2020-07-11 01:00:54

by Alex Shi

Subject: [PATCH v16 14/22] mm/thp: add tail pages into lru anyway in split_huge_page()

split_huge_page() must start with PageLRU(head), but lru bit *maybe*
cleared by isolate_lru_page, anyway the head still in lru list, since we
still held the lru_lock.

Signed-off-by: Alex Shi <[email protected]>
Cc: Kirill A. Shutemov <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Matthew Wilcox <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Mika Penttilä <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
mm/huge_memory.c | 15 +++++++++------
1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index d866b6e43434..4fe7b92c9330 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2348,15 +2348,18 @@ static void lru_add_page_tail(struct page *head, struct page *page_tail,
VM_BUG_ON_PAGE(PageLRU(page_tail), head);
lockdep_assert_held(&lruvec_pgdat(lruvec)->lru_lock);

- if (!list)
- SetPageLRU(page_tail);
-
- if (likely(PageLRU(head)))
- list_add_tail(&page_tail->lru, &head->lru);
- else if (list) {
+ if (list) {
/* page reclaim is reclaiming a huge page */
get_page(page_tail);
list_add_tail(&page_tail->lru, list);
+ } else {
+ /*
+ * Split start from PageLRU(head), but lru bit maybe cleared
+ * by isolate_lru_page, but head still in lru list, since we
+ * held the lru_lock.
+ */
+ SetPageLRU(page_tail);
+ list_add_tail(&page_tail->lru, &head->lru);
}
}

--
1.8.3.1

2020-07-17 09:34:10

by Alex Shi

Subject: Re: [PATCH v16 14/22] mm/thp: add tail pages into lru anyway in split_huge_page()

Add a VM_WARN_ON for tracking. and updated comments for the code.

Thanks

---
From f1381a1547625a6521777bf9235823d8fbd00dac Mon Sep 17 00:00:00 2001
From: Alex Shi <[email protected]>
Date: Fri, 10 Jul 2020 16:54:37 +0800
Subject: [PATCH v16 14/22] mm/thp: add tail pages into lru anyway in
split_huge_page()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Split_huge_page() must start with PageLRU(head), and we are holding the
lru_lock here. If the head was cleared lru bit unexpected, tracking it.

Signed-off-by: Alex Shi <[email protected]>
Cc: Kirill A. Shutemov <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Matthew Wilcox <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Mika Penttilä <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
mm/huge_memory.c | 16 ++++++++++------
1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index d866b6e43434..28538444197b 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2348,15 +2348,19 @@ static void lru_add_page_tail(struct page *head, struct page *page_tail,
VM_BUG_ON_PAGE(PageLRU(page_tail), head);
lockdep_assert_held(&lruvec_pgdat(lruvec)->lru_lock);

- if (!list)
- SetPageLRU(page_tail);
-
- if (likely(PageLRU(head)))
- list_add_tail(&page_tail->lru, &head->lru);
- else if (list) {
+ if (list) {
/* page reclaim is reclaiming a huge page */
get_page(page_tail);
list_add_tail(&page_tail->lru, list);
+ } else {
+ /*
+ * Split start from PageLRU(head), and we are holding the
+ * lru_lock.
+ * Do a warning if the head's lru bit was cleared unexpected.
+ */
+ VM_WARN_ON(!PageLRU(head));
+ SetPageLRU(page_tail);
+ list_add_tail(&page_tail->lru, &head->lru);
}
}

--
1.8.3.1

2020-07-20 08:52:05

by Kirill A. Shutemov

Subject: Re: [PATCH v16 14/22] mm/thp: add tail pages into lru anyway in split_huge_page()

On Fri, Jul 17, 2020 at 05:30:27PM +0800, Alex Shi wrote:
>
> Add a VM_WARN_ON for tracking. and updated comments for the code.
>
> Thanks
>
> ---
> From f1381a1547625a6521777bf9235823d8fbd00dac Mon Sep 17 00:00:00 2001
> From: Alex Shi <[email protected]>
> Date: Fri, 10 Jul 2020 16:54:37 +0800
> Subject: [PATCH v16 14/22] mm/thp: add tail pages into lru anyway in
> split_huge_page()
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
>
> Split_huge_page() must start with PageLRU(head), and we are holding the
> lru_lock here. If the head was cleared lru bit unexpected, tracking it.
>
> Signed-off-by: Alex Shi <[email protected]>

Reviewed-by: Kirill A. Shutemov <[email protected]>

--
Kirill A. Shutemov

2020-07-20 09:07:46

by Alex Shi

Subject: Re: [PATCH v16 14/22] mm/thp: add tail pages into lru anyway in split_huge_page()

?? 2020/7/20 ????4:49, Kirill A. Shutemov д??:
>>
>> Split_huge_page() must start with PageLRU(head), and we are holding the
>> lru_lock here. If the head was cleared lru bit unexpected, tracking it.
>>
>> Signed-off-by: Alex Shi <[email protected]>
> Reviewed-by: Kirill A. Shutemov <[email protected]>

Hi Kirill,

Millions thanks for review!

Alex