sunfish-adminandCursor
2d145c2878
Restore disable_configstore, EuiccDisabler and thermal CDEV guards, branding and rebuild/sync helpers; document device roles, laptop migration, and Gitea. Co-authored-by: Cursor <cursoragent@cursor.com>
59 lines
2.9 KiB
Markdown
59 lines
2.9 KiB
Markdown
# 007 parallel track — ReSukiSU / SusFS on Magisk `009` ROM
|
||
|
||
**Status (2026-10-10):** active **parallel** path. Not Magisk ship. Not deleted.
|
||
Docs under `007-*` stay; this file is the bridge to the **current** closure ROM.
|
||
|
||
**Why keep it:** Sept 2026 LAB bring-up **booted** ReSukiSU + SusFS v1.5.5 with
|
||
matched `boot`+`vendor` (`docs/007-lab-results-20260916.md`). Owner may want
|
||
to retry that stack on today’s Magisk `009` tree (Stage1/CVE/hardening +
|
||
userspace), especially for hide experiments (e.g. PairIP-class apps) without
|
||
throwing away Magisk ship.
|
||
|
||
## Parallelism rules
|
||
|
||
| Rule | Detail |
|
||
|------|--------|
|
||
| Ship zip | Magisk EXTENDROM preroot only (`009` bacon). Never enable `CONFIG_KSU` in the Magisk ship bacon. |
|
||
| Experiment | Separate `boot.img` (and usually matched vendor modules) built from the **same** LOS/`009` kernel tree **or** a worktree clone — flash LAB only. |
|
||
| Coexistence | Do **not** run Magisk + KernelSU in the same boot. A/B slots or keep last Magisk zip to restore. |
|
||
| GOLDEN | Never flash. |
|
||
| Spec | FR-016 optional recycle; `specs/007-kernelsu-sunfish/` remains valid standby. Sibling workspace `sunfish-a16-rom` may hold older KSU dirt — prefer re-integrating onto current `009` kernel. |
|
||
|
||
## What already worked (do not reinvent)
|
||
|
||
| Item | Result |
|
||
|------|--------|
|
||
| ReSukiSU + SusFS boot | PASS (matched vendor, SELinux `vendor_kernel_modules`) |
|
||
| Manager + `susfs_init` | PASS |
|
||
| Banking hide alone | **Not** enough (Swissquote/Yuh FAIL `-69` on that stack) |
|
||
| Boot-only SusFS flash | FAIL (module CRC / vibrator hang) — always match vendor `.ko` |
|
||
|
||
Runbook detail: [`007-kernelsu-runbook.md`](007-kernelsu-runbook.md).
|
||
Lab write-up: [`007-lab-results-20260916.md`](007-lab-results-20260916.md).
|
||
|
||
## Recommended trial on current ROM
|
||
|
||
1. Keep Magisk closure zip + both slots known-good.
|
||
2. On `rom-dev01`, from the **current** `009` kernel (not the Sept-only image):
|
||
- refresh ReSukiSU / SusFS to newest compatible (non-shallow; record
|
||
`KSU_VERSION` + manager APK — AGENTS.md upstream freshness).
|
||
- integrate per `007-kernelsu-runbook.md` (manual hooks + SusFS).
|
||
- build `bootimage` (+ vendor modules from same build).
|
||
3. Disable Magisk for that boot (KSU boot has no Magisk preroot).
|
||
4. Flash LAB: matched `boot` + vendor modules path from runbook; vbmeta
|
||
disable-verity as before if required for the experiment.
|
||
5. Install matching spoofed manager; smoke `sys.boot_completed`, SusFS version,
|
||
then optional banking / Wallapop probes.
|
||
6. Restore Magisk ship zip/slot when done.
|
||
|
||
Artifacts: `/srv/rom/artifacts/007-kernelsu/` (separate from
|
||
`009-axp-los23-port/`).
|
||
|
||
## Do-not
|
||
|
||
- Merge KSU into `scripts/009-bacon-axp-stage1-combo4-cve-hi-releasekeys.sh`
|
||
- Reuse Sept `boot-resukisu-*.img` on Oct Magisk userspace without matching
|
||
kernel/vendor ABI
|
||
- Treat bank PASS as already proven on KSU (it was not)
|
||
- Delete `007-*` docs or `specs/007-kernelsu-sunfish/`
|