Skip to content

FLOOR REPAIR: in-flight shared fill is charged as marginal until store — repair the classification boundary (design first, characterization is landed in docs/plans/required-floor-in-flight-shared-fill-findings.md) - #10018

Closed
gunbai-bot[bot] wants to merge 1 commit into
mainfrom
session/jolly-ibex-643

Conversation

@gunbai-bot

@gunbai-bot gunbai-bot Bot commented Sep 2, 2026

Copy link
Copy Markdown
Contributor

Auto-opened by session-dashboard for session jolly-ibex-643.
Pushing to session/jolly-ibex-643 advances this PR.

Worker attestation

Before flipping this PR to ready for review, confirm each item:

  • Title describes the change (not the session id or branch).
  • PR body summarises what and why (replace the TODO below).
  • Tests run: name the command (e.g. npm test, cargo test) and the result.
  • If this closes a work item, the body contains a Closes #N directive.
  • No commits on this branch are surprises (no fork/cherry-pick I did not make).
  • No secrets / credentials / large binaries staged.

Summary

TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.

Test plan

  • TODO: list the commands that ran (or "no tests changed; relied on CI") and the outcome.

@briansrls
briansrls marked this pull request as ready for review September 2, 2026 07:03
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Sep 2, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-09-02T07:12:45.420527Z ecd2c0e Draft marked ready
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@gunbai-bot

gunbai-bot Bot commented Sep 2, 2026

Copy link
Copy Markdown
Contributor Author

Duplicate of #10003, which was squash-merged as fb481ae.

The dashboard re-adopted session/jolly-ibex-643 after that lane was archived, re-opening its already-merged head ecd2c0e. Squash-merge means ecd2c0e is not an ancestor of main, so an ancestry check reads as unmerged — but the content is present. Verified by symbol rather than by ancestry: FillBudgetExceeded (4), CrossClaimFillFrame (4) and in_flight_cross_claim_fill (2) appear in identical counts in v1_interpreter.rs on main, and cli_run.rs plus the design doc are byte-identical. The 11-line difference in v1_interpreter.rs is main having advanced (#9886), not this branch carrying anything extra.

Closing rather than re-running review and CI on work that has already landed. The 3 failing checks here are the srv1 runner $HOME/euid rustup fault, which the operator is handling separately.

@gunbai-bot gunbai-bot Bot closed this Sep 2, 2026

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ecd2c0e858

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

}
E::ShellOutputLimitExceeded { .. } => WitnessRuntimeCause::ShellOutputLimitExceeded,
E::CallContractMismatch { .. } => WitnessRuntimeCause::CallContractMismatch,
E::FillBudgetExceeded { .. } => WitnessRuntimeCause::FillBudgetExceeded,

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Keep fill crossings on the budget-interruption path

When an admitted fill crosses the CPU ceiling, map_budget_error_to_witness_refusal leaves FillBudgetExceeded unchanged, so run_claim reaches this mapping and produces ClaimOutcome::RuntimeError rather than BudgetInterrupted. Consequently, required_floor_runner.rs routes an unenrolled crossing into failures and an expected-red crossing through RuntimeErrored; if that identity is already enrolled in floor_non_verdict, the latter can even satisfy the non-verdict roster instead of populating the always-blocking interrupted_before_verdict collection. Preserve a distinct fill cause while retaining the interruption/budget fields and blocking semantics.

Useful? React with 👍 / 👎.

Comment on lines +8864 to +8867
/// An admitted cross-claim producer was the active subject when the unchanged CPU safety
/// ceiling fired. The token makes the prospective-fill population countable without
/// treating first-touch order as intrinsic claim cost.
FillBudgetExceeded,

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Add the new cause to the modeled mirror

This enum is documented as a one-for-one mirror of gunbc.observation_ci_render.CiWitnessRuntimeCause, but the authoritative enum in dag/gunbc/observation_ci_render.dag:831-851 and its token function have no fill-budget arm. Rust can therefore emit cause=fill-budget-exceeded while the modeled renderer and consumers cannot construct or recognize the corresponding cause, leaving the mirror/parity contract incomplete; add the modeled variant and token alongside this Rust arm.

Useful? React with 👍 / 👎.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants