Re: [PATCH net-next 1/4] net: stmmac: socfpga: Agilex5 EMAC platform configuration

[G Thomas, Rohan]

Hi Russell,

Thanks for reviewing the patch.

On 10/29/2025 9:44 PM, Russell King (Oracle) wrote:
> On Wed, Oct 29, 2025 at 04:06:13PM +0800, Rohan G Thomas via B4 Relay wrote:
> > +static void socfpga_common_plat_dat(struct socfpga_dwmac *dwmac)
> > +{
> > +	struct plat_stmmacenet_data *plat_dat = dwmac->plat_dat;
> > +
> > +	plat_dat->bsp_priv = dwmac;
>
> Surely this is something which is always done? What's the point in
> moving this to a function that always needs to be called from the
> implementation specific setup_plat_dat() method?

Yes, for all the current platforms this is common.

> > +	plat_dat->fix_mac_speed = socfpga_dwmac_fix_mac_speed;
> > +	plat_dat->init = socfpga_dwmac_init;
> > +	plat_dat->pcs_init = socfpga_dwmac_pcs_init;
> > +	plat_dat->pcs_exit = socfpga_dwmac_pcs_exit;
> > +	plat_dat->select_pcs = socfpga_dwmac_select_pcs;
>
>  From what I can see in your patch series, these are never changed.
> So, I question the value of having this "common_plat_dat"
> initialisation function. Why not leave this code in
> socfpga_dwmac_probe(), and just move the initialisation of
> plat_dat->core_type and plat_dat->riwt_off ?

Agreed. Will keep it in socfpga_dwmac_probe() itself.

Best Regards,
Rohan