Patchwork [-V6,18/27] mm/THP: withdraw the pgtable after pmdp related operations

login
register
mail settings
Submitter Aneesh Kumar K.V
Date April 22, 2013, 10 a.m.
Message ID <1366624861-24948-19-git-send-email-aneesh.kumar@linux.vnet.ibm.com>
Download mbox | patch
Permalink /patch/238414/
State Superseded
Headers show

Comments

Aneesh Kumar K.V - April 22, 2013, 10 a.m.
From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>

For architectures like ppc64 we look at deposited pgtable when
calling pmdp_get_and_clear. So do the pgtable_trans_huge_withdraw
after finishing pmdp related operations.

Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
---
 mm/huge_memory.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
Andrea Arcangeli - April 22, 2013, 3:49 p.m.
Hi,

On Mon, Apr 22, 2013 at 03:30:52PM +0530, Aneesh Kumar K.V wrote:
> From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
> 
> For architectures like ppc64 we look at deposited pgtable when
> calling pmdp_get_and_clear. So do the pgtable_trans_huge_withdraw
> after finishing pmdp related operations.
> 
> Cc: Andrea Arcangeli <aarcange@redhat.com>
> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> ---
>  mm/huge_memory.c |    3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 84f3180..2a43782 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -1363,9 +1363,10 @@ int zap_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
>  		struct page *page;
>  		pgtable_t pgtable;
>  		pmd_t orig_pmd;
> -		pgtable = pgtable_trans_huge_withdraw(tlb->mm, pmd);
> +
>  		orig_pmd = pmdp_get_and_clear(tlb->mm, addr, pmd);
>  		tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
> +		pgtable = pgtable_trans_huge_withdraw(tlb->mm, pmd);
>  		if (is_huge_zero_pmd(orig_pmd)) {
>  			tlb->mm->nr_ptes--;
>  			spin_unlock(&tlb->mm->page_table_lock);

I think here a comment inline (not only in the commit msg) is in
order. Otherwise it's hard to imagine others to be aware of this arch
detail when they will read the code later. So it would be prone to
break later without a comment.

Thanks,
Andrea

Patch

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 84f3180..2a43782 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1363,9 +1363,10 @@  int zap_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
 		struct page *page;
 		pgtable_t pgtable;
 		pmd_t orig_pmd;
-		pgtable = pgtable_trans_huge_withdraw(tlb->mm, pmd);
+
 		orig_pmd = pmdp_get_and_clear(tlb->mm, addr, pmd);
 		tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
+		pgtable = pgtable_trans_huge_withdraw(tlb->mm, pmd);
 		if (is_huge_zero_pmd(orig_pmd)) {
 			tlb->mm->nr_ptes--;
 			spin_unlock(&tlb->mm->page_table_lock);