[PATCH v6 0/5] regulator: add the Will Semiconductor WL2868C camera PMIC

[Qianhui Xu]

This series was drafted with AI assistance; I reviewed and edited every
line myself, and I am responsible for its content, measurements and
decisions.  The commands behind the verification claims below are listed
in the series README that accompanies this posting.

The WL2868C is the camera PMIC on the Lenovo Legion Tab (TB322FC,
SM8750); its seven LDOs feed the rear camera, the front camera and the
VCM on i2c10.

Changes in v6:

 - Rebased onto current mainline (v7.3-20183-g05441d697f7f, tip of
   master on 2026-11-03).  v5 no longer applied: the board patch anchors
   its arch/arm64/boot/dts/qcom/Makefile hunk on the sm8650-qrd.dtb line,
   and sm8650-valve-deckard.dtb was inserted right there in between, so
   git am stopped on that file (Mark Brown: "This doesn't apply against
   current code, please check and resend").  v6 applies cleanly on the
   new base.

 - The SM8750 CCI patch is gone.  There is an identical CCI series in
   flight from Hangxiang Ma (Konrad: "There's an identical patchset in
   flight, please don't send competing series without a clear reason"),
   so this series no longer competes for those nodes.  The two
   observations sashiko made on the CCI patch (cci2 is not described; the
   sleep states change the pin function to gpio instead of keeping
   cci_i2c_sda/scl) belong to that series as well and are not addressed
   here.

 - The board compatible in arm/qcom.yaml is now a patch of its own
   instead of riding along with another binding (Conor asked for the
   split; checkpatch wants the same).

 - regulator binding: the paragraph that described which board supply
   comes from which LDO is gone from the binding description, it belongs
   to the board patch (Conor).  vsys-supply is added and at least one of
   vin7-supply and vsys-supply is now required, because both feed the
   internal control logic, which is powered from the higher of the two
   (Conor: "the higher of the two powers the internal control logic").
   The other inputs stay optional: they only feed LDOs, so an input may
   be left unconnected on a board that does not use those outputs, which
   the description now says.

 - driver: the enable register is written as a whole instead of doing a
   read-modify-write of SYS_EN plus one LDO bit.  The per-LDO bits can be
   preset by the firmware while SYS_EN is clear, so setting SYS_EN and
   the requested bit would switch the preset outputs on as well
   (sashiko, Medium).  The driver keeps the set of LDOs that consumers
   enabled and writes exactly those bits, so enabling one output cannot
   power another one.

 - The board DTS is split in two: a skeleton patch (compatible, chassis
   type, reserved GPIOs, board clocks) and the WL2868C regulators.  The
   rear sensor and its VCM are gone from this series together with the
   CCI controllers they hang off; they follow with the camera series
   that depends on the CCI nodes.  The board clocks stay: they are the
   XO, sleep and divided CXO clocks that sm8750.dtsi consumes, so every
   SM8750 board file has to provide them (Konrad asked why the board
   repeats what the binding example has; the binding example contains no
   clock, and these four are consumed by the SoC dtsi, as on
   sm8750-mtp.dts).

 - The VSYS input is described as a fixed regulator, as it is the
   always-on system rail.  VIN12, VIN34, VIN5 and VIN7 are fed by SPMI
   PMIC rails that this series does not describe yet, so they are left
   out and follow with the board series that adds those PMICs.

Verification, host-side only: every patch passes checkpatch.pl --strict
with 0 errors (the only two warnings are "MAINTAINERS need updating" for
the two new files, one of them being the driver's own entry added in
2/5); the series replays with plain git am on a pristine checkout of the
base commit, 5/5, and the result is the same tree as the branch it is
generated from; drivers/regulator/ builds with W=1 without warnings;
sm8750-lenovo-elden.dtb builds and dt-validate has nothing to report on
it; dt_binding_check passes for the new binding.

The regulators are still not exercised on hardware: the i2c10 probe, the
regulator registration and the LDO voltages come from our own downstream
board DTS, not from a measured boot of this series.

Qianhui Xu (5):
  regulator: dt-bindings: willsemi,wl2868c: add the WL2868C camera PMIC
  regulator: willsemi,wl2868c: add the WL2868C camera PMIC driver
  dt-bindings: arm: qcom: document the Lenovo Legion Y700 Gen4
  arm64: dts: qcom: sm8750-lenovo-elden: add the Lenovo Legion Y700 Gen4
  arm64: dts: qcom: sm8750-lenovo-elden: describe the WL2868C regulators

 .../devicetree/bindings/arm/qcom.yaml         |   1 +
 .../bindings/regulator/willsemi,wl2868c.yaml  | 104 +++++++++
 MAINTAINERS                                   |   7 +
 arch/arm64/boot/dts/qcom/Makefile             |   1 +
 .../boot/dts/qcom/sm8750-lenovo-elden.dts     | 129 +++++++++++
 drivers/regulator/Kconfig                     |  11 +
 drivers/regulator/Makefile                    |   1 +
 drivers/regulator/wl2868c-regulator.c         | 212 ++++++++++++++++++
 8 files changed, 466 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/regulator/willsemi,wl2868c.yaml
 create mode 100644 arch/arm64/boot/dts/qcom/sm8750-lenovo-elden.dts
 create mode 100644 drivers/regulator/wl2868c-regulator.c

-- 
2.53.0