Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions device/husky/NOTES.md
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,16 @@ Sources:
- `P/`: postmarketOS pmaports `device/testing` (2026-10-21).
- `G/`: Google's GPL kernel for this phone, `android-gs-shusky-6.1-android16` ([source.android.com: Build Pixel kernels](https://source.android.com/docs/setup/build/building-pixel-kernels)).

## 2026-11-15: Bluetooth, haptics, the torch and the gauge on merged `omamobile`, a third phone

Colin's phone, the third unit (Samsung 128 GB UFS 3.1, SKU G1MNW; [2026-11-08](#2026-11-08-a-third-phone-samsung-128-gb-ufs-31-still-locked)), unlocked since, with Google's CP3A.261005.005 on slot A and Omarchy on slot B. Kernel: `omamobile` at `82f31d3a7`, unchanged (`7.3.0-rc6-omamobile-push1-g82f31d3a7a35`, flashed with `bin/omamobile flash` on 2026-11-13). Firmware from the owner's own factory image with `tools/extract-vendor-firmware -d husky -s G1MNW` ([#722](https://github.com/derekross/omarchy-mobile/pull/722)), installed under `/usr/lib/firmware/updates`. `omamobile-device-husky` 0.1.13 from [#752](https://github.com/derekross/omarchy-mobile/pull/752) (0.1.12 plus the Bluetooth address step; merged 2026-11-15). Each item below was checked by a person at the phone.

- **Bluetooth** [V]: hci0 loads the patch (`BCM4398D0 G5SN_V17_HK3 REL BR:0169 FW:b8718197dc`), is powered at boot and takes the bootloader's address (#752). A pair of AirPods Pro paired from the Settings page, connected on PipeWire's `a2dp-sink` profile, and a quiet 440 Hz tone played through them was heard (2026-11-13). The same pairing and tone had also worked on 2026-11-12 on a test kernel with device-husky 0.1.10, before the address step. Calls over the Handsfree profile were not tried.
- **Haptics** [V]: `cs40l26-vibra 1-0043: Firmware 0x1800d4 rev 7.2.91 running` and its input device; two 500 ms `FF_RUMBLE` effects (strong 0x6000) uploaded through `EVIOCSFF` and played, and both were felt (2026-11-15).
- **Torch** [V]: `white:torch` (`lm3644` at `0-0063`, `led-max-microamp` 50000, `max_brightness` 17) set to 17 and back to 0 by sysfs: the LED lit and went out, seen twice (2026-11-15). One earlier on/off right after the haptics test wasn't noticed; the next two were.
- **Battery** [V]: `max170xx_battery` is there at boot with no overlay (the bus is on by default since [kernel #205](https://github.com/derekross/linux-omarchy-mobile/pull/205)): capacity, voltage, current and temperature read, and status `Charging` on the USB cable at about +0.8 to +0.9 A from a PC port with the charger driver not loaded. The published tree gives the gauge no battery model (its comment on `hsi2c_11` says so), and `charge_full_design` reads 3000000 against the pack's 4950 mAh, so the percentage is the gauge's default model's: it read 100 % at 4.328 V on 2026-11-11, which isn't comparable with a modelled gauge (#761). UPower acted on its level: on 2026-11-13 at 04:25, off the charger, its critical-battery action suspended the phone, and the phone didn't come back from that sleep ([SUSPEND-PLAN](SUSPEND-PLAN.md) draft, [#724](https://github.com/derekross/omarchy-mobile/pull/724)).
- **Also on this boot** [V]: Wi-Fi joined a 5 GHz network at -38 dBm; the GPU at 150-890 MHz through panthor; three cpufreq policies on schedutil; `/sys/power/mem_sleep` reads `[s2idle] deep`; `spm_lvl` 5; no NFC device; `running`, no failed units.

## 2026-11-13: display, GPU and CPU DVFS on merged `omamobile`

[Kernel #160](https://github.com/derekross/linux-omarchy-mobile/pull/160) (merged 2026-11-10) and [#157](https://github.com/derekross/linux-omarchy-mobile/pull/157) with [#158](https://github.com/derekross/linux-omarchy-mobile/pull/158) (merged 2026-11-13) booted from `omamobile` itself, at `82f31d3a7` (#158's last commit), on Brett's phone (Samsung UFS 3.1): `7.3.0-rc6-omamobile-push1-g82f31d3a7a35`, the repository's husky config, the pinned clang (`clang-r596125`), `modules-push`, slot B flashed with `bin/omamobile flash`. The initramfs was the one from the 2026-11-01 test set, unchanged (husky packs no modules into it), and `omamobile-device-husky` was 0.1.8, so the kernel was the only change on this boot. Before the flash, slot B's retry count read 2 after one boot, and entering fastboot with `reboot-to bootloader` didn't change it [V].
Expand Down
4 changes: 2 additions & 2 deletions device/husky/ROADMAP.md
Original file line number Diff line number Diff line change
Expand Up @@ -49,7 +49,7 @@ Status words as in DEVICES.md. "9a" is the Pixel 9a (`tegu`) work. "Needs" lists
| Modem | no | Lef, once a SIM is found ([Brett asked](https://github.com/derekross/omarchy-mobile/issues/1#issuecomment-6544100933)) | cheetah [MODEM-PLAN](../cheetah/MODEM-PLAN.md); 9a [MODEM-PLAN](../tegu/MODEM-PLAN.md) ("ONLINE with IPC, no SIM yet") | A SIM (a prepaid one is the plan; the phone's own SIM is a data-only eSIM); the baseband is `g5300i` [V: fastboot], the S5300 family like cheetah's [I]; the CP's PCIe link and the `cpif` driver; modem firmware from the owner's own factory image; the NV protections. **Lef has a private read-only backup of the modem and calibration partitions (checksummed, plus an encrypted second copy) [V: not in this repository]; [#449](https://github.com/derekross/omarchy-mobile/pull/449) (open) gives husky's descriptor the `NV_PARTS` so `tools/nv-backup` takes and checks it.** | Read cheetah's and the 9a's MODEM-PLAN M0 to M2 end to end. Stage 0 is read-only: partitions, and what a booted phone answers without starting the modem |
| NFC | testing (a test kernel with the node switched on by an overlay: the chip answers and reads tags through `nxp-nci`; kernel [#208](https://github.com/derekross/linux-omarchy-mobile/pull/208), [#209](https://github.com/derekross/linux-omarchy-mobile/pull/209), both drafts) | Lef | 9a [#105](https://github.com/derekross/linux-omarchy-mobile/pull/105), [#104](https://github.com/derekross/linux-omarchy-mobile/pull/104) | What we found: an ST ST54L combo, its NCI controller at 0x08 on `hsi2c_0` (the haptics bus of #200) with its IRQ on gpa7, the chip's enable line on gph1-6 and a reference clock from the PMU's NFC pad (register 0x3e84); its secure element is on `spi_12` and is never to be touched. The driver is `nxp-nci` (as the 9a), not `st-nci`. On a test kernel the chip answered the NCI power-on, detected an ISO-DEP card, and read a Type 2 tag with one NDEF record; without the clock pad (#209) the same chip still powered on but found no tag in one run. Not tested: suspend, the secure element, the power cost of the pad clock, a second phone. The 9a's NFC must not poll through s2idle (AGENTS.md) | Merge #208 and #209, then s2idle with the driver bound and the clock's power cost |
| Cameras, video decode | no (read from the stock device tree, not tested) | unassigned | 9a [#103](https://github.com/derekross/linux-omarchy-mobile/pull/103); cheetah [CAMERA-PLAN](../cheetah/CAMERA-PLAN.md) | The sensors, the MIPI receiver and the rest of the pipeline; large. What we found for video: the SoC's codecs are Samsung's MFC (HEVC, VP9; firmware loaded behind the S2MPU through the GSA image loader) and Google's BigWave for AV1; neither has a driver in any tree we use, so there is no `/dev/video*` and Chromium decodes YouTube on the CPU, which heats the phone [V: Lef's phone] | After the rest; for video, a read of Google's MFC and BigWave trees first |
| Suspend | testing (s2idle by hand and through `systemctl suspend` on Lef's phone: 8 successes, 0 failures on 2026-11-10, woken by Power; and two critical-battery suspends that didn't come back, one in `deep` before #656 and #244 ([SUSPEND-PLAN](SUSPEND-PLAN.md) §4)) | Colin (the runbook and its stages), with Lef's first runs | 9a [SLEEP-PLAN](../tegu/SLEEP-PLAN.md) (s2idle works, drain high); cheetah [SUSPEND-PLAN](../cheetah/SUSPEND-PLAN.md) | UFS hibern8 and system suspend ([UFS-PLAN](UFS-PLAN.md) U5 and U6); the "UFS at `spm_lvl` 5", "only s2idle" and "regulators late" gotchas. `/sys/power/mem_sleep` selects `deep` on the current kernels [V: Lef's phone, and Colin's on `husky-test-20261109-t1`]; omamobile-device-husky 0.1.10 makes systemd's Suspend use `s2idle`, but a test that writes `/sys/power/state` itself must still set `s2idle` first. Idle with the screen off draws about 0.5 A on Lef's phone, cause unknown; a CPU idle-state overlay and a power-domain overlay both reboot-looped the phone (recovered over fastboot), so neither is a safe next step [V: Lef's phone, test tree] | [SUSPEND-PLAN](SUSPEND-PLAN.md): H1 (`pm_test`) to H4 (ten cycles), attended and on the cable; UFS-PLAN U5 on the guarded initramfs boot |
| Suspend | testing (s2idle by hand and through `systemctl suspend` on Lef's phone, on a local kernel: 8 successes, 0 failures on 2026-11-10, woken by Power; and two critical-battery suspends that didn't come back, one in `deep` before #656 and #244 ([SUSPEND-PLAN](SUSPEND-PLAN.md) §4)) | Colin (the runbook and its stages), with Lef's first runs | 9a [SLEEP-PLAN](../tegu/SLEEP-PLAN.md) (s2idle works, drain high); cheetah [SUSPEND-PLAN](../cheetah/SUSPEND-PLAN.md) | UFS hibern8 and system suspend ([UFS-PLAN](UFS-PLAN.md) U5 and U6); the "UFS at `spm_lvl` 5", "only s2idle" and "regulators late" gotchas. `/sys/power/mem_sleep` selects `deep` on the current kernels [V: Lef's phone, and Colin's on `husky-test-20261109-t1`]; omamobile-device-husky 0.1.10 makes systemd's Suspend use `s2idle`, but a test that writes `/sys/power/state` itself must still set `s2idle` first. Idle with the screen off draws about 0.5 A on Lef's phone, cause unknown; a CPU idle-state overlay and a power-domain overlay both reboot-looped the phone (recovered over fastboot), so neither is a safe next step [V: Lef's phone, test tree] | [SUSPEND-PLAN](SUSPEND-PLAN.md): H1 (`pm_test`) to H4 (ten cycles), attended and on the cable; UFS-PLAN U5 on the guarded initramfs boot |
| Thermal, CPU DVFS | works on [kernel #158](https://github.com/derekross/linux-omarchy-mobile/pull/158) (merged 2026-11-13) | Brett | 9a [#87](https://github.com/derekross/linux-omarchy-mobile/pull/87) | #158 adds the CPU levels from the ECT and the CPU thermal zones; before it every core ran at the bootloader's rate. The turbo-mode marks are the 8 Pro's battery limiter caps, so only `zuma-husky.dts` points its CPUs at the tables (the 8a's caps are lower). On Lef's test tree cpufreq-dt runs three policies and the four ACPM TMU zones report; software video decode took the SoC to 86 C, so his phone runs lowered passive trips and CPU caps as a stopgap (local, not in a package) | A skin temperature source (the S2MPG15 ADC) and cpuidle, which needs the MCT timer first |
| USB host, Type-C | testing (role switch to host without VBUS, once) | Lef (with Charging) | cheetah [USB-C-PLAN](../cheetah/USB-C-PLAN.md); 9a [#124](https://github.com/derekross/linux-omarchy-mobile/pull/124) | The Type-C controller is the MAX77759 TCPC above. dwc3 switches to xHCI host and back with no VBUS ([issue #1](https://github.com/derekross/omarchy-mobile/issues/1#issuecomment-6562547503)); VBUS out is the gated part (draft [#282](https://github.com/derekross/omarchy-mobile/pull/282)) | As Charging |
| Time, persistent storage | partly | unassigned | none yet | No RTC: `systemd-timesyncd` sets the clock a few minutes after Wi-Fi is up, and the first-boot release-key problem is fixed (omarchy-mobile #188, closed). Root is on `userdata` behind the read-only guard; marking slot B successful is a manual, attended step decided on [#252](https://github.com/derekross/omarchy-mobile/issues/252); from 2026-11-01 to 2026-11-07 a local unit marked slot B on this phone; stopped pending #252, and nothing marks it on the 8 Pro until the boot-LUN guard change and the marking tool ([#483](https://github.com/derekross/omarchy-mobile/pull/483)) are in | Write a short plan |
Expand All @@ -59,7 +59,7 @@ Status words as in DEVICES.md. "9a" is the Pixel 9a (`tegu`) work. "Needs" lists

A proposal, not a verdict; Brett and Derek know the hardware better.

1. ~~Merge #148, #153, #154~~ (done 2026-10-31 and 2026-11-01), and the battery base (#163, #175, done). Brett's display, GPU, DVFS and Wi-Fi PRs (#160, #157, #158, and #156 through #187) landed 2026-11-09 to 2026-11-13. Next: #205 (the battery bus on by default), #199 and #200, the sound set (#203, #204, #450, #453).
1. ~~Merge #148, #153, #154~~ (done 2026-10-31 and 2026-11-01), and the battery base (#163, #175, done). Brett's display, GPU, DVFS and Wi-Fi PRs (#160, #157, #158, and #156 through #187) landed 2026-11-09 to 2026-11-13. #205 (the battery bus on by default), #199 (Bluetooth) and #200 (haptics and torch) merged too, and #450 and #453 landed. Next: the sound kernel PRs, #203 and #204.
2. **Battery, charging and USB-C** (Lef): the gauge and the charger work; the Type-C driver is a sink. Left: the model in the board file, the 92 % stop (#272), and host mode with VBUS out, which waits for the runbook review (#282) and a USB-C meter.
3. ~~**Display driver** (Brett)~~: #160 merged 2026-11-10. Open: the short black blink at the DRM master handover.
4. ~~**GPU, DVFS and thermal** (Brett)~~: #157 and #158 merged 2026-11-13. Next: a GPU load with the thermal zone watched, a skin temperature source and cpuidle.
Expand Down
Loading