Hi Lars-Peter, Mark
Thank you for your feedback
diff --git a/sound/soc/codecs/wm8510.c b/sound/soc/codecs/wm8510.c index 119ceac..2455606 100644 --- a/sound/soc/codecs/wm8510.c +++ b/sound/soc/codecs/wm8510.c @@ -651,6 +651,7 @@ static int wm8510_spi_remove(struct spi_device *spi) .probe = wm8510_spi_probe, .remove = wm8510_spi_remove, }; +module_spi_driver(wm8510_spi_driver); #endif /* CONFIG_SPI_MASTER */
#if IS_ENABLED(CONFIG_I2C) @@ -698,40 +699,9 @@ static int wm8510_i2c_remove(struct i2c_client *client) .remove = wm8510_i2c_remove, .id_table = wm8510_i2c_id, }; +module_i2c_driver(wm8510_i2c_driver); #endif
This wont work when the driver is built as a module as this creates multiple module init and exit entry points.
Ideally the driver would be split into submodules for i2c and spi, like we've already done for a few drivers.
Mark. I can do it, but what is your opinion ? Do nothing for i2c/spi driver, or use Lars-Peter's idea ? If former, just ignore i2c/spi driver patches. If later, I will send v2 patch.