msi-ec: add hwmon fan RPM monitoring for WMI1 hardware

[mohammad-abbas-mehdi]

Summary

Registers a hwmon device exposing read-only fan1_input / fan2_input on WMI1 hardware, reporting fan speed measured from the EC's tachometer.

What this improves over the current fan reporting

The driver today exposes fan speed as a percentage — the EC's commanded fan-curve step. That is the control target, not the fan's actual speed, and it becomes misleading whenever the EC bypasses its own curve. Cooler boost is the clear case: the fan spins up sharply, the percentage does not follow.

Measured on an MSI GV62 7RD:

	CPU fan	GPU fan
cooler boost off	~2,800 RPM	0 RPM (stopped)
cooler boost on	~6,700 RPM	~7,000 RPM

The commanded percentage barely moves across that transition. Measured vs. commanded is the crux: the existing percentage answers "what did the EC ask for"; this answers "what is the fan actually doing".

Why hwmon rather than a platform sysfs attribute

hwmon is the kernel-standard interface for this data, so sensors, node_exporter, desktop monitors and extensions like Vitals pick it up automatically with no msi-ec specific knowledge. It is the same interface msi-wmi-platform already uses on WMI2 — this brings parity to WMI1, which has no equivalent.

PR #475 (@Jiogo18) addresses the same gap via a custom platform sysfs attribute. This differs in exposing the data to the whole standard monitoring ecosystem rather than requiring per-tool knowledge of msi-ec internals.

What this does not do

• No fan control — the EC's closed-loop curve is untouched.
• No PWM write path — the channels are strictly read-only.
• No EC temperature channels — real temperatures come from coretemp/amdgpu, per @glpnk's point in Feature: expose CPU and GPU temperatures via Linux hwmon subsystem #739.

WMI2 coexistence

Where msi-wmi-platform is present on WMI2 hardware, fan channels are suppressed to avoid duplicate hwmon entries, aligned with the stated direction of migrating WMI2 logic there. If the GUID is absent on a WMI2 board, the channels register as a fallback.

Hardware validation

Validated on MSI GV62 7RD (MS-16J9, 16J9EMS1.112, WMI1).

A board configuration is included as the validation vehicle: fan and thermal addresses are confirmed on hardware; the remaining features are deliberately left unset pending confirmation. This is a partial bringup, not full device support.

Verified after a cold boot: the hwmon device registers, fan1_input/fan2_input track the tachometer rather than the commanded percentage, and cooler boost moves the reading as shown above.

EC dump — GV62 7RD (16J9EMS1.112)

|      | _0 _1 _2 _3 _4 _5 _6 _7 _8 _9 _a _b _c _d _e _f
|------+------------------------------------------------
| 0x0_ | 00 80 00 00 00 00 00 00 00 00 00 00 00 00 00 00  |................|
| 0x1_ | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  |................|
| 0x2_ | 00 00 00 00 00 00 00 00 00 00 00 80 c0 06 49 4b  |..............IK|
| 0x3_ | 02 05 00 0d 51 0a 05 00 54 0f 30 2a 88 01 c0 00  |....Q...T.0*....|
| 0x4_ | 98 08 34 00 86 0d e4 f9 06 07 f2 29 05 0c 0c 30  |..4........)...0|
| 0x5_ | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  |................|
| 0x6_ | 00 00 00 00 00 00 00 00 3d 0a 37 3d 43 49 4f 4f  |........=.7=CIOO|
| 0x7_ | 64 42 28 30 38 42 4c 56 56 00 0a 08 08 08 08 08  |dB(08BLVV.......|
| 0x8_ | 33 36 37 3c 41 46 4b 4b 63 00 00 36 3e 46 4e 56  |367<AFKKc..6>FNV|
| 0x9_ | 56 00 0a 07 07 07 07 07 02 16 7d 02 16 63 00 00  |V.........}..c..|
| 0xa_ | 31 36 4a 39 45 4d 53 31 2e 31 31 32 31 31 30 39  |16J9EMS1.1121109|
| 0xb_ | 32 30 31 37 31 35 3a 30 34 3a 32 38 5c 0d 00 00  |201715:04:28\...|
| 0xc_ | 00 07 31 00 00 9e 40 00 00 00 00 00 00 74 00 00  |..1...@......t..|
| 0xd_ | 00 00 40 00 73 00 01 00 00 00 00 80 00 00 00 00  |..@.s...........|
| 0xe_ | e2 02 00 80 10 00 00 00 00 00 0d 00 00 83 00 0d  |................|
| 0xf_ | 00 00 80 82 0c 00 0c 00 00 00 00 00 00 00 00 00  |................|

Notes

• Single commit, Signed-off-by present.
• MODULE_VERSION unchanged at 0.13 — no unilateral version bump.
• checkpatch.pl --strict: 0 errors, 0 warnings, 0 checks.

AI Disclosure (per project AI Policy)

• LLM: Claude (Anthropic) — Agent tool: Claude Code (CLI)
• AI role: driver code structure, code comments, and drafting of the commit message and this description.
• Contributor role: problem identification, EC register research, all architectural decisions, hardware measurement and validation on physical hardware, and line-by-line review of everything the agent produced.