Re: [Simulavr-devel] Howto UART, howto record pin behaviour

[ThomasK]

Hi Markus,

there are possibilities to fit your task. I use GDB very seldom for avr 
development. Mostly I use the script extensions, in my case python. For 
example for virtually running amforth :-) or the simulation a LIN 
network on AVR. But this needs too knowledge in python programming. 
Other script estension, which is possible, is TCL. But (my opinion) not 
often used and maybe you'll miss some parts there. (but this is possible 
to fix)

Of course, you can do the job also in C++ programming. But there you 
have a lot of work more to do. Advantage would be a higher performance.

If you want to use python, then a first start could be to look for the 
regression test suite and on the examples in examples directory. (but 
this is just a starter ... :-) )

> - Is installation of SimulAVR a requirement or is it fine to run it off
> the build location?

Yes, you can run it anywhere. Installation isn't necessary, but 
sometimes it's good to have a clean release independent on your simulavr 
workarea, where you, maybe, try something else!

> - Is SimulAVR reasonably cycle-accurate? Like if two pin changes are
> recorded 835 clock cycles apart, does this reasonably reliably mean the
> same happens on the real hardware at a time distance of 835 * 20 MHz =
> 41.75 us?

It should cycle-accurate! :-) But there are some exeptions, where it's 
not simple possible to fullfill all conditions. But a comparison between 
simulation and real hardware with logic analyser has showed up a good 
result. Timer output should be precise.

> - How would I set up a connection between the UART and a serial
> terminal/another application, using the (interrupt based) UART handling
> code of the firmware? UART interrupts can undoubtly have an influence on
> smooth operations on the real hardware, so this should be part of the
> simulation.

See above. There is also a UART simulation module for TCL extension, (in 
src/ui), maybe as a template for you. But hold in mind, that TCL 
extension isn't really bug-free.

> - I see the usage of special_output_port and similar things in
> http://www.nongnu.org/simulavr/intro.html#simple-example . Does this
> change the actual code, like in "this requires distinct compilations for
> the simulator and for the real hardware" or like in "this can change the
> firmwares' behaviour or timing"?[1]

This special output ports are ports, which not really exist. It's a good 
help to try or check some special cases, but useless, if you try to 
simulate production code! The mechanism there is to detect a port access 
and do something special, which is impossible in real world! So it 
changes your code and also timing.

> - Last, not least, how would I record I/O pin changes into a .vcd file?
> How would I define which pins are recorded?

Possible, see also timertests in regress directory. This could be a help 
to find out how.

cu, Thomas