Re: [PATCH v1 1/2] ufs: core: Configure only active lanes during link

[Bart Van Assche]

On 3/10/26 11:09 PM, palash.kambar@xxxxxxxxxxxxxxxx wrote:
> [ ... ]
> +static int ufshcd_get_connected_tx_lanes(struct ufs_hba *hba, u32 
*tx_lanes)
> +{
> +	return ufshcd_dme_get(hba,
> +			      UIC_ARG_MIB(PA_CONNECTEDTXDATALANES), tx_lanes);
> +}
> +
> +static int ufshcd_get_connected_rx_lanes(struct ufs_hba *hba, u32 *rx_lanes)
> +{
> +	return ufshcd_dme_get(hba,
> +			      UIC_ARG_MIB(PA_CONNECTEDRXDATALANES), rx_lanes);
> +}

The body of the above two functions is very short. Please remove these
functions and instead inline these function into their only caller.

> +static void ufshcd_validate_link_params(struct ufs_hba *hba)
> +{
> +	int val = 0;
> +
> +	if (ufshcd_get_connected_tx_lanes(hba, &val))
> +		return;

Shouldn't it be reported if ufshcd_get_connected_tx_lanes() fails?

> +	if (ufshcd_get_connected_rx_lanes(hba, &val))
> +		return;

Same question here - shouldn't it be reported if
ufshcd_get_connected_rx_lanes() fails?

Why does this patch only call dev_err() in case of a mismatch instead of 
adjusting hba->lanes_per_direction or making initialization fail?

Thanks,

Bart.