Add support for Lenovo LOQ 15IRX10 (R3CN)

[alfrix]

Installation type: github

make CC=clang LD=ld.lld && sudo make forcereloadmodule

Distribution: CachyOS
Model: LOQ: 15IRX10 / Type: 83JE
BIOS: R3CN41WW
CPU model: Intel Core i5 13450HX
GPU model: NVIDIA RTX 5060
Keyboard backlight: RGB
Light in lid or logo: no
Light at IO-Ports at back: no

Output of sudo dmidecode -t system

Getting SMBIOS data from sysfs.
SMBIOS 3.4 present.

Handle 0x0001, DMI type 1, 27 bytes
System Information
        Manufacturer: LENOVO
        Product Name: 83JE
        Version: LOQ 15IRX10
        Wake-up Type: Power Switch
        SKU Number: LENOVO_MT_83JE_BU_idea_FM_LOQ 15IRX10
        Family: LOQ 15IRX10

Handle 0x0024, DMI type 12, 5 bytes
System Configuration Options
        Option 1: ConfigOptions1
        Option 2: ConfigOptions2
        Option 3: ConfigOptions3

Handle 0x0027, DMI type 15, 29 bytes
System Event Log
        Area Length: 0 bytes
        Header Start Offset: 0x0000
        Header Length: 8192 bytes
        Data Start Offset: 0x2000
        Access Method: General-purpose non-volatile data functions
        Access Address: 0x0000
        Status: Valid, Not Full
        Change Token: 0x12345678
        Header Format: OEM-specific
        Supported Log Type Descriptors: 3
        Descriptor 1: POST memory resize
        Data Format 1: None
        Descriptor 2: POST error
        Data Format 2: POST results bitmap
        Descriptor 3: Log area reset/cleared
        Data Format 3: None

Handle 0x0034, DMI type 32, 11 bytes
System Boot Information
        Status: No errors detected

Output of sudo dmidecode -t bios:

Getting SMBIOS data from sysfs.
SMBIOS 3.4 present.

Handle 0x0000, DMI type 0, 26 bytes
BIOS Information
        Vendor: LENOVO
        Version: R3CN41WW
        Release Date: 09/18/2025
        Address: 0xE0000
        Runtime Size: 128 kB
        ROM Size: 48 MB
        Characteristics:
                PCI is supported
                BIOS is upgradeable
                BIOS shadowing is allowed
                Boot from CD is supported
                Selectable boot is supported
                EDD is supported
                Japanese floppy for NEC 9800 1.2 MB is supported (int 13h)
                Japanese floppy for Toshiba 1.2 MB is supported (int 13h)
                5.25"/360 kB floppy services are supported (int 13h)
                5.25"/1.2 MB floppy services are supported (int 13h)
                3.5"/720 kB floppy services are supported (int 13h)
                3.5"/2.88 MB floppy services are supported (int 13h)
                8042 keyboard services are supported (int 9h)
                CGA/mono video services are supported (int 10h)
                ACPI is supported
                USB legacy is supported
                BIOS boot specification is supported
                Targeted content distribution is supported
                UEFI is supported
        BIOS Revision: 1.41
        Firmware Revision: 1.41

Handle 0x0025, DMI type 13, 22 bytes
BIOS Language Information
        Language Description Format: Long
        Installable Languages: 8
                en|US|iso8859-1,0
                fr|FR|iso8859-1,0
                zh|TW|unicode,0
                ja|JP|unicode,0
                it|IT|iso8859-1,0
                es|ES|iso8859-1,0
                de|DE|iso8859-1,0
                pt|PT|iso8859-1,0
        Currently Installed Language: en|US|iso8859-1,0

Output of sudo cat /sys/kernel/debug/legion/fancurve:

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: 4245487636
temperature access method: 5
CPU temperature error: 0
CPU temperature: 50
CPU temperature EC error: 0
CPU temperature EC: 0
CPU temperature ACPI error: -5
CPU temperature ACPI: 0
CPU temperature WMI error: 0
CPU temperature WMI: 50
CPU temperature WMI2 error: 0
CPU temperature WMI2: 0
CPU temperature WMI3 error: 0
CPU temperature WMI3: 50
GPU temperature error: 0
GPU temperature: 46
GPU temperature EC error: 0
GPU temperature EC: 0
GPU temperature ACPI error: -5
GPU temperature ACPI: 0
GPU temperature WMI error: 0
GPU temperature WMI: 46
GPU temperature WMI2 error: 0
GPU temperature WMI2: 0
GPU temperature WMI3 error: 0
GPU temperature WMI3: 46
fan speed access method: 5
1 fanspeed error: 0
1 fanspeed: 1600
1 fanspeed EC error: 0
1 fanspeed EC: 0
1 fanspeed ACPI error: -5
1 fanspeed ACPI: 0
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: 1600
2 fanspeed error: 0
2 fanspeed: 1700
2 fanspeed EC error: 0
2 fanspeed EC: 0
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: 3
powermode EC error: 0
powermode EC: 98
powermode ACPI error: -5
powermode ACPI: 0
powermode WMI error: 0
powermode WMI: 3
has custom powermode: 1
ACPI rapidcharge error: 0
ACPI rapidcharge: 0
WMI backlight 2 state: 0
WMI backlight 3 state: -14
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: -22
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        0       0       0       0       0       0       0       50      0       0       0       0
3        11      11      28      28      0       0       40      54      0       0       0       0
3        16      16      40      40      0       0       50      58      0       0       0       0
3        17      17      43      43      0       0       54      62      0       0       0       0
3        19      19      48      48      0       0       58      66      0       0       0       0
3        21      21      53      53      0       0       62      70      0       0       0       0
3        24      24      61      61      0       0       66      74      0       0       0       0
3        28      28      71      71      0       0       70      78      0       0       0       0
3        30      30      76      76      0       0       74      82      0       0       0       0
3        33      33      84      84      0       0       78      86      0       0       0       0
=====================
Current fan curve in hardware (WMI; might be empty)
Fan curve current point id: 0
Fan curve points size: 0
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
=====================

dsl files from R3CN.zip

so i've added

diff --git a/kernel_module/legion-laptop.c b/kernel_module/legion-laptop.c
index 6d39a25..8304747 100644
--- a/kernel_module/legion-laptop.c
+++ b/kernel_module/legion-laptop.c
@@ -1006,6 +1006,25 @@ static const struct model_config model_nzcn = {
 	.ramio_size = 0x600
 };
 
+static const struct model_config model_r3cn = {
+	.registers = &ec_register_offsets_loq_v0,
+	.check_embedded_controller_id = true,
+	.embedded_controller_id = 0x8227,
+	.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_WMI2,
+	.access_method_fanspeed = ACCESS_METHOD_WMI3,
+	.access_method_temperature = ACCESS_METHOD_WMI3,
+	.access_method_fancurve = ACCESS_METHOD_EC3,
+	.access_method_fanfullspeed = ACCESS_METHOD_WMI3,
+	.acpi_check_dev = false,
+	.ramio_physical_start = 0xFE0B0400,
+	.ramio_size = 0x600
+};
+
 static const struct dmi_system_id denylist[] = { {} };
 
 static const struct dmi_system_id optimistic_allowlist[] = {
@@ -1381,6 +1400,15 @@ static const struct dmi_system_id optimistic_allowlist[] = {
 		},
 		.driver_data = (void *)&model_nzcn
 	},
+	{
+		// e.g. LOQ 15IRX10 (Intel 13450HX + RTX 5060)
+		.ident = "R3CN",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
+			DMI_MATCH(DMI_BIOS_VERSION, "R3CN"),
+		},
+		.driver_data = (void *)&model_r3cn
+	},
 	{}
 };
 

this is forced with master

forced with pr #347

those values seem closer, but also not correct, i had it in performance mode and i'm pretty sure that it should reach 4400 and fan speeds are asymmetric and not configurable independently on legion-space, so i expected to see that.

but at least i have some hope to see fan speeds with this module:

testing with some load it starts to show the expected fan asymmetry

[qquique]

The values that cat /sys/kernel/debug/legion/fancurve shows don't match what you have in the dsdt , maybe you ran it before applying the PR. Indeed the max rpm value can be 4100, 4400 and 5500 . These are some of the values from the tables :

						| FNT0 		| FNT1 	 | FNT2
Thermal Mode   			| 0x03 		| 0x03 	 | 0x03
FanId           		| One 		| One 	 | One
FanTableLen     		| 0x0A 		| 0x0A 	 | 0x0A
FanSpeeds            
   						| Zero   	| Zero   | Zero
   						| 0X044C 	| 0X044C | 0x04B0
						| 0X0640 	| 0x06A4 | 0x06A4
						| 0X06A4 	| 0x0834 | 0x0898
						| 0X0834 	| 0x0960 | 0x0A28
						| 0X0960 	| 0x0BB8 | 0x0CE4
						| 0X0BB8 	| 0x0CE4 | 0x0ED8
						| 0X0CE4 	| 0x1004 | 0x10CC
						| 0X0E10 	| 0x1130 | 0x11F8
						| 0X1004 	| 0x157C | 0x125C
CurrentFantMaxSpeed 	| 0X1004 	| 0x157C | 0x125C
DesignMaxSpeedNumber 	| 0X0A 		| 0x08 	 | 0x08

The interesting part is that that model can have a 10 values fancurve or 8 values 🤔 , this project module currently uses 10 values fixed.

Can you try again using the PR to see in which FNT table falls:

• echo 'balanced-performance' > /sys/firmware/acpi/platform-profile #set custom mode to modify values
• echo 3 > /sys/class/hwmon/hwmon<number_where_it_fall_device>/auto_points_defaults#use the performance default fan curve
• cat /sys/kernel/debug/legion/fancurve
• cat /sys/kernel/debug/legion/ecmemory | hexdump -C

Regarding the output of the gui I have a branch where I was just testing some kind of rounding, you can try it, I don't use the gui so I didn't made a PR for that.

[alfrix]

[root@15irx10-cachyos ~]# echo balanced-performance > /sys/firmware/acpi/platform_profile
[root@15irx10-cachyos ~]# echo 3 > /sys/class/hwmon/hwmon9/auto_points_defaults
[root@15irx10-cachyos ~]# cat /sys/kernel/debug/legion/fancurve
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: 4245487636
temperature access method: 5
CPU temperature error: 0
CPU temperature: 48
CPU temperature EC error: 0
CPU temperature EC: 0
CPU temperature ACPI error: -5
CPU temperature ACPI: 0
CPU temperature WMI error: 0
CPU temperature WMI: 48
CPU temperature WMI2 error: 0
CPU temperature WMI2: 0
CPU temperature WMI3 error: 0
CPU temperature WMI3: 48
GPU temperature error: 0
GPU temperature: 44
GPU temperature EC error: 0
GPU temperature EC: 0
GPU temperature ACPI error: -5
GPU temperature ACPI: 0
GPU temperature WMI error: 0
GPU temperature WMI: 44
GPU temperature WMI2 error: 0
GPU temperature WMI2: 0
GPU temperature WMI3 error: 0
GPU temperature WMI3: 44
fan speed access method: 5
1 fanspeed error: 0
1 fanspeed: 2100
1 fanspeed EC error: 0
1 fanspeed EC: 0
1 fanspeed ACPI error: -5
1 fanspeed ACPI: 0
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: 2100
2 fanspeed error: 0
2 fanspeed: 2400
2 fanspeed EC error: 0
2 fanspeed EC: 0
2 fanspeed ACPI error: 0
2 fanspeed ACPI: 2400
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: 2400
powermode access method: 3
powermode error: 0
powermode: 255
powermode EC error: 0
powermode EC: 98
powermode ACPI error: -5
powermode ACPI: 0
powermode WMI error: 0
powermode WMI: 255
has custom powermode: 1
ACPI rapidcharge error: 0
ACPI rapidcharge: 0
WMI backlight 2 state: 0
WMI backlight 3 state: -14
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: -22
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        0       0       0       0       0       0       0       50      0       0       0       0
3        11      11      28      28      0       0       40      57      0       0       0       0
3        16      16      40      40      0       0       53      64      0       0       0       0
3        17      17      43      43      0       0       60      71      0       0       0       0
3        21      21      53      53      0       0       68      85      0       0       0       0
3        24      24      61      61      0       0       81      88      0       0       0       0
3        30      30      76      76      0       0       85      93      0       0       0       0
3        33      33      84      84      0       0       90      95      0       0       0       0
3        36      36      91      91      0       0       92      97      0       0       0       0
3        36      36      91      91      0       0       94      100     0       0       0       0
=====================
Current fan curve in hardware (WMI; might be empty)
Fan curve current point id: 0
Fan curve points size: 0
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
=====================

[root@15irx10-cachyos ~]# cat /sys/kernel/debug/legion/ecmemory | hexdump -C
00000000  00 00 03 00 00 00 00 09  01 00 90 60 00 8d 5b 00  |...........`..[.|
00000010  00 00 00 00 00 00 00 00  22 41 01 28 40 62 00 02  |........"A.(@b..|
00000020  e6 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000030  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000060  00 80 16 52 1e 0b 00 00  00 00 00 00 00 00 00 00  |...R............|
00000070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000080  00 00 00 00 02 00 00 00  00 00 00 00 00 00 00 00  |................|
00000090  00 2c 2a 28 00 00 2d 00  00 00 00 00 00 00 00 00  |.,*(..-.........|
000000a0  01 00 00 c0 40 00 00 2c  23 00 00 08 00 2a 00 78  |....@..,#....*.x|
000000b0  2c 2c 27 00 23 00 2d 2a  00 00 00 0f 00 42 50 5b  |,,'.#.-*.....BP[|
000000c0  0a 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000000d0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
000000f0  00 00 00 00 00 00 00 00  00 00 00 ff ff 00 00 00  |................|
00000100  04 65 25 39 15 07 05 00  32 00 28 39 0b 35 40 10  |.e%9....2.(9.5@.|
00000110  3c 47 11 44 55 15 51 58  18 55 5d 1e 5a 5f 21 5c  |<G.DU.QX.U].Z_!\|
00000120  61 24 5e 64 24 00 00 00  00 00 00 00 00 00 00 00  |a$^d$...........|
00000130  00 15 15 04 65 2a 43 18  08 00 00 32 00 28 39 0e  |....e*C....2.(9.|
00000140  35 40 11 3c 47 14 44 55  18 51 58 1b 55 5d 21 5a  |5@.<G.DU.QX.U]!Z|
00000150  5f 24 5c 61 27 5e 64 27  00 00 00 00 00 00 00 00  |_$\a'^d'........|
00000160  00 00 00 00 18 18 00 00  00 00 00 00 00 00 00 00  |................|
00000170  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000200  01 01 01 01 00 00 11 00  00 00 00 00 00 00 00 00  |................|
00000210  02 0c 0e 11 15 15 15 15  15 2a 00 00 00 00 00 00  |.........*......|
00000220  5a a5 00 00 00 00 00 00  00 00 00 59 b1 00 00 00  |Z..........Y....|
00000230  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000240  01 23 00 00 00 00 00 20  04 00 02 00 00 00 00 00  |.#..... ........|
00000250  00 00 00 40 00 00 00 00  00 00 00 00 00 00 00 00  |...@............|
00000260  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000270  05 da 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000280  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000380  00 0b 00 00 00 48 07 00  00 00 00 00 00 00 00 00  |.....H..........|
00000390  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000400  47 00 00 00 00 01 00 00  5a a5 24 21 01 01 04 00  |G.......Z.$!....|
00000410  04 4a 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |.J..............|
00000420  01 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000430  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000470  00 00 01 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000480  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000490  00 00 00 80 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000004a0  21 21 05 00 00 00 00 00  00 00 00 00 00 00 00 00  |!!..............|
000004b0  c2 5d c9 63 03 00 00 00  00 00 00 00 00 00 00 00  |.].c............|
000004c0  21 62 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |!b..............|
000004d0  03 0e 3b 3b 0e 0e 00 00  00 00 00 00 00 00 00 00  |..;;............|
000004e0  f8 02 f8 02 00 00 5a 02  4c 01 00 00 4c 34 03 00  |......Z.L...L4..|
000004f0  00 0a 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000500  00 00 00 00 06 00 00 00  00 00 00 00 00 00 00 00  |................|
00000510  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000520  01 00 0a 00 00 58 53 00  00 00 00 00 00 00 00 00  |.....XS.........|
00000530  00 00 15 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000540  01 00 00 00 00 00 00 00  00 00 01 01 01 01 01 01  |................|
00000550  01 01 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000560  f6 00 f6 00 59 01 59 01  a1 01 a1 01 00 0e 03 64  |....Y.Y........d|
00000570  00 00 00 5f 5f 00 00 00  00 00 00 00 00 03 a2 a2  |...__...........|
00000580  03 00 00 00 00 00 00 30  00 00 00 00 00 00 00 00  |.......0........|
00000590  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000005a0  00 01 b1 03 00 00 60 00  00 00 00 00 80 00 00 00  |......`.........|
000005b0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 ff  |................|
000005c0  01 00 02 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000005d0  00 00 0a 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000005e0  00 00 00 00 d5 c9 00 00  00 00 00 00 00 00 00 00  |................|
000005f0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000600
[root@15irx10-cachyos ~]# 

this is with your pr and including the rounding commit (and my local whitelist)

bafe8b9 (HEAD -> loq-fancurves) fix pwm rounding
31be28d Add Lenovo LOQ 15IRX10
24779e5 (qquique/loq-fancurves) Use auto_points_defaults to select default fancurve from other powermodes when in custom mode.
4ffb9a0 Fixes for LZCN and NZCN models.
210170c Fixes LOQ model fancurves, new wmi method's

[qquique]

yep, that confirms that it falls in table FNT0 with 10 values for the fans, now you can use the default fancurves from the others power modes (quiet 1, balanced 2, performance 3 , extreme 224, custom 255) with the auto_points_defaults.

The last value is repeated because that is how it worked the Vantage App on Windows for the LZCN model, limiting the last range. I don't know if you still have access to that app on Windows to confirm, if is also limiting there, if not it can be just a change on this line :

fan_table.F00C = 0x09;

Thanks for testing.

[alfrix]

In the new models you need to use LegionSpace and not Vantage

Here i extracted the data from the LegionSpace app:
LOQ15IRX10_profiles.ods
or in image form:

[qquique]

with that data, and that the values are asymetric, the ecmemory shows that is using also the second vector of the FNT0 table, that is different to the old LZCN model that are the same values.

from dsl

Thermal Mode   			| 0x03 		| 0x03
FanId           		| One 		| 0x02
FanTableLen     		| 0x0A 		| 0x0A
FanSpeeds            
   						| Zero   	| Zero
   						| 0X044C 	| 0x0578
						| 0X0640 	| 0x064A
						| 0X06A4 	| 0x07D0
						| 0X0834 	| 0x0960
						| 0X0960 	| 0x0A8C
						| 0X0BB8 	| 0x0CE4
						| 0X0CE4 	| 0x0E10
						| 0X0E10 	| 0x0F3C
						| 0X1004 	| 0x1130
CurrentFantMaxSpeed 	| 0X1004 	| 0x1130
DesignMaxSpeedNumber 	| 0X0A 		| 0x0A

from ec

000000f0  00 00 00 00 00 00 00 00  00 00 00 ff ff 00 00 00  |................|
00000100  04 65 25 39 15 07 05 00  32 00 28 39 0b 35 40 10  |.e%9....2.(9.5@.|
dec                             0  50  0 40 57 11 53 64 16
00000110  3c 47 11 44 55 15 51 58  18 55 5d 1e 5a 5f 21 5c  |<G.DU.QX.U].Z_!\|
dec       60 71 17 68 85 21 81 88  24 85 93 30 90 95 33 92
00000120  61 24 5e 64 24 00 00 00  00 00 00 00 00 00 00 00  |a$^d$...........|
dec       97 36 94 100 36 
00000130  00 15 15 04 65 2a 43 18  08 00 00 32 00 28 39 0e  |....e*C....2.(9.|
dec                                       0 50  0 40 57 14
00000140  35 40 11 3c 47 14 44 55  18 51 58 1b 55 5d 21 5a  |5@.<G.DU.QX.U]!Z|
dec       53 64 17 60 71 20 68 85  24 81 88 27 85 93 33 90
00000150  5f 24 5c 61 27 5e 64 27  00 00 00 00 00 00 00 00  |_$\a'^d'........|
dec       95 36 92 97 39 94 100 39
00000160  00 00 00 00 18 18 00 00  00 00 00 00 00 00 00 00  |................|
00000170  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|

0  50  0 40 57 11 53 64 16 60 71 17 68 85 21 81 88  24 85 93 30 90 95 33 92 97 36 94 100 36
0  50  0 40 57 14 53 64 17 60 71 20 68 85 24 81 88  27 85 93 33 90 95 36 92 97 39 94 100 39

0 11 16 17 21 24 30 33 36 36
0 14 17 20 24 27 33 36 39 39

The PR code is assuming the same value :

static int ec_read_fancurve_loq(struct ecram *ecram,
				const struct model_config *model,
				struct fancurve *fancurve)
{
	...
		point->speed1 =
			ecram_read(ecram, model->registers->EXT_FAN1_BASE + (i * struct_offset));
		point->speed2 = point->speed1;

and because the ec offsets are different from LZCN/NZCN it will need to modify the v0 and create a new ec_registers v1 structure for that model for the EXT_FAN2_BASE.

for LZCN : 0xc509 + 6 + 16 + 16 + 1 = 0xc530
for R3CN : 0xc509 + 6 + 16 +16 + 8 + 5 = 0xc53c

static const struct ec_register_offsets ec_register_offsets_loq_v0 = {
...
	.EXT_FAN1_BASE = 0xC509, 
	.EXT_FAN2_BASE = 0xC530, 


static const struct ec_register_offsets ec_register_offsets_loq_v1 = {
...
	.EXT_FAN1_BASE = 0xC509, 
	.EXT_FAN2_BASE = 0xC53C,

and modify the model_r3cn to use the v1 and the c functions to use the EXT_FAN2_BASE . 😪 🛌

[alfrix]

compiled all the changes on my branch
https://github.com/alfrix/LenovoLegionLinux/tree/loq-2025

current reading of fan curves
silent

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        0       0       0       0       0       0       0       50      0       0       0       0
3        11      14      28      28      0       0       40      54      0       0       0       0
3        16      17      40      40      0       0       50      58      0       0       0       0
3        17      20      43      43      0       0       54      62      0       0       0       0
3        19      22      48      48      0       0       58      66      0       0       0       0
3        21      24      53      53      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       62      70      0       0       0       0
=====================

balanced

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        0       0       0       0       0       0       0       50      0       0       0       0
3        11      14      28      28      0       0       40      54      0       0       0       0
3        16      17      40      40      0       0       50      58      0       0       0       0
3        17      20      43      43      0       0       54      62      0       0       0       0
3        19      22      48      48      0       0       58      66      0       0       0       0
3        21      24      53      53      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       66      74      0       0       0       0
3        28      31      71      71      0       0       70      78      0       0       0       0
3        30      33      76      76      0       0       74      82      0       0       0       0
3        33      36      84      84      0       0       74      82      0       0       0       0
=====================

performance

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        0       0       0       0       0       0       0       50      0       0       0       0
3        11      14      28      28      0       0       40      54      0       0       0       0
3        16      17      40      40      0       0       50      58      0       0       0       0
3        17      20      43      43      0       0       54      62      0       0       0       0
3        19      22      48      48      0       0       58      66      0       0       0       0
3        21      24      53      53      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       66      74      0       0       0       0
3        28      31      71      71      0       0       70      78      0       0       0       0
3        30      33      76      76      0       0       74      82      0       0       0       0
3        33      36      84      84      0       0       78      86      0       0       0       0
=====================

custom

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        0       0       0       0       0       0       0       50      0       0       0       0
3        11      14      28      28      0       0       40      54      0       0       0       0
3        16      17      40      40      0       0       50      58      0       0       0       0
3        17      20      43      43      0       0       54      62      0       0       0       0
3        19      22      48      48      0       0       58      66      0       0       0       0
3        21      24      53      53      0       0       62      70      0       0       0       0
3        24      27      61      61      0       0       66      74      0       0       0       0
3        28      31      71      71      0       0       70      78      0       0       0       0
3        30      33      76      76      0       0       74      82      0       0       0       0
3        33      36      84      84      0       0       78      86      0       0       0       0
=====================

also the gui doesnt round the values properly

[qquique]

also the gui doesnt round the values properly

Maybe thats why I didnt make a PR of that at that time ..

your model is already on the PR .

EDIT:
I've tested and it correctly converts the PWM to RPM 🤔

$ cat /sys/class/hwmon/hwmon4/pwm1_auto_point1_pwm
35
$ cat /sys/class/hwmon/hwmon4/pwm2_auto_point1_pwm
40

return round(((self.get_fan_1_speed_pwm(point_id) * self.get_fan_1_max_rpm() + (100 * 255) - 1) // (100 * 255)) * 100 , ndigits=2)

>>> ((35 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
1400
>>> ((40 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
1600
>>>

your values :

>>> ((28 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
1100
>>> ((40 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
1600
>>> ((43 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
1700
>>> ((48 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
1900
>>> ((53 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
2100
>>> ((61 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
2400
>>> ((71 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
2800
>>> ((76 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
3000
>>> ((84 * 10000 + (100 * 255) -1) // (100 * 255)) * 100
3300

Only the second column in your cat /sys/kernel/debug/legion/fancurve pwm2 is wrong, is the same value as pwm1, should be different for your values, checking is wrong the old code not related to the PR :

static ssize_t fancurve_print_seqfile(const struct fancurve *fancurve,
				      struct seq_file *s)
{
...
		fancurve_get_speed_pwm(fancurve, i, 0, &speed_pwm1);
		fancurve_get_speed_pwm(fancurve, i, 0, &speed_pwm2);

should be 1 as in Line 5282

fancurve_get_speed_pwm(fancurve, i, 1, &speed_pwm2);

[BaraaLazkani]

I also have the LOQ 15IRX10 (83JE) with BIOS R3CN42WW (latest from Dec 25, 2025).

Issue:

• GPU power limited to 35W on Ubuntu 24.04 (nvidia-smi shows 35W cap)
• Windows works fine at 100W with Legion Space
• This severely impacts ML training performance (3x slower)

My setup:

• Distribution: Ubuntu 24.04
• Kernel: 6.14.0-37-generic
• CPU: Intel Core Ultra 5 125H
• GPU: NVIDIA RTX 5050 Laptop
• BIOS: R3CN42WW (newer than yours - Dec 25, 2025)

Background:
I'm an AI/ML engineer, not a Linux expert. My work involves training deep learning models, so the 35W→100W unlock would be critical for my workflow. I'd be very willing to test the patch, but I'll need detailed step-by-step instructions on how to apply it.

@alfrix - Could you provide clear instructions for someone not experienced with kernel module patching? Happy to test and provide logs/feedback!

Thanks for working on LOQ support!

[alfrix]

@BaraaLazkani

I'm an AI/ML engineer, not a Linux expert. My work involves training deep learning models, so the 35W→100W unlock would be critical for my workflow. I'd be very willing to test the patch, but I'll need detailed step-by-step instructions on how to apply it.

you don't really need the kernel module or LLL at all, just change the profile with fn+q
performance (red) will give you 85w+15w if the cpu has margin, so it reches 100w

custom profile
to get 100w base you need to configure the custom profile and enable the option to be in the fn+q rotation in legion-space (windows), but in my testing this will not give you 115w (100w base + cpu to gpu power.)
i just use the custom profile to lower the max thermal boundaries, i dont like my cpu going more than 90c when compiling, because the whole laptop starts to feel hot, and durability may be impacted, remember that motherboard components are affected too.

Off-topic:
the last 15w it's like 2-3fps difference, it definitively tapers off, 55w to 85w improves quite a bit, but the last 15w just doesnt do a lot
for llm usage yi coder chat 9b on gpu:
quiet: 46tk/s 60w
balanced: 50tk/s 70w
performance: 60tk/s 100w
power was measured with nvidia-smi and you can check my spreadsheet on the above where you can see that power is ctgp+cpu to gpu, again custom profile is equal to performance because it seems to be limited to 100w it doesn sum the cpu to gpu power.

so i've tried with gpt-oss-20b mixed with 14/24 layers on the gpu
quiet: 9.5tk/s
balanced: 22tk/s
performance: 23tk/s
the gpu in this case always used around 20w so it's more cpu limited, and the cpu power limits played a bigger role

again a custom profile with all the settings of performance but a manual tgp of 100w achieved exactly the same results as the performance profile.

test have been done on cachy-os with the nvidia-590 driver, everything worked as expected

performance profile seems to be balanced to sustain a 165w load, it makes sense since the charger for some models is 170w, it also makes sense that the bios provides a "battery level protection" which will lower the overall power budget to avoid discharging the battery, since you need to account the screen, kbd lights, ssd, motherboard, etc
so to sustain the full power for long periods you would probably need the 245w charger

TLDR: you don't need the module, switching profiles with fn+q works as expected and i think there is no difference between custom with manual 100w tgp and performance that can also reach it if the cpu is not overloaded.

[BaraaLazkani]

Hi @alfrix,

I have the same LOQ 15IRX10 (83JE) and I'm experiencing the issue you mentioned, but with some differences.

My setup:

• Model: LOQ 15IRX10 (83JE)
• BIOS: R3CN42WW (Dec 25, 2025 - newer than yours)
• Distribution: Ubuntu 24.04 LTS
• Kernel: 6.14.0-37-generic
• CPU: Intel Core Ultra 5 125H
• GPU: NVIDIA RTX 5050 Laptop

The problem:
You mentioned Fn+Q should work without the kernel module, but it doesn't for me:

• Fn+Q changes the LED color (hardware responds)
• But nvidia-smi stays stuck at 35W regardless of mode
• /sys/firmware/acpi/platform_profile doesn't exist
• ideapad_laptop is loaded, but dmesg shows: ideapad_acpi VPC2004:00: DYTC interface is not available

For me I don't care about the extra 15, I just want when I toggle to performance mode ot the custom mode I have set in legion space to get 90-100 watt for my gpu. not capped to 35 watt. I had RTX 3050 4vram before and it was going faster than the RTX5050 8vram with this cap limit.

Questions:

1. Could the newer BIOS (R3CN42WW vs your R3CN41WW) cause different behavior?
2. Do you have any idea how could I solve this issue?

Thanks for your work on this! Having LOQ support would help many of us doing ML work on Linux.

[alfrix]

Could the newer BIOS (R3CN42WW vs your R3CN41WW) cause different behavior?

i would need to test it, but i doubt it

Do you have any idea how could I solve this issue?

maybe the issue is ubuntu or a config, i would try cachyOS if i were you, a live-usb should suffice for testing

[BaraaLazkani]

I'm afriad I can't use sth other than ubuntu because I work on ROS 2 Jazzy which is supported very much by only ubuntu 24.04
@alfrix

[alfrix]

I'm afriad I can't use sth other than ubuntu because I work on ROS 2 Jazzy which is supported very much by only ubuntu 24.04 @alfrix

even if you can't use cachyOS, there is value in testing, you could determine if there is an issue in the OS or in the BIOS, if cachyOS does not exhibit this behavior of locking to 35w, then it's definitively something in ubuntu.
if it does, then you may try resetting the bios or if that doesn't work then reflash it.
again the profiles should work regardless of having LLL or not
also testing that windows is performing normally with the power setting would be worth it
if it's really down to ubuntu only, then you will have to explore possible new drivers (580-590) on it maybe there is a backports repo, idk, i don't use ubuntu since the 10 or around that version

[BaraaLazkani]

@alfrix

Attempt 1: Initial Module Load with Corrected EC ID

Changes Made

Modified ~/LenovoLegionLinux/kernel_module/legion-laptop.c to add R3CN BIOS support with EC chip 0x5508:

static const struct model_config model_r3cn = {
    .registers = &ec_register_offsets_loq_v0,
    .check_embedded_controller_id = true,
    .embedded_controller_id = 0x5508,  // Changed from 0x8227
    .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_WMI2,
    .access_method_fanspeed = ACCESS_METHOD_WMI3,
    .access_method_temperature = ACCESS_METHOD_WMI3,
    .access_method_fancurve = ACCESS_METHOD_EC3,
    .access_method_fanfullspeed = ACCESS_METHOD_WMI3,
    .acpi_check_dev = false,
    .ramio_physical_start = 0xFE0B0400,
    .ramio_size = 0x600
};

Added to allowlist:

{
    .ident = "R3CN",
    .matches = {
        DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
        DMI_MATCH(DMI_BIOS_VERSION, "R3CN"),
    },
    .driver_data = (void *)&model_r3cn
},

Commands Executed

cd ~/LenovoLegionLinux/kernel_module
make clean && make
sudo rmmod legion-laptop
sudo insmod legion-laptop.ko
sudo dmesg | tail -30

dmesg Output (Successful Load)

[  735.028509] legion PNP0C09:00: Read embedded controller ID 0x5508
[  735.028509] legion PNP0C09:00: Creating debugfs interface
[  735.028518] legion_laptop: Creating sysfs interface
[  735.028527] legion_laptop: Creating hwmon interface
[  735.028586] legion_laptop: Creating platform profile support
[  735.028604] legion_laptop: Init WMI driver support
[  735.028642] legion_wmi 887B54E3-DDDC-4B2C-8B88-68A26A8835D0: Register after probing for WMI.
[  735.028659] legion_wmi BFD42481-AEE3-4501-A107-AFB68425C5F8: Register after probing for WMI.
[  735.028666] legion_wmi D062906B-12D4-4510-999D-4831EE80E985: Register after probing for WMI.
[  735.028678] legion_wmi BFD42481-AEE3-4502-A107-AFB68425C5F8: Register after probing for WMI.
[  735.028684] legion_wmi BC72A435-E8C1-4275-B3E2-D8B8074ABA59: Register after probing for WMI.
[  735.028691] legion_wmi 10AFC6D9-EA8B-4590-A2E7-1CD3C84BB4B1: Register after probing for WMI.
[  735.028697] legion_wmi D320289E-8FEA-41E0-86F9-611D83151B5F: Register after probing for WMI.
[  735.028718] legion_laptop: Init keyboard backlight LED driver
[  735.028954] legion PNP0C09:00: Led platform::kbd_backlight renamed to platform::kbd_backlight_1 due to name collision
[  735.029028] legion_laptop: Init Y-Logo LED driver
[  735.029046] legion_laptop: Error WMI call for reading brightness: expected a value between 0 and 1, but got 3
[  735.029047] legion_laptop: Error reading brightness for light: 3
[  735.029048] legion PNP0C09:00: Failed to init Y-Logo LED driver. Skipping ...
[  735.029048] legion_laptop: Init IO-Port LED driver
[  735.029065] legion_laptop: Error WMI call for reading brightness: expected a value between 1 and 2, but got 3
[  735.029065] legion_laptop: Error reading brightness for light: 5
[  735.029066] legion PNP0C09:00: Failed to init IO-Port LED driver. Skipping ...
[  735.029067] legion PNP0C09:00: legion_laptop loaded for this device

Result

• ✅ Module loaded successfully
• ✅ EC chip 0x5508 recognized
• ✅ Platform profile support created
• ℹ️ Y-Logo and IO-Port LED errors are normal (LOQ doesn't have these LEDs)

---

Attempt 2: Test Platform Profile Functionality

Commands Executed

ls /sys/firmware/acpi/platform_profile
cat /sys/firmware/acpi/platform_profile_choices
cat /sys/firmware/acpi/platform_profile
nvidia-smi

Output

/sys/firmware/acpi/platform_profile
quiet balanced balanced-performance performance
quiet

nvidia-smi Output

+-----------------------------------------------------------------------------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 5050 ...    Off |   00000000:01:00.0  On |                  N/A |
| N/A   45C    P8              5W /   35W |     235MiB /   8151MiB |      0%      Default |
+-----------------------------------------+------------------------+----------------------+

Result

• ✅ Platform profiles accessible
• ❌ GPU still at 35W cap

---

Attempt 3: Switch to Performance Mode

Commands Executed

echo performance | sudo tee /sys/firmware/acpi/platform_profile
nvidia-smi

Output

performance

nvidia-smi Output

| N/A   45C    P8              5W /   35W |     244MiB /   8151MiB |      0%      Default |

Result

• ❌ GPU power cap remained at 35W despite mode change to performance

---

Attempt 4: Test Fn+Q Hardware Key

Commands Executed

# Pressed Fn+Q multiple times while monitoring
cat /sys/firmware/acpi/platform_profile

Output (cycling through modes)

balanced-performance
performance
balanced-performance
quiet
balanced
balanced-performance

nvidia-smi Output

| N/A   38C    P8              4W /   35W |     230MiB /   8151MiB |      4%      Default |

Result

• ✅ Fn+Q correctly cycles modes
• ❌ GPU power cap unchanged

---

Attempt 5: Check dmesg for Power Mode Commands

Commands Executed

sudo dmesg | grep -i "legion\|ec\|power" | tail -50

Key Output Lines

[  735.028198] legion_laptop: Creating RAM access to embedded controller
[  735.028212] legion_laptop: Successfully mapped embedded controller: 0xfe0b0400 (in RAM)/0x0 (in EC) to virtual 0x00000000d155a950
[  735.028509] legion PNP0C09:00: Read embedded controller ID 0x5508
[  735.028586] legion_laptop: Creating platform profile support
[  865.551182] legion_laptop: Fan event: legion type: 2;  acpi type: 1 (1=integer)
[  876.914292] legion_laptop: Set powermode
[  914.927761] legion_laptop: Set powermode

Result

• ✅ Module sending "Set powermode" commands
• ✅ EC responding with fan events
• ❌ GPU power not changing

---

Attempt 6: Test Under GPU Load

Commands Executed

echo performance | sudo tee /sys/firmware/acpi/platform_profile

python3 << 'EOF'
import torch
import time
if torch.cuda.is_available():
    device = torch.device('cuda')
    x = torch.randn(5000, 5000, device=device)
    for i in range(50):
        y = torch.matmul(x, x)
        time.sleep(0.1)
EOF

nvidia-smi Output During Load

| N/A   43C    P8              10W /  35W |     239MiB /   8151MiB |      4%      Default |

Result

• ❌ GPU used ~10W but cap remained at 35W (should jump to 100W in performance mode)

---

Attempt 7: Analysis of EC Register Configuration

Commands Executed

grep -n "ec_register_offsets_loq" legion-laptop.c | head -20
sed -n '367,400p' legion-laptop.c

Current Configuration (ec_register_offsets_loq_v0)

static const struct ec_register_offsets ec_register_offsets_loq_v0 = {
    .ECHIPID1 = 0x2000,
    .ECHIPID2 = 0x2001,
    .ECHIPVER = 0x2002,
    .ECDEBUG = 0x2003,
    .EXT_FAN_CUR_POINT = 0xC5a0,
    .EXT_FAN_POINTS_SIZE = 0xC5a0, // constant 0
    .EXT_FAN1_BASE = 0xC530,
    .EXT_FAN2_BASE = 0xC530, // same rpm as cpu
    .EXT_FAN_ACC_BASE = 0xC5a0, // not found yet
    .EXT_FAN_DEC_BASE = 0xC5a0, // not found yet
    .EXT_CPU_TEMP = 0xC52F,
    .EXT_CPU_TEMP_HYST = 0xC5a0, // not found yet
    .EXT_GPU_TEMP = 0xC531,
    .EXT_GPU_TEMP_HYST = 0xC5a0, // not found yet
    // ... many more registers marked as 0xC5a0 (not found yet)
    .EXT_POWERMODE = 0xc41D,  // <-- Key register for power control
    // ...
};

Comparison with Legion Configurations

sed -n '235,267p' legion-laptop.c  # Legion v0
sed -n '268,300p' legion-laptop.c  # Legion v1

Power mode register addresses:

• Legion v0: EXT_POWERMODE = 0xc420
• Legion v1: EXT_POWERMODE = 0xc41D (same as LOQ v0)
• LOQ v0: EXT_POWERMODE = 0xc41D

Result

• Many LOQ v0 registers unmapped (0xC5a0), suggesting incomplete reverse engineering
• LOQ v0 shares powermode address with Legion v1 but not Legion v0

---

Attempt 8: Switch to Legion v1 Register Configuration

Hypothesis

LOQ v0 has many unmapped registers. Legion v1 uses same powermode address (0xc41D) but has complete register mappings. Perhaps complete mapping will work better.

Changes Made

sed -i '1009s/ec_register_offsets_loq_v0/ec_register_offsets_v1/' legion-laptop.c

Changed line ~1009 from:

.registers = &ec_register_offsets_loq_v0,

To:

.registers = &ec_register_offsets_v1,

Commands Executed

make clean && make
sudo rmmod legion-laptop
sudo insmod legion-laptop.ko
echo performance | sudo tee /sys/firmware/acpi/platform_profile
nvidia-smi

dmesg Output

[ 1692.436830] legion PNP0C09:00: Read embedded controller ID 0x5508
[ 1692.437024] legion_laptop: Init WMI driver support
[ 1692.437208] legion_laptop: Init keyboard backlight LED driver
[ 1692.437330] legion PNP0C09:00: Led platform::kbd_backlight renamed to platform::kbd_backlight_1 due to name collision
[ 1692.437458] legion_laptop: Init Y-Logo LED driver
[ 1692.437549] legion_laptop: Init IO-Port LED driver
[ 1692.437635] legion PNP0C09:00: legion_laptop loaded for this device

nvidia-smi Output

| N/A   47C    P5              9W /   35W |     288MiB /   8151MiB |      5%      Default |

Result

• ❌ GPU power cap still 35W
• ⚠️ Keyboard backlight stopped working (incorrect register addresses for EC 0x5508)
• ℹ️ No Y-Logo/IO-Port LED errors this time (Legion v1 expects them, doesn't complain when missing)

---

Summary of Findings

What Works ✅

1. ✅ Module loads successfully with EC chip 0x5508
2. ✅ Platform profile interface created (/sys/firmware/acpi/platform_profile)
3. ✅ Fn+Q correctly cycles through power modes (quiet/balanced/balanced-performance/performance)
4. ✅ "Set powermode" commands being sent by module (visible in dmesg)
5. ✅ EC responding to module (fan events detected)
6. ✅ No system crashes or instability during testing

What Doesn't Work ❌

1. ❌ GPU power cap remains at 35W in all power modes
2. ❌ Power cap doesn't change under GPU load
3. ❌ Legion v1 register configuration causes keyboard backlight to fail (wrong memory addresses)

---

Technical Analysis

Root Cause

EC chip 0x5508 (Core Ultra 5 125H variant) uses different EC register memory addresses than:

• EC chip 0x8227 (alfrix's Core i5-13450HX variant)
• Legion laptop EC chips (v0 and v1)

Evidence

1. Module successfully sends power mode commands (dmesg shows "Set powermode")
2. EC responds to other commands (fan events working)
3. Power mode interface works (Fn+Q cycles correctly)
4. But: GPU power limit doesn't change → powermode commands reaching wrong memory address

Current Memory Configuration

// From model_r3cn configuration
.memoryio_physical_ec_start = 0xC400,
.memoryio_size = 0x300,
.ramio_physical_start = 0xFE0B0400,
.ramio_size = 0x600

// Power mode register (from ec_register_offsets_loq_v0)
.EXT_POWERMODE = 0xc41D,

These addresses work for EC 0x8227 but not EC 0x5508.

Why Different EC Chips?

User	CPU	GPU	EC Chip
alfrix	Core i5-13450HX	RTX 5060	0x8227
This report	Core Ultra 5 125H	RTX 5050	0x5508

Hypothesis: Different CPU generations (13th gen vs Core Ultra) use different EC firmware with different register layouts.

---

Questions for Maintainers

1. Is EC chip 0x5508 a known variant? Different from 0x8227 (alfrix's chip) on same laptop model (83JE)

2. How to determine correct EXT_POWERMODE register address for EC 0x5508? Current 0xc41D doesn't control GPU power

3. Should we dump ACPI/EC tables? What's the safest method to find correct register addresses?

4. Are memory base addresses correct? Using 0xC400 and 0xFE0B0400 from alfrix's configuration

5. Does Core Ultra 5 125H use different EC firmware than Core i5-13450HX? Both are LOQ 15IRX10 (83JE) but different CPU generations

---

Environment Details

DMI Information

DMI_SYS_VENDOR: LENOVO
DMI_PRODUCT_NAME: 83JE
DMI_BIOS_VERSION: R3CN42WW

EC Details

• EC Chip ID: 0x5508 (read from address 0x2000-0x2001)
• EC Chip Version: (read from address 0x2002)

Memory Mapping (current)

• EC RAM: 0xFE0B0400 → Virtual 0x00000000d155a950
• EC Memory IO: 0xC400
• Memory IO Size: 0x300
• RAM IO Size: 0x600

System Software

• Kernel: 6.14.0-37-generic
• Ubuntu: 24.04
• NVIDIA Driver: 580.95.05
• CUDA Version: 13.0

Dependencies Installed

make gcc linux-headers-6.14.0-37-generic build-essential git lm-sensors wget
python3-pyqt6 python3-yaml python3-pip python3-argcomplete dkms openssl mokutil

---

Tested Configurations Summary

Test	Register Set	EXT_POWERMODE	GPU Power Cap	Keyboard Backlight
1	ec_register_offsets_loq_v0	0xc41D	❌ 35W	✅ Works
2	ec_register_offsets_v1	0xc41D	❌ 35W	❌ Broken

Not tested yet:

• ec_register_offsets_v0 (EXT_POWERMODE = 0xc420)
• Custom register configuration for EC 0x5508

---

Request for Next Steps

What diagnostic information would help identify correct register addresses for EC 0x5508? Happy to:

• Dump ACPI tables
• Test different register address configurations
• Monitor EC memory during Windows operation (if Windows-side tools exist)
• Test experimental configurations
• Provide any additional hardware/software information needed

The module framework works perfectly; we just need the correct memory map for this EC variant.

[qquique]

what is the output of : systemctl status nvidia-powerd.service and nvidia-smi -q -d POWER ?