* [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-07 22:58 ` j.glisse
0 siblings, 0 replies; 12+ messages in thread
From: j.glisse @ 2013-05-07 22:58 UTC (permalink / raw)
To: linux-mm; +Cc: linux-kernel, Jerome Glisse
From: Jerome Glisse <jglisse@redhat.com>
Do not increase page count if FOLL_GET is not set. None of the
current user can trigger the issue because none of the current
user call __get_user_pages with both the pages array ptr non
NULL and the FOLL_GET flags non set in other word all caller
of __get_user_pages that don't set the FOLL_GET flags also call
with pages == NULL.
v2: Do not use get_page_foll. Improved comment.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
---
mm/hugetlb.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index ca9a7c6..32f323b 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -2981,7 +2981,9 @@ long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma,
same_page:
if (pages) {
pages[i] = mem_map_offset(page, pfn_offset);
- get_page(pages[i]);
+ if (flags & FOLL_GET) {
+ get_page(pages[i]);
+ }
}
if (vmas)
--
1.7.11.7
^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-07 22:58 ` j.glisse
0 siblings, 0 replies; 12+ messages in thread
From: j.glisse @ 2013-05-07 22:58 UTC (permalink / raw)
To: linux-mm; +Cc: linux-kernel, Jerome Glisse
From: Jerome Glisse <jglisse@redhat.com>
Do not increase page count if FOLL_GET is not set. None of the
current user can trigger the issue because none of the current
user call __get_user_pages with both the pages array ptr non
NULL and the FOLL_GET flags non set in other word all caller
of __get_user_pages that don't set the FOLL_GET flags also call
with pages == NULL.
v2: Do not use get_page_foll. Improved comment.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
---
mm/hugetlb.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index ca9a7c6..32f323b 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -2981,7 +2981,9 @@ long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma,
same_page:
if (pages) {
pages[i] = mem_map_offset(page, pfn_offset);
- get_page(pages[i]);
+ if (flags & FOLL_GET) {
+ get_page(pages[i]);
+ }
}
if (vmas)
--
1.7.11.7
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
2013-05-07 22:58 ` j.glisse
@ 2013-05-08 0:47 ` KOSAKI Motohiro
-1 siblings, 0 replies; 12+ messages in thread
From: KOSAKI Motohiro @ 2013-05-08 0:47 UTC (permalink / raw)
To: Jerome Glisse; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
> From: Jerome Glisse <jglisse@redhat.com>
>
> Do not increase page count if FOLL_GET is not set. None of the
> current user can trigger the issue because none of the current
> user call __get_user_pages with both the pages array ptr non
> NULL and the FOLL_GET flags non set in other word all caller
> of __get_user_pages that don't set the FOLL_GET flags also call
> with pages == NULL.
Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
{
(snip)
VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-08 0:47 ` KOSAKI Motohiro
0 siblings, 0 replies; 12+ messages in thread
From: KOSAKI Motohiro @ 2013-05-08 0:47 UTC (permalink / raw)
To: Jerome Glisse; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
> From: Jerome Glisse <jglisse@redhat.com>
>
> Do not increase page count if FOLL_GET is not set. None of the
> current user can trigger the issue because none of the current
> user call __get_user_pages with both the pages array ptr non
> NULL and the FOLL_GET flags non set in other word all caller
> of __get_user_pages that don't set the FOLL_GET flags also call
> with pages == NULL.
Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
{
(snip)
VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
2013-05-08 0:47 ` KOSAKI Motohiro
@ 2013-05-08 0:51 ` Jerome Glisse
-1 siblings, 0 replies; 12+ messages in thread
From: Jerome Glisse @ 2013-05-08 0:51 UTC (permalink / raw)
To: KOSAKI Motohiro; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 8:47 PM, KOSAKI Motohiro
<kosaki.motohiro@gmail.com> wrote:
> On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
>> From: Jerome Glisse <jglisse@redhat.com>
>>
>> Do not increase page count if FOLL_GET is not set. None of the
>> current user can trigger the issue because none of the current
>> user call __get_user_pages with both the pages array ptr non
>> NULL and the FOLL_GET flags non set in other word all caller
>> of __get_user_pages that don't set the FOLL_GET flags also call
>> with pages == NULL.
>
> Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
Yes but this allow pages != NULL and FOLL_GET not set and as i said
there is no such user of that yet and this is exactly what i was
trying to use.
Cheers,
Jerome
>
> long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
> {
> (snip)
> VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-08 0:51 ` Jerome Glisse
0 siblings, 0 replies; 12+ messages in thread
From: Jerome Glisse @ 2013-05-08 0:51 UTC (permalink / raw)
To: KOSAKI Motohiro; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 8:47 PM, KOSAKI Motohiro
<kosaki.motohiro@gmail.com> wrote:
> On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
>> From: Jerome Glisse <jglisse@redhat.com>
>>
>> Do not increase page count if FOLL_GET is not set. None of the
>> current user can trigger the issue because none of the current
>> user call __get_user_pages with both the pages array ptr non
>> NULL and the FOLL_GET flags non set in other word all caller
>> of __get_user_pages that don't set the FOLL_GET flags also call
>> with pages == NULL.
>
> Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
Yes but this allow pages != NULL and FOLL_GET not set and as i said
there is no such user of that yet and this is exactly what i was
trying to use.
Cheers,
Jerome
>
> long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
> {
> (snip)
> VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
2013-05-08 0:51 ` Jerome Glisse
@ 2013-05-08 2:41 ` KOSAKI Motohiro
-1 siblings, 0 replies; 12+ messages in thread
From: KOSAKI Motohiro @ 2013-05-08 2:41 UTC (permalink / raw)
To: Jerome Glisse; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 8:51 PM, Jerome Glisse <j.glisse@gmail.com> wrote:
> On Tue, May 7, 2013 at 8:47 PM, KOSAKI Motohiro
> <kosaki.motohiro@gmail.com> wrote:
>> On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
>>> From: Jerome Glisse <jglisse@redhat.com>
>>>
>>> Do not increase page count if FOLL_GET is not set. None of the
>>> current user can trigger the issue because none of the current
>>> user call __get_user_pages with both the pages array ptr non
>>> NULL and the FOLL_GET flags non set in other word all caller
>>> of __get_user_pages that don't set the FOLL_GET flags also call
>>> with pages == NULL.
>>
>> Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
>
> Yes but this allow pages != NULL and FOLL_GET not set and as i said
> there is no such user of that yet and this is exactly what i was
> trying to use.
Why? The following bug_on inhibit both case.
>> VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-08 2:41 ` KOSAKI Motohiro
0 siblings, 0 replies; 12+ messages in thread
From: KOSAKI Motohiro @ 2013-05-08 2:41 UTC (permalink / raw)
To: Jerome Glisse; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 8:51 PM, Jerome Glisse <j.glisse@gmail.com> wrote:
> On Tue, May 7, 2013 at 8:47 PM, KOSAKI Motohiro
> <kosaki.motohiro@gmail.com> wrote:
>> On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
>>> From: Jerome Glisse <jglisse@redhat.com>
>>>
>>> Do not increase page count if FOLL_GET is not set. None of the
>>> current user can trigger the issue because none of the current
>>> user call __get_user_pages with both the pages array ptr non
>>> NULL and the FOLL_GET flags non set in other word all caller
>>> of __get_user_pages that don't set the FOLL_GET flags also call
>>> with pages == NULL.
>>
>> Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
>
> Yes but this allow pages != NULL and FOLL_GET not set and as i said
> there is no such user of that yet and this is exactly what i was
> trying to use.
Why? The following bug_on inhibit both case.
>> VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
2013-05-08 2:41 ` KOSAKI Motohiro
@ 2013-05-08 14:33 ` Jerome Glisse
-1 siblings, 0 replies; 12+ messages in thread
From: Jerome Glisse @ 2013-05-08 14:33 UTC (permalink / raw)
To: KOSAKI Motohiro; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 10:41 PM, KOSAKI Motohiro
<kosaki.motohiro@gmail.com> wrote:
> On Tue, May 7, 2013 at 8:51 PM, Jerome Glisse <j.glisse@gmail.com> wrote:
>> On Tue, May 7, 2013 at 8:47 PM, KOSAKI Motohiro
>> <kosaki.motohiro@gmail.com> wrote:
>>> On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
>>>> From: Jerome Glisse <jglisse@redhat.com>
>>>>
>>>> Do not increase page count if FOLL_GET is not set. None of the
>>>> current user can trigger the issue because none of the current
>>>> user call __get_user_pages with both the pages array ptr non
>>>> NULL and the FOLL_GET flags non set in other word all caller
>>>> of __get_user_pages that don't set the FOLL_GET flags also call
>>>> with pages == NULL.
>>>
>>> Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
>>
>> Yes but this allow pages != NULL and FOLL_GET not set and as i said
>> there is no such user of that yet and this is exactly what i was
>> trying to use.
>
> Why? The following bug_on inhibit both case.
Yes i get lost on the double negation, but still my patch is correct
and i am not using gup but follow_hugetlb_page directly and i run into
the issue. My patch does not change the behavior for current user,
just fix assumption new user might have when not setting the FOLL_GET
flags.
>>> VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-08 14:33 ` Jerome Glisse
0 siblings, 0 replies; 12+ messages in thread
From: Jerome Glisse @ 2013-05-08 14:33 UTC (permalink / raw)
To: KOSAKI Motohiro; +Cc: linux-mm, LKML, Jerome Glisse
On Tue, May 7, 2013 at 10:41 PM, KOSAKI Motohiro
<kosaki.motohiro@gmail.com> wrote:
> On Tue, May 7, 2013 at 8:51 PM, Jerome Glisse <j.glisse@gmail.com> wrote:
>> On Tue, May 7, 2013 at 8:47 PM, KOSAKI Motohiro
>> <kosaki.motohiro@gmail.com> wrote:
>>> On Tue, May 7, 2013 at 6:58 PM, <j.glisse@gmail.com> wrote:
>>>> From: Jerome Glisse <jglisse@redhat.com>
>>>>
>>>> Do not increase page count if FOLL_GET is not set. None of the
>>>> current user can trigger the issue because none of the current
>>>> user call __get_user_pages with both the pages array ptr non
>>>> NULL and the FOLL_GET flags non set in other word all caller
>>>> of __get_user_pages that don't set the FOLL_GET flags also call
>>>> with pages == NULL.
>>>
>>> Because, __get_user_pages() doesn't allow pages==NULL and FOLL_GET is on.
>>
>> Yes but this allow pages != NULL and FOLL_GET not set and as i said
>> there is no such user of that yet and this is exactly what i was
>> trying to use.
>
> Why? The following bug_on inhibit both case.
Yes i get lost on the double negation, but still my patch is correct
and i am not using gup but follow_hugetlb_page directly and i run into
the issue. My patch does not change the behavior for current user,
just fix assumption new user might have when not setting the FOLL_GET
flags.
>>> VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
2013-05-08 14:33 ` Jerome Glisse
@ 2013-05-08 23:41 ` KOSAKI Motohiro
-1 siblings, 0 replies; 12+ messages in thread
From: KOSAKI Motohiro @ 2013-05-08 23:41 UTC (permalink / raw)
To: Jerome Glisse; +Cc: linux-mm, LKML, Jerome Glisse
>> Why? The following bug_on inhibit both case.
>
> Yes i get lost on the double negation, but still my patch is correct
> and i am not using gup but follow_hugetlb_page directly and i run into
> the issue. My patch does not change the behavior for current user,
> just fix assumption new user might have when not setting the FOLL_GET
> flags.
I have no idea. I haven't seen your new use case.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2
@ 2013-05-08 23:41 ` KOSAKI Motohiro
0 siblings, 0 replies; 12+ messages in thread
From: KOSAKI Motohiro @ 2013-05-08 23:41 UTC (permalink / raw)
To: Jerome Glisse; +Cc: linux-mm, LKML, Jerome Glisse
>> Why? The following bug_on inhibit both case.
>
> Yes i get lost on the double negation, but still my patch is correct
> and i am not using gup but follow_hugetlb_page directly and i run into
> the issue. My patch does not change the behavior for current user,
> just fix assumption new user might have when not setting the FOLL_GET
> flags.
I have no idea. I haven't seen your new use case.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2013-05-08 23:42 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-05-07 22:58 [PATCH] mm: honor FOLL_GET flag in follow_hugetlb_page v2 j.glisse
2013-05-07 22:58 ` j.glisse
2013-05-08 0:47 ` KOSAKI Motohiro
2013-05-08 0:47 ` KOSAKI Motohiro
2013-05-08 0:51 ` Jerome Glisse
2013-05-08 0:51 ` Jerome Glisse
2013-05-08 2:41 ` KOSAKI Motohiro
2013-05-08 2:41 ` KOSAKI Motohiro
2013-05-08 14:33 ` Jerome Glisse
2013-05-08 14:33 ` Jerome Glisse
2013-05-08 23:41 ` KOSAKI Motohiro
2013-05-08 23:41 ` KOSAKI Motohiro
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.