Re: [PATCH] [v2] i3c, iio: fix i3c driver dependencies

From: Alexandre Belloni <alexandre.belloni@bootlin.com>
To: Arnd Bergmann <arnd@kernel.org>
Cc: "Guenter Roeck" <linux@roeck-us.net>,
	"Jonathan Cameron" <jic23@kernel.org>,
	"Arnd Bergmann" <arnd@arndb.de>,
	"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
	"Andy Shevchenko" <andriy.shevchenko@intel.com>,
	"Frank Li" <Frank.Li@nxp.com>,
	"Adrian Fluturel" <fluturel.adrian@gmail.com>,
	"Carlos Song" <carlos.song@nxp.com>,
	"David Lechner" <dlechner@baylibre.com>,
	"Nuno Sá" <nuno.sa@analog.com>,
	"Andy Shevchenko" <andy@kernel.org>,
	"Jean Delvare" <jdelvare@suse.de>,
	linux-hwmon@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-i3c@lists.infradead.org, linux-iio@vger.kernel.org
Subject: Re: [PATCH] [v2] i3c, iio: fix i3c driver dependencies
Date: Wed, 4 Feb 2026 17:33:21 +0100	[thread overview]
Message-ID: <20260204163321cfaa320e@mail.local> (raw)
In-Reply-To: <20260202160543.3654499-1-arnd@kernel.org>

On 02/02/2026 17:04:46+0100, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
> 
> All combined i2c/i3c drivers appear to suffer from the same link
> time problem when CONFIG_I3C is set to 'm':
> 
> arm-linux-gnueabi-ld: drivers/iio/magnetometer/mmc5633.o: in function `mmc5633_i3c_driver_init':
> mmc5633.c:(.init.text+0x30): undefined reference to `i3c_driver_register_with_owner'
> 
> This was previously fixed several times by marking individual
> drivers as 'depends on I2C; depends on I3C || !I3C', but this gets
> tedious and is somewhat confusing.
> 
> Add a Kconfig symbol 'I3C_OR_I2C' to help replace those dependencies,
> and use this in all the existing drivers that had already fixed it
> as well as the new mmc5633 driver.
> 
> Fixes: 6e5f6bf2e3f0 ("iio: magnetometer: Add mmc5633 sensor")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> v2: restore accidentally deleted lines
> ---
>  drivers/hwmon/Kconfig            |  6 ++----
>  drivers/i3c/Kconfig              | 12 ++++++++++++
>  drivers/iio/magnetometer/Kconfig |  2 +-
>  drivers/misc/amd-sbi/Kconfig     |  3 +--
>  4 files changed, 16 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
> index 41c381764c2b..ecfba861f66d 100644
> --- a/drivers/hwmon/Kconfig
> +++ b/drivers/hwmon/Kconfig
> @@ -1493,8 +1493,7 @@ config SENSORS_LM73
>  
>  config SENSORS_LM75
>  	tristate "National Semiconductor LM75 and compatibles"
> -	depends on I2C
> -	depends on I3C || !I3C
> +	depends on I3C_OR_I2C
>  	select REGMAP_I2C
>  	select REGMAP_I3C if I3C
>  	help
> @@ -2392,8 +2391,7 @@ config SENSORS_TMP103
>  
>  config SENSORS_TMP108
>  	tristate "Texas Instruments TMP108"
> -	depends on I2C
> -	depends on I3C || !I3C
> +	depends on I3C_OR_I2C
>  	select REGMAP_I2C
>  	select REGMAP_I3C if I3C
>  	help
> diff --git a/drivers/i3c/Kconfig b/drivers/i3c/Kconfig
> index 30a441506f61..626c54b386d5 100644
> --- a/drivers/i3c/Kconfig
> +++ b/drivers/i3c/Kconfig
> @@ -22,3 +22,15 @@ menuconfig I3C
>  if I3C
>  source "drivers/i3c/master/Kconfig"
>  endif # I3C
> +
> +config I3C_OR_I2C

I'm fine with this but I wanted to mention that we have RTC_I2C_AND_SPI
the "and" being because it is from the point of view of the device (the
device supports both bus) while we have I3C_OR_I2C, "or" being about the
bus (the device can be on any bus).

So I guess at some point we are going to need a combination of
I3C_OR_I2C and SPI and this may be weird to have both OR and AND.

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com