[Hardware Support] Legion Pro 7 Gen 10 16IAX10H (83F5) - Working with force=1

[alstergee]

Hello,

SYSTEM INFORMATION

Model: Lenovo Legion Pro 7 16IAX10H
DMI Product Name: 83F5
BIOS Version: Q7CN44WW
CPU: Intel Core Ultra 9 275HX (Arrow Lake, 24-core)
GPU: NVIDIA GeForce RTX 5090 Laptop
RAM: 32GB DDR5
OS: Ubuntu 25.10 (Plucky Puffin)
Kernel: 6.17.0-8-generic

SUMMARY

Successfully got LenovoLegionLinux working on the Legion Pro 7 Gen 10 (16IAX10H) with Intel Arrow Lake CPU and RTX 5090. The module loads and functions correctly with force=1 parameter. Requesting addition of model ID 83F5 to the allowlist.

WHAT WORKS

• Battery Conservation Mode (tested, functional)
• Fan Speed Control
• Fn Lock Status
• Hybrid Mode Status
• Monitor Mode
• hwmon Interface (creates legion_hwmon)
• Platform Profile Support
• WMI Driver Support
• Keyboard Backlight LED
• Y-Logo LED

WHAT DOESN'T WORK

• IO-Port LED - Initialization fails with error:
  "Error WMI call for reading brightness: expected a value between 1 and 2, but got 0"
  This is non-critical; everything else functions normally.

KERNEL LOG (dmesg)

legion_laptop: Loading legion_laptop
legion PNP0C09:00: legion_laptop platform driver probing
legion PNP0C09:00: Read identifying information: DMI_SYS_VENDOR: LENOVO; DMI_PRODUCT_NAME: 83F5; DMI_BIOS_VERSION:Q7CN44WW
legion_laptop: Read identifying information: DMI_SYS_VENDOR: LENOVO; DMI_PRODUCT_NAME: 83F5; DMI_BIOS_VERSION:Q7CN44WW
legion PNP0C09:00: Using configuration for system: GKCN
legion PNP0C09:00: Skipping ACPI _STA check
legion_laptop: Creating RAM access to embedded controller
legion_laptop: Successfully mapped embedded controller: 0xfe00d400 (in RAM)/0x0 (in EC) to virtual 0x000000005b237bf5
legion PNP0C09:00: Read embedded controller ID 0x5508
legion PNP0C09:00: Skipped checking embedded controller id
legion PNP0C09:00: Creating debugfs interface
legion_laptop: Creating sysfs interface
legion_laptop: Creating hwmon interface
legion_laptop: Creating platform profile support
legion_laptop: Init WMI driver support
legion_wmi 10AFC6D9-EA8B-4590-A2E7-1CD3C84BB4B1: Register after probing for WMI.
legion_wmi D320289E-8FEA-41E0-86F9-611D83151B5F: Register after probing for WMI.
legion_laptop: Init keyboard backlight LED driver
legion_laptop: Init Y-Logo LED driver
legion_laptop: Init IO-Port LED driver
legion_laptop: Error WMI call for reading brightness: expected a value between 1 and 2, but got 0
legion_laptop: Error reading brightness for light: 5
legion PNP0C09:00: Failed to init IO-Port LED driver. Skipping ...
legion PNP0C09:00: legion_laptop loaded for this device

TECHNICAL DETAILS

• Uses configuration: GKCN
• Embedded controller ID: 0x5508
• sysfs path: /sys/devices/pci0000:00/0000:00:1f.0/PNP0C09:00/VPC2004:00/conservation_mode

REQUEST

Please add DMI Product Name 83F5 (Legion Pro 7 16IAX10H) to the allowlist. The module works well with force=1 and the only issue is the IO-Port LED driver which gracefully fails without affecting other functionality.

Happy to provide any additional information or testing if needed.

[gluceri]

Hello @alstergee
I own a Lenovo Legion Pro 7 16AFR10H (AMD 9955HX). Can you confirm that using the GKCN profile temperature and fan speed values are correct?
The 16AFR10H uses 0x5508 chip too, but I got wrong values for temperature and fan speeds. Things get better changing

.access_method_fanspeed = ACCESS_METHOD_EC,
.access_method_temperature = ACCESS_METHOD_EC,
.access_method_fancurve = ACCESS_METHOD_EC,

to

.access_method_fanspeed = ACCESS_METHOD_WMI3,
.access_method_temperature = ACCESS_METHOD_WMI3,
.access_method_fancurve = ACCESS_METHOD_WMI3,

Furthermore I am not sure about the correctness of the ram and memory related parameters

System Information

Model: Lenovo Legion Pro 7 16AFR10H
DMI Product Name: 83RU
BIOS Version: SMCN14WW
CPU: AMD Ryzen 9955HX
GPU: NVIDIA GeForce RTX 5070Ti Laptop
RAM: 64GB DDR5
OS: AerchLinux up-to-date
Kernel: 6.18.3-custom (6.18.3 + audio patch)

GKCN configuration

With this config I have obviously wrong values for temperatures and fan speed

DMESG

[ 6895.544843] legion_laptop: Loading legion_laptop
[ 6895.544904] legion PNP0C09:00: legion_laptop platform driver probing
[ 6895.544906] legion PNP0C09:00: Read identifying information: DMI_SYS_VENDOR: LENOVO; DMI_PRODUCT_NAME: 83RU; DMI_BIOS_VERSION:SMCN14WW
[ 6895.544907] legion_laptop: Read identifying information: DMI_SYS_VENDOR: LENOVO; DMI_PRODUCT_NAME: 83RU; DMI_BIOS_VERSION:SMCN14WW
[ 6895.544909] legion PNP0C09:00: is_denied: 0; is_allowed: 1; do_load_by_list: 1; do_load: 1
[ 6895.544910] legion PNP0C09:00: Using configuration for system: SMCN
[ 6895.544959] legion PNP0C09:00: ACPI CFG: 2081289482
[ 6895.544959] legion_laptop: Creating RAM access to embedded controller
[ 6895.544973] legion_laptop: Successfully mapped embedded controller: 0xfe00d400 (in RAM)/0x0 (in EC) to virtual 0x00000000826e604b
[ 6895.545155] legion PNP0C09:00: Read embedded controller ID 0x5508
[ 6895.545156] legion PNP0C09:00: Creating debugfs interface
[ 6895.545162] legion_laptop: Creating sysfs interface
[ 6895.545172] legion_laptop: Creating hwmon interface
[ 6895.545213] legion_laptop: Creating platform profile support
[ 6895.545231] legion_laptop: Init WMI driver support
[ 6895.545262] legion_wmi 10AFC6D9-EA8B-4590-A2E7-1CD3C84BB4B1-20: Register after probing for WMI.
[ 6895.545272] legion_wmi D320289E-8FEA-41E0-86F9-611D83151B5F-21: Register after probing for WMI.
[ 6895.545289] legion_laptop: Init keyboard backlight LED driver
[ 6895.545979] legion_laptop: Init Y-Logo LED driver
[ 6895.546597] legion_laptop: Init IO-Port LED driver
[ 6895.546609] legion_laptop: Error WMI call for reading brightness: expected a value between 1 and 2, but got 0
[ 6895.546610] legion_laptop: Error reading brightness for light: 5
[ 6895.546610] legion PNP0C09:00: Failed to init IO-Port LED driver. Skipping ...
[ 6895.546611] legion PNP0C09:00: legion_laptop loaded for this device

Sensors

I only reports interesting values via sensors legion_hwmon-isa-0000

legion_hwmon-isa-0000
Adapter: ISA adapter
Fan 1:           23135 RPM  (max = 10000 RPM)  <==== WRONG
Fan 2:           18020 RPM  (max = 10000 RPM)  <==== WRONG
CPU Temperature:  +80.0°C  <==== WRONG
GPU Temperature:  +87.0°C  <==== WRONG
IC Temperature:   +55.0°C  
pwm1:                 N/A  (mode = N/A)

Fan curve

EC Chip ID: 5508
EC Chip Version: 2b0
legion_laptop features: fancurve powermode platformprofile platformprofilenotify minifancurve fancurve_pmw_speed fancurve_rpm_speed
legion_laptop ec_readonly: 0
ACPI CFG error: 0
ACPI CFG: 2081289482
temperature access method: 1
CPU temperature error: 0
CPU temperature: 80
CPU temperature EC error: 0
CPU temperature EC: 80
CPU temperature ACPI error: 0
CPU temperature ACPI: 55
CPU temperature WMI error: 0
CPU temperature WMI: 0
CPU temperature WMI2 error: 0
CPU temperature WMI2: 0
CPU temperature WMI3 error: 0
CPU temperature WMI3: 55
GPU temperature error: 0
GPU temperature: 87
GPU temperature EC error: 0
GPU temperature EC: 87
GPU temperature ACPI error: 0
GPU temperature ACPI: 45
GPU temperature WMI error: 0
GPU temperature WMI: 0
GPU temperature WMI2 error: 0
GPU temperature WMI2: 0
GPU temperature WMI3 error: 0
GPU temperature WMI3: 45
fan speed access method: 1
1 fanspeed error: 0
1 fanspeed: 23135
1 fanspeed EC error: 0
1 fanspeed EC: 23135
1 fanspeed ACPI error: 0
1 fanspeed ACPI: 1800
1 fanspeed WMI error: 0
1 fanspeed WMI: 0
1 fanspeed WMI2 error: 0
1 fanspeed WMI2: 0
1 fanspeed WMI3 error: 0
1 fanspeed WMI3: 1800
2 fanspeed error: 0
2 fanspeed: 18020
2 fanspeed EC error: 0
2 fanspeed EC: 18020
2 fanspeed ACPI error: 0
2 fanspeed ACPI: 1800
2 fanspeed WMI error: 0
2 fanspeed WMI: 0
2 fanspeed WMI2 error: 0
2 fanspeed WMI2: 0
2 fanspeed WMI3 error: 0
2 fanspeed WMI3: 1800
powermode access method: 3
powermode error: 0
powermode: 2
powermode EC error: 0
powermode EC: 64
powermode ACPI error: -5
powermode ACPI: 0
powermode WMI error: 0
powermode WMI: 2
has custom powermode: 1
ACPI rapidcharge error: 0
ACPI rapidcharge: 1
WMI backlight 2 state: 0
WMI backlight 3 state: 0
WMI light IO port: 0
WMI light Y logo/lid: 0
EC minifancurve feature enabled: 1
EC minifancurve on cool: error
EC lockfancontroller error: 0
EC lockfancontroller: false
fanfullspeed error: 0
fanfullspeed: 0
fanfullspeed EC error: 0
fanfullspeed EC: 0
Max speed for fancurve: 10000
Current fan curve in hardware:
Fan curve current point id: 0
Fan curve points size: 10
u(speed_of_unit)|speed1[u]|speed2[u]|speed1[pwm]|speed2[pwm]|acceleration|deceleration|cpu_min_temp|cpu_max_temp|gpu_min_temp|gpu_max_temp|ic_min_temp|ic_max_temp
3        64      0       163     163     0       0       0       1       0       0       0       60
3        0       0       0       0       0       0       0       2       0       0       0       0
3        129     0       328     328     0       0       0       3       0       0       0       0
3        2       0       5       5       0       0       0       4       0       0       0       0
3        54      0       137     137     0       0       0       5       0       0       0       0
3        0       1       0       0       0       0       0       6       0       0       0       0
3        0       0       0       0       0       0       0       7       0       0       0       0
3        0       0       0       0       0       0       0       8       0       0       0       0
3        0       0       0       0       0       0       0       9       0       0       0       0
3        0       0       0       0       0       0       0       10      0       0       0       0
=====================
Current fan curve in hardware (WMI; might be empty)
Fan curve current point id: 0
Fan curve points size: 10
u(speed_of_unit)|speed1[u]|speed2[u]|speed1[pwm]|speed2[pwm]|acceleration|deceleration|cpu_min_temp|cpu_max_temp|gpu_min_temp|gpu_max_temp|ic_min_temp|ic_max_temp
1        1       0       2       2       0       0       0       0       0       0       0       0
1        2       0       5       5       0       0       0       0       0       0       0       0
1        3       0       7       7       0       0       0       0       0       0       0       0
1        4       0       10      10      0       0       0       0       0       0       0       0
1        5       0       12      12      0       0       0       0       0       0       0       0
1        6       0       15      15      0       0       0       0       0       0       0       0
1        7       0       17      17      0       0       0       0       0       0       0       0
1        8       0       20      20      0       0       0       0       0       0       0       0
1        9       0       22      22      0       0       0       0       0       0       0       0
1        10      0       25      25      0       0       0       0       0       0       0       0
=====================

CUSTOM configuration

This is derived from GKCN changing the .access_method_fanspeed, .access_method_temperature, .access_method_fancurve

static const struct model_config model_smcn = {
	.registers = &ec_register_offsets_v0,
	.check_embedded_controller_id = true,
	.embedded_controller_id = 0x5508,
	.memoryio_physical_ec_start = 0xC400,
	.memoryio_size = 0x300,
	.has_minifancurve = true,
	.has_custom_powermode = true,
	.access_method_powermode = ACCESS_METHOD_WMI,
	.access_method_keyboard = ACCESS_METHOD_WMI,
	.access_method_fanspeed = ACCESS_METHOD_WMI,    <===== CHANGED
	.access_method_temperature = ACCESS_METHOD_WMI, <===== CHANGED
	.access_method_fancurve = ACCESS_METHOD_WMI,    <===== CHANGED
	.access_method_fanfullspeed = ACCESS_METHOD_WMI,
	.acpi_check_dev = true,
	.ramio_physical_start = 0xFE00D400,
	.ramio_size = 0x600
};

Still not sure about the correctness of the other values

DMESG

[ 7626.839637] legion_laptop: Loading legion_laptop
[ 7626.839731] legion PNP0C09:00: legion_laptop platform driver probing
[ 7626.839733] legion PNP0C09:00: Read identifying information: DMI_SYS_VENDOR: LENOVO; DMI_PRODUCT_NAME: 83RU; DMI_BIOS_VERSION:SMCN14WW
[ 7626.839734] legion_laptop: Read identifying information: DMI_SYS_VENDOR: LENOVO; DMI_PRODUCT_NAME: 83RU; DMI_BIOS_VERSION:SMCN14WW
[ 7626.839736] legion PNP0C09:00: is_denied: 0; is_allowed: 1; do_load_by_list: 1; do_load: 1
[ 7626.839737] legion PNP0C09:00: Using configuration for system: SMCN
[ 7626.839811] legion PNP0C09:00: ACPI CFG: 2081289482
[ 7626.839812] legion_laptop: Creating RAM access to embedded controller
[ 7626.839834] legion_laptop: Successfully mapped embedded controller: 0xfe00d400 (in RAM)/0x0 (in EC) to virtual 0x000000006fc6f611
[ 7626.840008] legion PNP0C09:00: Read embedded controller ID 0x5508
[ 7626.840009] legion PNP0C09:00: Creating debugfs interface
[ 7626.840018] legion_laptop: Creating sysfs interface
[ 7626.840029] legion_laptop: Creating hwmon interface
[ 7626.840079] legion_laptop: Creating platform profile support
[ 7626.840107] legion_laptop: Init WMI driver support
[ 7626.840144] legion_wmi 10AFC6D9-EA8B-4590-A2E7-1CD3C84BB4B1-20: Register after probing for WMI.
[ 7626.840154] legion_wmi D320289E-8FEA-41E0-86F9-611D83151B5F-21: Register after probing for WMI.
[ 7626.840179] legion_laptop: Init keyboard backlight LED driver
[ 7626.840775] legion_laptop: Init Y-Logo LED driver
[ 7626.841495] legion_laptop: Init IO-Port LED driver
[ 7626.841510] legion_laptop: Error WMI call for reading brightness: expected a value between 1 and 2, but got 0
[ 7626.841512] legion_laptop: Error reading brightness for light: 5
[ 7626.841513] legion PNP0C09:00: Failed to init IO-Port LED driver. Skipping ...
[ 7626.841514] legion PNP0C09:00: legion_laptop loaded for this device

Sensors

I only reports interesting values via sensors legion_hwmon-isa-0000

legion_hwmon-isa-0000
Adapter: ISA adapter
Fan 1:           1700 RPM  (max = 10000 RPM)
Fan 2:           1700 RPM  (max = 10000 RPM)
CPU Temperature:  +56.0°C  
GPU Temperature:  +46.0°C  
IC Temperature:   +55.0°C  
pwm1:                 N/A  (mode = N/A)

Fan curve

EC Chip ID: 5508
EC Chip Version: 2b0
legion_laptop features: fancurve powermode platformprofile platformprofilenotify minifancurve fancurve_pmw_speed fancurve_rpm_speed
legion_laptop ec_readonly: 0
ACPI CFG error: 0
ACPI CFG: 2081289482
temperature access method: 5
CPU temperature error: 0
CPU temperature: 55
CPU temperature EC error: 0
CPU temperature EC: 80
CPU temperature ACPI error: 0
CPU temperature ACPI: 55
CPU temperature WMI error: 0
CPU temperature WMI: 0
CPU temperature WMI2 error: 0
CPU temperature WMI2: 0
CPU temperature WMI3 error: 0
CPU temperature WMI3: 55
GPU temperature error: 0
GPU temperature: 45
GPU temperature EC error: 0
GPU temperature EC: 87
GPU temperature ACPI error: 0
GPU temperature ACPI: 45
GPU temperature WMI error: 0
GPU temperature WMI: 0
GPU temperature WMI2 error: 0
GPU temperature WMI2: 0
GPU temperature WMI3 error: 0
GPU temperature WMI3: 45
fan speed access method: 5
1 fanspeed error: 0
1 fanspeed: 1700
1 fanspeed EC error: 0
1 fanspeed EC: 23135
1 fanspeed ACPI error: 0
1 fanspeed ACPI: 1700
1 fanspeed WMI error: 0
1 fanspeed WMI: 0
1 fanspeed WMI2 error: 0
1 fanspeed WMI2: 0
1 fanspeed WMI3 error: 0
1 fanspeed WMI3: 1700
2 fanspeed error: 0
2 fanspeed: 1700
2 fanspeed EC error: 0
2 fanspeed EC: 18020
2 fanspeed ACPI error: 0
2 fanspeed ACPI: 1700
2 fanspeed WMI error: 0
2 fanspeed WMI: 0
2 fanspeed WMI2 error: 0
2 fanspeed WMI2: 0
2 fanspeed WMI3 error: 0
2 fanspeed WMI3: 1700
powermode access method: 3
powermode error: 0
powermode: 2
powermode EC error: 0
powermode EC: 64
powermode ACPI error: -5
powermode ACPI: 0
powermode WMI error: 0
powermode WMI: 2
has custom powermode: 1
ACPI rapidcharge error: 0
ACPI rapidcharge: 1
WMI backlight 2 state: 0
WMI backlight 3 state: 0
WMI light IO port: 0
WMI light Y logo/lid: 0
EC minifancurve feature enabled: 1
EC minifancurve on cool: error
EC lockfancontroller error: 0
EC lockfancontroller: false
fanfullspeed error: 0
fanfullspeed: 0
fanfullspeed EC error: 0
fanfullspeed EC: 0
Max speed for fancurve: 10000
Current fan curve in hardware:
Fan curve current point id: 0
Fan curve points size: 10
u(speed_of_unit)|speed1[u]|speed2[u]|speed1[pwm]|speed2[pwm]|acceleration|deceleration|cpu_min_temp|cpu_max_temp|gpu_min_temp|gpu_max_temp|ic_min_temp|ic_max_temp
1        1       0       2       2       0       0       0       0       0       0       0       0
1        2       0       5       5       0       0       0       0       0       0       0       0
1        3       0       7       7       0       0       0       0       0       0       0       0
1        4       0       10      10      0       0       0       0       0       0       0       0
1        5       0       12      12      0       0       0       0       0       0       0       0
1        6       0       15      15      0       0       0       0       0       0       0       0
1        7       0       17      17      0       0       0       0       0       0       0       0
1        8       0       20      20      0       0       0       0       0       0       0       0
1        9       0       22      22      0       0       0       0       0       0       0       0
1        10      0       25      25      0       0       0       0       0       0       0       0
=====================
Current fan curve in hardware (WMI; might be empty)
Fan curve current point id: 0
Fan curve points size: 10
u(speed_of_unit)|speed1[u]|speed2[u]|speed1[pwm]|speed2[pwm]|acceleration|deceleration|cpu_min_temp|cpu_max_temp|gpu_min_temp|gpu_max_temp|ic_min_temp|ic_max_temp
1        1       0       2       2       0       0       0       0       0       0       0       0
1        2       0       5       5       0       0       0       0       0       0       0       0
1        3       0       7       7       0       0       0       0       0       0       0       0
1        4       0       10      10      0       0       0       0       0       0       0       0
1        5       0       12      12      0       0       0       0       0       0       0       0
1        6       0       15      15      0       0       0       0       0       0       0       0
1        7       0       17      17      0       0       0       0       0       0       0       0
1        8       0       20      20      0       0       0       0       0       0       0       0
1        9       0       22      22      0       0       0       0       0       0       0       0
1        10      0       25      25      0       0       0       0       0       0       0       0
=====================

[ChaoticSi1ence]

@gluceri — I maintain a fork that already has proper WMI3 support for the Gen 10 Legion Pro 7 family. I just pushed a model_smcn config for your 83RU (BIOS prefix SMCN), so the module should load without force=1.

What's in the fork (relevant to your machine)

• model_smcn config with ACCESS_METHOD_WMI3 for fan speed, temperature, and fan curve — exactly what you found fixes the wrong readings from EC direct access
• ramio_physical_start = 0xFE00D400 — correct for your AMD platform (the Intel Q7CN variant uses 0xFE0B0400)
• 3-fan support (num_fans = 3) — the Gen 10 Pro 7 chassis has CPU, GPU, and auxiliary fans. If your machine only has 2, the 3rd fan will just read 0 RPM harmlessly
• wmi_dryrun module parameter — gates all WMI write calls (fan curve, fullspeed, power mode) so you can safely test read-only operations first
• access_method_keyboard = ACCESS_METHOD_NO_ACCESS — the keyboard backlight on this chassis is controlled via USB HID, not WMI. Setting it to ACCESS_METHOD_WMI would hit WMAF light_id 0x05 (IO-Port), which returns a zeroed buffer and causes the "expected 1-2, got 0" error you're seeing. With NO_ACCESS, that error goes away entirely
• has_minifancurve = false — you noted EC minifancurve returns an error; this disables it cleanly

Differences from your proposed config

Field	Your proposal	model_smcn	Reason
access_method_keyboard	ACCESS_METHOD_WMI	ACCESS_METHOD_NO_ACCESS	USB HID keyboard, WMI causes IO-Port LED error
has_minifancurve	true	false	EC minifancurve returns error on this hardware
num_fans	(default=2)	3	Same chassis as Q7CN, 3rd fan is auxiliary
acpi_check_dev	true	false	Safer default; _STA behavior unverified on SMCN

How to test

git clone https://github.com/ChaoticSi1ence/LenovoLegionLinux.git
cd LenovoLegionLinux/kernel_module

# Build
make

# Install (as root)
sudo insmod legion-laptop.ko wmi_dryrun=1

# Check dmesg — should show "Using configuration for system: SMCN"
dmesg | grep legion

# Verify sensors
cat /sys/class/hwmon/hwmon*/name  # find legion_hwmon
cat /sys/class/hwmon/hwmon*/fan1_input
cat /sys/class/hwmon/hwmon*/fan2_input
cat /sys/class/hwmon/hwmon*/fan3_input  # 3rd fan — may be 0 if not present
cat /sys/class/hwmon/hwmon*/temp1_input
cat /sys/class/hwmon/hwmon*/temp2_input
cat /sys/class/hwmon/hwmon*/temp3_input

With wmi_dryrun=1, all write operations (fan curve, power mode, fullspeed) are logged but not executed, so it's safe to poke around. Once you're satisfied the reads are correct, you can reload without wmi_dryrun to enable writes.

Let me know how it goes — especially whether fan3_input reports anything and whether the IO-Port LED error is gone.

[Ozzbuzzz69]

How did you get the fan speed control working i cant seem to make it work at all, also not the custom profile, then i have to be dual booting windows for that to happen

[alstergee]

@gluceri — I can't speak to the AMD variant directly since mine is the Intel model (83F5 / Q7CN / Arrow Lake). On my machine the GKCN config with EC direct access works fine — temps and fan speeds read correctly via hwmon. Sounds like @ChaoticSi1ence's fork with the WMI3 access methods is exactly what you need for the 83RU/SMCN. The EC memory mapping is the same base address (0xFE00D400) but the access method difference makes sense given different firmware implementations between the Intel and AMD variants.

@Ozzbuzzz69 — No dual boot needed. Here's what I did:

1. Install the module via DKMS

Follow the install instructions from the LenovoLegionLinux repo. On my system it installed as legion-laptop 0.0.19 via DKMS.

2. Force-load it (Gen 10 isn't on the allowlist yet)

# Auto-load on boot
echo "legion-laptop" | sudo tee /etc/modules-load.d/legion-laptop.conf

# Force=1 bypasses the allowlist check
echo "options legion-laptop force=1" | sudo tee /etc/modprobe.d/legion-laptop.conf

Then either reboot or sudo modprobe legion-laptop.

3. Fan speed / power profiles work through platform_profile

Once the module is loaded you get 4 profiles with no Windows needed:

# See available profiles
cat /sys/firmware/acpi/platform_profile_choices
# quiet balanced balanced-performance performance

# Check current
cat /sys/firmware/acpi/platform_profile

# Switch to performance
echo performance | sudo tee /sys/firmware/acpi/platform_profile

Most desktop environments (GNOME, KDE) will also pick these up in their power settings GUI.

4. What works on my system (83F5, Intel Ultra 9 275HX, RTX 5090)

• Fan speed monitoring via sensors (legion_hwmon) — CPU fan, GPU fan, CPU/GPU/IC temps
• Platform profiles (quiet / balanced / balanced-performance / performance)
• Conservation mode (battery charge limit)
• Fn Lock, keyboard backlight, Y-logo LED
• The only thing that doesn't work is the IO-Port LED (non-critical cosmetic error) and the max-fan-speed toggle throws unexpected register values

Important: you'll need to share your exact model number (cat /sys/class/dmi/id/product_name) and BIOS version (cat /sys/class/dmi/id/bios_version) — the config that works depends heavily on which variant you have (Intel vs AMD, which BIOS prefix). If you're on the AMD side, check out @ChaoticSi1ence's fork above.

[Ozzbuzzz69]

okay that worked for me i finally got the purple light (custom profile)
another question i have, how do you then control your fan speed like how do you maunally set the rpm or make custom power profiles as in how do you set the overclocking values and such, because it confuses me a bit

also thank you so much for the reply you are the absolute GOAT <3

also i have the intel variant and this is the
Product name:
83F5
and
bios version:
Q7CN45WW

[alstergee]

USB-C PD Charging Fix: Blacklist ucsi_acpi

Follow-up on my Legion Pro 7 16IAX10H (83F5, Q7CN44WW BIOS, kernel 6.18.0).

Problem

USB-C Power Delivery charging was completely broken. Symptoms:

• Chargers would connect briefly then drop after 1-3 minutes
• voltage_now always read 0 in UCSI sysfs
• usb_power_delivery_revision on partner reported 0.0 instead of 3.0
• No source-capabilities PDO objects ever appeared in sysfs
• Some chargers (Shargeek 100W) never established a PD contract at all — only 1.4W
• Failed PD negotiations would wedge the UCSI PPM, requiring a full reboot to recover
• The ACPI deferred workqueue showed CPU hogging warnings (acpi_os_execute_deferred hogged CPU for >10000us), possibly from legion-laptop WMI3 calls starving UCSI event processing

Root Cause

The ucsi_acpi kernel driver (UCSI = USB Type-C Connector System software Interface) was actively interfering with PD negotiation. The Lenovo EC's UCSI PPM (Platform Policy Manager) implementation doesn't properly populate PD data (GET_PDOS returns empty, GET_CONNECTOR_STATUS returns incomplete RDO). This is the same class of bug as the ThinkPad T14 Gen 5 (fwupd/firmware-lenovo#521, Lenovo ticket LO-4169).

Fix

Blacklist ucsi_acpi so the EC's PD controller negotiates independently:

echo "blacklist ucsi_acpi" | sudo tee /etc/modprobe.d/blacklist-ucsi.conf

Results

Charger	With ucsi_acpi	Without ucsi_acpi
Shargeek 100W	1.4W (no PD contract)	40-55W ✅
Anker 65W	~5W (5V default only)	Expected full power
Barrel jack 330W	70W ✅	70W ✅

The PD controller in the EC works perfectly fine on its own — the UCSI driver was the problem. You lose sysfs PD reporting (/sys/class/typec/, /sys/class/power_supply/ucsi-source-psy-*) but it was all broken/inaccurate anyway.

Relevant to other Gen 10 owners

If you have a Legion Gen 10 (Q7CN, Q7CM, SMCN configs) and USB-C charging doesn't work or drops out, try blacklisting ucsi_acpi. This is likely a Lenovo EC firmware issue across the Gen 10 line.

[alstergee]

Hey @Ozzbuzzz69 — I have the same laptop (Pro 7 16IAX10H, 83F5) on BIOS Q7CN44WW, so pretty close to your Q7CN45WW. Here's what I've found so far:

Fan Control

The platform profile interface works and is the primary way to control fan behavior:

# Check available profiles
cat /sys/firmware/acpi/platform_profile_choices
# quiet balanced balanced-performance performance

# Set a profile
echo performance | sudo tee /sys/firmware/acpi/platform_profile

• quiet — fans spin down to 0 RPM (fully passive cooling, will thermal throttle under sustained load)
• balanced / balanced-performance — normal behavior, fans ramp as needed
• performance — fans start at ~1600/1700 RPM baseline and ramp aggressively under load

Unfortunately, that's about as far as fan control goes on this model right now. The WMI3 fan curve interface is exposed via hwmon (10-point autopoints for PWM), and you can write PWM values to it, but all temperature thresholds read back as 0. So custom fan curves don't actually function — looks like a bug in the WMI3 implementation on Q7CN firmware. Direct manual RPM control (pwm1_mode) returns EPERM, so that's a no-go too.

TL;DR: Platform profiles are the only working fan control mechanism on this model for now.

Custom Power Profiles

The 4 platform profiles listed above are what's available. If you want one to persist across reboots, a simple systemd oneshot service does the trick:

# /etc/systemd/system/legion-performance.service
[Unit]
Description=Set platform profile on boot
After=multi-user.target

[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/bin/bash -c 'echo performance > /sys/firmware/acpi/platform_profile'

[Install]
WantedBy=multi-user.target

Then sudo systemctl enable legion-performance.service.

Overclocking

Haven't tested or attempted any overclocking on this model, so I can't offer guidance there.

Bonus: USB-C PD Charging Fix

If you're seeing broken USB-C PD charging (charger connects briefly then drops to trickle charge), blacklist ucsi_acpi:

echo "blacklist ucsi_acpi" | sudo tee /etc/modprobe.d/blacklist-ucsi.conf

There's a bug in the EC firmware's UCSI PPM implementation — it returns empty PDO data and actively sabotages PD negotiation. Same issue as ThinkPad T14 Gen 5 (Lenovo tracker LO-4169). Blacklisting the kernel module lets the EC's PD controller negotiate directly without interference. Went from ~1W to 40-55W on a Shargeek 250W charger.

System Temps

CPU and GPU temps are readable through standard hwmon — coretemp for CPU, nvidia for GPU. No special configuration needed:

# Find temp sensors
for h in /sys/class/hwmon/hwmon*/; do echo "$(cat ${h}name): $(cat ${h}temp1_input 2>/dev/null | awk '{printf "%.1f°C", $1/1000}')"; done

[alstergee]

Full Status Report — Legion Pro 7 16IAX10H (83F5), BIOS Q7CN44WW

Went through every control surface the module exposes on this model. Here's where things stand.

What Works

Feature	How	Notes
Platform Profiles	/sys/firmware/acpi/platform_profile	quiet / balanced / balanced-performance / performance — all switch instantly
Fan Speed Reading	fan1_input, fan2_input via hwmon	Reads correctly. fan3 not exposed despite 3-fan hardware
System Temps	legion_hwmon, coretemp, nvme, spd5118, iwlwifi	CPU/GPU/board temps via legion_hwmon; per-core via coretemp; NVMe, RAM, WiFi all readable
RGB LEDs (Spectrum)	Custom USB HID tool (048d:c197)	Full per-zone color/effect control via 960-byte HID Feature Reports. Wrote a Python CLI + web UI for it
Y Logo LED	/sys/class/leds/platform::ylogo/brightness	On/off (0 or 1)
Kbd Backlight (basic)	/sys/class/leds/platform::kbd_backlight/brightness	3 levels: 0/1/2. This is the white-only WMI backlight, separate from Spectrum RGB
Screen Backlight	/sys/class/backlight/intel_backlight	0–496 range, works fine
Fn Lock	ideapad_acpi	Read/write works
Fn Hotkeys	ideapad extra buttons → input events	Volume, brightness, etc. all generate proper key events
Conservation Mode	ideapad_acpi	Caps charge at ~60% — writable
Battery Stats	power_supply/BAT0	status, capacity, power_now all readable
USB-C PD Charging	Requires blacklisting ucsi_acpi	See below

What's Broken

Feature	Problem
Custom Fan Curves	PWM values write successfully, but ALL temperature thresholds read back as 0. The EC never triggers fan speed changes because it has no temp triggers. WMI3 bug on Q7CN firmware.
Manual Fan RPM	pwm1_mode returns EPERM on write — can't switch between auto/manual
Fan 3	Not exposed despite 3-fan hardware (num_fans = 3 in config)
Fan Curve Accel/Decel	All values read 0 — same underlying WMI3 issue
Rapid Charge	Not exposed by ideapad_acpi on this model

Mode Switching — One Knob Controls Everything

There's no way to independently control fan speed vs system power. The platform profile is a single knob:

• quiet = low power + fans at 0 RPM (passive, will throttle)
• balanced = normal power + fans ramp as needed
• performance = full power + fans at ~1600/1700 RPM baseline, ramp under load

Custom fan curves would allow decoupling these if the WMI3 temp thresholds worked. But they don't.

---

Divergent Code / Patches

1. Kernel Module — model_q7cn Config

The fork's closest model was model_jiah (Gen 10 AMD), which had access_method_keyboard = ACCESS_METHOD_NO_ACCESS and access_method_fancurve = ACCESS_METHOD_NO_ACCESS. Created a new model_q7cn config that enables both:

// legion-laptop.c — model_q7cn config
// Legion Pro 7 16IAX10H (2025) - Model 83F5 - Core Ultra 9 275HX, RTX 5090, 3 fans
static const struct model_config model_q7cn = {
    .registers = &ec_register_offsets_v0,
    .check_embedded_controller_id = true,
    .embedded_controller_id = 0x5508,
    .memoryio_physical_ec_start = 0xC400,
    .memoryio_size = 0x300,
    .num_fans = 3,
    .has_minifancurve = false,
    .has_custom_powermode = true,
    .access_method_powermode = ACCESS_METHOD_WMI,
    .access_method_keyboard = ACCESS_METHOD_WMI,       // was NO_ACCESS in model_jiah
    .access_method_fanspeed = ACCESS_METHOD_WMI3,
    .access_method_temperature = ACCESS_METHOD_WMI3,
    .access_method_fancurve = ACCESS_METHOD_WMI3,       // was NO_ACCESS in model_jiah
    .access_method_fanfullspeed = ACCESS_METHOD_WMI,
    .acpi_check_dev = false,
    .ramio_physical_start = 0xFE0B0400,
    .ramio_size = 0x600
};

Why: ACCESS_METHOD_WMI for keyboard enables the 3-level white backlight via sysfs (Spectrum RGB is separate, over USB). ACCESS_METHOD_WMI3 for fancurve exposes the 10-point autopoints — they write OK but temps read back as 0, so the curves don't actually function.

2. USB-C PD Charging Fix — Blacklist ucsi_acpi

# /etc/modprobe.d/blacklist-ucsi.conf
blacklist ucsi_acpi

Why: The Lenovo EC firmware's UCSI PPM implementation is broken on this model. GET_PDOS returns empty data, GET_CONNECTOR_STATUS returns incomplete RDO. The kernel's ucsi_acpi driver constantly queries this broken mailbox, which actively interferes with PD negotiation. Symptom: USB-C charger connects briefly at full wattage, then drops to trickle charge (~1W) within 30-60 seconds.

With ucsi_acpi blacklisted, the EC's PD controller negotiates directly without interference. Went from ~1W to 40-55W sustained on a 250W USB-C charger. Same bug as ThinkPad T14 Gen 5 (Lenovo tracker LO-4169, fwupd/firmware-lenovo#521).

Trade-off: You lose typec sysfs enumeration (no /sys/class/typec/port*), but since the data was garbage anyway, you lose nothing useful.

3. Boot Persistence — Systemd Services

Performance profile on boot:

# /etc/systemd/system/legion-performance.service
[Unit]
Description=Set Legion performance profile and aggressive fans on boot
After=legion-laptop.service dkms.service
After=sys-firmware-acpi-platform_profile.device

[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/bin/bash -c 'echo performance > /sys/firmware/acpi/platform_profile'

[Install]
WantedBy=multi-user.target

4. Spectrum RGB — USB HID Protocol

Gen 10 Legion uses ITE 8258 (048d:c197) for RGB lighting. Control is via 960-byte HID Feature Reports, not through the kernel module. Wrote a Python CLI and web UI that talk directly to the HID device. Full per-zone control with static, breathing, wave, and other effects. Happy to share the source if anyone wants it.

---

Questions for the Community

Has anyone with Q7CN / 83F5 (or the AMD variant 83RU) solved any of these?

1. Fan curve temperature thresholds always reading 0 — Is this a WMI3 implementation issue specific to Gen 10, or are we missing something in how the curve data is packed/unpacked? The PWM values write and read back fine, it's only the temps that are zeroed.

2. pwm1_mode EPERM — Anyone got manual fan control working on this generation? The WMI call behind it seems to reject the request entirely.

3. Fan 3 not exposed — The config says num_fans = 3 and the hardware has 3 fans, but only fan1/fan2 appear in hwmon. Is there a different access path for the third fan?

4. Any way to decouple power profile from fan behavior? — Right now platform_profile is the only working lever and it ties performance + fan speed together.

Any insights appreciated — still digging through the WMI ACPI tables to see if there's something we're missing.

[Ozzbuzzz69]

ah thats a bummer thanks a lot for the reply tho, i hope someone finds a solution to it:)

[alstergee]

Update: DSDT Deep Dive + Custom Mode (Purple) Analysis

Decompiled the ACPI DSDT on the 83F5 (Q7CN44WW) and mapped out exactly how custom mode and fan curves work at the firmware level. Sharing everything here in case it helps others.

Fan Curve Registers (F9F0-F9F9) Are Custom-Mode Only

The WMI fan table (WMAB methods 5/6) reads/writes EC registers F9F0-F9F9 — 10 bytes at EC offsets 0x80-0x89. These store fan speed percentages only (0-100). There are no temperature thresholds in the WMI protocol at all — the remaining 40 bytes of the 88-byte read buffer are hardcoded zeros in the DSDT.

Critically, F9F0-F9F9 are only the active fan curve when the EC is in Custom mode (SPMO=3). For Quiet/Balanced/Performance, the EC uses its own internal firmware-hardcoded curves. We confirmed this by reading the WMI fan table in all three modes — identical values come back, but actual fan RPMs differ:

Mode	Fan RPMs	WMI Curve Values
Performance	1600/1700	0, 12, 14, 20, 23, 25, 29, 34, 40, 45
Quiet	900/900	(same)
Balanced	300/0	(same)

The per-mode fan behavior is entirely EC firmware, not the WMI table.

Custom Mode (Purple / 0xFF / SPMO=3) — Full DSDT Analysis

Fn+Q cycles through modes via EC register SPMO. The ACPI _Q13 handler reads the new value and syncs software state. Custom mode is SPMO=3, and it's unique:

Register	Custom Mode	Other Modes
SPMO	3	0/1/2/4
CTMB (Custom Turbo Mode Boost)	1	0
SSFM (Set Smart Fan Mode flag)	1 (on switch)	1 (on switch)
EXTM (Extreme flag)	0	0 (except Extreme=1)
Requires AC?	Yes	Quiet/Balanced: No

When you enter Custom mode:

1. SSFM=1 tells the EC to re-read the fan curve
2. CTMB=1 enables custom CPU power limit tuning (SPL/SPPT/FPPT via EC registers CCP1, CCP2, F5E0, F5E2, F5E4)
3. F9F0-F9F9 becomes the active fan curve
4. F9FA controls LED color and fan curve selection bucket

If AC is removed while in Custom mode, the EC drops to Balanced (SPMO=0) and caches ASMC=2 to restore Custom when AC returns.

The FSTM Bug — Linux Driver vs Windows LenovoLegionToolkit

The WMI fan table write buffer (WMAB method 6) has a header byte FSTM at offset 0x00. The ACPI DSDT reads this field:

CreateByteField (Arg2, Zero, FSTM)

The Linux legion-laptop driver sends FSTM = 0 (via memset). Windows LenovoLegionToolkit always sends FSTM = 1. This is likely a table validity/mode flag. With FSTM=0, the EC may ignore or misinterpret the fan table.

We've patched the driver to set FSTM=1 but haven't tested it yet (we need to do this carefully because an earlier experiment with custom mode + wrong values caused a thermal shutdown — see below).

Mainline lenovo-wmi-gamezone Driver Conflict

Derek Clark's mainline WMI drivers (landed in 6.17) are compiled and loaded on our 6.18 kernel, but lenovo_wmi_gamezone cannot bind because legion_laptop grabs GUID 887B54E3-DDDC-4B2C-8B88-68A26A8835D0 first. The mainline driver supports PLATFORM_PROFILE_CUSTOM natively — if it could bind, custom mode would just work via echo custom > /sys/firmware/acpi/platform_profile.

We're working on patching legion_laptop to expose proper PLATFORM_PROFILE_CUSTOM instead of hijacking balanced-performance for custom mode.

Warning: What NOT To Do

We accidentally caused a thermal shutdown by entering custom mode and writing 100% fan speed to all 10 curve points. A flat curve kills the fans. The EC appears to interpret a flat curve (all same values) as invalid and stops the fans entirely. Combined with FSTM=0, this bricked the fan controller until reboot.

Safe fan curves should:

• Be monotonically increasing (ramp up, not flat)
• Stay within the EC's expected range (raw percentages 0-100)
• Have reasonable minimums at higher points (LLT uses minimum [0,0,0,0,0,0,0,1,3,5] on its 0-10 scale)
• Always be tested with a finger on the power button

LECR Commands (LED/EC Register Communication)

The DSDT reveals LECR as the EC command interface:

• LECR(0xD0, 1, 1, 2) — Apply fan curve from F9F0-F9F9 (called after WMAB method 6 writes)
• LECR(0xD0, 1, 2, 2) — Signal power bucket changed (called after F9FA write)
• LECR(0xDA, ...) — Y-Logo / power LED control (color, brightness)
• LECR(0xD1, zone, level, 2) — LED zone brightness

Next Steps

1. Testing FSTM=1 fix with sane fan curve values in Custom mode
2. Capturing Windows LegionSpace WMI calls to see exactly what it sends when applying a custom profile
3. Verifying fan RPM changes after writing curve in custom mode
4. Building a proper safety layer — fan curve validation + RPM watchdog

Has anyone with Gen 10 (Q7CN/83F5, SMCN/83RU, or similar) gotten custom fan curves actually working? Or captured WMI traces from LegionSpace? The DSDT decompilation is available if anyone wants to compare against their model.

[ChaoticSi1ence]

@alstergee — Incredible DSDT work. I've incorporated your findings into the fork. Here's what changed based on your analysis:

Changes pushed

1. FSTM=1 fix — The fan curve write buffer now sets buffer[0x00] = 1 (the table validity flag). You nailed it: the driver was sending FSTM=0 via memset, which likely caused the EC to ignore custom fan curves entirely. This may be the missing piece for making custom fan curves actually work in Custom mode.

2. Keyboard backlight enabled — Changed access_method_keyboard from ACCESS_METHOD_NO_ACCESS to ACCESS_METHOD_WMI for both model_q7cn and model_smcn. This enables the 3-level white WMI backlight you confirmed working (/sys/class/leds/platform::kbd_backlight/brightness, levels 0/1/2). The per-zone RGB (048d:c197 USB HID) is completely separate and unaffected.

3. Fan curve monotonicity validation — Added a hard reject (-EINVAL) for non-monotonic fan curve writes. Based on your finding that a flat curve caused thermal shutdown by making the EC stop the fans. The check runs always (including in wmi_dryrun mode) to catch errors during testing.

4. WMI3 fan curve architecture documented — Added comprehensive code comments based on your DSDT analysis: speed-only percentages (no temp thresholds), Custom mode only (SPMO=3), LECR(0xD0,1,1,2) apply command, and the flat curve danger.

5. USB-C PD fix documented — Added your ucsi_acpi blacklist fix to the README with credit.

Addressing your questions

Fan 3 not exposed — This is because GKCN/model_v0 doesn't set num_fans (defaults to 0), so the hwmon visibility function hides fan3_input. The fork's model_q7cn config sets num_fans = 3, which makes fan3_input and fan3_label visible. When you test the fork, fan3 should appear. The read path dispatches fan_id=2 to OtherMethodFeature_FAN_SPEED_4 (0x04030004) — the non-sequential fan ID is already handled.

Fan curve temperature thresholds always 0 — Your DSDT analysis confirms this is by design, not a bug. The WMI3 SET_TABLE protocol only carries speed percentages. The remaining bytes in the read buffer are hardcoded zeros in the DSDT. The hwmon autopoints expose temperature columns, but they're meaningless on Gen 10 WMI3 hardware.

pwm1_mode EPERM — This goes through lockfancontroller, which writes EC register 0xC4AB via port I/O (not WMI). The EC may not support this on Gen 10 firmware. Still investigating.

ramio_physical_start — I verified our Q7CN45WW machine maps at 0xFE0B0400 (visible in the Unmapping embedded controller dmesg line on rmmod). Your 0xFE00D400 came from GKCN/model_v0's hardcoded value, not from your hardware. Could you check your actual address? After sudo rmmod legion_laptop, look for:

dmesg | grep "Unmapping embedded"

This tells us the real address for Q7CN44WW. It likely matches ours (0xFE0B0400) since it's the same 83F5 hardware.

For @Ozzbuzzz69

You have the same laptop (83F5, Q7CN45WW). The fork has a proper model_q7cn config so you don't need force=1. It also exposes the 3rd fan, has wmi_dryrun safety mode, and now includes the FSTM fix for custom fan curves. Build instructions are in the README.