pluto# uname -a
NetBSD pluto 1.0A NetBSD 1.0A (PLUTO) #0: Tue May  2 19:22:06 EDT 1995     ken@pluto:/usr/src/sys/arch/i386/compile/PLUTO i386

supped Tue May 2.

[captured using serial console]
pluto% ./pcomm
panic: npxinit: masked
Stopped at      _Debugger+0x4:  leave
db> t
_Debugger(f8114a54,f81923f2,f7bffcd4,f7bffe4c,f7bffcd4) at _Debugger+0x4
_panic(f81923f2,f7bffcdc,f81922ed,f7bffd1c,f8100a67) at _panic+0x3b
_npxinit(f7bffd1c,f8100a67,f8700010,f8720010,f7bffe4c) at _npxinit+0x33
_npxdna(f8700010,f8720010,f7bffe4c,f8701200,f7bffd1c) at _npxdna+0x2d
_Xdna(a000,f8701200,f7bffe4c,e57e0,f819d8fc) at _Xdna+0x1b
_cpu_coredump(f86e6700,f8701200,f8727400,f7bffe4c,b) at _cpu_coredump+0x39
_coredump(f86e6700) at _coredump+0x2cf
_sigexit(f86e6700,b,106,f86e6700,239e6) at _sigexit+0x2b
_postsig(b) at _postsig+0x92
_trap() at _trap+0x5b0
--- trap (number 6) ---
0x239e6:
db> show registers
es                0x10
ds                0x10
edi         0xf7bffe4c  _kstack+0x1e4c
esi         0xf8701200  _end+0x54d684
ebp         0xf7bffcb0  _kstack+0x1cb0
ebx              0x100
edx              0x2fa
ecx         0xf8114a5f  _logioctl+0xd3
eax                  0
eip         0xf817ee0c  _Debugger+0x4
cs          0xf7bf0008
eflags           0x282
esp         0xf7bffcb0  _kstack+0x1cb0
ss                0x10
_Debugger+0x4:  leave
db>
[end of transcript]

It's hard to repeat this one exactly.  I would run my hacked-up
version of pcomm compiled with ncurses 1.9 (happed with both -static
and my libncurses.so.0.1), it would hit a segmentation fault in some
routine in ncurses and coredump() would panic.

I fixed the ncurses problem (it was a usage error caused by a bad
#ifdef) and of course the panic goes away, but it's really still out
there.