On 7/1/24 04:36, Quentin Schulz wrote:
Hi Guenter, On 6/28/24 5:13 PM, Guenter Roeck wrote:The driver only supports a single chip, so an enum to determine the chip type is unnecessary. Drop it. Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx> --- drivers/hwmon/amc6821.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/hwmon/amc6821.c b/drivers/hwmon/amc6821.c index 042e2044de7b..ebffc9393c3d 100644 --- a/drivers/hwmon/amc6821.c +++ b/drivers/hwmon/amc6821.c @@ -36,8 +36,6 @@ module_param(pwminv, int, 0444); static int init = 1; /*Power-on initialization.*/ module_param(init, int, 0444); -enum chips { amc6821 }; - #define AMC6821_REG_DEV_ID 0x3D #define AMC6821_REG_COMP_ID 0x3E #define AMC6821_REG_CONF1 0x00 @@ -943,7 +941,7 @@ static int amc6821_probe(struct i2c_client *client) } static const struct i2c_device_id amc6821_id[] = { - { "amc6821", amc6821 }, + { "amc6821", 0 },amc6821_id being a global variable, its content should be initialized by the compiler if omitted, and the default value of kernel_ulong_t (aka unsigned long) is 0. So I think we could just remove the second parameter there. Doesn't hurt to keep it though and it seems there's a mix of implicit and explicit initialization among the i2c kernel drivers, so with that said:
I may use that inconsistently myself. I am never sure if I want to say "keep default" or "explicitly state that the value is 0". I'll keep it as-is.
Reviewed-by: Quentin Schulz <quentin.schulz@xxxxxxxxx>
Thanks, Guenter
