Hi,
On 1/11/23 17:38, Jiaxun Yang wrote:
>
>
> 在2022年11月9日十一月 下午8:59，Hans de Goede写道：
>> Hi All,
>>
>> I'm emailing you all because you have written patches or
>> reported bugs related to the ideapad-laptop touchpad
>> handling the past.
>
> Hi all,
>
> Sorry for chime in this old thread, but I'm able to get some input from a
> Lenovo engineer.
>
> Quoting him:
>
> "For newer Lenovo laptops we use TPRD/TPWR methods under touchpad's ACPI I2C HID
> device to sync touchpad state with EC. TPRD will return current touchpad state known
> by EC and OS can use TPWR to switch EC's touchpad state. This state will be used by EC
> to control LEDs and touchpad power saving signals."
>
> According to my understanding we only need replace read write to VPCCMD_R_TOUCHPAD
> with calling TPRD/TPWR methods to get all new ideapads work.
>
> As per my reverse engineering on ASLs this method actually reads a flag from EC's
> LPC memory space and it do work on some early ideapads (The earliest one I can
> trace is Ideapad 320-15ISK which is released on 2018).
>
> I'm going to try to implement it in kernel. Though I haven't decide which part of
> driver should handle this, as those methods are under ACPI I2C HID device perhaps
> we should put this function under i2c-hid-acpi driver. However as the method is very
> ideated specific, putting in ideapad-acpi seems more reasonable... Any thoughts?
This really should be handled in ideapad-acpi, but that means figuring out a way
to find the touchpad device in ideapad-apci. For starters you can search for
ACPI devs which match:
static const struct acpi_device_id i2c_hid_acpi_match[] = {
{"ACPI0C50", 0 },
{"PNP0C50", 0 },
{ },
};
And then check if they have the TPRD/TPWR methods. This does risk piking
another device then the actual touchpad though, most likely a touchscreen
(in case where the ACPI tables accidentally also have TPRD/TPWR methods
on the touchscreen).
Once you have find the right ACPI device, we can call the methods
(when found) from:
ideapad_sync_touchpad_state()
Note that only ever reads the touchpad state...
> He also told me how to get VPC2004 device's BIOS interface version, though the
> differences between versions remains unclear to me. I think we can expose it
> in dmesg and debugfs to help with future debugging and hopefully reduce the amount
> of DMI quirks.
Yes reading + logging (in dmesg) the BIOS interface version would be good.
Thank you for all your work on this!
Regards,
Hans
