Re: [PATCH for-4.8 V2 03/10] powerpc/mm/radix: Add radix_set_pte to use in early init

[Nicholas Piggin <npiggin@gmail.com>]

On Sat, 23 Jul 2016 14:42:36 +0530
"Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com> wrote:

> We want to use the static key based feature check in set_pte_at. Since
> we call radix__map_kernel_page early in boot before jump label is
> initialized we can't call set_pte_at there. Add radix__set_pte for the
> same.
> 
> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> ---
>  arch/powerpc/mm/pgtable-radix.c | 23 ++++++++++++++++++++++-
>  1 file changed, 22 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/powerpc/mm/pgtable-radix.c
> b/arch/powerpc/mm/pgtable-radix.c index 003ff48a11b6..6d2eb76b508e
> 100644 --- a/arch/powerpc/mm/pgtable-radix.c
> +++ b/arch/powerpc/mm/pgtable-radix.c
> @@ -39,6 +39,27 @@ static __ref void *early_alloc_pgtable(unsigned
> long size) 
>  	return pt;
>  }
> +/*
> + * set_pte stores a linux PTE into the linux page table.
> + */
> +static void radix__set_pte(struct mm_struct *mm, unsigned long addr,
> pte_t *ptep,
> +			   pte_t pte)
> +{
> +	/*
> +	 * When handling numa faults, we already have the pte marked
> +	 * _PAGE_PRESENT, but we can be sure that it is not in hpte.
> +	 * Hence we can use set_pte_at for them.
> +	 */
> +	VM_WARN_ON(pte_present(*ptep) && !pte_protnone(*ptep));
> +
> +	/*
> +	 * Add the pte bit when tryint set a pte
> +	 */
> +	pte = __pte(pte_val(pte) | _PAGE_PTE);
> +
> +	/* Perform the setting of the PTE */
> +	radix__set_pte_at(mm, addr, ptep, pte, 0);
> +}
>  
>  int radix__map_kernel_page(unsigned long ea, unsigned long pa,
>  			  pgprot_t flags,
> @@ -102,7 +123,7 @@ int radix__map_kernel_page(unsigned long ea,
> unsigned long pa, }
>  
>  set_the_pte:
> -	set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
> flags));
> +	radix__set_pte(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
> flags)); smp_wmb();

What we have in existing code is set_pte_at() function that adds
the _PAGE_PTE bit, then calls __set_pte_at(), which calls radix or hash
version of __set_pte_at().

Now we also have radix__set_pte(), which has the function of the
set_pte_at(), which is starting to confuse the naming convention.
The new function is a radix-only set_pte_at(), rather than the
radix implementation that gets called via set_pte().

set_pte_at_radix()? That kind of sucks too, though. It might be better
if the radix/hash variants were called __radix__set_pte_at(), and this
new function was called radix__set_pte_at().

Thanks,
Nick