* [PATCH] mm: page_idle_get_page() does not need zone_lru_lock
@ 2016-12-06 5:55 Hugh Dickins
2016-12-07 8:32 ` Vlastimil Babka
2016-12-07 11:08 ` Vladimir Davydov
0 siblings, 2 replies; 4+ messages in thread
From: Hugh Dickins @ 2016-12-06 5:55 UTC (permalink / raw)
To: Andrew Morton; +Cc: Vladimir Davydov, Pavel Emelyanov, linux-mm
Rechecking PageLRU() after get_page_unless_zero() may have value, but
holding zone_lru_lock around that serves no useful purpose: delete it.
Signed-off-by: Hugh Dickins <hughd@google.com>
---
mm/page_idle.c | 4 ----
1 file changed, 4 deletions(-)
--- 4.9-rc8/mm/page_idle.c 2016-10-02 16:24:33.000000000 -0700
+++ linux/mm/page_idle.c 2016-12-05 19:44:32.646625435 -0800
@@ -30,7 +30,6 @@
static struct page *page_idle_get_page(unsigned long pfn)
{
struct page *page;
- struct zone *zone;
if (!pfn_valid(pfn))
return NULL;
@@ -40,13 +39,10 @@ static struct page *page_idle_get_page(u
!get_page_unless_zero(page))
return NULL;
- zone = page_zone(page);
- spin_lock_irq(zone_lru_lock(zone));
if (unlikely(!PageLRU(page))) {
put_page(page);
page = NULL;
}
- spin_unlock_irq(zone_lru_lock(zone));
return page;
}
--
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] 4+ messages in thread
* Re: [PATCH] mm: page_idle_get_page() does not need zone_lru_lock
2016-12-06 5:55 [PATCH] mm: page_idle_get_page() does not need zone_lru_lock Hugh Dickins
@ 2016-12-07 8:32 ` Vlastimil Babka
2016-12-07 11:08 ` Vladimir Davydov
1 sibling, 0 replies; 4+ messages in thread
From: Vlastimil Babka @ 2016-12-07 8:32 UTC (permalink / raw)
To: Hugh Dickins, Andrew Morton; +Cc: Vladimir Davydov, Pavel Emelyanov, linux-mm
On 12/06/2016 06:55 AM, Hugh Dickins wrote:
> Rechecking PageLRU() after get_page_unless_zero() may have value, but
> holding zone_lru_lock around that serves no useful purpose: delete it.
>
> Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
> ---
>
> mm/page_idle.c | 4 ----
> 1 file changed, 4 deletions(-)
>
> --- 4.9-rc8/mm/page_idle.c 2016-10-02 16:24:33.000000000 -0700
> +++ linux/mm/page_idle.c 2016-12-05 19:44:32.646625435 -0800
> @@ -30,7 +30,6 @@
> static struct page *page_idle_get_page(unsigned long pfn)
> {
> struct page *page;
> - struct zone *zone;
>
> if (!pfn_valid(pfn))
> return NULL;
> @@ -40,13 +39,10 @@ static struct page *page_idle_get_page(u
> !get_page_unless_zero(page))
> return NULL;
>
> - zone = page_zone(page);
> - spin_lock_irq(zone_lru_lock(zone));
> if (unlikely(!PageLRU(page))) {
> put_page(page);
> page = NULL;
> }
> - spin_unlock_irq(zone_lru_lock(zone));
> return page;
> }
>
>
> --
> 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>
>
--
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] 4+ messages in thread
* Re: [PATCH] mm: page_idle_get_page() does not need zone_lru_lock
2016-12-06 5:55 [PATCH] mm: page_idle_get_page() does not need zone_lru_lock Hugh Dickins
2016-12-07 8:32 ` Vlastimil Babka
@ 2016-12-07 11:08 ` Vladimir Davydov
2016-12-07 13:07 ` Vlastimil Babka
1 sibling, 1 reply; 4+ messages in thread
From: Vladimir Davydov @ 2016-12-07 11:08 UTC (permalink / raw)
To: Hugh Dickins; +Cc: Minchan Kim, Andrew Morton, Pavel Emelyanov, linux-mm
Hello,
On Mon, Dec 05, 2016 at 09:55:10PM -0800, Hugh Dickins wrote:
> Rechecking PageLRU() after get_page_unless_zero() may have value, but
> holding zone_lru_lock around that serves no useful purpose: delete it.
IIRC this lock/unlock was added on purpose, by request from Minchan. It
serves as a barrier that guarantees that all page fields (specifically
->mapping in case of anonymous pages) have been properly initialized by
the time we pass it to rmap_walk(). Here's a reference to the thread
where this problem was discussed:
http://lkml.kernel.org/r/<20150430082531.GD21771@blaptop>
>
> Signed-off-by: Hugh Dickins <hughd@google.com>
> ---
>
> mm/page_idle.c | 4 ----
> 1 file changed, 4 deletions(-)
>
> --- 4.9-rc8/mm/page_idle.c 2016-10-02 16:24:33.000000000 -0700
> +++ linux/mm/page_idle.c 2016-12-05 19:44:32.646625435 -0800
> @@ -30,7 +30,6 @@
> static struct page *page_idle_get_page(unsigned long pfn)
> {
> struct page *page;
> - struct zone *zone;
>
> if (!pfn_valid(pfn))
> return NULL;
> @@ -40,13 +39,10 @@ static struct page *page_idle_get_page(u
> !get_page_unless_zero(page))
> return NULL;
>
> - zone = page_zone(page);
> - spin_lock_irq(zone_lru_lock(zone));
> if (unlikely(!PageLRU(page))) {
> put_page(page);
> page = NULL;
> }
> - spin_unlock_irq(zone_lru_lock(zone));
> return page;
> }
>
--
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] 4+ messages in thread
* Re: [PATCH] mm: page_idle_get_page() does not need zone_lru_lock
2016-12-07 11:08 ` Vladimir Davydov
@ 2016-12-07 13:07 ` Vlastimil Babka
0 siblings, 0 replies; 4+ messages in thread
From: Vlastimil Babka @ 2016-12-07 13:07 UTC (permalink / raw)
To: Vladimir Davydov, Hugh Dickins
Cc: Minchan Kim, Andrew Morton, Pavel Emelyanov, linux-mm
On 12/07/2016 12:08 PM, Vladimir Davydov wrote:
> Hello,
>
> On Mon, Dec 05, 2016 at 09:55:10PM -0800, Hugh Dickins wrote:
>> Rechecking PageLRU() after get_page_unless_zero() may have value, but
>> holding zone_lru_lock around that serves no useful purpose: delete it.
>
> IIRC this lock/unlock was added on purpose, by request from Minchan. It
> serves as a barrier that guarantees that all page fields (specifically
> ->mapping in case of anonymous pages) have been properly initialized by
> the time we pass it to rmap_walk(). Here's a reference to the thread
> where this problem was discussed:
>
> http://lkml.kernel.org/r/<20150430082531.GD21771@blaptop>
Um OK, but then there should have been a comment explaining why the lock
is there :/
Vlastimil
>>
>> Signed-off-by: Hugh Dickins <hughd@google.com>
>> ---
>>
>> mm/page_idle.c | 4 ----
>> 1 file changed, 4 deletions(-)
>>
>> --- 4.9-rc8/mm/page_idle.c 2016-10-02 16:24:33.000000000 -0700
>> +++ linux/mm/page_idle.c 2016-12-05 19:44:32.646625435 -0800
>> @@ -30,7 +30,6 @@
>> static struct page *page_idle_get_page(unsigned long pfn)
>> {
>> struct page *page;
>> - struct zone *zone;
>>
>> if (!pfn_valid(pfn))
>> return NULL;
>> @@ -40,13 +39,10 @@ static struct page *page_idle_get_page(u
>> !get_page_unless_zero(page))
>> return NULL;
>>
>> - zone = page_zone(page);
>> - spin_lock_irq(zone_lru_lock(zone));
>> if (unlikely(!PageLRU(page))) {
>> put_page(page);
>> page = NULL;
>> }
>> - spin_unlock_irq(zone_lru_lock(zone));
>> return page;
>> }
>>
>
> --
> 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>
>
--
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] 4+ messages in thread
end of thread, other threads:[~2016-12-07 13:08 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-12-06 5:55 [PATCH] mm: page_idle_get_page() does not need zone_lru_lock Hugh Dickins
2016-12-07 8:32 ` Vlastimil Babka
2016-12-07 11:08 ` Vladimir Davydov
2016-12-07 13:07 ` Vlastimil Babka
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox