Re: gEDA-user: iverilog parse bug?

[John Sheahan <jrsheahan@optushome.com.au>]

try the same thing with cver.
for me, iverilog delay and SDF backannotation processing is not quite 
there yet at gate level.
Does a great job with  RTL though.
john


Laurin Blacken wrote:

> This may have already been reported as a bug, but I'll send this in 
> just in case.
>
> I'm trying to simulate a placed and routed xilinx design from which 
> the xilinx tool has generated a gate level module.
> Iverilog is having trouble with the xilinx sim primitives that get 
> instantiated..
>
>
> iverilog -y /usr/local/xilinx/verilog/src/simprims zsimtest.v
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:17: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:17: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:18: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:18: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:19: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:19: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:20: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_TRI.v:20: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_INV.v:17: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_INV.v:17: warning: choosing 
> typ expression.
> /usr/local/xilinx/verilog/src/simprims/X_FF.v:37: parse error
> /usr/local/xilinx/verilog/src/simprims/X_FF.v:31: error: syntax error 
> in specify block
> zsimtest.v:14: error: Unable to bind wire/reg/memory `ztest' in 
> `zsimtest'
> 1 error(s) during elaboration.
> ...
>
> It appears to me that there is a problem with the specify block?
>
> I've got icarus running on 3 different platforms: RH Linux, Cygwin and 
> Mac OSX -same thing on all three.
>
> My goal is to simulate the gate level fpga module with  back-annotated 
> sdf info.
>
> I can send the .v files for the simprims in question if you don't 
> already have them. I can send the gate level module and tbench if you 
> think that would be necessary.
>
> Is it something I'm doing or is it a bug?
> Is there a work around?
>
> Laurin Blacken