Page 1 of 1
Tracing in log some address with Ostrich Moates 2.0 and Nist
Posted: Wed Aug 26, 2009 6:55 pm
by RomChip200
I need to trace some specific adresses of variables, being accessed.
I know Ostrich Moates 2.0 is able to do that. How is it integrated in Nistune ?
Can I re-use the Map tracing feature (I dunno when this comes in play during logging) or trace region ...
Posted: Thu Aug 27, 2009 2:01 am
by Matt
you can trace it inside the hex editor. open it up and the cursor will float around all map areas currently displayed
or you can define specific address file regions and it will trace those also (it traces all defined maps in the address file)
the implementation with h/w maptrace does have trouble when you are trying to write the moates during tracing. the start/stop/start of the trace stream to write emulator bytes eventually can kill itself. i've still got that on my bug list to investigate and fix
Posted: Thu Aug 27, 2009 2:05 am
by RomChip200
Let me clarify.
1) I don't writre the moates during tracing.
2) I don't want to highlight the adresses being accessed in real time BUT I need the values contained in these adresses in real time and more, in the Consult log recorder too to be able to make the link between TP, AFR, timing (...) and the value of these variables for each sample.
I need to get the values to track the knock stuff.
Matt wrote:you can trace it inside the hex editor. open it up and the cursor will float around all map areas currently displayed
or you can define specific address file regions and it will trace those also (it traces all defined maps in the address file)
the implementation with h/w maptrace does have trouble when you are trying to write the moates during tracing. the start/stop/start of the trace stream to write emulator bytes eventually can kill itself. i've still got that on my bug list to investigate and fix
Posted: Thu Aug 27, 2009 2:11 am
by Matt
Moates will only return the addresses that are being accessed. It will sample a collection of addresses in a given time, compare them against the window we specify and then return those to us
We then float an overlay cursor around the hex/map window for the addresses accessed
There is no way to know the value of a traced address with moates
You need a logic analyser for that (this is how I trace through code whilst ECU is running and make a full disassembly including address and data used (including RAM data))
Posted: Thu Aug 27, 2009 3:19 am
by RomChip200
There's another way to do it with Nistune
:
Today, when logging to Consult, you specify the registers you want to log.
You could add custom/user adress values (the user enters them) and these adresses would be read regularly with classical C9 command and recorded in the log file as TP, timing ... are.
It sounds good, no ?
Posted: Thu Aug 27, 2009 8:39 pm
by Matt
What about the Consult RAM trace window? That does the same thing - displays the register address values you are interested in. I used that a lot for finding new registers with the type 1 boards during development phase
Posted: Thu Aug 27, 2009 8:42 pm
by RomChip200
Yes, good first way but how to include the values read in the CSV file !?
With Consult RAM trace window, I'm obliged to have my eyes focused on the display all the time.
Matt wrote:What about the Consult RAM trace window? That does the same thing - displays the register address values you are interested in. I used that a lot for finding new registers with the type 1 boards during development phase
Posted: Thu Aug 27, 2009 8:50 pm
by Matt
if you want logging, then for Type 2+ ecus you just hack the register area in the ROM file upload with an emulator and then you can log it
yeah wont be adding logging from the RAM trace window
Posted: Thu Aug 27, 2009 8:52 pm
by RomChip200
Will do that then, no problem
Matt wrote:if you want logging, then for Type 2+ ecus you just hack the register area in the ROM file upload with an emulator and then you can log it
yeah wont be adding logging from the RAM trace window