oxidecomputer / oxidecomputer/dendrite

Fetching a transceiver's part info shouldn't fail silently

Open
#133 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Rust
Stars
20
Forks
3
Avg merge
3d 11h
Merged PRs (30d)
4

Description

In #104, a customer link failed to come all the way up, because the part-specific overrides for the transceiver's equalization settings were never applied. The underlying issue was likely #111, where we incorrectly marked a transceiver as unsupported too aggressively. Because of that, we failed to fetch the transceiver's MPN, and thus failed to look up the alternate equalization settings.

That's all fine in some sense, because of the link-reconciliation process in Dendrite, that continually tries to apply all the link settings, including these equalization settings. But we have no actual smoking gun proving #111 is the source of the error, because we silently fail in this case. Here, for example:

https://github.com/oxidecomputer/dendrite/blob/6ba23e71121c196e1e3c4e0621ba7a6f046237c7/dpd/src/link.rs#L1735-L1750

We're fetching the MPN, and later using it, but none of that code logs the fact that the transceiver couldn't be identified. We should do that, to ensure that this case doesn't pass silently in the future.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start in dpd/src/link.rs around lines 1735-1750, where the transceiver MPN is fetched and later used. Trace the identification failure path and review nearby logging conventions. Done means a failed transceiver identification is recorded instead of passing silently, while preserving the existing reconciliation behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
networking
Issue type
Bug
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
48/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.