lgy@blake.acs.washington.edu (Laurence Yaffe) (10/06/89)
The RC-2030 (i.e., a system w/o any display) that I received last week
has a strange problem which is preventing me from allowing the system console
(a.k.a. /dev/tty1) to be used for terminal access by normal users. Typing a
Control-X (at any time) causes the appearance of the prompt:
Enter cmd (help):
If you type "help", the following rather surprising menu appears:
help .......... display help screen
r .......... read iop memory
R .......... read R2000 memory
mr .......... read map register
w .......... write iop memory
W .......... write R2000 memory
mw .......... write map register
set .......... set/display variables
echo .......... display text
i .......... input from port
o .......... output from port
loop .......... repeat command (num or '~')
F .......... fill r2000 memory
go .......... set iop address
FMA .......... fill with addresses
IMA .......... fill with inverted addresses
scsi .......... dump scsi info
I have not experimented heavily with what can be done at this point - I doubt
that doing so is in the best interest of maintaining a healthy system. This
behavior is not observed on /dev/tty0, on network logins, or on the console
display of an RS-2030. This response to ^X occurs regardless of whether or
not you are logged in. Not surprisingly, the response is not captured by the
'script' process.
Can anyone explain what's going on? And, most importantly, how can this
screwball behavior be disabled?
--
Laurence G. Yaffe Internet: lgy@newton.phys.washington.edu
University of Washington Bitnet: yaffe@uwaphast.bitnetroe@sobmips.UUCP (r.peterson) (10/08/89)
From article <3940@blake.acs.washington.edu>, by lgy@blake.acs.washington.edu (Laurence Yaffe): > > The RC-2030 (i.e., a system w/o any display) [ ... ] > Control-X (at any time) [on /dev/tty1] causes: > > Enter cmd (help): > > If you type "help", the following rather surprising menu appears: > > help .......... display help screen > r .......... read iop memory > R .......... read R2000 memory > mr .......... read map register > w .......... write iop memory > W .......... write R2000 memory > mw .......... write map register > set .......... set/display variables > echo .......... display text > i .......... input from port > o .......... output from port > loop .......... repeat command (num or '~') > F .......... fill r2000 memory > go .......... set iop address > FMA .......... fill with addresses > IMA .......... fill with inverted addresses > scsi .......... dump scsi info > The R[CS]2030 have an I/O processor whose RS232 I/O is performed over tty1. This is why, for example, the power-on diagnostic test output is directed to tty1, not 0 - most of the tests run in the IOP. So, you are talking the the debug monitor code in the IOP itself. Any you are correct - messing about here is a good way to crash machines. A quick look around does not show how to shut this off - I'll look around and see what I can find. -- Roe Peterson {attcan,mcgill-vision,telly}!sobeco!roe
mohan@mips.COM (Ram Mohan Vedula) (10/12/89)
In article <1989Oct7.190837.12240@sobmips.UUCP> roe@sobmips.UUCP (r.peterson) writes: >From article <3940@blake.acs.washington.edu>, by lgy@blake.acs.washington.edu (Laurence Yaffe): >> >> The RC-2030 (i.e., a system w/o any display) [ ... ] >> Control-X (at any time) [on /dev/tty1] causes: >> >> Enter cmd (help): >> >> If you type "help", the following rather surprising menu appears: >> >> help .......... display help screen >> r .......... read iop memory >> R .......... read R2000 memory >> mr .......... read map register >> w .......... write iop memory >> W .......... write R2000 memory >> mw .......... write map register >> set .......... set/display variables >> echo .......... display text >> i .......... input from port >> o .......... output from port >> loop .......... repeat command (num or '~') >> F .......... fill r2000 memory >> go .......... set iop address >> FMA .......... fill with addresses >> IMA .......... fill with inverted addresses >> scsi .......... dump scsi info >> > >The R[CS]2030 have an I/O processor whose RS232 I/O is performed over >tty1. This is why, for example, the power-on diagnostic test output >is directed to tty1, not 0 - most of the tests run in the IOP. > >So, you are talking the the debug monitor code in the IOP itself. Any >you are correct - messing about here is a good way to crash machines. > >A quick look around does not show how to shut this off - I'll look around >and see what I can find. >-- > Roe Peterson > {attcan,mcgill-vision,telly}!sobeco!roe Press 'RETURN' to get out of the IOP menu. -- Mohan Vedula MIPS Computer Systems, Inc. {ames,decwrl,pyramid}!mips!mohan 750 Arques Ave. mohan@mips.com Sunnyvale, CA 94086 (408) 720-2946