Re: [PATCH net-next v5 4/6] net: dsa: motorcomm: introduce mib_init to control MIB polling initialization

[Kyle Switch <kyle.switch@motor-comm.com>]

On 9/4/26 21:45, Andrew Lunn wrote:
> On Fri, Sep 04, 2026 at 05:54:14PM +0800, Kyle Switch wrote:
>> Add a mib_init flag to indicate whether MIB polling has been
>> successfully initialized. This prevents the driver from attempting
>> to clean up MIB-related resources on removal for switch families
>> that do not support MIB polling, which would otherwise trigger
>> errors or crashes.
> This one i'm not convinced about. At some point, you are going to make
> initialisation of MIBs conditional. Why not use that condition to
> control removal as well a setup? Why do you need both a conditional
> and this pp->mib_init state flag?
>
>> @@ -4771,6 +4772,7 @@ static void yt921x_mdio_remove(struct mdio_device *mdiodev)
>>   	for (size_t i = ARRAY_SIZE(priv->ports); i-- > 0; ) {
>>   		struct yt921x_port *pp = &priv->ports[i];
>>   
>> +		if (pp->mib_init)
>>   			disable_delayed_work_sync(&pp->mib_read);
> This looks odd. When adding an if (), you would expect the next line
> to require an additional indent. Was the indentation already wrong?
Ans: In the next version, we will modify patches 3/6 and 4/6 of this series.
>      Andrew
>
> ---
> pw-bot: cr