[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7e476fd2-5818-c395-cdf2-00b5229c1a73@intel.com>
Date: Fri, 6 Oct 2017 16:10:31 -0700
From: Dave Hansen <dave.hansen@...el.com>
To: "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
Andrew Morton <akpm@...ux-foundation.org>
Cc: linux-mm@...ck.org, linux-kernel@...r.kernel.org,
linux-api@...r.kernel.org, Michal Hocko <mhocko@...nel.org>
Subject: Re: [PATCHv2 1/2] mm: Introduce wrappers to access mm->nr_ptes
On 10/05/2017 03:14 AM, Kirill A. Shutemov wrote:
> +++ b/arch/sparc/mm/hugetlbpage.c
> @@ -396,7 +396,7 @@ static void hugetlb_free_pte_range(struct mmu_gather *tlb, pmd_t *pmd,
>
> pmd_clear(pmd);
> pte_free_tlb(tlb, token, addr);
> - atomic_long_dec(&tlb->mm->nr_ptes);
> + mm_dec_nr_ptes(tlb->mm);
> }
If we're going to go replace all of these, I wonder if we should start
doing it more generically.
mm_dec_nr_pgtable(PGTABLE_PTE, tlb->mm)
or even:
mm_dec_nr_pgtable(PGTABLE_LEVEL1, tlb->mm)
Instead of having a separate batch of functions for each level.
Powered by blists - more mailing lists