CoverageIDType in Alert Results table

[David]

I’m working on a testbench using OSVVM and exploring how to track coverage properly. I’ve been using scoreboards and functional coverage, and I’m trying to understand how the CoverageIDType fits into the Alert Results.

When I track coverage of a signal using CoverageIDType, the Alert Results table always shows it with a pass value of 0. Then the coverage is shown per bin hit in the breakdown below.

Can you help me understand how the CoverageIDType is supposed to be used (and why its appearing in the results summary) please?

[Jim Lewis]

Hi David,
Your pictures did not post for some reason. So I will give a more general perspective.

When a CoverageID is created, a corresponding AlertLogID is also created. If the ID, created with NewID, it is used for tracking errors – such as when an error bin is encountered. This ID is also used to determine what level an error signals at: FAILURE, ERROR, or ERROR + No printing.

If the ID, created with NewReqID (Beta feature of 2025.02 – documentation just posted to dev branch – meaning I updated it after your post and before mine), it is used for tracking errors and treats coverage bins as requirements. The requirements goal is the sum of the number of coverage bins. The requirements count is the sum of the number of coverage bins with passing coverage goals.

The AlertLog report does end up with lots of extra stuff in it. In the call to NewID / NewReqID, ReportMode can be changed to either DISABLED (do not show in summary table) or NONZERO (only show in summary table if non-zero). AlertLogIDs report up, so even if there is an ERROR, it will be reported in the next higher level – which worst case is the DEFAULT bin.

For FIFOs, I always use ReportMode DISABLED in the call to NewID. FIFOs have one FAILURE, pop with an empty FIFO. In this case, the test ends with FAILURE from the FIFO and then the report. Printing that twice is not necessary.

Cheers,
Jim

[Author]

Posts