thermal: surgically excise power-down timeout#2585
Merged
Conversation
Presently, the thermal loop will power down the system if any component remains above its "critical" threshold for 60 seconds. We believe this behavior to be excessively conservative, and unnecessary, as an uncontrollable thermal runaway event will eventually result in something tripping its "power-down" threshold and shutting the system down immediately. This commit makes the smallest possible change to the thermal loop to remove the timeout. This allows us to operate at the edge of our thermal envelope for as long as is necessary, while still ensuring that the fans go full-throttle when we hit 90 Funny AMD Temperature Units, or when anything else hits its crit threshold. The rest of the control loop behavior should be identical.
hawkw
commented
Jul 2, 2026
Comment on lines
-1440
to
-1443
| } else if now_ms > *start_time + self.overheat_timeout_ms { | ||
| // If blasting the fans hasn't cooled us down in this amount | ||
| // of time, then something is terribly wrong - abort! | ||
| self.transition_to_uncontrollable(now_ms) |
Collaborator
|
this tentatively looks okay to me but I don't have the full context so I'll let others review/comment |
bcantrill
approved these changes
Jul 2, 2026
bcantrill
left a comment
bcantrill
left a comment
Collaborator
There was a problem hiding this comment.
Looks great -- thanks for updating the surrounding comments as well.
hawkw
added a commit
that referenced
this pull request
Jul 2, 2026
Presently, the thermal loop will power down the system if any component remains above its "critical" threshold for 60 seconds. We believe this behavior to be excessively conservative, and unnecessary, as an uncontrollable thermal runaway event will eventually result in something tripping its "power-down" threshold and shutting the system down immediately. This commit makes the smallest possible change to the thermal loop to remove the timeout. This allows us to operate at the edge of our thermal envelope for as long as is necessary, while still ensuring that the fans go full-throttle when we hit 90 Funny AMD Temperature Units, or when anything else hits its crit threshold. The rest of the control loop behavior should be identical.
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
3 participants
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.
Presently, the thermal loop will power down the system if any component remains above its "critical" threshold for 60 seconds. We believe this behavior to be excessively conservative, and unnecessary, as an uncontrollable thermal runaway event will eventually result in something tripping its "power-down" threshold and shutting the system down immediately.
This commit makes the smallest possible change to the thermal loop to remove the timeout. This allows us to operate at the edge of our thermal envelope for as long as is necessary, while still ensuring that the fans go full-throttle when we hit 90 Funny AMD Temperature Units, or when anything else hits its crit threshold. The rest of the control loop behavior should be identical.