Repository navigation
aurora-sep: retry Ctrl-C in the survey stop test - #2160
RyanTheTide merged 1 commit into
Conversation
(cherry picked from commit 796e02859d18ce3ac6c787d97fdbff221dc2e76c)
Codex Review Summary: ❌ FailedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
|
GREEN on eed0c6e Two-round review. Round 1 at 796e02859d: content GO (the +11/-2 survey-stop test change is byte-identical to omacom #443's diff, author kept, honest-state lines complete), one finding: parent a7dd26e is behind current main. Round 2 re-check at eed0c6e: rebased onto current aurora main dcce0e7 (base rule satisfied), test file bytes identical (sha256 db47c5ad917e), cherry-pick trailer now present (cherry picked from 796e02859d18), author preserved, and the test ran green on the submitter's host (1 passed, 67 deselected). Text confirmed: cause-not-established framing with the INT trap / bash-signal-deferral as the most-likely hypothesis, retry clearly stated as a workaround rather than a fix, same assertions as before (exit 130, "interrupted" line on stderr, no leftover output). Supersedes omacom/linux-aurora#443. Host-side test only, never part of a kernel build - READY (non-draft) state is appropriate. GREEN on eed0c6e. |
|
The account paying for this security review has reached its Codex usage limits. The payer can check the Codex usage dashboard. For personal accounts, using credits requires enabling “Use credits for security reviews” in Code review settings. If you do not manage the paying account, contact this repository's admins. |
9bb8bc5
into
aurora-silicon:integration/aurora
[Joshua's agent] aurora-sep: retry Ctrl-C in the survey stop test
The M3 survey-stop test failed in CI with
AssertionError: 0 != 130(emptystderr): the check that the survey dies on Ctrl-C saw exit 0 instead. The
survey sometimes runs to the end after one SIGINT; cause not established, most
likely the signal arriving while bash defers it (the INT trap is installed
before any load starts and is never reset, so it is not a missing handler).
The test now retries the SIGINT up to 4 times with a short wait each — what a
user does when the first Ctrl-C seems lost — then falls back to the long wait.
Same assertions as before: exit 130, the "interrupted" line on stderr, no
leftover output. The retry is a stated workaround, not a fix for the cause.
Honest state: host-side test change only (tools/aurora-sep/tests/test_m3_max.py,
+11/-2), no kernel code touched. Ran green on the submitter's host with the
pick applied (test_ctrl_c_stops_the_load: 1 passed, 67 deselected). Old CI on
the omacom PR was all green; no maintainer review there. This move supersedes
omacom/linux-aurora#443 (same commit, rebased onto today's
aurora-silicon/linux integration/aurora). Never booted as part of a kernel
build — it is a host test, run with pytest on Linux.