Re: [Tinycc-devel] Stack based target machine

[Nick Kelsey]

Hi Anton,

Very interesting.

Have you looked at GNU cgen?
It is a related concept in some ways - you write a script to describe 
the architecture and it generates a big chunk of the assembler port 
(GAS) and generates the simulator code for running c-torture simulation 
(via GDB I think). The catch - the script is a Scheme script and 
compiled with Guile... completely unreadable :-)

With your application I can see two big advantages:
1) The script is highly readable - huge plus
2) The host VM code will integrate well with other applications (ie not 
targeted only for use with building a new GNU toolchain).
A little bit of generic wrapper code and the VM could be linked to 
provide a simulator that fits the GNU testbed model.
The question would be how practical/interesting/useful would it be to 
extend to include a description of the assembler language and getting it 
to generate the bulk of the assembler port?

On the subject of TCC am making progress - TCC is simple and elegant 
once you get your head around a few things. I will post back once the 
core of the backend is in a more useful state. Looking at generated code 
there are some obvious situations where adding another instruction to 
the instruction set will help... I am sure there are lots more that are 
not so obvious... if you are interested I would welcome your help.

Nick

Anton Ertl wrote:

> *This message was transferred with a trial version of CommuniGate(tm) Pro*
> Nick Kelsey wrote:
>  
>
> > Unfortunately they have insufficient resources to run TCC natively, so I 
> > am looking at compiling to a virtual machine.
> >    
>
> In November I wrote here about using Vmgen
> <http://www.complang.tuwien.ac.at/anton/vmgen/> to build a stack-based
> VM for tcc.  For such a project I would volunteer to do the work
> having to do with Vmgen, but I would need to cooperate with someone
> who finds his way around tcc and can deal with the object file
> format.
>
> For your project, Vmgen would not help you for the VM interpreter on
> the embedded platform, but if you also want an interpreter on the
> server side for debugging the programs or the tcc port, it might still
> be worth it.
>
>  
>
> > I am leaning towards the idea of a stack based VM to keep things simple 
> > on the VM side (ie no general purpose registers thus no register 
> > decoding and addressing required by the host cpu). In reality the 
> > hardest aspect will be the compiler port so that will dictate the VM 
> > architecture more than anything else.
> >    
>
> Yes.  I am more a believer in stack-based VMs myself, but in your case
> a register-based VM might be beneficial; however, it's probably best
> to do what fits easiest in tcc.
>
> - anton
>
>
> _______________________________________________
> Tinycc-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/tinycc-devel
>
>