Calculating the Checksum

[David A. Cooley]

Hi Don,
The checksum is a hex number that is arrived at from the data in the
eprom... Changing a bit changes the checksum.  The data is checksummed, and
the checksum value is also burnt into the chip... The ECM verifies that the
data checksum and the stored checksums match to ensure the data is "good".
Not sure what the calc is to arrive at the checksum though.
Later,
DAve

-----Original Message-----
From: Don.F.Broadus at ucm.com <Don.F.Broadus at ucm.com>
To: diy_efi at efi332.eng.ohio-state.edu <diy_efi at efi332.eng.ohio-state.edu>
Date: Friday, September 11, 1998 3:40 PM
Subject: RE: Calculating the Checksum


> Could you please explain  checksum for me.  I thought checksum was a
>hex value
> That was used to verify if an EPROM had the correct program  data.
>
>
>Thanks  Don
>
>
>
>
>
> -----Original Message-----
> From: Roger Heflin [SMTP:rah at horizon.hit.net]
> Sent: Friday, September 11, 1998 12:55 PM
> To: diy_efi at efi332.eng.ohio-state.edu
> Subject: Re: Calculating the Checksum
>
>
>
> A couple of questions.  I have a 93 Z28  (32kb) eprom.  I am working
>on the
> disassembly.   I have put the eprom in memory at several locations,
>the
> location that seems to work is $8000, with it there, the code seems
>to
> be writing to things listed as ram in my databooks, any only reading
>from
> things that are known to be rom.   Is this a resonable method to
>use?
>
> Another thing, I assume the checksum code is in the build in rom,
>and not
> in the rom I have, since you really can't validate a checksum with
>code
> you don't even know if works.
>
> Mostly right now I am just decoding what looks like initialization
>stuff.
> I assume that probably the block learns/integrators would be
>initialized to
> 128. I also would guess that alot of the other things would be
>initialized to
> 0.  So far things look pretty simple, but I am only into the simple
>stuff.
>
> I have a 68hc11 disassembler program, a detailed 6809 instruction
>book and
> decoder card, and a 1988 vintage motorola single chip microcomputer
>data
> book that lists quite a few variants that all use pretty much the
>same
> instruction set with a few additions, and extra ram/rom/portio/ad
>varing
> depending on what was specified.    I would guess that GM is using a
>
> custom one of these chips, probably with a burned into the chip rom,
>for
> default operation.
>
> Roger
>
> On Fri, 11 Sep 1998, Georg Faustmann wrote:
>
> > If the system is a Delco unit, then the Checksum on/off byte may
>be at
> > $0008, set this to $AA, the checksum (stored at $0006,$0007)
>should be the
> > total of all the bytes from $0008 to then end of the chip.
> >
> > In some cases the address of the checksum switch could be offset,
>eg $3008,
> > $8008.
> >
> > Notes on how to find checksums:-
> >
> > If you had access to an emulator with a Romwatch (real time trace)
>type
> > function where the actual bytes being read at that moment in time
>are
> > highlighted, then you can sit there and watch where the bytes
>start to be
> > sequentially accessed by the CPU (usually every 10 seconds or so),
>you can
> > then scroll down until you find out where they stop being
>accessed, and
> > will give you the start and end point of the checksum. Then add
>these range
> > of numbers up, and look for a 16 bit value that is the last 16
>bits of this
> > sum, and search for this number outside of the checksum range,
>because it
> > can't be in the checksum area.
> >
> > Georg
> >
>