| Age | Commit message (Collapse) | Author | Files | Lines | |
|---|---|---|---|---|---|
| 2026-04-17 | Tweak trap loop detector heuristic. (#1666) | Prashanth Mundkur | 1 | -15/+24 | |
| The previous heuristic did not suffice for a few ACT4 tests where the trap count grew slowly but steadily over the xret count. Adjust by introducing an instruction retired callback and resetting the loop detector when a sufficient number of instructions have retired since the last trap. This still catches tight trap loops but allows the above tests to complete successfully. --------- Co-authored-by: Alex Richardson <alexrichardson@google.com> | |||||
| 2026-04-14 | Detect potentially infinite trap loops. (#1656) | Prashanth Mundkur | 1 | -0/+40 | |
| This helps in automated tested environments where an infinite loop caused by a failing test can cause issues. An `xret_callback` is used in combination with the existing `trap_callback` to detect trap loops. The `trap_callback` is now called after the state changes due to the trap are visible. | |||||
