2021-05-11 13:50:56

by Miaohe Lin

Subject: [PATCH v3 2/5] mm/huge_memory.c: use page->deferred_list

Now that we can represent the location of ->deferred_list instead of
->mapping + ->index, make use of it to improve readability.

Reviewed-by: Yang Shi <[email protected]>
Reviewed-by: David Hildenbrand <[email protected]>
Signed-off-by: Miaohe Lin <[email protected]>
---
mm/huge_memory.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 63ed6b25deaa..76ca1eb2a223 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2868,7 +2868,7 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
spin_lock_irqsave(&ds_queue->split_queue_lock, flags);
/* Take pin on all head pages to avoid freeing them under us */
list_for_each_safe(pos, next, &ds_queue->split_queue) {
- page = list_entry((void *)pos, struct page, mapping);
+ page = list_entry((void *)pos, struct page, deferred_list);
page = compound_head(page);
if (get_page_unless_zero(page)) {
list_move(page_deferred_list(page), &list);
@@ -2883,7 +2883,7 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
spin_unlock_irqrestore(&ds_queue->split_queue_lock, flags);

list_for_each_safe(pos, next, &list) {
- page = list_entry((void *)pos, struct page, mapping);
+ page = list_entry((void *)pos, struct page, deferred_list);
if (!trylock_page(page))
goto next;
/* split_huge_page() removes page from list on success */
--
2.23.0

2021-05-11 23:09:11

by Matthew Wilcox (Oracle)

Subject: Re: [PATCH v3 2/5] mm/huge_memory.c: use page->deferred_list

On Tue, May 11, 2021 at 09:48:54PM +0800, Miaohe Lin wrote:
> Now that we can represent the location of ->deferred_list instead of
> ->mapping + ->index, make use of it to improve readability.
>
> Reviewed-by: Yang Shi <[email protected]>
> Reviewed-by: David Hildenbrand <[email protected]>
> Signed-off-by: Miaohe Lin <[email protected]>
> ---
> mm/huge_memory.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 63ed6b25deaa..76ca1eb2a223 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -2868,7 +2868,7 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
> spin_lock_irqsave(&ds_queue->split_queue_lock, flags);
> /* Take pin on all head pages to avoid freeing them under us */
> list_for_each_safe(pos, next, &ds_queue->split_queue) {
> - page = list_entry((void *)pos, struct page, mapping);
> + page = list_entry((void *)pos, struct page, deferred_list);
> page = compound_head(page);

This is an equivalent transformation, but it doesn't really go far
enough. I think you want something like this:

struct page *page, *next;

list_for_each_entry_safe(page, next, &ds_queue->split_queue,
deferred_list) {
struct page *head = page - 1;
... then use head throughout ...
}

2021-05-12 02:03:19

by Miaohe Lin

Subject: Re: [PATCH v3 2/5] mm/huge_memory.c: use page->deferred_list

On 2021/5/12 7:03, Matthew Wilcox wrote:
> On Tue, May 11, 2021 at 09:48:54PM +0800, Miaohe Lin wrote:
>> Now that we can represent the location of ->deferred_list instead of
>> ->mapping + ->index, make use of it to improve readability.
>>
>> Reviewed-by: Yang Shi <[email protected]>
>> Reviewed-by: David Hildenbrand <[email protected]>
>> Signed-off-by: Miaohe Lin <[email protected]>
>> ---
>> mm/huge_memory.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
>> index 63ed6b25deaa..76ca1eb2a223 100644
>> --- a/mm/huge_memory.c
>> +++ b/mm/huge_memory.c
>> @@ -2868,7 +2868,7 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
>> spin_lock_irqsave(&ds_queue->split_queue_lock, flags);
>> /* Take pin on all head pages to avoid freeing them under us */
>> list_for_each_safe(pos, next, &ds_queue->split_queue) {
>> - page = list_entry((void *)pos, struct page, mapping);
>> + page = list_entry((void *)pos, struct page, deferred_list);
>> page = compound_head(page);
>
> This is an equivalent transformation, but it doesn't really go far
> enough. I think you want something like this:
>
> struct page *page, *next;
>
> list_for_each_entry_safe(page, next, &ds_queue->split_queue,
> deferred_list) {
> struct page *head = page - 1;
> ... then use head throughout ...
> }
>

Many thanks for your time and reminder. list_for_each_entry_safe is equivalent
to list_for_each_safe + list_entry and there is many places using list_for_each_safe
+ list_entry, so I think it's ok to keep the code as it is.
Thanks again. :)

> .
>