dmk, Wed May  3 16:49:50 PDT 1995

This is really an RFE for the OS.   Once booted the OS handles the the
keyboard and serial console input. This should be reassiged to
kernel/other, or another os-related category.

Once the OS has taken over the console duering the boot sequence, if it
notices a STOP/A on the keyboard or a BREAK on the serial port console,
the zs driver calls a platform dependent routine named
abort_sequence_enter(), whose default action is to call debug_enter()
on desktops and on servers checks if the key switch is in the secure
position, and if it is, it ignores the abort-sequence, otherwise it
calls debug_enter().  debug_enter() drops you into the prom or the
standalone debugger (kadb) if you booted using a standalone debugger.

A trivial workaround is to use adb to no-op the call to debug_enter in
abort_sequence_enter.  The RFE is to add an /etc/system settable
variable to disable the abort-sequence, tested in
abort_sequence_enter. On servers the /etc/system variable
would override the key position test. (if either one is set, ignore
the abort sequence).

You'd still need to have a console present during the initial boot
sequence, since the prom handles console input until the
kernel gets past consconfig().

dmk, Mon Jul 22 16:59:12 PDT 1996

I integrated the OS side of this, PSARC/1995/439, into 2.6 build 10 (?).
Once the OS takes over control of the input devices, BREAK and L1/A
sequences can be disabled via new options in kbd(1) and are read
and used during the boot sequence from a new /etc/default/kbd file.


[ichiban: 11/4/97]

This bugfix has been integrated to 5.6 by dmk. I am closing this bug as
integrated.

167 munchkin(ichiban)% pwd
/ws/ref-sunos5.6/usr/src/uts/common/io

SCCS/s.kbd.c:

D 1.105 97/03/14 01:14:00 robinson      139 138 00014/00014/02214
Lint and cstyle
  :
[snip]
  :
D 1.99.1.1      96/06/06 18:51:46 dmk   134 132 00011/00001/02205
PSARC/1995/439 - abort-sequence-disable


 See also: 4038312

Hook 1: 
        Hook 2: 
        Hook 3: 
        Hook 4: 
        Hook 5: 
        Hook 6:

Submitter:                             blw Date: May  3 1995 12:00AM
        Dispatch operator:                 bugtraq Date: May  3 1995 12:00AM
        Acceptor:                         bugfixer Date: Jul 22 1996 12:00AM
        Evaluator:                             dmk Date: Jul 22 1996 12:00AM
        Commit operator:                   ichiban Date: Nov  4 1997 11:43AM
        Fix operator:                      ichiban Date: Nov  4 1997 11:43AM
        Integrating operator:              ichiban Date: Nov  4 1997 11:43AM
        Verify operator:                           Date: 
        Closeout operator:                         Date:

Customer:
        Company: Sun Microsystems Computer Corporation
        Employee: Brian Wong
        User Role: N, novalue
        User Type: I, Internal (SMI) Customer
        Release: 2.22
        Hardware Version: generic
        O/S version (unbundled products): s495_20
        SO Number: 
        Sun Contact: blw
        Contact Name: 
        Contact Mailaddr: 
    Customer:
        Company: Sun Belgium
        Employee: Herman Baumers
        User Role: N, novalue
        User Type: E, External Customer
        Release: 2.22
        Hardware Version: sun4u
        O/S version (unbundled products): 5.5
        SO Number: 5443837
        Sun Contact: hb@belgium
        Contact Name: 
        Contact Mailaddr: 
    Customer:
        Company: WSI CORPORATION
        Employee: Paul J. Bayer
        User Role: T, Technical Support
        User Type: E, External Customer
        Release: 5.5.1
        Hardware Version: sun4u
        O/S version (unbundled products): 5.5.1
        SO Number: 2803161
        Sun Contact: dpalko@east
        Contact Name: 
        Contact Mailaddr: 

 Escalation(s):

Some mechanism required to prevent power cycle on
console from dropping system into PROM.

SunWeb Home SunWeb Search SunSolve Home Simple Search

Sun Proprietary/Confidential: Internal Use Only
Feedback to SunSolve Team