2023-04-21 16:31:51

by Marius Hoch

[permalink] [raw]
Subject: [PATCH] ACPI: scan: Ignore MAGN0001

The LSM303D on the Lenovo Yoga Tablet 2 series is present
as both ACCL0001 and MAGN0001. As we can only ever register an
i2c client for one of them, ignore MAGN0001.

Currently this errors:
i2c i2c-4: Failed to register i2c client MAGN0001:00 at 0x1d (-16)

Tested on a Lenovo Yoga Tablet 2 1051-F.

Signed-off-by: Marius Hoch <[email protected]>
---
drivers/acpi/scan.c | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
index 0c6f06abe3f4..d21c411e9719 100644
--- a/drivers/acpi/scan.c
+++ b/drivers/acpi/scan.c
@@ -1730,6 +1730,12 @@ static bool acpi_device_enumeration_by_parent(struct acpi_device *device)
*/
{"BCM4752", },
{"LNV4752", },
+ /*
+ * The LSM303D on the Lenovo Yoga Tablet 2 series is present
+ * as both ACCL0001 and MAGN0001. As we can only ever register an
+ * i2c client for one of them, ignore MAGN0001.
+ */
+ {"MAGN0001", },
{}
};


base-commit: cb0856346a60fe3eb837ba5e73588a41f81ac05f
--
2.40.0


2023-04-21 18:10:51

by Hans de Goede

[permalink] [raw]
Subject: Re: [PATCH] ACPI: scan: Ignore MAGN0001

Hi Marius,

On 4/21/23 18:27, Marius Hoch wrote:
> The LSM303D on the Lenovo Yoga Tablet 2 series is present
> as both ACCL0001 and MAGN0001. As we can only ever register an
> i2c client for one of them, ignore MAGN0001.
>
> Currently this errors:
> i2c i2c-4: Failed to register i2c client MAGN0001:00 at 0x1d (-16)
>
> Tested on a Lenovo Yoga Tablet 2 1051-F.
>
> Signed-off-by: Marius Hoch <[email protected]>

By adding the HID to the ignore_serial_bus_ids[] list, you are
causing the ACPI code to instead instantiate a platform_device
for it.

This will successfully silence the error you are trying to silence,
but I don't think it really is what you intent / want to do ?

IMHO it would be better to add a NOT_PRESENT_ENTRY_HID() entry
to the override_status_ids[] table in drivers/acpi/x86/utils.c,
this will override the return of _STA, causing neither a
platform_device not an i2c_client to get instantiated for
the duplicate MAGN0001 device.

(please keep the comment above the entry why the entry is there)

Regards,

Hans







> ---
> drivers/acpi/scan.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> index 0c6f06abe3f4..d21c411e9719 100644
> --- a/drivers/acpi/scan.c
> +++ b/drivers/acpi/scan.c
> @@ -1730,6 +1730,12 @@ static bool acpi_device_enumeration_by_parent(struct acpi_device *device)
> */
> {"BCM4752", },
> {"LNV4752", },
> + /*
> + * The LSM303D on the Lenovo Yoga Tablet 2 series is present
> + * as both ACCL0001 and MAGN0001. As we can only ever register an
> + * i2c client for one of them, ignore MAGN0001.
> + */
> + {"MAGN0001", },
> {}
> };
>
>
> base-commit: cb0856346a60fe3eb837ba5e73588a41f81ac05f

2023-04-22 16:48:42

by Marius Hoch

[permalink] [raw]
Subject: Re: [PATCH] ACPI: scan: Ignore MAGN0001

Hi Hans,

On 21/04/2023 20:07, Hans de Goede wrote:
> Hi Marius,
>
> On 4/21/23 18:27, Marius Hoch wrote:
>> The LSM303D on the Lenovo Yoga Tablet 2 series is present
>> as both ACCL0001 and MAGN0001. As we can only ever register an
>> i2c client for one of them, ignore MAGN0001.
>>
>> Currently this errors:
>> i2c i2c-4: Failed to register i2c client MAGN0001:00 at 0x1d (-16)
>>
>> Tested on a Lenovo Yoga Tablet 2 1051-F.
>>
>> Signed-off-by: Marius Hoch <[email protected]>
> By adding the HID to the ignore_serial_bus_ids[] list, you are
> causing the ACPI code to instead instantiate a platform_device
> for it.
>
> This will successfully silence the error you are trying to silence,
> but I don't think it really is what you intent / want to do ?
It is not, no. Given we will only ever care for ACCL0001, we can
entirely skip MAGN0001.
>
> IMHO it would be better to add a NOT_PRESENT_ENTRY_HID() entry
> to the override_status_ids[] table in drivers/acpi/x86/utils.c,
> this will override the return of _STA, causing neither a
> platform_device not an i2c_client to get instantiated for
> the duplicate MAGN0001 device.
That sounds much better, thanks for the pointer!
I've send a new patch ([email protected]).

Cheers,
Marius
>
> (please keep the comment above the entry why the entry is there)
>
> Regards,
>
> Hans
>
>
>
>
>
>
>
>> ---
>> drivers/acpi/scan.c | 6 ++++++
>> 1 file changed, 6 insertions(+)
>>
>> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
>> index 0c6f06abe3f4..d21c411e9719 100644
>> --- a/drivers/acpi/scan.c
>> +++ b/drivers/acpi/scan.c
>> @@ -1730,6 +1730,12 @@ static bool acpi_device_enumeration_by_parent(struct acpi_device *device)
>> */
>> {"BCM4752", },
>> {"LNV4752", },
>> + /*
>> + * The LSM303D on the Lenovo Yoga Tablet 2 series is present
>> + * as both ACCL0001 and MAGN0001. As we can only ever register an
>> + * i2c client for one of them, ignore MAGN0001.
>> + */
>> + {"MAGN0001", },
>> {}
>> };
>>
>>
>> base-commit: cb0856346a60fe3eb837ba5e73588a41f81ac05f