feat(fan): read hwmon directly — dual-fan support and simulated device tests

[vietanhdev]

Dual-fan ThinkPads (P1, P15, X1 Extreme) had a second fan the app could not see. /proc/acpi/ibm/fan has a speed: field and no speed2:, so hwmon is the only way to read it.

Verified live on a P1 Gen 4i: two tachometers at 3813 and 3515 RPM, one control channel.

The structural point

Tachometers and control channels are not one-to-one. thinkpad_acpi writes the same level to both fans, so a dual-fan machine has 2 tachs and 1 PWM. Code assuming one fan per channel is wrong on every P-series — and anything offering per-fan control would be lying about the hardware.

Replaces the iwlwifi_1-virtual-0

Adapter: Virtual device
temp1: +46.0°C

thinkpad-isa-0000
Adapter: ISA adapter
fan1: 3486 RPM
fan2: 3238 RPM
CPU: +74.0°C
GPU: +58.0°C
temp3: +63.0°C
temp4: +1.0°C
temp5: +61.0°C
temp6: +56.0°C
temp7: +49.0°C
temp8: N/A
pwm1: 128%

ucsi_source_psy_USBC000:001-isa-0000
Adapter: ISA adapter
in0: 0.00 V (min = +0.00 V, max = +0.00 V)
curr1: 3.00 A (max = +0.00 A)

BAT0-acpi-0
Adapter: ACPI interface
in0: 16.67 V
power1: 0.00 W

coretemp-isa-0000
Adapter: ISA adapter
Package id 0: +77.0°C (high = +100.0°C, crit = +100.0°C)
Core 0: +70.0°C (high = +100.0°C, crit = +100.0°C)
Core 1: +70.0°C (high = +100.0°C, crit = +100.0°C)
Core 2: +66.0°C (high = +100.0°C, crit = +100.0°C)
Core 3: +64.0°C (high = +100.0°C, crit = +100.0°C)
Core 4: +58.0°C (high = +100.0°C, crit = +100.0°C)
Core 5: +77.0°C (high = +100.0°C, crit = +100.0°C)
Core 6: +70.0°C (high = +100.0°C, crit = +100.0°C)
Core 7: +74.0°C (high = +100.0°C, crit = +100.0°C)

ucsi_source_psy_USBC000:002-isa-0000
Adapter: ISA adapter
in0: 0.00 V (min = +0.00 V, max = +0.00 V)
curr1: 0.00 A (max = +0.00 A)

nvme-pci-0400
Adapter: PCI adapter
Composite: +57.9°C (low = -273.1°C, high = +81.8°C)
(crit = +84.8°C)
Sensor 1: +57.9°C (low = -273.1°C, high = +65261.8°C)
Sensor 2: +58.9°C (low = -273.1°C, high = +65261.8°C)

acpitz-acpi-0
Adapter: ACPI interface
temp1: +74.0°C subprocess

Three problems with shelling out:

• Needed lm-sensors installed, so temperatures silently vanished without it
• Hardcoded the chip name thinkpad-isa-0000
• Double-counted the first fan — procfs contributed "Fan1", sensors contributed "fan1", so the UI listed one fan twice

Dropping it also removed the last use of the regex crate, now gone from the tree.

Simulated devices

The captured profile covers exactly one machine. tests/synthetic_hardware.rs builds sysfs trees in temp dirs for thirteen shapes and states — the ones most likely to break precisely because nobody develops on them:

Shape	Why it matters
single-fan ThinkPad	the common T/X case
four-fan	reader must not cap at what we own
non-ThinkPad laptop	a dell_smm fan and coretemp still read
no hwmon at all	every container
fan_control=1 absent	the most common broken state — differs from working by exactly three commands: lines
stopped fan at 0 RPM	a reading, not a missing sensor; dropping it hides the most safety-relevant state
pwm1_enable=0	means FULL SPEED on thinkpad_acpi, not off — reading it as "disabled" inverts the most aggressive setting
malformed sysfs	empty, non-numeric, out-of-range
chip with no name file	not an hwmon chip
stable ordering	UI must not reshuffle between polls

A real gap in the earlier fixture

The capture script resolved /sys/class/hwmon symlinks to their /sys/devices targets and stored only those — so the data was present but undiscoverable, because code walks /sys/class/hwmon, which was absent. It mirrors the class view now (as real directories; symlinks would point outside the profile at the real machine). Profile regenerated, redaction re-verified.

Suite: 85 → 99.

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Updated (UTC)
❌ Deployment failed View logs	thinkutils	59a5969	Jul 19 2026, 03:48 PM

[vietanhdev]

Superseded — the PPA source build landed on main via #26.