In reply to @VlsiYJ:
Just sharing my thoughts here :
- The idea of checking “connections” with assertions is laborious.
- The performance is going to suffer, as you will have way too many assertions.
- You may end up making the same mistake in the assertion, that you may be doing when connecting the signals. Assertions work best when the “coder” and “assertion writer” are two different people. Having two people for verifying just connectivity isn’t worth it :) A peer review may be faster and optimal.
- In most of my designs, if I have messed up a connection, the test/simulation will fail in some unexpected way, and can be debugged to a incorrect connection. If your test doesn’t fail with incorrect connections then it implies either your stimulus isn’t good enough, or the incorrect connection doesn’t really matter for functionality.
- Protocol checkers/assertions are usually sufficient to identify issue with the DUT not driving/responding to AXI signals correctly. The point here is that higher level assertions at the monitor/DUT interface are more efficient, and give you more ROI.
Disclaimer, just my thoughts. Your exact usecase may be different and may warrant assertions for connections.