Re: [sv-ec] covergroup sample method

From: Michael Burns <michael.burns_at_.....>
Date: Fri Oct 12 2007 - 16:39:14 PDT
The problem is that we want to trigger coverage from within assertions using 
data in assertion local variables. I don't believe we are allowed to declare 
covergroups within assertions.

--Mike

David Scott wrote:
> Mike,
> 
> What's wrong with overriding using construction of specific covergroup 
> instances?  For example:
> 
>  covergroup cg (ref int x);
>    coverpoint x;
>  endgroup
>  ...
>  cg cv_myLocalInt = new(myLocalInt);
>  ...
>  cg cv_someOtherInt = new(someOtherInt);
> 
> If you want to have the data for both of these merged together, the 
> "merge_instances" option that we just approved -- Mantis 1897 -- could 
> suffice.  The only difference from your proposal is that you have to 
> keep around multiple instances of the covergroup and use the appropriate 
> one for your desired "override".
> 
> -- David S
> 
> 
> Michael Burns wrote:
>>
>> Hi folks,
>>
>> We've found a need in our methodology for an extension to the 
>> covergroup sample method. Basically, we'd like to be able to call the 
>> sample method on a covergroup with overrides so that we can specify 
>> the data we want covered in the call, rather than just sampling 
>> whatever the covergroup is connected to. This allows us to gather data 
>> in local variables and dump it into the covergroup to be covered. This 
>> is useful in assertion-based methodologies and perhaps elsewhere.
>>
>> There is a detailed proposal in Freescale, but I don't have it in hand 
>> yet. I wanted to bring this up now so we have a chance to consider it 
>> in Monday's meeting if possible.
>>
>> The general idea is that if you had a covergroup like this:
>>
>>   covergroup cg (ref int x);
>>     coverpoint x;
>>   endgroup
>>
>> you could explicitly specify data to be sampled like so:
>>
>>   cg.sample(.x(myLocalInt));
>>
>> where the args to sample() are overriding the formal arguments of the 
>> covergroup.
>>
>> Has this need been addressed in some other way I'm not aware of? Does 
>> this seem like a feasible approach?
>>
>> --Mike
>>
>>
> 


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Fri Oct 12 16:40:07 2007

This archive was generated by hypermail 2.1.8 : Fri Oct 12 2007 - 16:40:15 PDT