[dvsim] Make block-level regression directory timestamps sane and introduce last good link
- Dominant language
- SystemVerilog
- Stars
- 3.6k
- Forks
- 1.1k
- Avg merge
- 2d 22h
- Merged PRs (30d)
- 141
Description
### Description
As was highlighted in https://github.com/lowRISC/opentitan/pull/19281 the per block results for a particular regression all have different timestamps for the directory containing each set of results, instead a single regression run should use the same timestamp across all the directories.
Additionally it would be helpful if along with 'latest' we had some 'last_good' set of results. We'd need to define what last good actually means but in particular we don't want it to include regression runs where nothing has occurred because some infrastructure issue has broken the regression.
Contributor guide
Research direction
Start by reading the issue context and the related pull request 19281 to understand how dvsim currently names per-block regression result directories. Clarify what qualifies as a “last good” regression, especially for infrastructure failures, before identifying the relevant result-management entry points; done means shared timestamps and an agreed, reliable last_good result link.
Written by the indexing model from the issue text.
Assessment
- Domain
- testing-qa, tooling
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100