On Mon, Oct 14, 2024 at 03:40:51PM +0200, Andrew Lunn wrote:
> On Sun, Oct 13, 2024 at 11:27:39PM +0200, Jan Petrous via B4 Relay wrote:
> > From: "Jan Petrous (OSS)" <jan.petrous@xxxxxxxxxxx>
> >
> > The helper rgmii_clock() implemented Russel's hint during stmmac
> > glue driver review:
> >
> > > We seem to have multiple cases of very similar logic in lots of stmmac
> > > platform drivers, and I think it's about time we said no more to this.
> > > So, what I think we should do is as follows:
> > >
> > > add the following helper - either in stmmac, or more generically
> > > (phylib? - in which case its name will need changing.)
> > >
> > > static long stmmac_get_rgmii_clock(int speed)
> > > {
> > > switch (speed) {
> > > case SPEED_10:
> > > return 2500000;
> > >
> > > case SPEED_100:
> > > return 25000000;
> > >
> > > case SPEED_1000:
> > > return 125000000;
> > >
> > > default:
> > > return -ENVAL;
> > > }
> > > }
> > >
> > > Then, this can become:
> > >
> > > long tx_clk_rate;
> > >
> > > ...
> > >
> > > tx_clk_rate = stmmac_get_rgmii_clock(speed);
> > > if (tx_clk_rate < 0) {
> > > dev_err(gmac->dev, "Unsupported/Invalid speed: %d\n", speed);
> > > return;
> > > }
> > >
> > > ret = clk_set_rate(gmac->tx_clk, tx_clk_rate);
> >
> > Suggested-by: Russell King (Oracle) <rmk+kernel@xxxxxxxxxxxxxxx>
> > Signed-off-by: Jan Petrous (OSS) <jan.petrous@xxxxxxxxxxx>
>
> But of an unusual commit message, but it does explain the "Why?".
I will reformulate description in v4.
>
> Reviewed-by: Andrew Lunn <andrew@xxxxxxx>
>
Thanks.
> >
> > +/**
> > + * rgmii_clock - map link speed to the clock rate
> > + * @speed: link speed value
> > + *
> > + * Description: maps RGMII supported link speeds
> > + * into the clock rates.
> > + */
>
> A Returns: line would be nice.
will add it in v4.
Thanks
/Jan
