On Tue, 28 Dec 2021 16:03:45 -0800
Randy Dunlap <rdunlap@xxxxxxxxxxxxx> wrote:
> Provide a simple implementation of clk_set_parent() in the lantiq
> subarch so that callers of it will build without errors.
>
> Fixes these build errors:
>
> ERROR: modpost: "clk_set_parent" [sound/soc/jz4740/snd-soc-jz4740-i2s.ko] undefined!
> ERROR: modpost: "clk_set_parent" [sound/soc/atmel/snd-soc-atmel-i2s.ko] undefined!
>
> Fixes: 171bb2f19ed6 ("MIPS: Lantiq: Add initial support for Lantiq SoCs")
> Signed-off-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> Reported-by: kernel test robot <lkp@xxxxxxxxx>
> --to=linux-mips@xxxxxxxxxxxxxxx --cc="John Crispin <john@xxxxxxxxxxx>" --cc="Jonathan Cameron <jic23@xxxxxxxxxx>" --cc="Russell King <linux@xxxxxxxxxxxxxxx>" --cc="Andy Shevchenko <andy.shevchenko@xxxxxxxxx>" --cc=alsa-devel@xxxxxxxxxxxxxxxx --to="Thomas Bogendoerfer <tsbogend@xxxxxxxxxxxxxxxx>"
I guess hitting this sooner or later was inevitable given the recent fix
of adding clk_get_parent() to these two. It might be worth adding a
WARN_ON() or similar to this function though so we know if anything
ever actually hits it at runtime. Seems like at least some of the
other similar stubs do so (omap1 for example).
With or without a warning, LGTM
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> ---
> arch/mips/lantiq/clk.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> --- linux-next-20211224.orig/arch/mips/lantiq/clk.c
> +++ linux-next-20211224/arch/mips/lantiq/clk.c
> @@ -164,6 +164,12 @@ struct clk *clk_get_parent(struct clk *c
> }
> EXPORT_SYMBOL(clk_get_parent);
>
> +int clk_set_parent(struct clk *clk, struct clk *parent)
> +{
> + return 0;
> +}
> +EXPORT_SYMBOL(clk_set_parent);
> +
> static inline u32 get_counter_resolution(void)
> {
> u32 res;
