On Tue, 29 Nov 2022, Bjorn Helgaas wrote:
> [+cc Takayuki, thanks for your report!]
>
> On Wed, Nov 04, 2020 at 11:02:23PM +0100, Lars Povlsen wrote:
>> The mcp2221 driver GPIO output handling has has several issues.
>>
>> * A wrong value is used for the GPIO direction.
>>
>> * Wrong offsets are calculated for some GPIO set value/set direction
>>   operations, when offset is larger than 0.
>>
>> This has been fixed by introducing proper manifest constants for the
>> direction encoding, and using 'offsetof' when calculating GPIO
>> register offsets.
>>
>> The updated driver has been tested with the Sparx5 pcb134/pcb135
>> board, which has the mcp2221 device with several (output) GPIO's.
>>
>> Signed-off-by: Lars Povlsen <lars.povlsen@microchip.com>
>> ---
>>  drivers/hid/hid-mcp2221.c | 48 +++++++++++++++++++++++++++++++--------
>>  1 file changed, 39 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/hid/hid-mcp2221.c b/drivers/hid/hid-mcp2221.c
>> index 0d27ccb55dd9..4211b9839209 100644
>> --- a/drivers/hid/hid-mcp2221.c
>> +++ b/drivers/hid/hid-mcp2221.c
>
>> +/* MCP GPIO get command layout */
>> +struct mcp_get_gpio {
>> +	u8 cmd;
>> +	u8 dummy;
>> +	struct {
>> +		u8 direction;
>> +		u8 value;
>> +	} gpio[MCP_NGPIO];
>> +} __packed;
>
> This bug report: https://bugzilla.kernel.org/show_bug.cgi?id=216736
> suggests that direction and value may be reversed here.
>
> Mentioning here in case nobody actively monitors the bugzilla.
>
> Bjorn
>

I can confirm that the order is reversed when compared to both the data sheet
and what I can observe in the hid reports.
The patch from the linked bugzilla bug fixes the issue for me.

> --- a/drivers/hid/hid-mcp2221.c
> +++ b/drivers/hid/hid-mcp2221.c
> @@ -74,8 +74,8 @@ struct mcp_get_gpio {
>  	u8 cmd;
>  	u8 dummy;
>  	struct {
> -		u8 direction;
>  		u8 value;
> +		u8 direction;
>  	} gpio[MCP_NGPIO];
>  } __packed;

Tested-by: Sven Zühlsdorf <sven.zuehlsdorf@vigem.de>