Subject: RE: DirectC C-Layer: open arrays and abstract access
From: Kevin Cameron x3251 (Kevin.Cameron@nsc.com)
Date: Fri Dec 20 2002 - 12:27:30 PST
> From Joao.Geada@synopsys.com Fri Dec 20 11:51:47 2002
>
> Kevin said:
> > That's what's in my proposal. My working assumption is that access to new data types has to be available
> > through VPI (and/or PLI) anyway, so you might as well use the same mechanism for argument access from
> > called from SV - it makes your code more reusable and avoids having to spec two interfaces.
>
> But this is an assumption and I do not believe this assumption holds for SV
> 3.1
>
> Please note that there were no volunteers to the task of defining the new VPI model to encompass all
> the new stuff in SystemVerilog *and* no donation to the committee providing the ground work for this.
>
> We have to go with what we have, and in what we have VPI has no means of
> access to:
> - complex types (structs, unions)
> - all the primitive C types added to SV (char, byte, int, short, ...)
> - classes
> - any dynamic variable
> - etc
>
> Joao
That's fine, but Andrzej's proposal is new (not a donation). We should probably
revisit the requirements and decide if/when VPI access is going to be defined.
Since I'm not an expert on VPI I was leaving the definition of the VPI calls
for later, and just concentrating on the call linkage and non-abstract access
stuff.
There are objects which could be passed to C which are already covered by VPI,
are you proposing using something other than VPI for accessing them? If not,
shouldn't what we add look similar?
Kev.
This archive was generated by hypermail 2b28 : Fri Dec 20 2002 - 12:28:47 PST