Repository navigation
Conversation
b7r6
force-pushed
the
fix/kill-revoked-revalidates
branch
2 times, most recently
from
October 7, 2026 20:31
33adde4 to
7f89f6f
Compare
…n, not timestamp kill_revoked_operations snapshots (worker, operation) pairs under the scheduler mutex, then confirms revocation against the state manager lock-free. If the operation is declined/requeued and re-dispatched to the same worker inside that window, the final locked re-guard used to pass on the fresh entry and the kill landed on the new, legitimate dispatch. An earlier version of this fix carried the entry's dispatched_at timestamp as the dispatch's identity, but a timestamp is not an identity: dispatched_at has second resolution (duration_since(UNIX_EPOCH).as_secs()), so a requeue-and-redispatch inside one wall-clock second carries the same dispatched_at and the stale kill still lands. Review caught the collision; the regression test now pins both dispatches to the same second and fails on the timestamp scheme. Each dispatch now carries a dispatch_generation: a scheduler-wide u64 incremented under the scheduler mutex at send time, stored on the running_action_infos entry. The sweep snapshots the generation and worker_notify_kill_operation only sends the kill if the entry under the lock still carries it; a re-dispatch always gets a strictly greater generation, so a stale kill can never match. The counter is scheduler-wide rather than per-worker: a per-worker counter would restart when a worker re-registers under the same id, letting a stale snapshot from the old incarnation match a fresh dispatch on the new one. dispatched_at stays as a timestamp for the acknowledgement timeout math. Also registers tests/kill_revoked_redispatch_race_test.rs in the Bazel rust_test_suite, which had not picked it up. Provenance: adversarial review fleet; incompleteness found by a follow-up review of TraceMachina#2838. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
b7r6
force-pushed
the
fix/kill-revoked-revalidates
branch
from
October 8, 2026 17:51
7f89f6f to
2760207
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What and why
Lane A stability fix from the findings catalog (see
nativelink-findings-2026-10-05.csvon the
meta/findings-catalog-2026-10-05branch), rebased onto OSS main@bf4b3aa855. This fork PR exists to validate the change on CI that is
representative of upstream PR lanes (public read-only cache, same endpoints)
before any upstream submission.
How was this verified?
The branch carries its own regression tests (fails-without where the fix
class allows it — see the branch's test commits). This PR's purpose is the
full representative CI run; green here rehearses the exact checks an
upstream PR would face.
Risk
Fork-internal validation PR; nothing lands anywhere by merging this. The
underlying fix's risk assessment happens at upstream-submission time.
AI assistance
Agent-assisted pipeline (Claude) for rebase and PR mechanics under my
direction; the fix itself carries verification in its tests.