Re: [PATCH v3 08/14] x86, dax, libnvdimm: move wb_cache_pmem() to libnvdimm

[Christoph Hellwig <hch@lst.de>]

> +void clean_cache_range(void *addr, size_t size);
>  
>  static inline int
>  __copy_from_user_inatomic_nocache(void *dst, const void __user *src,
> diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c
> index f42d2fd86ca3..baa80ff29da8 100644
> --- a/arch/x86/lib/usercopy_64.c
> +++ b/arch/x86/lib/usercopy_64.c
> @@ -85,7 +85,7 @@ copy_user_handle_tail(char *to, char *from, unsigned len)
>   * instruction. Note that @size is internally rounded up to be cache
>   * line size aligned.
>   */
> -static void clean_cache_range(void *addr, size_t size)
> +void clean_cache_range(void *addr, size_t size)

Can you keep clean_cache_range private please?  Just add
arch_wb_cache_pmem to usercopy_64.c just behind it so that the
compiler can tail-call and export that instead.

> --- a/drivers/nvdimm/pmem.h
> +++ b/drivers/nvdimm/pmem.h
> @@ -4,6 +4,13 @@
>  #include <linux/types.h>
>  #include <linux/pfn_t.h>
>  #include <linux/fs.h>
> +#include <asm/pmem.h>
> +
> +#ifndef CONFIG_ARCH_HAS_PMEM_API
> +static inline void arch_wb_cache_pmem(void *addr, size_t size)
> +{
> +}
> +#endif

And our normal Linux style would be to have this linux linux/pmem.h,
which should always be included for the asm version.