Hey Rob,

I've rebuilt my SBBS docker image and on buster it core dumps. On stretch it seems to run fine.

I did re-build it without RELEASE=1, however, it doesnt seem to have created debugging systems (from the gdb output) - but here is a backtrace:

Reading symbols from exec/sbbs...(no debugging symbols found)...done.
[New LWP 21]
[New LWP 14]
[New LWP 19]
[New LWP 20]
[New LWP 18]
[New LWP 15]
[New LWP 16]
[New LWP 17]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by `sbbs'.
Program terminated with signal SIGILL, Illegal instruction.
#0 0x00007f058492a4a0 in aclConsistent.isra.1.part.2 () from /opt/sbbs/exec/libsbbs.so
[Current thread is 1 (Thread 0x7f057a7fc700 (LWP 21))]
(gdb) bt
#0 0x00007f058492a4a0 in aclConsistent.isra.1.part.2 () from /opt/sbbs/exec/libsbbs.so
#1 0x00007f058492b1e1 in initAttributeACL () from /opt/sbbs/exec/libsbbs.so #2 0x00007f05848f0629 in krnlBeginInit () from /opt/sbbs/exec/libsbbs.so
#3 0x00007f058485df14 in initCryptlib () from /opt/sbbs/exec/libsbbs.so
#4 0x00007f0584858109 in cryptInit () from /opt/sbbs/exec/libsbbs.so
#5 0x00007f058484ab99 in internal_do_cryptInit () from /opt/sbbs/exec/libsbbs.so
#6 0x00007f0583be4759 in __pthread_once_slow (once_control=0x7f0584c72728 <crypt_init_once>, init_routine=0x7f058484ab90 <internal_do_cryptInit>)
at pthread_once.c:116
#7 0x00007f058484b177 in do_cryptInit () from /opt/sbbs/exec/libsbbs.so
#8 0x00007f058484b1f3 in get_ssl_cert () from /opt/sbbs/exec/libsbbs.so
#9 0x00007f0584253249 in web_server () from /opt/sbbs/exec/libwebsrvr.so #10 0x00007f0583bdd4a4 in start_thread (arg=0x7f057a7fc700) at pthread_create.c:456
#11 0x00007f0583082d0f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:97

My build line is:
make USE_DOSEMU=1 NO_X=1 JSINCLUDE=/usr/include/js JSLIB=mozjs185 SBBSDIR=/opt/sbbs install

Any ideas?

Looks like a cryptlib issue. Perhaps it's the input (ctrl/cryptlib.key or ctrl/ssl.cert) - maybe try moving those files as a test? Or maybe cryptlib just doesn't like something about that environment.

digital man

Synchronet/BBS Terminology Definition #53:
PETSCII = PET Standard Code of Information Interchange (a.k.a. CBM ASCII) Norco, CA WX: 72.3øF, 70.0% humidity, 0 mph WSW wind, 0.00 inches rain/24hrs ---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net

Re: Coredump on Buster
By: Digital Man to Alterego on Wed Sep 18 2019 09:39 am

Looks like a cryptlib issue. Perhaps it's the input (ctrl/cryptlib.key or ctrl/ssl.cert) - maybe try moving those files as a test? Or maybe cryptlib just doesn't like something about that environment.

BTW: I tried with both my existing SBBS data, and a "new" environment (ie: no existing data) - both core dump.

So in the later case, the initial start up would have created the key and cert right?

I also rebuilt using buster as a base (no libmoz - so differeint make options) and still a coredump on startup.
...ëîå*

... Peace was the way. Kirk, The City on the Edge of Forever, stardate unknown

---
þ Synchronet þ Alterant | an SBBS in Docker on Pi!

Re: Coredump on Buster
By: Alterego to Digital Man on Thu Sep 19 2019 06:05 am

Re: Coredump on Buster
By: Digital Man to Alterego on Wed Sep 18 2019 09:39 am

Looks like a cryptlib issue. Perhaps it's the input (ctrl/cryptlib.key or ctrl/ssl.cert) - maybe try moving those files as a test? Or maybe cryptlib just doesn't like something about that environment.

BTW: I tried with both my existing SBBS data, and a "new" environment (ie: no existing data) - both core dump.

So in the later case, the initial start up would have created the key and cert right?

Yes.

I also rebuilt using buster as a base (no libmoz - so differeint make options) and still a coredump on startup.

As a work-around, you could try disabling all SSH and TSL support.

digital man

Synchronet/BBS Terminology Definition #26:
FF = Form Feed (ASCII 12, Ctrl-L)
Norco, CA WX: 84.4øF, 42.0% humidity, 4 mph WSW wind, 0.00 inches rain/24hrs ---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net

Re: Coredump on Buster
By: Digital Man to Alterego on Wed Sep 18 2019 02:00 pm

As a work-around, you could try disabling all SSH and TSL support.

So I dont believe I have anything TLS enabled - I did disable SSH and winner.

So cryptlib needs some fixes?

I can test stuff pretty easily here if want to try stuff to debug this... (it'll be in docker...)
...ëîå*

... Curiosity killed the cat, but for a while I was a suspect.

---
þ Synchronet þ Alterant | an SBBS in Docker on Pi!

Re: Coredump on Buster
By: Alterego to Digital Man on Thu Sep 19 2019 03:29 pm

Re: Coredump on Buster
By: Digital Man to Alterego on Wed Sep 18 2019 02:00 pm

As a work-around, you could try disabling all SSH and TSL support.

So I dont believe I have anything TLS enabled - I did disable SSH and winner.

So cryptlib needs some fixes?

Perhaps.

I can test stuff pretty easily here if want to try stuff to debug this... (it'll be in docker...)

Okay. Cryptlib has its own maintainers and support mailing lists if you want to persue getting it fixed. https://www.cs.auckland.ac.nz/~pgut001/cryptlib/

digital man

This Is Spinal Tap quote #13:
Nigel Tufnel: You can't really dust for vomit.
Norco, CA WX: 63.5øF, 75.0% humidity, 0 mph S wind, 0.00 inches rain/24hrs
---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net