Hi,

* Cory Maccarrone <darkstar6262@gmail.com> [091212 17:35]:
> This change implements the clocks, platform driver, and register
> information necessary to use the spi100k bus with OMAP 7xx systems.
> 
> The clocks added are dummy clocks because, although we're pretty
> sure there are clocks used for SPI, all current booting methods result
> in proper operation without the enabling of any other clocks.

Looks like the driver is in mainline, so let's try to make it working..

> --- a/arch/arm/mach-omap1/devices.c
> +++ b/arch/arm/mach-omap1/devices.c
> @@ -196,6 +198,37 @@ void __init omap1_init_mmc(struct omap_mmc_platform_data **mmc_data,
>  
>  /*-------------------------------------------------------------------------*/
>  
> +/* OMAP7xx SPI support */
> +#if defined(CONFIG_SPI_OMAP_100K) || defined(CONFIG_SPI_OMAP_100K_MODULE)
> +
> +struct platform_device omap_spi1 = {
> +	.name           = "omap1_spi100k",
> +	.id             = 1,
> +};
> +
> +struct platform_device omap_spi2 = {
> +	.name           = "omap1_spi100k",
> +	.id             = 2,
> +};
> +
> +static void omap_init_spi100k(void)
> +{
> +	omap_spi1.dev.platform_data = ioremap(OMAP7XX_SPI1_BASE, 0x7ff);
> +	BUG_ON(!omap_spi1.dev.platform_data);
> +
> +	omap_spi2.dev.platform_data = ioremap(OMAP7XX_SPI2_BASE, 0x7ff);
> +	BUG_ON(!omap_spi2.dev.platform_data);
> +
> +	platform_device_register(&omap_spi1);
> +	platform_device_register(&omap_spi2);
> +}

I've updated your patch to just check the return value of ioremap instead
of BUG_ON. I've also merged it with the mux patch, see below. Can you
please try it out and make sure everything is OK?

Regards,

Tony