True coverage: a goal of verification

Abstract

There are a number of RTL coverage tools on the market today that essentially tells you only that a set of signals has been toggled by a particular diagnostic test. This is useful in showing what areas of the RTL design are definitely not covered by the diagnostic test but tells you very little about the set of signals that have been toggled. In an extreme case a diagnostic test may fail to fail when anyone of these signals are in error. The following is a strategy for examining the coverage indicated by a commercial coverage testing software package and obtaining a truer picture of a diagnostic test's real coverage. This concept is extended to a full regression test suite.