Closed arichardson closed 1 year ago
Should be ready for review now.
Includes #219 to ensure the CI can pass.
Rebased.
Looks Good to me. My only question is whether we should propagate the "became-unrepresentable" information in incoffset when the source cap is invalid. I don't have a specific case for which it is wrong though.
Ah that is a good point - probably not too useful if it's already untagged.
Test failures are timeouts on the ASAN builds. Merging.
Catches up with https://github.com/CTSRD-CHERI/sail-cheri-riscv/pull/56
Verified by running TestRIG against sail. While fuzzing against sail I also found some additional divergences that I'll upload fixes for in separate PRs.