Re: [linux-sunxi] Re: [PATCH 04/10] net: stmmac: sunxi platfrom extensionsfor GMAC in Allwinner A20 SoC's

From: srinivas kandagatla
Date: Thu Jan 02 2014 - 08:17:39 EST

Hi Chen,

On 24/12/13 03:27, Chen-Yu Tsai wrote:
> Srinivas,
> Let's keep platform data as of_data, so SoC compatibles can pass
> hardware feature flags for cores that don't support auto-detection.

I understand your concern here, But making platform_data as of_data
would just open a wide door for other hacky stuff too. So other glue
drivers would just use this interface instead, ignoring standard
interfaces. Also there would be issues on who takes the priority if the
parameters are specified in multiple places.

Can't this field/s be one of the variable in the struct stmmac_of_data
rather than reusing platform data?

> We can adjust the callbacks as you suggested, and I added a .free
> to complement .setup. Driver private data and platform data are
> off limits to the callbacks. My version of the callbacks:
> void (*fix_mac_speed)(void *priv, unsigned int speed);
> void (*bus_setup)(void __iomem *ioaddr);

In reply to your question in last email:
bus_setup this is something very specific to ST which configures the bus
parameters of AMBA-to-STBUS converter. This register falls inside the
memory map of stmmac.

> void *(*setup)(struct platform_device *pdev);
> void (*free)(struct platform_device *pdev, void *priv);
> int (*init)(struct platform_device *pdev, void *priv);
> void (*exit)(struct platform_device *pdev, void *priv);

The callbacks to Ok to me, unless Peppe has more comments on this.


> Cheers,
> ChenYu

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at