Re: [sv-ac] SVA: Issue with coverage // need for a "full_paths" coverage?

You would need the same for and if one of sides always passed vacuously

From: Ben Cohen [mailto:hdlcohen@gmail.com]
Sent: Wednesday, March 21, 2012 03:46 PM
To: sv-ac@eda-stds.org <sv-ac@eda-stds.org>
Subject: [sv-ac] SVA: Issue with coverage // need for a "full_paths" coverage?

Before I write a mantis, I would like to hear your comments on the following issue.
If a property succeeds only with the LHS of an or in my testcases, there could be a case where the RHS is false if it were exercised, but my testcases have not exercised the RHS. My concern is the thoroughness of the testcases. When I write coverage, how can I be assured that all sequences or properties were covered?
If I have the following:
cover property(a |-> seq1 or seq2);
assert property(a |-> seq1 or seq2);
  I am not guaranteed that my testcases have covered both seq1 and seq2 even if the assertion is covered. This is because if seq1 succeeds, coverage s true.
With vacuity in something like
cover property(a |-> (c |-> seq1) or (d->seq2)) ;
I have the same issue.
That puts a penalty on the writers of coverage code to have to split the cover statements to the cover of sub-sequences or sub-properties.
It would be nice to have a qualifier that ensures that a successful coverage means that all sub-sequences or sub-properties vere covered.
For example:
cover property(full_paths(a |-> seq1 or seq2));
cover property(full_paths(a |-> prop1 or prop2));
Ben Cohen

--
This message has been scanned for viruses and
dangerous content by MailScanner<http://www.mailscanner.info/>, and is
believed to be clean.

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.