On some x86 Bay Trail tablets which shipped with Android as factory OS,
the DSDT is so broken that the PMIC needs to be manually instantiated by
the special x86-android-tablets.ko "fixup" driver for cases like this.

Add an i2c_device_id table so that the driver can match on manually
instantiated i2c_client-s (which lack an ACPI fwnode to match on).

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/mfd/intel_soc_pmic_crc.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/mfd/intel_soc_pmic_crc.c b/drivers/mfd/intel_soc_pmic_crc.c
index 876d017f74fe..ba22458ccb84 100644
--- a/drivers/mfd/intel_soc_pmic_crc.c
+++ b/drivers/mfd/intel_soc_pmic_crc.c
@@ -259,12 +259,19 @@ static const struct acpi_device_id crystal_cove_acpi_match[] = {
 };
 MODULE_DEVICE_TABLE(acpi, crystal_cove_acpi_match);
 
+static const struct i2c_device_id crystal_cove_i2c_match[] = {
+	{ "intel-crystal-cove" },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, crystal_cove_i2c_match);
+
 static struct i2c_driver crystal_cove_i2c_driver = {
 	.driver = {
 		.name = "crystal_cove_i2c",
 		.pm = pm_sleep_ptr(&crystal_cove_pm_ops),
 		.acpi_match_table = crystal_cove_acpi_match,
 	},
+	.id_table = crystal_cove_i2c_match,
 	.probe = crystal_cove_i2c_probe,
 	.remove = crystal_cove_i2c_remove,
 	.shutdown = crystal_cove_shutdown,
-- 
2.47.0

On Fri, Oct 25, 2024 at 10:37:12AM +0200, Hans de Goede wrote:
> On some x86 Bay Trail tablets which shipped with Android as factory OS,
> the DSDT is so broken that the PMIC needs to be manually instantiated by
> the special x86-android-tablets.ko "fixup" driver for cases like this.
> 
> Add an i2c_device_id table so that the driver can match on manually
> instantiated i2c_client-s (which lack an ACPI fwnode to match on).

...

> +static const struct i2c_device_id crystal_cove_i2c_match[] = {
> +	{ "intel-crystal-cove" },

Why this can't be "crystal_cove_i2c"?

> +	{ }
> +};

...

>  	.driver = {
>  		.name = "crystal_cove_i2c",
>  		.pm = pm_sleep_ptr(&crystal_cove_pm_ops),
>  		.acpi_match_table = crystal_cove_acpi_match,
>  	},

-- 
With Best Regards,
Andy Shevchenko

Hi Andy,

On 25-Oct-24 3:27 PM, Andy Shevchenko wrote:
> On Fri, Oct 25, 2024 at 10:37:12AM +0200, Hans de Goede wrote:
>> On some x86 Bay Trail tablets which shipped with Android as factory OS,
>> the DSDT is so broken that the PMIC needs to be manually instantiated by
>> the special x86-android-tablets.ko "fixup" driver for cases like this.
>>
>> Add an i2c_device_id table so that the driver can match on manually
>> instantiated i2c_client-s (which lack an ACPI fwnode to match on).
> 
> ...
> 
>> +static const struct i2c_device_id crystal_cove_i2c_match[] = {
>> +	{ "intel-crystal-cove" },
> 
> Why this can't be "crystal_cove_i2c"?

It can be any string as long as it is unique. Typically this will
be vendor-model-name though and having i2c in there is a bit weird
since this is used for the modalias, which gets prefixed with
"i2c:" already.

Therefor I would prefer to keep this as is. But if you have
a strong preference I can change this for v2.

Please let me know how you want to proceed with this patch.

Regards,

Hans



> 
>> +	{ }
>> +};
> 
> ...
> 
>>  	.driver = {
>>  		.name = "crystal_cove_i2c",
>>  		.pm = pm_sleep_ptr(&crystal_cove_pm_ops),
>>  		.acpi_match_table = crystal_cove_acpi_match,
>>  	},
>

On Sat, Oct 26, 2024 at 01:42:36PM +0200, Hans de Goede wrote:
> On 25-Oct-24 3:27 PM, Andy Shevchenko wrote:
> > On Fri, Oct 25, 2024 at 10:37:12AM +0200, Hans de Goede wrote:

...

> >> +static const struct i2c_device_id crystal_cove_i2c_match[] = {
> >> +	{ "intel-crystal-cove" },
> > 
> > Why this can't be "crystal_cove_i2c"?
> 
> It can be any string as long as it is unique. Typically this will
> be vendor-model-name though and having i2c in there is a bit weird
> since this is used for the modalias, which gets prefixed with
> "i2c:" already.
> 
> Therefor I would prefer to keep this as is. But if you have
> a strong preference I can change this for v2.
> 
> Please let me know how you want to proceed with this patch.

I think that I would like to have the same name there and in the .driver below.
If you think we need align these across PMIC MFD drivers, I'm fine with that as
well.

> >> +	{ }
> >> +};

...

> >>  	.driver = {
> >>  		.name = "crystal_cove_i2c",

At least two options based on the existing code:
"Crystal Cove PMIC" or "intel_soc_pmic_crc". I'm also
not against other one as long as it's done for all
PMIC MFD drivers.

> >>  		.pm = pm_sleep_ptr(&crystal_cove_pm_ops),
> >>  		.acpi_match_table = crystal_cove_acpi_match,
> >>  	},

-- 
With Best Regards,
Andy Shevchenko