core-v-verif
core-v-verif copied to clipboard
openhwgroup
Functional coverage needs to cover issue highlighted in 1193
Task Outcome
Check if the ISA DVplan and functional coverage covers the following special case, expand it if it doesn't.
The case where the core is woken up from a locally enabled, but globally disabled interrupt, where the given interrupt is a 1 cycle high value, needs to be covered. The case is currently know to show up rarely in random generated instruction streams, and we need to make sure we have checked.
Background information
See core-v-verif 1193 for details
Minimum Completion Criteria
An update to the coverage model that checks for the special case is PRed.
Nominal Completion Criteria
Consideration of, and addition of new coverage goals, for special cases related to the above special case.
Hi, I made the following modifications to this task. First, the sentence
Check if the functional coverage covers the following special case...
was amended to:
Check if the ISA DVplan and functional coverage covers the following special case...
Second, the Completion Criteria was expanded. This is a very interesting case and we can thank the Gods of Constrained-Random Stimulus that we hit it. Issue #1193 is a hint that we should be looking for other special cases related to this one.
I am looking for an opinion about which DVplan should be updated. It seems that the Interrupts DVplan is the natural choice, but a case could be made for putting this in the generic ISA DVplans.
