Neil asks:
> On the subject of halts...
>
> Does anyone know where these things are documented?
> I know that I can often get an idea from the support database by looking
> for a similar halt, but is there a single repository with all halts
> documented.
At least some of these things are documented in the 5.0 version of
MPE/iX Error Messages Manual Volume II. They are located in Chapter 30
entitled System Startup, Shutdown, and Configuration.
This description of the error codes seen on the front panel LEDs divides them
into classes based on the leading hexadecimal digits thereof. The pertinent
entry for bf00 is:
-------------------------------------------------------------------
| | | |
| Err. | Code | Description |
| | | |
-------------------------------------------------------------------
| | | |
| B | Bxxx | Operating system software error. |
| | | |
-------------------------------------------------------------------
The beginning of the description for class error B is:
Class Error B: Initialization Messages
When the system halts, the reason for the halt can be seen on the SPU
control panel status display (hex display). The first number is the
source of the halt call, and following it are numbers of the form 0nxx
where n is a number that starts at 1 and increments. The reason can be
interpreted by stringing the xx values together to form a number.
When the monitor is the source of the halt, the first number is B000.
System abort uses B007. The following codes relate the cause of the
error.
Well, what can I say! These are not exactly initialization messages.
They aren't even the beginning of the end. More like the end of the end.
Don't be confused by the system abort comment. The error codes that follow
these paragraphs constitute a fairly complete listing for the monitor halts.
I've come across a few more, but I don't know if they are actually in use.
>
> Regards
>
> Neil
>
>
> > -----Original Message-----
> > From: HP-3000 Systems Discussion [SMTP:[log in to unmask]] On
> > Behalf Of Stan Sieler
> > Sent: Tuesday, February 25, 1997 11:08 PM
> > To: [log in to unmask]
> > Subject: Re: FLT error
> >
> > Donna writes:
> > > under the heading of 'it could never happen to me....' :-} i got the
> > > dreaded early morning phone call telling me my 3000 (918) had gone
> > > down. the response center wasn't able to tell me anything about the
> > > series of error messages i was getting, so i thought i ask here. in
> > > the bottom left-hand corner of the console, repeating in a cycle,
> > the
> > > system was reporting:
> > > flt bf00
> > > flt 0182
> > > flt 0200
> > > flt dead
> > >
> > > the best the rc could tell me was they thought the o/s had panicked,
> > > although they didn't know why, and had shut the system down. this
> > > happened around 6:30am and there were 2 users logged on - so the
> > > system load was probably (real) light. tia - d
> >
> >
> > Well, BF00 means "low level halt" (the "B" and "00"), presumably by
> > processor in slot $F.
Actually, Stan is right. The general form is Bxyy, where x is the module
number (using the parlance of _lockup2, the bugger that drives the LEDs) of
the cpu requesting this lock up and yy is the "machine halt code" being
reported. yy == 0 is a monitor halt or "low level halt" as Stan calls it.
The other machine halt code mentioned in the error message manual is the one
for system abort (namely, yy == 07). It should be mentioned that CM Sudden
Deaths use yy == 0b. One hardly ever sees other machine halt codes.
Several years ago I did see an ICS Stack Overflow (yy == 03) during system
testing.
> >
> > But, most all low-level halts I'm familiar with have codes between
> > $0001 and
> > $00fb. Yours is $8200 (the FLT 0182, 0200 is read as: here's a
> > 16-bit number...part 01 is $82, and part 02 (the low order part) is
> > $00).
_lockup2 optionally displays none, one, or two bytes worth of additional
information followed by dead at display intervals of 1.28 seconds each.
> >
> > If the number had been 0100 0282, or just 0182 (and no 02xx value),
_lockup2 supposedly suppressed leading zeros from being displayed, so that
just 0182 would be displayed.
> > then
> > that would be low-level halt $82, which would mean something like
> > "see section 5 of the processor ACD (Architecture Control Document) to
> > see what trap $82 - $80 (or, $2) is".
That would be a Power Failure Interrupt.
But, according to what Donna saw, the error code is 8200, not 82. The error
codes listed for the monitor halts in the error messages manual go as high
as 131. In reviewing the _lockup2 code, I came across some as high as 3a1,
but nothing close to 8200, so I don't have a clue as to what happened here.
> >
> > Did you take a memory dump?
> >
> > --
> > Stan Sieler
> > [log in to unmask]
> > http://www.allegro.com/sieler.htm
> > l
Joe Searle -- [log in to unmask]
The above are my personal observations and do not represent those of
my employer.
|
