Re: [PATCH] HID: i2c-hid-of: Fall back to ACPI _DSM when hid-descr-addr is missing

[Dmitry Torokhov]

On Fri, May 29, 2026 at 05:00:37PM +0200, Hans de Goede wrote:
> Hi,
> 
> On 29-May-26 14:16, 谢致邦 (XIE Zhibang) wrote:
> > Before commit b33752c30023 ("HID: i2c-hid: Reorganize so ACPI and OF are
> > separate modules"), the unified i2c-hid driver handled both PNP0C50 ACPI
> > devices and hid-over-i2c OF devices. After the split, devices with _HID
> > "PRP0001" and _DSD compatible "hid-over-i2c" are only probed by
> > i2c_hid_of, which requires "hid-descr-addr" in the _DSD. Some devices,
> > for example the Lenovo KaiTian N60d and Inspur CP300L3, provide the HID
> > descriptor address only through the _DSM method. Fall back to the _DSM
> > call when the property is absent.
> > 
> > Fixes: b33752c30023 ("HID: i2c-hid: Reorganize so ACPI and OF are separate modules")
> > Signed-off-by: 谢致邦 (XIE Zhibang) <Yeking@xxxxxxxxx>
> 
> Thank you for the new patch, this is an interesting approach and better
> then the modalias magic from the previous version.
> 
> Note I'm not the i2c-hid maintainer, with that said I think this should
> be acceptable. But currently it duplicates the _DSM handling code and
> that should be fixed.
> 
> I think this should be changed to a series of 3 patches:
> 
> 1. Move the i2c_hid_acpi_blacklist handling out of
>    i2c_hid_acpi_get_descriptor() into i2c_hid_acpi_probe()
>    to above the devm_kzalloc() call.
> 
> 2. Move i2c_hid_acpi_get_descriptor() to a generic
>    int i2c_hid_acpi_get_descriptor(struct device *dev)
>    helper in drivers/hid/i2c-hid/i2c-hid-core.c .
>    Wrapped in #ifdef CONFIG_ACPI and with a static inline
>    stub in drivers/hid/i2c-hid/i2c-hid.h when CONFIG_ACPI
>    is not set, e.g. in drivers/hid/i2c-hid/i2c-hid.h add:
> 
> #ifdef CONFIG_ACPI
> 	int i2c_hid_acpi_get_descriptor_address(struct device *dev);
> #else
> 	static inline int i2c_hid_acpi_get_descriptor_address(struct device *dev)
> 	{
> 		return -ENODEV;
> 	}
> #endif
> 
> 3. Modify i2c-hid-of.c to try i2c_hid_acpi_get_descriptor_address() as
> fallback for the missing "hid-descr-addr" property.  Please also add
> a comment in the code explaining that this fallback is about ACPI I2C-hid
> devices which use a "PRP0001" ACPI _HID with an "hid-over-i2c" compatible.

I think we should also stick a big fat warning if _DSM succeeds in that
code branch: hopefully manufacturers will notice and fix the firmware on
new devices.

Thanks.

-- 
Dmitry