[Bug 221774] New: hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[Bug 221774] New: hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221774

            Bug ID: 221774
           Summary: hid-multitouch: report ID mismatch check breaks
                    touchpad palm detection on ASUS ROG Flow Z13 (2025)
                    GZ302EA (0b05:1a30)
           Product: Drivers
           Version: 2.5
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: USB
          Assignee: drivers_usb@kernel-bugs.kernel.org
          Reporter: mayhemandcoffee@gmail.com
        Regression: No

Created attachment 310482
  --> https://bugzilla.kernel.org/attachment.cgi?id=310482&action=edit
touch registers while typing

Component: Drivers / Input Devices (or USB, whichever's closest)

Kernel Version: 7.1.4-arch1-1 (regression); confirmed working on 6.18.39-1-lts

Description:

On the ASUS ROG Flow Z13 (2025), model GZ302EA, the detachable keyboard cover's
touchpad (USB 0b05:1a30) loses its onboard palm-detection capability on kernels
containing commit e716edafedad4952fe3a4a273d2e039a84e8681a ("HID: multitouch:
Check to ensure report responses match the request").

dmesg shows:

hid-multitouch 0003:0B05:1A30.0004: Returned feature report did not match the
request

Effect: the touchpad falls back to a legacy single-touch-style report — missing
ABS_MT_SLOT, ABS_MT_TOOL_TYPE, and multi-finger BTN_TOOL_* capabilities —
instead of the full Precision Touchpad protocol. This means palm rejection
stops working entirely in userspace (libinput has no tool-type/slot data to
reject a palm contact with).

Working baseline: on linux-lts 6.18.39 (predates the commit), evtest on the
touchpad node shows the full protocol, including ABS_MT_TOOL_TYPE correctly
reporting value 2 (MT_TOOL_PALM) when a palm contact is detected, and 0
(MT_TOOL_FINGER) for normal touches.

Suspected root cause: in mt_feature_mapping(), the vendor-specific usage
0xff0000c5 triggers a mt_get_feature() call described in the driver source as
"Retrieve the Win8 blob once to enable some devices." This device's firmware
appears to respond to that specific GET_FEATURE request with a mismatched
report ID, causing the new check to goto free and abort before the device gets
switched into its full-capability reporting mode.

Steps to reproduce:

Boot a kernel containing commit e716edafedad4952fe3a4a273d2e039a84e8681a
sudo evtest on the touchpad's /dev/input/eventN node
Observe missing ABS_MT_SLOT / ABS_MT_TOOL_TYPE in the capabilities list, and
dmesg | grep hid-multitouch showing the mismatch error

Requested fix: a vendor/product-ID based quirk in hid-multitouch.c (similar to
the existing MT_CLS_ASUS handling) to bypass or tolerate the report-ID mismatch
for 0b05:1a30, restoring the device to full-capability mode.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221774] hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221774

The Linux kernel's regression tracker (Thorsten Leemhuis) (regressions@leemhuis.info) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |regressions@leemhuis.info

--- Comment #1 from The Linux kernel's regression tracker (Thorsten Leemhuis) (regressions@leemhuis.info) ---
* Would be good if you could check if latest mainlin is still affected
* Have you tried if reverting the change helps?
* Can you please attach a dmesg output from an affected kernel.
* This likely needs to be forwarded by mail. Can I CC you on that? This would
expose your email address to the world.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221774] hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221774

Lovekesh Solanki (lovekeshsolanki00@gmail.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lovekeshsolanki00@gmail.com

--- Comment #2 from Lovekesh Solanki (lovekeshsolanki00@gmail.com) ---
It is possible the commit e716edafedad causes mt_get_feature() to reject the
Win8 response due to firmware returning mismatched report Id. So
hid_report_raw_event() is never reached and device doesn't switch into
Precision mode.
Rather than reverting the commit, which would drop the ID check for every
device, it may be possible to add a quirk for 0b05:1a30 so only this model
tolerates the mismatch.

I have prepared a patch which does that, but I don't have the hardware so I
cannot confirm.
Would it be possible for someone to test it? I'm not sure how to send it here
I'm not familiar with the process.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221774] hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221774

--- Comment #3 from mayhemandcoffee@gmail.com ---
(In reply to Lovekesh Solanki from comment #2)
> It is possible the commit e716edafedad causes mt_get_feature() to reject the
> Win8 response due to firmware returning mismatched report Id. So
> hid_report_raw_event() is never reached and device doesn't switch into
> Precision mode.
> Rather than reverting the commit, which would drop the ID check for every
> device, it may be possible to add a quirk for 0b05:1a30 so only this model
> tolerates the mismatch.
> 
> I have prepared a patch which does that, but I don't have the hardware so I
> cannot confirm.
> Would it be possible for someone to test it? I'm not sure how to send it
> here I'm not familiar with the process.

I have the affected hardware (ASUS ROG Flow Z13 2025, keyboard touchpad VID:PID
0b05:1a30) and I'm happy to test your patch. Please attach it or send it in
whatever format is easiest.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221774] hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221774

--- Comment #4 from Lovekesh Solanki (lovekeshsolanki00@gmail.com) ---
Created attachment 310490
  --> https://bugzilla.kernel.org/attachment.cgi?id=310490&action=edit
HID: multitouch: Add report ID mismatch quirk for ASUS ROG Z13 Folio

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221774] hid-multitouch: report ID mismatch check breaks touchpad palm detection on ASUS ROG Flow Z13 (2025) GZ302EA (0b05:1a30)

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221774

--- Comment #5 from mayhemandcoffee@gmail.com ---
(In reply to Lovekesh Solanki from comment #4)
> Created attachment 310490 [details]
> HID: multitouch: Add report ID mismatch quirk for ASUS ROG Z13 Folio

Hi,

I tested the patch on my ASUS ROG Flow Z13 (2025) GZ302EA.

The patch applied cleanly and I rebuilt the Arch Linux 7.1.4 kernel with it. I
verified that the changes are present in drivers/hid/hid-multitouch.c,
including the new MT_QUIRK_IGNORE_FEATURE_ID_MISMATCH, the
MT_CLS_ASUS_ROG_Z13_FOLIO class, and the new USB device entry.

After booting the patched kernel:

hid-multitouch binds to the touchpad (0003:0B05:1A30.0004)
libinput list-devices is unchanged

The touchpad still reports:

Capabilities: pointer gesture
Disable-w-typing: n/a
I don't observe any behavioral changes compared to the unpatched kernel.

So while the patch applies successfully and the kernel boots normally, it does
not appear to restore the expected Precision Touchpad behavior on my device.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.