PinePhone Pro component support matrix

This is the PinePhone Pro per-component reference index. The phone is a Rockchip RK3399S system with a few dozen distinct devices on board — some on-SoC, some on i2c, some on SDIO, some on USB. Every one has its own page with identity (part number, bus, GPIO/IRQ), driver status, our-tree and upstream pointers, and open work.

The PineTab2 now has its own component matrix so the tablet work can grow without muddying the phone’s support map. The repo root HARDWARE.md is the source-of-truth manifest with TODOs; this index is the navigable, status-pilled view. The hardware reference appendix is a flat manifest table organized for cross-essay lookup.

Counts

Out of 45 components catalogued:

• ● working 30 working on hardware
• ◐ partial 9 partial — driver loads but flaky, disabled, or incomplete
• ○ blocked 0 blocked on a missing piece elsewhere
• · not started 6 not started

(Wave-1 + wave-2 implementation runs of 2026-04-30 moved gd25lq128 from blocked → partial, and rk-efuse / magnetometer / vibrator / sgm3140 from not-started → partial. The 2026-05-02/03 PD bring-up arc additionally promoted RK818 (full USB-PD sink → battery charging), RK3399 TSADC (live cpu/gpu temps, later corrected to use hardware QSEL rather than trim), AF8133J magnetometer (reset-gpio polarity fix → real X/Y/Z reads), and SGM3140 (/dev/led/flash torch + flash strobe verified by hand) from partial → working. A 2026-05-03 status pass also promoted MPU-6500, STK3311, vibrator, and RK3399 eFuse — all have working drivers exposing their full kernel-side feature set; “partial” was actually about userspace policy that lives outside each driver, not about driver completeness. See the per-component pages for what’s left to wire into userspace nicely. TSADC moved back to partial on 2026-05-05 after live overheat evidence exposed a readout bug; the follow-up fixed sensor reporting with RK3399 hardware QSEL and kernel #195 proved guarded code-domain comparator IRQs. It remains partial because hardware TSHUT is still disarmed. The power-domain controller also moved back to partial: the sysctl-only driver round-tripped RGA on kernel #144, but PINEPHONE_PRO disables it after a warm-reboot boot-refusal correlation.)

Status wording is intentionally conservative. partial usually means “candidate code and a verifier exist, but live behaviour is flaky or only one mode is verified.” The not-yet-phone-tested set is: GD25LQ128 (no SPI-NOR driver attached → /dev/flash0 absent; the bench tooling assumes a flashctl(8) that doesn’t exist in FreeBSD), PL330 DMA (driver wired in kernel config, no consumer yet exercises it), and rk_power_domain (driver exists but is disabled until the attach/boot issue is isolated). Treat their pages as compile-ready / bench-pending until bench:phone or the component-specific verifier captures real PinePhone Pro output.

SoC and memory

Clocks, pinctrl, DMA, power, IOMMU

SoC peripherals

USB

Display

Audio

Power

Wireless and modem

Sensors

Input and indicators

Cameras

Notes on cataloguing

The PineTab2 (RK3566) shares the family — same PMIC family (RK817 vs RK818), same touchscreen vendor, same DSI panel topology — but is catalogued separately under PineTab2 hardware and PineTab2 components because the SoC and several peripherals differ.

A few component pages flag specific discrepancies between the upstream PinePhone Pro DTS, our schematic notes, and live driver behaviour:

• FUSB302 — mainline/local DTS place the controller at &i2c4 address 0x22; FreeBSD logs the same 7-bit address as shifted address 0x44 on iicbus3. Older notes mixed this with i2c0.
• SYR828 — labelled vdd_gpu in the upstream DTS, not “A53 little cluster VDD” as some of our older notes claimed.
• AP6255 — the WiFi side has host-side SDIO IRQ delivery after the Rockchip bit-24 mask fix, and the SDIO bridge now requests a post-discovery run clock so the card leaves probe speed. The open work is boot/idle polish and larger transfer receipts; see the finishing plan §WiFi.

When in doubt, the per-component page wins over the index. The index exists to find pages, not to summarize them — every one-line cell here is reductive.

Related references

• PinePhone Pro hardware reference — board-specific phone manifest.
• Hardware reference manifest — flat table view of both target boards with bus addresses and pin assignments.
• PineTab2 component matrix — tablet-side equivalent.
• PinePhone Pro schematic support map — board-level functional grouping.
• Cross-driver audit (2026-04-30) — Linux/OpenBSD reference comparison and root-cause hypotheses for the open wedges.
• Repo root HARDWARE.md — TODOs and effort estimates per component.
• Pine64 PinePhone Pro wiki — canonical board-level component list.