Re: [PATCH] HID: lenovo: suppress spurious F23 key events on ThinkPad T14s gen6/7 snapdragon

["Mark Pearson" <mpearson-lenovo@squebb.ca>]

On Tue, Nov 10, 2026, at 4:36 AM, Hans de Goede wrote:
> Hi,
>
> On 10-Nov-26 01:49, Mark Pearson wrote:
>> Hi Hans
>> 
>> On Mon, Nov 9, 2026, at 4:43 PM, Hans de Goede wrote:
>>> The Lenovo ThinkPad T14s gen6/7 snapdragon version I2C-HID keyboard sends
>>> spurirous F23 key-events when the Fn key is pressed/released. Leading to
>> 
>> Minor - typo (spurirous)
>
> Ack.
>
>>> e.g. garbage ESC-sequences in the terminal when doing FN key combos.
>>>
>>> A hwdb entry was added to map these to KEY_UNKNOWN [1], but this also maps
>>> the F23 part of the left-meta + left-shift + F23 combo send by the co-pilot
>>> key to KEY_UNKNOWN making that key unusable.
>>>
>>> Add a raw_event filter which removes any 0x72 codes from the 6 scancodes
>>> reported in the standard HID keyboard boot protocol report, but only when
>>> the report's modifier byte is not 0x0a to preserve the F23 press event
>>> when the co-pilot key is pressed.
>>>
>> Interesting - I'd missed that.
>> 
>> Should I be flagging this to the FW team? I think I'm unlikely to get anything fixed at this point realistically - so don't want to block the patch...but if there's something we can learn from this for maybe the next generation?
>
> For future hw it would be nice if just pressing Fn by itself would not send
> any spurious keypresses. Or at least not on the I2C-HID side. If this is
> an EC event instead we can just filter it in the EC driver.
>

Ack. Thanks - I'll let them know.

>
>>> Link: https://github.com/systemd/systemd/pull/44733 [1]
>>> Signed-off-by: Hans de Goede <johannes.goede@oss.qualcomm.com>
>>> ---
>>>  drivers/hid/hid-ids.h    |  1 +
>>>  drivers/hid/hid-lenovo.c | 21 +++++++++++++++++++++
>>>  2 files changed, 22 insertions(+)
>>>
>>> diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h
>>> index 994e8bc69924..70d18d32fe3c 100644
>>> --- a/drivers/hid/hid-ids.h
>>> +++ b/drivers/hid/hid-ids.h
>>> @@ -473,6 +473,7 @@
>>>  #define USB_DEVICE_ID_EDIFIER_QR30	0xa101	/* EDIFIER Hal0 2.0 SE */
>>>
>>>  #define USB_VENDOR_ID_ELAN		0x04f3
>>> +#define I2C_DEVICE_ID_ELAN_LENOVO_T14S_KEYBOARD	0x000d
>>>  #define USB_DEVICE_ID_SB974D            0x0400
>>>  #define USB_DEVICE_ID_TOSHIBA_CLICK_L9W	0x0401
>>>  #define USB_DEVICE_ID_HP_X2		0x074d
>>> diff --git a/drivers/hid/hid-lenovo.c b/drivers/hid/hid-lenovo.c
>>> index 30bd3f794243..e997d0cf2672 100644
>>> --- a/drivers/hid/hid-lenovo.c
>>> +++ b/drivers/hid/hid-lenovo.c
>>> @@ -812,6 +812,22 @@ static int lenovo_raw_event(struct hid_device *hdev,
>>>  		      size >= 4 && report->id == 0x03))
>>>  		return lenovo_raw_event_TP_X12_tab(hdev, get_unaligned_le32(data));
>>>
>>> +	/*
>>> +	 * Lenovo ThinkPad T14s gen6/7 snapdragon version filter out spurious
>>> +	 * F23 events when Fn key is pressed. F23 cannot be unmapped because of
>>> +	 * the copilot key sending left-meta + left-shift + F23. Filter report 8
>>> +	 * (keyboard) scancode 0x72 (F23) when the modifier byte is not 0x0a.
>>> +	 */
>>> +	if (unlikely(hdev->product == I2C_DEVICE_ID_ELAN_LENOVO_T14S_KEYBOARD &&
>>> +		     size == 9 && data[0] == 8 && data[1] != 0x0a)) {
>>> +		for (int i = 3; i < 9; i++)
>>> +			if (data[i] == 0x72) {
>>> +				memmove(&data[i], &data[i + 1], 8 - i);
>>> +				data[8] = 0;
>>> +				break;
>>> +			}
>>> +	}
>>> +
>>>  	return 0;
>>>  }
>>>
>>> @@ -1393,6 +1409,10 @@ static int lenovo_probe(struct hid_device *hdev,
>>>  {
>>>  	int ret;
>>>
>>> +	/* Make the T14s  keyboard trackpoint a separate input-device */
>>> +	if (hdev->product == I2C_DEVICE_ID_ELAN_LENOVO_T14S_KEYBOARD)
>>> +		hdev->quirks |= HID_QUIRK_INPUT_PER_APP;
>>> +
>> 
>> Shouldn't this be in a separate commit?
>
> No this "quirk" is the default behavior of the hid-generic driver. This driver
> now will take over control from the hid-generic driver because of the device-id
> addition below.
>
> So setting this quirk preserves the existing default behavior.
>
Ah - my bad. Thanks for the explain.

>> 
>>>  	ret = hid_parse(hdev);
>>>  	if (ret) {
>>>  		hid_err(hdev, "hid_parse failed\n");
>>> @@ -1558,6 +1578,7 @@ static const struct hid_device_id 
>>> lenovo_devices[] = {
>>>  	{ HID_USB_DEVICE(USB_VENDOR_ID_IBM, 
>>> USB_DEVICE_ID_IBM_SCROLLPOINT_800DPI_OPTICAL_PRO) },
>>>  	{ HID_USB_DEVICE(USB_VENDOR_ID_LENOVO, 
>>> USB_DEVICE_ID_LENOVO_SCROLLPOINT_OPTICAL) },
>>>  	{ HID_USB_DEVICE(USB_VENDOR_ID_LENOVO, 
>>> USB_DEVICE_ID_LENOVO_TP10UBKBD) },
>>> +	{ HID_I2C_DEVICE(USB_VENDOR_ID_ELAN, 
>>> I2C_DEVICE_ID_ELAN_LENOVO_T14S_KEYBOARD) },
>>>  	/*
>>>  	 * Note bind to the HID_GROUP_GENERIC group, so that we only bind to 
>>> the keyboard
>>>  	 * part, while letting hid-multitouch.c handle the touchpad and 
>>> trackpoint.
>>> -- 
>>> 2.56.0
>> 

Reviewed-by: Mark Pearson <mpearson-lenovo@squebb.ca>
Mark