Re: [PATCH net-next v2 6/9] net: dsa: microchip: dcb: add special handling for KSZ88X3 family

[Simon Horman <horms@kernel.org>]

On Wed, Apr 03, 2024 at 11:29:02AM +0200, Oleksij Rempel wrote:

...

> diff --git a/drivers/net/dsa/microchip/ksz8795.c b/drivers/net/dsa/microchip/ksz8795.c
> index 0ead198bacb2c..a520352d5b93e 100644
> --- a/drivers/net/dsa/microchip/ksz8795.c
> +++ b/drivers/net/dsa/microchip/ksz8795.c
> @@ -179,6 +179,21 @@ static int ksz8_port_queue_split(struct ksz_device *dev, int port, int queues)
>  	return ksz_prmw8(dev, port, reg_2q, mask_2q, data_2q);
>  }
>  
> +int ksz8_all_queues_split(struct ksz_device *dev, int queues)
> +{
> +	struct dsa_switch *ds = dev->ds;
> +	const struct dsa_port *dp;
> +	int ret;
> +
> +	dsa_switch_for_each_port(dp, ds) {
> +		ret = ksz8_port_queue_split(dev, dp->index, queues);
> +		if (ret)
> +			return ret;
> +	}
> +
> +	return ret;

nit: Probably it can't occur, but if dsa_switch_for_each_port() iterates
     zero times then ret is used uninitialised here.

     In any case, perhaps it is slightly nicer to reduce the scope
     of ret to inside the loop, and simply return 0 here.

> +}
> +
>  void ksz8_r_mib_cnt(struct ksz_device *dev, int port, u16 addr, u64 *cnt)
>  {
>  	const u32 *masks;

...