On Fri, Dec 11, 2020 at 01:19:16AM +0000, Andre Przywara wrote:
> +	SUNXI_PIN(SUNXI_PINCTRL_PIN(G, 10),
> +		  SUNXI_FUNCTION(0x0, "gpio_in"),
> +		  SUNXI_FUNCTION(0x1, "gpio_out"),
> +		  SUNXI_FUNCTION(0x2, "h_i2s2"),	/* MCLK */
> +		  SUNXI_FUNCTION(0x3, "clock"),		/* X32KFOUT */
> +		  SUNXI_FUNCTION_IRQ_BANK(0x6, 5, 10)),	/* PG_EINT10 */

It looks like there's no i2s vs h_i2s issue anymore? Maybe we can just
do s/h_i2s/i2s

> +	SUNXI_PIN(SUNXI_PINCTRL_PIN(I, 3),
> +		  SUNXI_FUNCTION(0x0, "gpio_in"),
> +		  SUNXI_FUNCTION(0x1, "gpio_out"),
> +		  SUNXI_FUNCTION(0x2, "emac0"),		/* ERXD0 */
> +		  SUNXI_FUNCTION(0x3, "dmic"),		/* DATA2 */
> +		  SUNXI_FUNCTION(0x4, "h_i2s0"),	/* DO0 */
> +		  SUNXI_FUNCTION(0x5, "h_i2s0_a"),	/* DI1 */
> +		  SUNXI_FUNCTION_IRQ_BANK(0x6, 7, 3)),	/* PI_EINT3 */

The way we dealt with this for the A100 is that we put the function as a
suffix which is more natural, so h_i2s0_dout0 and h_i2s0_din1

Maxime