port-i386/52525: "i82489_icr_wait: busy" panic on entry to ddb

[gson%gson.org@localhost (Andreas Gustafsson)]

>Number:         52525
>Category:       port-i386
>Synopsis:       "i82489_icr_wait: busy" panic on entry to ddb
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    port-i386-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Sep 04 17:40:01 +0000 2017
>Originator:     Andreas Gustafsson
>Release:        NetBSD-current, source date 2017.07.21.02.51.12
>Organization:

>Environment:
System: NetBSD
Architecture: i386
Machine: i386
>Description:

Trying to debug a system lock-up, I tried to enter DDB using the
keyboard.  This caused an immediate panic and crash dump.  Examining
the dump with gdb shows:

  (gdb) target kvm netbsd.2.core
  0xc011dd95 in maybe_dump (howto=260) at /bracket/prod/current/src/sys/arch/i386/i386/machdep.c:708
  708     /bracket/prod/current/src/sys/arch/i386/i386/machdep.c: No such file or directory.
  (gdb) where
  #0  0xc011dd95 in maybe_dump (howto=260) at /bracket/prod/current/src/sys/arch/i386/i386/machdep.c:708
  #1  cpu_reboot (howto=howto@entry=260, bootstr=bootstr@entry=0x0) at /bracket/prod/current/src/sys/arch/i386/i386/machdep.c:729
  #2  0xc092c540 in vpanic (fmt=fmt@entry=0xc0f59aa9 "i82489_icr_wait: busy", ap=ap@entry=0xd83a8cf0 "") at /bracket/prod/current/src/sys/kern/subr_prf.c:342
  #3  0xc092c5ca in panic (fmt=fmt@entry=0xc0f59aa9 "i82489_icr_wait: busy") at /bracket/prod/current/src/sys/kern/subr_prf.c:258
  #4  0xc015e40f in i82489_icr_wait () at /bracket/prod/current/src/sys/arch/x86/x86/lapic.c:746
  #5  0xc015e461 in i82489_ipi (vec=240, target=786432, dl=0) at /bracket/prod/current/src/sys/arch/x86/x86/lapic.c:818
  #6  0xc011a110 in db_suspend_others () at /bracket/prod/current/src/sys/arch/i386/i386/db_interface.c:153
  #7  kdb_trap (type=type@entry=1, code=code@entry=0, regs=regs@entry=0xd83a8e1c) at /bracket/prod/current/src/sys/arch/i386/i386/db_interface.c:227
  #8  0xc011f992 in trap (frame=0xd83a8e1c) at /bracket/prod/current/src/sys/arch/i386/i386/trap.c:314
  #9  0xc0116f5f in alltraps ()
  #10 0xd83a8e1c in ?? ()
  #11 0xc07a4a05 in internal_command (ksym2=<optimized out>, ksym=<optimized out>, type=<synthetic pointer>, sc=0x6) at /bracket/prod/current/src/sys/dev/wscons/wskbd.c:1591
  #12 wskbd_translate (id=0xc12bb960 <wskbd_console_data>, type=type@entry=2, value=value@entry=41) at /bracket/prod/current/src/sys/dev/wscons/wskbd.c:1719
  #13 0xc07a4cf8 in wskbd_input (dev=0xc202a108, type=2, value=41) at /bracket/prod/current/src/sys/dev/wscons/wskbd.c:684
  #14 0xc03a73fe in ukbd_decode (sc=0xc20e0c48, ud=<optimized out>) at /bracket/prod/current/src/sys/dev/usb/ukbd.c:884
  #15 0xc0910b6f in callout_softclock (v=0x0) at /bracket/prod/current/src/sys/kern/kern_timeout.c:738
  #16 0xc0904a49 in softint_execute (l=<optimized out>, s=2, si=0xd831d074) at /bracket/prod/current/src/sys/kern/kern_softint.c:589
  #17 softint_dispatch (pinned=0xc1d7bd20, s=2) at /bracket/prod/current/src/sys/kern/kern_softint.c:871
  #18 0xc0100ffb in Xsoftintr ()
  #19 0xc1d7bd20 in ?? ()
  Backtrace stopped: previous frame inner to this frame (corrupt stack?)

I see that another user has reported a similar problem in 

  http://mail-index.netbsd.org/port-i386/2015/10/27/msg003401.html

The machine in case is an old Compaq Deskpro EN.  It's now back in
production use, so I can't easily test patches.

>How-To-Repeat:

Attempt to enter ddb on a Compaq Deskpro EN running NetBSD/i386.

>Fix: