Repository navigation
arm64: dts: apple: t6040-j616s: Shed CPU power at the GPU firmware limit - #511
Closed
RyanTheTide wants to merge 1 commit into
Closed
RyanTheTide wants to merge 1 commit into
RyanTheTide wants to merge 1 commit into
Conversation
The GPU zones shed CPU power from 105 C, a pure backstop. But the J616s fans give no extra headroom under Linux: the SMC keeps them in its own mode (3), turns them on only at minimum speed (seen from 93-99 C die on a cold machine, from about 60 C once it is heat-soaked) and refuses an OS write of manual mode. The GPU firmware itself starts limiting at 101 C, and conformance runs reached 103.8 C. Start the GPU zones' passive trip at that same 101 C, so that once the GPU is at its own limit the CPU clusters give up power to it instead of both being limited separately. The critical trip stays at 118 C. Assisted-by: LLM Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
Contributor
Author
|
Note: the gpubench/CTS comparison for this change is pending; the measurement agent will run it. The J616s test kernel fan2 carries this commit together with the fan-control driver change. |
Contributor
Author
|
Final head for batching: 07f6411 (t-3207 batch1). The gpubench/CTS comparison is pending with the measurement agent. New context from Ryan's Fallout 4 session: the SMC does ramp the fans above minimum on its own, about 100 rpm per 15 s, to 2374/2557 rpm at around 62 W, and the CPU die peaked at 108 C before the fans caught up. That supports shedding CPU power once the GPU reaches its own 101 C limit. |
Contributor
Author
|
Landed in integration/aurora through #533 (J616s consolidation, 2026-10-28). |
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
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.
This PR moves the J616s GPU thermal zones' passive trip from 105 C to 101 C. That is the temperature at which the GPU firmware starts limiting itself. It is a DT-only change: no driver, parameter or config change.
Why:
fan_controlpath gets-EIOonF0Md = 1, and public prior art for M4 says this needs theFtstunlock, which was not tried.The critical trip stays at 118 C, and the CPU zones are unchanged.
Not measured yet. Before merge, run gpubench or CTS against the current DT to see what it costs on GPU-heavy loads. The only cost is CPU frequency while a GPU die sensor is at or above 101 C.
Evidence:
reports/j616s-sensors-20261027/FANS.md.🤖 Generated with Claude Code