i2c: acpi: Dell XPS 17 9720 ELAN 04F3:311C touchpad (VEN_04F3) needs 100 kHz like DLL0945
Subsystems (4)
A message that is not in the archive (sent off-list or not yet received): CYQ1dYF3ZUj0L_VRzN490SfuHRaulkT_y5NWcIEIl8OtpIgWgXyOlwEyUqZD_uz9bmVBXM2uLOo8bkyzcR-qCxNObstCqFUlK6lUYXK5euE=@proton.me
Hi,
Quoted 24 lines · On Fri, Nov 06, 2026 at 05:48:02PM +0000, [email protected] wrote:
Hi Mika, Andi,
The ELAN 04F3:311C touchpad in the Dell XPS 17 9720 intermittently becomes sluggish and sticky for several seconds to minutes, then recovers. I dual boot with Windows and I do not experience this issue on Windows, only on Linux and across multiple distros (Linux mint, Ubuntu and
Pop_os). I'm not the only one who experiences this issue; it seems to be widespread among Dell XPS users. This looks like the excessive-smoothing issue fixed for the Precision 5560 by commit0b7c9528facd(“i2c: Force DLL0945 touchpad i2c freq to 100khz”). The Precision 5560 uses the same 04F3:311C touchpad. On the XPS 17 9720, the firmware exposes it with_HID“VEN_04F3” instead of “DLL0945”, so the existing quirk doesn’t apply.Forcing the bus to 100 kHz fixes it here. The DSDT already supports 100 kHz:
_SB.PC00.I2C1.TPD0._INI sets SPED from the TPDS NVS variable (0 = 100 kHz, 1 = 400 kHz, 2 = 1 MHz), and this firmware uses TPDS = 1. To test it, I loaded an SSDT through configfs that sets _SB.PC00.I2C1.TPD0.SPED =0x186A0and rebound i2c_designware.1. After 5 days of use, including on AC power, the lag has not come back. Without the override, it returns as randomly as before.System:
• Dell XPS 17 9720, BIOS 1.41.0
• Touchpad: VEN_04F3:00 04F3:311C on i2c_designware.1 (_SB.PC00.I2C1.TPD0),
Quoted 7 lines
touchpad firmware 0x000b• Kernel: 7.2.15-76070215-generic
Adding “VEN_04F3” to
i2c_acpi_force_100khz_device_idswould fix this machine. However,VEN_04F3looks like a generic vendor ID rather than a model-specific one like DLL0945, so it may need to be limited to this model with a DMI match. I’m happy to test a patch.
That VEN_04F3 is not even a valid ACPI/PNP ID. Have you checked if there is a BIOS upgrade for your system?
It's odd that Dell would use identifier like this as they typically follow these pretty well. Anyways if there is no BIOS upgrade (or if it does not change anything) then I don't see much other option than adding a quirk. I do see there is one quirk added already for GPIO side on Dell system with VEN_0488 entry (drivers/gpio/gpiolib-acpi-quirks.c).
In that case we would need output of 'sudo dmidecode' to be able to add the DMI strings.
Hi Mika,
Thank you for the response (I honestly wasn't expecting anyone to respond for this niche issue---so a welcome surprise).
Yes my BIOS version (v1.41.0) is up to date. Currently, there are no updates available through Windows or from fwupd. I have received BIOS updates many times via Windows since purchasing this computer in 2022 and have had 0 change to this problem.
For what it’s worth, the DSDT has a valid default ID. TPD0 starts with Name (ITPN, “DELL0AFF”), and _INI then overwrites it from the BIOS-provided TPHD variable, which yields “VEN_04F3”.
Please find attached the output of 'sudo dmidecode'.
Thank you for helping me fix this.
Quoted 47 lines · Mika Westerberg
Hi,
Quoted 24 lines · On Fri, Nov 06, 2026 at 05:48:02PM +0000, [email protected] wrote:
Hi Mika, Andi,
The ELAN 04F3:311C touchpad in the Dell XPS 17 9720 intermittently becomes sluggish and sticky for several seconds to minutes, then recovers. I dual boot with Windows and I do not experience this issue on Windows, only on Linux and across multiple distros (Linux mint, Ubuntu and
Pop_os). I'm not the only one who experiences this issue; it seems to be widespread among Dell XPS users. This looks like the excessive-smoothing issue fixed for the Precision 5560 by commit0b7c9528facd(“i2c: Force DLL0945 touchpad i2c freq to 100khz”). The Precision 5560 uses the same 04F3:311C touchpad. On the XPS 17 9720, the firmware exposes it with_HID“VEN_04F3” instead of “DLL0945”, so the existing quirk doesn’t apply.Forcing the bus to 100 kHz fixes it here. The DSDT already supports 100 kHz:
_SB.PC00.I2C1.TPD0._INI sets SPED from the TPDS NVS variable (0 = 100 kHz, 1 = 400 kHz, 2 = 1 MHz), and this firmware uses TPDS = 1. To test it, I loaded an SSDT through configfs that sets _SB.PC00.I2C1.TPD0.SPED =0x186A0and rebound i2c_designware.1. After 5 days of use, including on AC power, the lag has not come back. Without the override, it returns as randomly as before.System:
• Dell XPS 17 9720, BIOS 1.41.0
• Touchpad: VEN_04F3:00 04F3:311C on i2c_designware.1 (_SB.PC00.I2C1.TPD0),Quoted 7 lines
touchpad firmware 0x000b• Kernel: 7.2.15-76070215-generic
Adding “VEN_04F3” to
i2c_acpi_force_100khz_device_idswould fix this machine. However,VEN_04F3looks like a generic vendor ID rather than a model-specific one like DLL0945, so it may need to be limited to this model with a DMI match. I’m happy to test a patch.That
VEN_04F3is not even a valid ACPI/PNP ID. Have you checked if there is a BIOS upgrade for your system?It's odd that Dell would use identifier like this as they typically follow these pretty well. Anyways if there is no BIOS upgrade (or if it does not change anything) then I don't see much other option than adding a quirk. I do see there is one quirk added already for GPIO side on Dell system with
VEN_0488entry (drivers/gpio/gpiolib-acpi-quirks.c).In that case we would need output of 'sudo dmidecode' to be able to add the DMI strings.
Hi,
Quoted 11 lines · On Mon, Nov 09, 2026 at 08:57:13PM +0000, [email protected] wrote:
Hi Mika,
Thank you for the response (I honestly wasn't expecting anyone to respond for this niche issue---so a welcome surprise).
Yes my BIOS version (v1.41.0) is up to date. Currently, there are no updates available through Windows or from fwupd. I have received BIOS updates many times via Windows since purchasing this computer in 2022 and have had 0 change to this problem.
For what it’s worth, the DSDT has a valid default ID. TPD0 starts with Name (ITPN, “DELL0AFF”), and
_INIthen overwrites it from the BIOS-provided TPHD variable, which yields “VEN_04F3”.Please find attached the output of 'sudo dmidecode'.
Thank you for helping me fix this.
Can you try the below patch?
drivers/i2c/i2c-core-acpi.c +24 -1
@@ -7,6 +7,7 @@#include <linux/acpi.h> #include <linux/device.h> +#include <linux/dmi.h> #include <linux/err.h> #include <linux/i2c.h> #include <linux/list.h>
@@ -356,6 +357,17 @@ static const struct acpi_device_id i2c_acpi_force_400khz_device_ids[] = { { } }; +static const struct dmi_system_id i2c_acpi_dell_xps_17_9720_dmi_ids[] = { + { + .matches = { +DMI_MATCH (DMI_SYS_VENDOR ,"Dell Inc." ), +DMI_MATCH (DMI_PRODUCT_NAME ,"XPS 17 9720" ), +DMI_MATCH (DMI_BIOS_VERSION ,"1.41.0" ), + }, + }, + { } +}; +static const struct acpi_device_id i2c_acpi_force_100khz_device_ids[] = {/* * When a 400KHz freq is used on this model of ELAN touchpad in Linux,
@@ -377,6 +389,14 @@ static const struct acpi_device_id i2c_acpi_force_100khz_device_ids[] = { { .id = "ELAN0678" }, { .id = "ELAN06FA" }, { .id = "ELAN1300" }, + { +/* +* Some Dell XPS 17 9720 is not using standard ACPI IDs +* so we need to match also via DMI strings. +*/ + .id = "VEN_04F3", + .driver_data = (kernel_ulong_t)i2c_acpi_dell_xps_17_9720_dmi_ids, + }, { } };
@@ -385,6 +405,7 @@ static acpi_status i2c_acpi_lookup_speed(acpi_handle handle, u32 level, {struct i2c_acpi_lookup *lookup = data;struct acpi_device *adev =acpi_fetch_acpi_dev (handle); +const struct acpi_device_id *id;if (!adev ||i2c_acpi_do_lookup (adev, lookup))return AE_OK ;
@@ -398,7 +419,9 @@ static acpi_status i2c_acpi_lookup_speed(acpi_handle handle, u32 level,if (acpi_match_device_ids (adev, i2c_acpi_force_400khz_device_ids) ==0 ) lookup->force_speed =I2C_MAX_FAST_MODE_FREQ ; -if (acpi_match_device_ids (adev, i2c_acpi_force_100khz_device_ids) ==0 ) + id =acpi_match_acpi_device (i2c_acpi_force_100khz_device_ids, adev); +if (id && (!id->driver_data || +dmi_check_system ((const struct dmi_system_id *)id->driver_data))) lookup->force_speed =I2C_MAX_STANDARD_MODE_FREQ ;return AE_OK ;
Hi Mika,
Tested on top of v7.4-rc2 (067c5c24cef2) on a Dell XPS 17 9720, BIOS 1.41.0, without the SSDT override.
The quirk is applied at boot.
I've been using my laptop all day and the touchpad lag has not come back so far including on AC power. On the unpatched kernel it returns as before.
Interestingly enough, I did use an off-brand AC charger for my laptop and the touchpad lag returned on the patched and unpatched Linux kernels AND Windows. When I used the Dell AC charger, the lag was gone on the patched kernel and Windows, but remained on the unpatched kernel. Perhaps the root cause of this problem, on Linux relative to Windows is that Windows aggressively filters noise without smoothing? Not sure, just a thought
One question about the DMI match: does it need DMI_BIOS_VERSION? The VEN_04F3 ID check already makes the quirk stop applying if a future BIOS switches to a proper ACPI ID. The 400 kHz problem also seems to be in the touchpad itself, since the same 04F3:311C part is affected on the Precision 5560. Matching only on sys_vendor and product_name would also cover users on older BIOS versions, and any newer BIOS that still reports VEN_04F3.
Dell also ships the same BIOS image for the Precision 5770, so it likely exposes the same ID, though I can't test that model.
Thanks for the quick patch!
Quoted 82 lines · Mika Westerberg
Hi,
Quoted 11 lines · On Mon, Nov 09, 2026 at 08:57:13PM +0000, [email protected] wrote:
Hi Mika,
Thank you for the response (I honestly wasn't expecting anyone to respond for this niche issue---so a welcome surprise).
Yes my BIOS version (v1.41.0) is up to date. Currently, there are no updates available through Windows or from fwupd. I have received BIOS updates many times via Windows since purchasing this computer in 2022 and have had 0 change to this problem.
For what it’s worth, the DSDT has a valid default ID. TPD0 starts with Name (ITPN, “DELL0AFF”), and
_INIthen overwrites it from the BIOS-provided TPHD variable, which yields “VEN_04F3”.Please find attached the output of 'sudo dmidecode'.
Thank you for helping me fix this.
Can you try the below patch?
drivers/i2c/i2c-core-acpi.c +24
-1@@ -7,6 +7,7 @@#include <linux/acpi.h> #include <linux/device.h> +#include <linux/dmi.h> #include <linux/err.h> #include <linux/i2c.h> #include <linux/list.h> @@ -356,6 +357,17 @@ static const struct acpi_device_id i2c_acpi_force_400khz_device_ids[] = { { } }; +static const struct dmi_system_id i2c_acpi_dell_xps_17_9720_dmi_ids[] = { + { + .matches = { +DMI_MATCH (DMI_SYS_VENDOR ,"Dell Inc." ), +DMI_MATCH (DMI_PRODUCT_NAME ,"XPS 17 9720" ), +DMI_MATCH (DMI_BIOS_VERSION ,"1.41.0" ), + }, + }, + { } +}; +static const struct acpi_device_id i2c_acpi_force_100khz_device_ids[] = {/* * When a 400KHz freq is used on this model of ELAN touchpad in Linux, @@ -377,6 +389,14 @@ static const struct acpi_device_id i2c_acpi_force_100khz_device_ids[] = { { .id = "ELAN0678" }, { .id = "ELAN06FA" }, { .id = "ELAN1300" }, + { +/* +* Some Dell XPS 17 9720 is not using standard ACPI IDs +* so we need to match also via DMI strings. +*/ + .id = "VEN_04F3", + .driver_data = (kernel_ulong_t)i2c_acpi_dell_xps_17_9720_dmi_ids, + }, { } };@@ -385,6 +405,7 @@ static acpi_status i2c_acpi_lookup_speed(acpi_handle handle, u32 level, {struct i2c_acpi_lookup *lookup = data;struct acpi_device *adev =acpi_fetch_acpi_dev (handle); +const struct acpi_device_id *id;if (!adev ||i2c_acpi_do_lookup (adev, lookup))return AE_OK ;@@ -398,7 +419,9 @@ static acpi_status i2c_acpi_lookup_speed(acpi_handle handle, u32 level,if (acpi_match_device_ids (adev, i2c_acpi_force_400khz_device_ids) ==0 ) lookup->force_speed =I2C_MAX_FAST_MODE_FREQ ; -if (acpi_match_device_ids (adev, i2c_acpi_force_100khz_device_ids) ==0 ) + id =acpi_match_acpi_device (i2c_acpi_force_100khz_device_ids, adev); +if (id && (!id->driver_data || +dmi_check_system ((const struct dmi_system_id *)id->driver_data))) lookup->force_speed =I2C_MAX_STANDARD_MODE_FREQ ;return AE_OK ;
Hi,
Quoted 8 lines · On Thu, Nov 12, 2026 at 02:18:35AM +0000, [email protected] wrote:
Hi Mika,
Tested on top of v7.4-rc2 (
067c5c24cef2) on a Dell XPS 17 9720, BIOS 1.41.0, without the SSDT override.The quirk is applied at boot.
I've been using my laptop all day and the touchpad lag has not come back so far including on AC power. On the unpatched kernel it returns as before.
Thanks for testing!
Quoted 1 line
Interestingly enough, I did use an off-brand AC charger for my laptop and the touchpad lag returned on the patched and unpatched Linux kernels AND Windows. When I used the Dell AC charger, the lag was gone on the patched kernel and Windows, but remained on the unpatched kernel. Perhaps the root cause of this problem, on Linux relative to Windows is that Windows aggressively filters noise without smoothing? Not sure, just a thought
Indeed it can very well be. Someone should really investigate this properly by hooking I2C analyzer to the bus to see what is going on there. However, that is not trivial to do with form factor devices.
Quoted 3 lines
One question about the DMI match: does it need
DMI_BIOS_VERSION? TheVEN_04F3ID check already makes the quirk stop applying if a future BIOS switches to a proper ACPI ID. The 400 kHz problem also seems to be in the touchpad itself, since the same 04F3:311C part is affected on the Precision 5560. Matching only onsys_vendorandproduct_namewould also cover users on older BIOS versions, and any newer BIOS that still reportsVEN_04F3.Dell also ships the same BIOS image for the Precision 5770, so it likely exposes the same ID, though I can't test that model.
I agree we can drop the BIOS version. I will send the updated patch soon.