cvs commit: src/sys/alpha/alpha support.s src/sys/i386/i386 swtch.s src/sys/kern kern_shutdown.c src/sys/sys systm.h

[Bruce Evans]

On Sun, 22 Feb 2004, Dag-Erling [iso-8859-1] Sm=F8rgrav wrote:

> Jun Kuriyama <kuriyama at imgsrc.co.jp> writes:
> > I checked something around backtrace().  A db_print_backtrace() is
> > depending on DDB, but as Bruce said, DDB is dangerous for default
> > feature for -stable release.
>
> No, just make sure DDB_UNATTENDED is also the default.

DDB_UNATTENDED doesn't work right yet.  Its only effect is to clear
the debugger_on_panic variable, which controls entries to the debugger
on panics and on fatal traps.  The following entries to the debugger
are not controlled:
- someone attending the console and hitting the Debugger hotkey.
- stray Debugger() calls in source code.  There is now a non-stray one
  in watchdog_fire().
- stray breakpoints in source code or left over from previous debugging
  session.
- stray trace traps, perhaps from trace flags set in previous debugging
  session (these can be deep in stacks and hard to kill).

Bruce