[PATCH] HID: multitouch: add a quirk to skip set inputmode for 2 new laptops

we have 2 latest Thinkpad laptops, the synaptics trackpoint module is
connected to i8042 bus and the synaptics touchpad module is on the i2c
bus. The trackpoint is driven by input/mouse/trackpoint.c and the
touchpad is driven by hid-multitouch.c.

They all work well independently, but if users press any buttons of
trackpoint and meanwhile move finger on the touchpad, the touchpad
can't work, the i2c bus can't generate interrupts even. That is to say
the touchpad can't work with trackpoint together, once trackpoint
works, the touchpad stops working.

The current hid driver parses the device descriptor and selects the
hid-multitouch.c and applies the MT_CLS_WIN_8 to the touchpad, I found
this issue begins to happen after the driver sets the
MT_INPUTMODE_TOUCHPAD to the device, If skipping to set it, the
touchpad work well and doesn't have that issue, even after suspend and
resume, the touchpad still work well.

This touchpad module doesn't support multi inputmodes, and its init
mode is set by BIOS already, it is safe to skip to set it again in
the kernel driver.

Signed-off-by: Hui Wang <redacted>
---
 drivers/hid/hid-multitouch.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/drivers/hid/hid-multitouch.c b/drivers/hid/hid-multitouch.c
index 0743ef51d3b2..3a658c4b4b05 100644
--- a/drivers/hid/hid-multitouch.c
+++ b/drivers/hid/hid-multitouch.c

@@ -70,6 +70,7 @@ MODULE_LICENSE("GPL");
 #define MT_QUIRK_WIN8_PTP_BUTTONS	BIT(18)
 #define MT_QUIRK_SEPARATE_APP_REPORT	BIT(19)
 #define MT_QUIRK_FORCE_MULTI_INPUT	BIT(20)
+#define MT_QUIRK_SKIP_SET_INPUTMODE	BIT(21)
 
 #define MT_INPUTMODE_TOUCHSCREEN	0x02
 #define MT_INPUTMODE_TOUCHPAD		0x03

@@ -191,6 +192,7 @@ static void mt_post_parse(struct mt_device *td, struct mt_application *app);
 #define MT_CLS_EXPORT_ALL_INPUTS		0x0013
 /* reserved					0x0014 */
 #define MT_CLS_WIN_8_FORCE_MULTI_INPUT		0x0015
+#define MT_CLS_WIN_8_SKIP_SET_INPUTMODE		0x0016
 
 /* vendor specific classes */
 #define MT_CLS_3M				0x0101

@@ -283,6 +285,15 @@ static const struct mt_class mt_classes[] = {
 			MT_QUIRK_WIN8_PTP_BUTTONS |
 			MT_QUIRK_FORCE_MULTI_INPUT,
 		.export_all_inputs = true },
+	{ .name = MT_CLS_WIN_8_SKIP_SET_INPUTMODE,
+		.quirks = MT_QUIRK_ALWAYS_VALID |
+			MT_QUIRK_IGNORE_DUPLICATES |
+			MT_QUIRK_HOVERING |
+			MT_QUIRK_CONTACT_CNT_ACCURATE |
+			MT_QUIRK_STICKY_FINGERS |
+			MT_QUIRK_WIN8_PTP_BUTTONS |
+			MT_QUIRK_SKIP_SET_INPUTMODE,
+		.export_all_inputs = true },
 
 	/*
 	 * vendor specific classes

@@ -1419,6 +1430,8 @@ static bool mt_need_to_apply_feature(struct hid_device *hdev,
 
 	switch (usage->hid) {
 	case HID_DG_INPUTMODE:
+		if (cls->quirks & MT_QUIRK_SKIP_SET_INPUTMODE)
+			return false;
 		/*
 		 * Some elan panels wrongly declare 2 input mode features,
 		 * and silently ignore when we set the value in the second

@@ -2058,6 +2071,11 @@ static const struct hid_device_id mt_devices[] = {
 		HID_DEVICE(BUS_I2C, HID_GROUP_MULTITOUCH_WIN_8,
 			USB_VENDOR_ID_SYNAPTICS, 0xce09) },
 
+	/* Synaptics devices */
+	{ .driver_data = MT_CLS_WIN_8_SKIP_SET_INPUTMODE,
+		HID_DEVICE(BUS_I2C, HID_GROUP_MULTITOUCH_WIN_8,
+			USB_VENDOR_ID_SYNAPTICS, 0xce57) },
+
 	/* TopSeed panels */
 	{ .driver_data = MT_CLS_TOPSEED,
 		MT_USB_DEVICE(USB_VENDOR_ID_TOPSEED2,

-- 
2.25.1

Hi


2021. január 7., csütörtök 12:27 keltezéssel, Hui Wang írta:

we have 2 latest Thinkpad laptops, the synaptics trackpoint module is
connected to i8042 bus and the synaptics touchpad module is on the i2c
bus. The trackpoint is driven by input/mouse/trackpoint.c and the
touchpad is driven by hid-multitouch.c.

They all work well independently, but if users press any buttons of
trackpoint and meanwhile move finger on the touchpad, the touchpad
can't work, the i2c bus can't generate interrupts even. That is to say
the touchpad can't work with trackpoint together, once trackpoint
works, the touchpad stops working.

The current hid driver parses the device descriptor and selects the
hid-multitouch.c and applies the MT_CLS_WIN_8 to the touchpad, I found
this issue begins to happen after the driver sets the
MT_INPUTMODE_TOUCHPAD to the device, If skipping to set it, the
touchpad work well and doesn't have that issue, even after suspend and
resume, the touchpad still work well.

This touchpad module doesn't support multi inputmodes, and its init
mode is set by BIOS already, it is safe to skip to set it again in
the kernel driver.
[...]

Assuming I'm not missing anything obvious, a windows precision touchpad
should revert back to reporting events via its HID compliant mouse collection
when power cycled or reset by the host[1].

Doesn't that imply that the events will be reported via the mouse collection
in this case, and that the multi-touch capabilities of the device will not
be usable? Or am I missing something?


[1]: https://docs.microsoft.com/en-us/windows-hardware/design/component-guidelines/windows-precision-touchpad-required-hid-top-level-collections#input-mode-feature-report


Regards,
Barnabás Pőcze

On 1/7/21 8:42 PM, Barnabás Pőcze wrote:

Hi


2021. január 7., csütörtök 12:27 keltezéssel, Hui Wang írta:

quoted

we have 2 latest Thinkpad laptops, the synaptics trackpoint module is
connected to i8042 bus and the synaptics touchpad module is on the i2c
bus. The trackpoint is driven by input/mouse/trackpoint.c and the
touchpad is driven by hid-multitouch.c.

They all work well independently, but if users press any buttons of
trackpoint and meanwhile move finger on the touchpad, the touchpad
can't work, the i2c bus can't generate interrupts even. That is to say
the touchpad can't work with trackpoint together, once trackpoint
works, the touchpad stops working.

The current hid driver parses the device descriptor and selects the
hid-multitouch.c and applies the MT_CLS_WIN_8 to the touchpad, I found
this issue begins to happen after the driver sets the
MT_INPUTMODE_TOUCHPAD to the device, If skipping to set it, the
touchpad work well and doesn't have that issue, even after suspend and
resume, the touchpad still work well.

This touchpad module doesn't support multi inputmodes, and its init
mode is set by BIOS already, it is safe to skip to set it again in
the kernel driver.
[...]

Assuming I'm not missing anything obvious, a windows precision touchpad
should revert back to reporting events via its HID compliant mouse collection
when power cycled or reset by the host[1].

Doesn't that imply that the events will be reported via the mouse collection
in this case, and that the multi-touch capabilities of the device will not
be usable? Or am I missing something?


[1]: https://docs.microsoft.com/en-us/windows-hardware/design/component-guidelines/windows-precision-touchpad-required-hid-top-level-collections#input-mode-feature-report

Yes, you are right, after applying this patch, the events will be 
reported through mouse collection and some features specific to touchpad 
collection like multi-touch are lost. So this patch is not a correct 
solution, please ignore this patch.

And I found If using the hid-rmi.c to drive this touchpad, it will work 
well, the touchpad and trackpoint could work well together and the 
touchpad has multi-touch capability. But because the report descriptor 
contains an Input item which has the usage HID_DG_CONTACTID, the 
hid->group is set to HID_GROUP_MULTITOUCH, the hid-multitouch.c is 
selected to drive this touchpad.

Is it Okay to do the change below and let hid-rmi.c drive this touchpad?

diff --git a/drivers/hid/hid-rmi.c b/drivers/hid/hid-rmi.c
index 311eee599ce9..58a7529499f5 100644
--- a/drivers/hid/hid-rmi.c
+++ b/drivers/hid/hid-rmi.c

@@ -759,6 +759,7 @@ static const struct hid_device_id rmi_id[] = {
         { HID_USB_DEVICE(USB_VENDOR_ID_PRIMAX, 

USB_DEVICE_ID_PRIMAX_REZEL) },
         { HID_USB_DEVICE(USB_VENDOR_ID_SYNAPTICS, 
USB_DEVICE_ID_SYNAPTICS_ACER_SWITCH5),
                 .driver_data = RMI_DEVICE_OUTPUT_SET_REPORT },
+       { HID_I2C_DEVICE(USB_VENDOR_ID_SYNAPTICS, 0xce57) },
         { HID_DEVICE(HID_BUS_ANY, HID_GROUP_RMI, HID_ANY_ID, HID_ANY_ID) },
         { }
  };

Thanks,

Hui.

Regards,
Barnabás Pőcze

This turns out to be a firmware issue, synaptics provided a new firmware 
for the touchpad, and we upgraded the firmware for the touchpad under 
windows (we only have windows firmware upgrade utility), now the 
touchpad could work with trackpoint together under Linux.

Thanks.

On 1/9/21 8:35 PM, Hui Wang wrote: