2017-06-21 21:50:34

by Nitin Gupta

Subject: [PATCH] sparc64: Fix gup_huge_pmd

The function assumes that each PMD points to head of a
huge page. This is not correct as a PMD can point to
start of any 8M region with a, say 256M, hugepage. The
fix ensures that it points to the correct head of any PMD
huge page.

Signed-off-by: Nitin Gupta <[email protected]>
---
arch/sparc/mm/gup.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/arch/sparc/mm/gup.c b/arch/sparc/mm/gup.c
index cd0e32b..9116a6f 100644
--- a/arch/sparc/mm/gup.c
+++ b/arch/sparc/mm/gup.c
@@ -80,6 +80,8 @@ static int gup_huge_pmd(pmd_t *pmdp, pmd_t pmd, unsigned long addr,
refs = 0;
head = pmd_page(pmd);
page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
+ if (PageTail(head))
+ head = compound_head(head);
do {
VM_BUG_ON(compound_head(page) != head);
pages[*nr] = page;
--
2.9.2

2017-06-22 10:54:21

by Julian Calaby

Subject: Re: [PATCH] sparc64: Fix gup_huge_pmd

Hi Nitin,

On Thu, Jun 22, 2017 at 7:50 AM, Nitin Gupta <[email protected]> wrote:
> The function assumes that each PMD points to head of a
> huge page. This is not correct as a PMD can point to
> start of any 8M region with a, say 256M, hugepage. The
> fix ensures that it points to the correct head of any PMD
> huge page.
>
> Signed-off-by: Nitin Gupta <[email protected]>
> ---
> arch/sparc/mm/gup.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/arch/sparc/mm/gup.c b/arch/sparc/mm/gup.c
> index cd0e32b..9116a6f 100644
> --- a/arch/sparc/mm/gup.c
> +++ b/arch/sparc/mm/gup.c
> @@ -80,6 +80,8 @@ static int gup_huge_pmd(pmd_t *pmdp, pmd_t pmd, unsigned long addr,
> refs = 0;
> head = pmd_page(pmd);
> page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
> + if (PageTail(head))
> + head = compound_head(head);

Stupid question: shouldn't this go before the page calculation?

> do {
> VM_BUG_ON(compound_head(page) != head);
> pages[*nr] = page;

Thanks,

--
Julian Calaby

Email: [email protected]
Profile: http://www.google.com/profiles/julian.calaby/

2017-06-22 14:38:00

by Nitin Gupta

Subject: Re: [PATCH] sparc64: Fix gup_huge_pmd

Hi Julian,


On 6/22/17 3:53 AM, Julian Calaby wrote:
> On Thu, Jun 22, 2017 at 7:50 AM, Nitin Gupta <[email protected]> wrote:
>> The function assumes that each PMD points to head of a
>> huge page. This is not correct as a PMD can point to
>> start of any 8M region with a, say 256M, hugepage. The
>> fix ensures that it points to the correct head of any PMD
>> huge page.
>>
>> Signed-off-by: Nitin Gupta <[email protected]>
>> ---
>> arch/sparc/mm/gup.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/arch/sparc/mm/gup.c b/arch/sparc/mm/gup.c
>> index cd0e32b..9116a6f 100644
>> --- a/arch/sparc/mm/gup.c
>> +++ b/arch/sparc/mm/gup.c
>> @@ -80,6 +80,8 @@ static int gup_huge_pmd(pmd_t *pmdp, pmd_t pmd, unsigned long addr,
>> refs = 0;
>> head = pmd_page(pmd);
>> page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
>> + if (PageTail(head))
>> + head = compound_head(head);
> Stupid question: shouldn't this go before the page calculation?

No, it should be after page calculation: First, 'head' points to base of
the PMD page, then 'page' points to an offset within that page. Finally,
we make sure that head variable points to head of the compound page
which contains the addr.

I think confusion comes from the use of 'head' for pointing to a
non-head page. So, maybe it would be more clear to write that part
of the function this way:

page = pmd_page(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
head = compound_head(page);

Thanks,
Nitin

2017-06-22 14:45:38

by Julian Calaby

Subject: Re: [PATCH] sparc64: Fix gup_huge_pmd

Hi Nitin,

On Fri, Jun 23, 2017 at 12:37 AM, Nitin Gupta <[email protected]> wrote:
> Hi Julian,
>
>
> On 6/22/17 3:53 AM, Julian Calaby wrote:
>>
>> On Thu, Jun 22, 2017 at 7:50 AM, Nitin Gupta <[email protected]>
>> wrote:
>>>
>>> The function assumes that each PMD points to head of a
>>> huge page. This is not correct as a PMD can point to
>>> start of any 8M region with a, say 256M, hugepage. The
>>> fix ensures that it points to the correct head of any PMD
>>> huge page.
>>>
>>> Signed-off-by: Nitin Gupta <[email protected]>
>>> ---
>>> arch/sparc/mm/gup.c | 2 ++
>>> 1 file changed, 2 insertions(+)
>>>
>>> diff --git a/arch/sparc/mm/gup.c b/arch/sparc/mm/gup.c
>>> index cd0e32b..9116a6f 100644
>>> --- a/arch/sparc/mm/gup.c
>>> +++ b/arch/sparc/mm/gup.c
>>> @@ -80,6 +80,8 @@ static int gup_huge_pmd(pmd_t *pmdp, pmd_t pmd,
>>> unsigned long addr,
>>> refs = 0;
>>> head = pmd_page(pmd);
>>> page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
>>> + if (PageTail(head))
>>> + head = compound_head(head);
>>
>> Stupid question: shouldn't this go before the page calculation?
>
>
> No, it should be after page calculation: First, 'head' points to base of
> the PMD page, then 'page' points to an offset within that page. Finally,
> we make sure that head variable points to head of the compound page
> which contains the addr.

Thanks for the explanation, that makes a bit more sense to me.

> I think confusion comes from the use of 'head' for pointing to a
> non-head page. So, maybe it would be more clear to write that part
> of the function this way:
>
> page = pmd_page(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
> head = compound_head(page);

More verbose variable names would help too. =)

Thanks,

--
Julian Calaby

Email: [email protected]
Profile: http://www.google.com/profiles/julian.calaby/