Overview
ImperasTS-MMU is part of the ImperasTS family of RISC-V verification test suites. In the available evidence, it is grouped with ImperasTS-PMP and ImperasTS-ePMP as a set of directed suites for virtual memory and protection features.[1]
Verification role
ImperasTS-MMU is described as a directed test suite intended to target areas where random stimulus often leaves coverage gaps.[2] This positions it as a complement to constrained-random verification flows: random testing can expose broad behavior, while directed TS-MMU tests can be added to close specific virtual-memory coverage holes.
Targeted MMU scenarios
The evidence gives a specific example in which coverage analysis found weak points in Sv39 and Sv48 page table walks. Adding TS-MMU tests then exposed a subtle ordering issue in TLB flush logic.[3]
This supports the following technical characterization:
- Sv39 targeting: TS-MMU can be applied to coverage gaps involving Sv39 page table walks.[3]
- Sv48 targeting: TS-MMU can be applied to coverage gaps involving Sv48 page table walks.[3]
- Page-table-walk focus: The example explicitly identifies page table walks as a weak area that TS-MMU tests helped address.[3]
- TLB flush exposure: The example reports that TS-MMU tests exposed an ordering issue in TLB flush logic.[3]
Configuration
The test suites for vector, MMU, PMP, and ePMP are described as configurable to match the user's RISC-V processor.[4]
Use in a hybrid verification flow
The evidence describes a hybrid coverage-closure flow in which constrained-random sweeps are followed by functional coverage analysis, after which directed tests such as TS-MMU can be used to target weak spots.[5] In this role, ImperasTS-MMU contributes targeted closure for virtual-memory behavior rather than replacing random stimulus.
[1]: ImperasTS-MMU/PMP/ePMP are listed as directed suites for virtual memory and protection features in evidence chunk b73d6860-3caf-4eb7-811e-eda1693f60f3.
[2]: The evidence states that these directed suites efficiently target areas where random stimulus often leaves gaps in chunk b73d6860-3caf-4eb7-811e-eda1693f60f3.
[3]: The evidence states that weak points in Sv39 and Sv48 page table walks were addressed by adding TS-MMU tests, which exposed a subtle ordering issue in TLB flush logic, in chunk b73d6860-3caf-4eb7-811e-eda1693f60f3.
[4]: The evidence states that the test suites for vector, MMU, PMP, and ePMP are configured to match the user's RISC-V processor in chunk b73d6860-3caf-4eb7-811e-eda1693f60f3.
[5]: The evidence describes a hybrid flow combining constrained-random sweeps, functional coverage analysis, and directed test closure in chunk b73d6860-3caf-4eb7-811e-eda1693f60f3.