State of the emulator

State of the emulator

Emulator bug list

  • The 'crash-to-bce-die' bug. Observed rarely, not readily reproducible.
  • C compiler seems excessively fragile.
  • 'Corrected master directory UID in header' messages.
  • pxss notify timeout (very rare, not readily reproducible)
  • Large card decks corrupted.
  • quota leak.
  • GTSS TSS (gtss)
  • act_ctl_: discrepancy
  • MPC poll code fails

MPC poll code fails as tape MPC does not grok read memory command. See mpc poll status] for details.

  • disk pack mount fails

See [] for details

  • disk special interrupt lost

See [] for details

Won't fix

  • cob crashes — can not reproduce

Can't reproduce; can't find original notes.

  • FNP does not grok negotiated echo. — won't fix

See echnego status for details

  • cpu_test — won't fix
  • cpu_test 51 seg_no_write fails under certain circumstances; I believe that this is bug in the test program, not the emulator.
M-> [auto-input] admin

r 16:43 43.286 4626

M-> [auto-input] test_cpu -sof -sel 49 51

Test 49   <cmpc_a>

Test 51   <sreg_no_write>

r 16:44 44.104 53

M-> [auto-input] test_cpu -sof -sel 50 51

Test 50   <cmpc_b>

Test 51   <sreg_no_write>

r 16:45 56.143 7

M-> [auto-input] test_cpu -sof -sel 49 50 51

Test 49   <cmpc_a>

Test 50   <cmpc_b>

Test 51   <sreg_no_write>

Error:  fault_tag_1 by sreg_no_write$|1774 
referencing stack_4|40040 (in process dir)
Ascii data where pointer expected.
r 16:46 113.985 39 level 2

See [] for details

  • BCE reports boot_delta exceeded — Works As Designed; field as Multics issue

BCE reports boot_delta exceeded, but the boot_delta field of the clok card is blank, and AM81 implies that blank disables the boot delta check.

'init_clocks.pl1' does not have a check for 'boot_delta == 0', so the message would be expected.

  • alm -list fails — Multics bug
alm bootload_tape_label
ALM 8.14
r 19:03 2.377 29

alm bootload_tape_label -list
ALM 8.14
alm: Phase error while generating the object segment
alm: Fatal error.  Translation aborted. >user_dir_dir>SysEng>Anthony>fromSource

See []

Resolved issues

  • Multics Video System — fixed
  • long lines — fixed

If, starting from a 132x50 terminal window (that is, 132 columns) and invoking emacs with:
emacs -ttp vt132 -ll 132 -pl 50
And you type all the way to character position 132 and then type beyond, eventually the system stops responding (the FNP stops echoing). And your process is stuck, and nothing further happens. Only recourse is to disconnect the telnet connection and reconnect.

  • Kermit has high error rate (might be FNP emulator issue).—fixed
  • Pascal compiler generates bad code — fixed
  • Multiple SCUs — fixed
  • MR12.3 install crashes when # of SCU > 1 — fixed
  • emacs does not work — fixed
  • zyre IPC fails intermittently; beacon stops on CPU. — fixed
  • bind_fnp command fails with error message. —fixed

Script test_bind_fnp.ini…

Multics MR12.5 - 01/05/90  0202.6 pdt Fri
M-> [auto-input] admin

r 02:03 37.521 4624

M-> [auto-input] create_dir >user_dir_dir>SysAdmin>a>mcs.7.6c

r 02:03 1.165 17

M-> [auto-input] cwd >user_dir_dir>SysAdmin>a>mcs.7.6c

r 02:03 0.974 12

M-> [auto-input] ted

M-> [auto-input] r >ldd>mcs>info>site_mcs.bind_fnp 

M-> [auto-input] w site_mcs.bind_fnp

M-> [auto-input] q

r 02:03 3.346 73

M-> [auto-input] ac x ([segs >ldd>mcs>o>*.archive -absp])

r 02:03 5.214 91

M-> [auto-input] bind_fnp site_mcs -list

bind_fnp:  Trace macro at dia_man.objdk|2404 does not begin with TRA instruction
bind_fnp: Binding unsuccessful
r 02:03 12.369 69
  • bce/go — fixed in c40bb192

The ring 4 command 'bce' suspends Multics and invokes bce; the bce 'go' command should restart Multics, but just stays in bce.

  • C compiler crashes — fixed in cd2ad6 (master) 871949 (0fnp)

See [] for details

  • read_mail crashes — fixed

See [] for details

  • set_acl crashes the Initializer — fixed

See []

  • 12.5 upgrade fails — fixed


  • Kermit does not work — fixed

Note: Kermit works with 0fnp, but not master.

  • Pascal programs don't run — fixed in Alpha 2.0 RC 5

See []

  • tape_nstd_ fails //— fixed in 5a589c99, d93ffbf0
io_call attach test tape_nstd_ foo
Tape foo,blk=2800 will be mounted with no write ring.
Tape foo,blk=2800 mounted on drive tapa_01 with no write ring.
r 10:39 4.285 3

io_call open test sqi
r 10:39 0.016 0

io_call read test -sm r1
io_call: Segment >user_dir_dir>SysEng>Anthony>r1 created.

Error:  out_of_bounds at nstd_dim$nstd_read|4166
referencing nstd_sdb2_|776000 (in process dir)
Attempt to access beyond end of segment.

status nstd_sdb2_

names:    nstd_sdb2_
type:               segment
date used:          08/25/93  1042.4 pdt Wed
date modified:      08/25/93  1042.4 pdt Wed
branch modified:    08/25/93  1039.4 pdt Wed
bit count:          0
records used:       255
mode:               rw

I wonder is there is a modulus bug in the emulator out-of-bounds checking ?

  • pl1 compiler error — fixed in Alpha2 .0 RC 5
ac x >ldd>sss>s>bound_command_loop_.s.archive print_ready_message_.pl1

pl1 -ot -list print_ready_message_.pl1
PL/1 33e

A conversion error occurred when 1e-6 was converted to its internal
SOURCE:   vcpu = (data.virtual_cpu - cpu) * 1e-6;
pl1: An error of severity 3 has occurred.
pl1: Translation failed. print_ready_message_.pl1

Misc. odd behaviors


Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License