Skip to content

tools: prepare J613 RTC alarm wiring for trial Stage 2 - #157

Merged
DjDeveloperr merged 1 commit into
mainfrom
fix/j613-rtc-alarm
Nov 11, 2026
Merged

DjDeveloperr merged 1 commit into
mainfrom
fix/j613-rtc-alarm

Conversation

@satchlj

@satchlj satchlj commented Nov 11, 2026

Copy link
Copy Markdown
Member

GPT-6.1 Sol on @satchlj's behalf

Add an opt-in offline J613 RTC-alarm DTB preparation tool for Aura's existing Stage 2 composer. It supplies the three PMIC cells, SPMI IRQ2, wakeup-source and the SPMI provider's explicit AIC IRQ437 parent. Without interrupt-parent, SPMI stays in polling mode and the RTC waits indefinitely for its IRQ domain despite otherwise valid alarm resources.

Aura main consumes external kernel DTBs rather than owning the Linux DTS files. This tool produces a separate DTB and input/output SHA256 receipt for that composition path. It refuses in-place output, existing output/receipt, wrong board/providers, disabled resources and existing/partial wiring. It does not update an installed loader or enable deep sleep.

Hardware: J613/T8122, kernel 7.3.0-aurora73-upstream-2-ARCH, qualified 30.9 / source 112007b70271361ed5c9270b144732f427c6feb8. Timed wake needed this alarm wiring carried by a one-shot trial boot loader. Five awake alarms passed; alarm-armed pm_test freezer/devices/platform returned in 5.164/6.649/6.699 s. Real s2idle returned in 60.050, 60.432, 60.684, 300.869 s and 1800.789 s, each on RTC IRQ93 with self-disarm. Internal display/backlight, MTP, both USB controllers, Wi-Fi, audio, hardware GPU and Touch ID service gate checked after return. External-display s2idle remains owed. Battery-energy delta was 0 uWh under a non-discharging powered condition; external supply energy/total sleep power was not measured.

Validation: both J613 board outputs match every resolved property of the hardware-tested RTC-only DTBs. The extended supplier audit rejects the former broken wiring and accepts the correction. Aura PMP boot-data check passed; Python suite 651 passed, 17 skipped (two archive-environment failures passed after restoring Git metadata/removing AppleDouble sidecars). Existing-output and repeat-application refusal checks preserve input hashes.

Credits: Eryk Wieliczko's alarm/SPMI approach omacom/linux-aurora#274 and Samuel Carreno's companion board wiring omacom/linux-aurora#285; explicit interrupt-parent follows Linux's working t6030.dtsi example.

Signed-off-by: Satya Benson satya@benson.earth

Signed-off-by: Satya Benson <satya@benson.earth>

@DjDeveloperr DjDeveloperr left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Codex (GPT-6.1 Sol), reviewing on @DjDeveloperr’s behalf.

No blocking findings for this opt-in offline tool. It verifies the J613/T8122 providers/layout and rejects in-place, pre-existing output/receipt and partial wiring before producing a new DTB plus hash receipt. The alarm's SPMI/AIC parent and NVMEM cells match the documented qualified RTC-only wiring. All exact-head checks passed. Merge is limited to preparation tooling: it does not install a loader, enable deep sleep or qualify external-display suspend/total sleep power. Posted timed-wake evidence remains attributed; no hardware was touched during this review.

@DjDeveloperr
DjDeveloperr merged commit 60dea2d into main Nov 11, 2026
10 checks passed
@RyanTheTide
RyanTheTide deleted the fix/j613-rtc-alarm branch November 16, 2026 05:20
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.

2 participants