* [PATCH v3] mm: migrate: record the mlocked page status to remove unnecessary lru drain
@ 2023-10-21 4:33 Baolin Wang
2023-10-23 0:54 ` Huang, Ying
2023-10-23 13:16 ` Zi Yan
0 siblings, 2 replies; 3+ messages in thread
From: Baolin Wang @ 2023-10-21 4:33 UTC (permalink / raw)
To: akpm
Cc: mgorman, hughd, vbabka, ying.huang, ziy, fengwei.yin,
baolin.wang, linux-mm, linux-kernel
When doing compaction, I found the lru_add_drain() is an obvious hotspot
when migrating pages. The distribution of this hotspot is as follows:
- 18.75% compact_zone
- 17.39% migrate_pages
- 13.79% migrate_pages_batch
- 11.66% migrate_folio_move
- 7.02% lru_add_drain
+ 7.02% lru_add_drain_cpu
+ 3.00% move_to_new_folio
1.23% rmap_walk
+ 1.92% migrate_folio_unmap
+ 3.20% migrate_pages_sync
+ 0.90% isolate_migratepages
The lru_add_drain() was added by commit c3096e6782b7 ("mm/migrate:
__unmap_and_move() push good newpage to LRU") to drain the newpage to LRU
immediately, to help to build up the correct newpage->mlock_count in
remove_migration_ptes() for mlocked pages. However, if there are no mlocked
pages are migrating, then we can avoid this lru drain operation, especailly
for the heavy concurrent scenarios.
So we can record the source pages' mlocked status in migrate_folio_unmap(),
and only drain the lru list when the mlocked status is set in migrate_folio_move().
In addition, the page was already isolated from lru when migrating, so checking
the mlocked status is stable by folio_test_mlocked() in migrate_folio_unmap().
After this patch, I can see the hotpot of the lru_add_drain() is gone:
- 9.41% migrate_pages_batch
- 6.15% migrate_folio_move
- 3.64% move_to_new_folio
+ 1.80% migrate_folio_extra
+ 1.70% buffer_migrate_folio
+ 1.41% rmap_walk
+ 0.62% folio_add_lru
+ 3.07% migrate_folio_unmap
Meanwhile, the compaction latency shows some improvements when running
thpscale:
base patched
Amean fault-both-1 1131.22 ( 0.00%) 1112.55 * 1.65%*
Amean fault-both-3 2489.75 ( 0.00%) 2324.15 * 6.65%*
Amean fault-both-5 3257.37 ( 0.00%) 3183.18 * 2.28%*
Amean fault-both-7 4257.99 ( 0.00%) 4079.04 * 4.20%*
Amean fault-both-12 6614.02 ( 0.00%) 6075.60 * 8.14%*
Amean fault-both-18 10607.78 ( 0.00%) 8978.86 * 15.36%*
Amean fault-both-24 14911.65 ( 0.00%) 11619.55 * 22.08%*
Amean fault-both-30 14954.67 ( 0.00%) 14925.66 * 0.19%*
Amean fault-both-32 16654.87 ( 0.00%) 15580.31 * 6.45%*
Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
---
Changes from v2:
- Use BIT() to define macros.
- Simplify handing page_was_mapped and page_was_mlocked variables with
introducing old_page_state variable.
Changes from v1:
- Use separate flags in __migrate_folio_record() to avoid to pack flags
in each call site per Ying.
---
mm/migrate.c | 48 +++++++++++++++++++++++++++++-------------------
1 file changed, 29 insertions(+), 19 deletions(-)
diff --git a/mm/migrate.c b/mm/migrate.c
index 125194f5af0f..35a88334bb3c 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1027,22 +1027,28 @@ union migration_ptr {
struct anon_vma *anon_vma;
struct address_space *mapping;
};
+
+enum {
+ PAGE_WAS_MAPPED = BIT(0),
+ PAGE_WAS_MLOCKED = BIT(1),
+};
+
static void __migrate_folio_record(struct folio *dst,
- unsigned long page_was_mapped,
+ unsigned long old_page_state,
struct anon_vma *anon_vma)
{
union migration_ptr ptr = { .anon_vma = anon_vma };
dst->mapping = ptr.mapping;
- dst->private = (void *)page_was_mapped;
+ dst->private = (void *)old_page_state;
}
static void __migrate_folio_extract(struct folio *dst,
- int *page_was_mappedp,
+ int *old_page_state,
struct anon_vma **anon_vmap)
{
union migration_ptr ptr = { .mapping = dst->mapping };
*anon_vmap = ptr.anon_vma;
- *page_was_mappedp = (unsigned long)dst->private;
+ *old_page_state = (unsigned long)dst->private;
dst->mapping = NULL;
dst->private = NULL;
}
@@ -1103,7 +1109,7 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
{
struct folio *dst;
int rc = -EAGAIN;
- int page_was_mapped = 0;
+ int old_page_state = 0;
struct anon_vma *anon_vma = NULL;
bool is_lru = !__folio_test_movable(src);
bool locked = false;
@@ -1157,6 +1163,8 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
folio_lock(src);
}
locked = true;
+ if (folio_test_mlocked(src))
+ old_page_state |= PAGE_WAS_MLOCKED;
if (folio_test_writeback(src)) {
/*
@@ -1206,7 +1214,7 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
dst_locked = true;
if (unlikely(!is_lru)) {
- __migrate_folio_record(dst, page_was_mapped, anon_vma);
+ __migrate_folio_record(dst, old_page_state, anon_vma);
return MIGRATEPAGE_UNMAP;
}
@@ -1232,11 +1240,11 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
VM_BUG_ON_FOLIO(folio_test_anon(src) &&
!folio_test_ksm(src) && !anon_vma, src);
try_to_migrate(src, mode == MIGRATE_ASYNC ? TTU_BATCH_FLUSH : 0);
- page_was_mapped = 1;
+ old_page_state |= PAGE_WAS_MAPPED;
}
if (!folio_mapped(src)) {
- __migrate_folio_record(dst, page_was_mapped, anon_vma);
+ __migrate_folio_record(dst, old_page_state, anon_vma);
return MIGRATEPAGE_UNMAP;
}
@@ -1248,7 +1256,8 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
if (rc == -EAGAIN)
ret = NULL;
- migrate_folio_undo_src(src, page_was_mapped, anon_vma, locked, ret);
+ migrate_folio_undo_src(src, old_page_state & PAGE_WAS_MAPPED,
+ anon_vma, locked, ret);
migrate_folio_undo_dst(dst, dst_locked, put_new_folio, private);
return rc;
@@ -1261,12 +1270,12 @@ static int migrate_folio_move(free_folio_t put_new_folio, unsigned long private,
struct list_head *ret)
{
int rc;
- int page_was_mapped = 0;
+ int old_page_state = 0;
struct anon_vma *anon_vma = NULL;
bool is_lru = !__folio_test_movable(src);
struct list_head *prev;
- __migrate_folio_extract(dst, &page_was_mapped, &anon_vma);
+ __migrate_folio_extract(dst, &old_page_state, &anon_vma);
prev = dst->lru.prev;
list_del(&dst->lru);
@@ -1287,10 +1296,10 @@ static int migrate_folio_move(free_folio_t put_new_folio, unsigned long private,
* isolated from the unevictable LRU: but this case is the easiest.
*/
folio_add_lru(dst);
- if (page_was_mapped)
+ if (old_page_state & PAGE_WAS_MLOCKED)
lru_add_drain();
- if (page_was_mapped)
+ if (old_page_state & PAGE_WAS_MAPPED)
remove_migration_ptes(src, dst, false);
out_unlock_both:
@@ -1322,11 +1331,12 @@ static int migrate_folio_move(free_folio_t put_new_folio, unsigned long private,
*/
if (rc == -EAGAIN) {
list_add(&dst->lru, prev);
- __migrate_folio_record(dst, page_was_mapped, anon_vma);
+ __migrate_folio_record(dst, old_page_state, anon_vma);
return rc;
}
- migrate_folio_undo_src(src, page_was_mapped, anon_vma, true, ret);
+ migrate_folio_undo_src(src, old_page_state & PAGE_WAS_MAPPED,
+ anon_vma, true, ret);
migrate_folio_undo_dst(dst, true, put_new_folio, private);
return rc;
@@ -1799,12 +1809,12 @@ static int migrate_pages_batch(struct list_head *from,
dst = list_first_entry(&dst_folios, struct folio, lru);
dst2 = list_next_entry(dst, lru);
list_for_each_entry_safe(folio, folio2, &unmap_folios, lru) {
- int page_was_mapped = 0;
+ int old_page_state = 0;
struct anon_vma *anon_vma = NULL;
- __migrate_folio_extract(dst, &page_was_mapped, &anon_vma);
- migrate_folio_undo_src(folio, page_was_mapped, anon_vma,
- true, ret_folios);
+ __migrate_folio_extract(dst, &old_page_state, &anon_vma);
+ migrate_folio_undo_src(folio, old_page_state & PAGE_WAS_MAPPED,
+ anon_vma, true, ret_folios);
list_del(&dst->lru);
migrate_folio_undo_dst(dst, true, put_new_folio, private);
dst = dst2;
--
2.39.3
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v3] mm: migrate: record the mlocked page status to remove unnecessary lru drain
2023-10-21 4:33 [PATCH v3] mm: migrate: record the mlocked page status to remove unnecessary lru drain Baolin Wang
@ 2023-10-23 0:54 ` Huang, Ying
2023-10-23 13:16 ` Zi Yan
1 sibling, 0 replies; 3+ messages in thread
From: Huang, Ying @ 2023-10-23 0:54 UTC (permalink / raw)
To: Baolin Wang
Cc: akpm, mgorman, hughd, vbabka, ziy, fengwei.yin, linux-mm, linux-kernel
Baolin Wang <baolin.wang@linux.alibaba.com> writes:
> When doing compaction, I found the lru_add_drain() is an obvious hotspot
> when migrating pages. The distribution of this hotspot is as follows:
> - 18.75% compact_zone
> - 17.39% migrate_pages
> - 13.79% migrate_pages_batch
> - 11.66% migrate_folio_move
> - 7.02% lru_add_drain
> + 7.02% lru_add_drain_cpu
> + 3.00% move_to_new_folio
> 1.23% rmap_walk
> + 1.92% migrate_folio_unmap
> + 3.20% migrate_pages_sync
> + 0.90% isolate_migratepages
>
> The lru_add_drain() was added by commit c3096e6782b7 ("mm/migrate:
> __unmap_and_move() push good newpage to LRU") to drain the newpage to LRU
> immediately, to help to build up the correct newpage->mlock_count in
> remove_migration_ptes() for mlocked pages. However, if there are no mlocked
> pages are migrating, then we can avoid this lru drain operation, especailly
> for the heavy concurrent scenarios.
>
> So we can record the source pages' mlocked status in migrate_folio_unmap(),
> and only drain the lru list when the mlocked status is set in migrate_folio_move().
> In addition, the page was already isolated from lru when migrating, so checking
> the mlocked status is stable by folio_test_mlocked() in migrate_folio_unmap().
>
> After this patch, I can see the hotpot of the lru_add_drain() is gone:
> - 9.41% migrate_pages_batch
> - 6.15% migrate_folio_move
> - 3.64% move_to_new_folio
> + 1.80% migrate_folio_extra
> + 1.70% buffer_migrate_folio
> + 1.41% rmap_walk
> + 0.62% folio_add_lru
> + 3.07% migrate_folio_unmap
>
> Meanwhile, the compaction latency shows some improvements when running
> thpscale:
> base patched
> Amean fault-both-1 1131.22 ( 0.00%) 1112.55 * 1.65%*
> Amean fault-both-3 2489.75 ( 0.00%) 2324.15 * 6.65%*
> Amean fault-both-5 3257.37 ( 0.00%) 3183.18 * 2.28%*
> Amean fault-both-7 4257.99 ( 0.00%) 4079.04 * 4.20%*
> Amean fault-both-12 6614.02 ( 0.00%) 6075.60 * 8.14%*
> Amean fault-both-18 10607.78 ( 0.00%) 8978.86 * 15.36%*
> Amean fault-both-24 14911.65 ( 0.00%) 11619.55 * 22.08%*
> Amean fault-both-30 14954.67 ( 0.00%) 14925.66 * 0.19%*
> Amean fault-both-32 16654.87 ( 0.00%) 15580.31 * 6.45%*
>
> Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
LGTM, Thanks!
Reviewed-by: "Huang, Ying" <ying.huang@intel.com>
> ---
> Changes from v2:
> - Use BIT() to define macros.
> - Simplify handing page_was_mapped and page_was_mlocked variables with
> introducing old_page_state variable.
>
> Changes from v1:
> - Use separate flags in __migrate_folio_record() to avoid to pack flags
> in each call site per Ying.
> ---
> mm/migrate.c | 48 +++++++++++++++++++++++++++++-------------------
> 1 file changed, 29 insertions(+), 19 deletions(-)
>
> diff --git a/mm/migrate.c b/mm/migrate.c
> index 125194f5af0f..35a88334bb3c 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -1027,22 +1027,28 @@ union migration_ptr {
> struct anon_vma *anon_vma;
> struct address_space *mapping;
> };
> +
> +enum {
> + PAGE_WAS_MAPPED = BIT(0),
> + PAGE_WAS_MLOCKED = BIT(1),
> +};
> +
> static void __migrate_folio_record(struct folio *dst,
> - unsigned long page_was_mapped,
> + unsigned long old_page_state,
> struct anon_vma *anon_vma)
> {
> union migration_ptr ptr = { .anon_vma = anon_vma };
> dst->mapping = ptr.mapping;
> - dst->private = (void *)page_was_mapped;
> + dst->private = (void *)old_page_state;
> }
>
> static void __migrate_folio_extract(struct folio *dst,
> - int *page_was_mappedp,
> + int *old_page_state,
> struct anon_vma **anon_vmap)
> {
> union migration_ptr ptr = { .mapping = dst->mapping };
> *anon_vmap = ptr.anon_vma;
> - *page_was_mappedp = (unsigned long)dst->private;
> + *old_page_state = (unsigned long)dst->private;
> dst->mapping = NULL;
> dst->private = NULL;
> }
> @@ -1103,7 +1109,7 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
> {
> struct folio *dst;
> int rc = -EAGAIN;
> - int page_was_mapped = 0;
> + int old_page_state = 0;
> struct anon_vma *anon_vma = NULL;
> bool is_lru = !__folio_test_movable(src);
> bool locked = false;
> @@ -1157,6 +1163,8 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
> folio_lock(src);
> }
> locked = true;
> + if (folio_test_mlocked(src))
> + old_page_state |= PAGE_WAS_MLOCKED;
>
> if (folio_test_writeback(src)) {
> /*
> @@ -1206,7 +1214,7 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
> dst_locked = true;
>
> if (unlikely(!is_lru)) {
> - __migrate_folio_record(dst, page_was_mapped, anon_vma);
> + __migrate_folio_record(dst, old_page_state, anon_vma);
> return MIGRATEPAGE_UNMAP;
> }
>
> @@ -1232,11 +1240,11 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
> VM_BUG_ON_FOLIO(folio_test_anon(src) &&
> !folio_test_ksm(src) && !anon_vma, src);
> try_to_migrate(src, mode == MIGRATE_ASYNC ? TTU_BATCH_FLUSH : 0);
> - page_was_mapped = 1;
> + old_page_state |= PAGE_WAS_MAPPED;
> }
>
> if (!folio_mapped(src)) {
> - __migrate_folio_record(dst, page_was_mapped, anon_vma);
> + __migrate_folio_record(dst, old_page_state, anon_vma);
> return MIGRATEPAGE_UNMAP;
> }
>
> @@ -1248,7 +1256,8 @@ static int migrate_folio_unmap(new_folio_t get_new_folio,
> if (rc == -EAGAIN)
> ret = NULL;
>
> - migrate_folio_undo_src(src, page_was_mapped, anon_vma, locked, ret);
> + migrate_folio_undo_src(src, old_page_state & PAGE_WAS_MAPPED,
> + anon_vma, locked, ret);
> migrate_folio_undo_dst(dst, dst_locked, put_new_folio, private);
>
> return rc;
> @@ -1261,12 +1270,12 @@ static int migrate_folio_move(free_folio_t put_new_folio, unsigned long private,
> struct list_head *ret)
> {
> int rc;
> - int page_was_mapped = 0;
> + int old_page_state = 0;
> struct anon_vma *anon_vma = NULL;
> bool is_lru = !__folio_test_movable(src);
> struct list_head *prev;
>
> - __migrate_folio_extract(dst, &page_was_mapped, &anon_vma);
> + __migrate_folio_extract(dst, &old_page_state, &anon_vma);
> prev = dst->lru.prev;
> list_del(&dst->lru);
>
> @@ -1287,10 +1296,10 @@ static int migrate_folio_move(free_folio_t put_new_folio, unsigned long private,
> * isolated from the unevictable LRU: but this case is the easiest.
> */
> folio_add_lru(dst);
> - if (page_was_mapped)
> + if (old_page_state & PAGE_WAS_MLOCKED)
> lru_add_drain();
>
> - if (page_was_mapped)
> + if (old_page_state & PAGE_WAS_MAPPED)
> remove_migration_ptes(src, dst, false);
>
> out_unlock_both:
> @@ -1322,11 +1331,12 @@ static int migrate_folio_move(free_folio_t put_new_folio, unsigned long private,
> */
> if (rc == -EAGAIN) {
> list_add(&dst->lru, prev);
> - __migrate_folio_record(dst, page_was_mapped, anon_vma);
> + __migrate_folio_record(dst, old_page_state, anon_vma);
> return rc;
> }
>
> - migrate_folio_undo_src(src, page_was_mapped, anon_vma, true, ret);
> + migrate_folio_undo_src(src, old_page_state & PAGE_WAS_MAPPED,
> + anon_vma, true, ret);
> migrate_folio_undo_dst(dst, true, put_new_folio, private);
>
> return rc;
> @@ -1799,12 +1809,12 @@ static int migrate_pages_batch(struct list_head *from,
> dst = list_first_entry(&dst_folios, struct folio, lru);
> dst2 = list_next_entry(dst, lru);
> list_for_each_entry_safe(folio, folio2, &unmap_folios, lru) {
> - int page_was_mapped = 0;
> + int old_page_state = 0;
> struct anon_vma *anon_vma = NULL;
>
> - __migrate_folio_extract(dst, &page_was_mapped, &anon_vma);
> - migrate_folio_undo_src(folio, page_was_mapped, anon_vma,
> - true, ret_folios);
> + __migrate_folio_extract(dst, &old_page_state, &anon_vma);
> + migrate_folio_undo_src(folio, old_page_state & PAGE_WAS_MAPPED,
> + anon_vma, true, ret_folios);
> list_del(&dst->lru);
> migrate_folio_undo_dst(dst, true, put_new_folio, private);
> dst = dst2;
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v3] mm: migrate: record the mlocked page status to remove unnecessary lru drain
2023-10-21 4:33 [PATCH v3] mm: migrate: record the mlocked page status to remove unnecessary lru drain Baolin Wang
2023-10-23 0:54 ` Huang, Ying
@ 2023-10-23 13:16 ` Zi Yan
1 sibling, 0 replies; 3+ messages in thread
From: Zi Yan @ 2023-10-23 13:16 UTC (permalink / raw)
To: Baolin Wang
Cc: akpm, mgorman, hughd, vbabka, ying.huang, fengwei.yin, linux-mm,
linux-kernel
[-- Attachment #1: Type: text/plain, Size: 3081 bytes --]
On 21 Oct 2023, at 0:33, Baolin Wang wrote:
> When doing compaction, I found the lru_add_drain() is an obvious hotspot
> when migrating pages. The distribution of this hotspot is as follows:
> - 18.75% compact_zone
> - 17.39% migrate_pages
> - 13.79% migrate_pages_batch
> - 11.66% migrate_folio_move
> - 7.02% lru_add_drain
> + 7.02% lru_add_drain_cpu
> + 3.00% move_to_new_folio
> 1.23% rmap_walk
> + 1.92% migrate_folio_unmap
> + 3.20% migrate_pages_sync
> + 0.90% isolate_migratepages
>
> The lru_add_drain() was added by commit c3096e6782b7 ("mm/migrate:
> __unmap_and_move() push good newpage to LRU") to drain the newpage to LRU
> immediately, to help to build up the correct newpage->mlock_count in
> remove_migration_ptes() for mlocked pages. However, if there are no mlocked
> pages are migrating, then we can avoid this lru drain operation, especailly
> for the heavy concurrent scenarios.
>
> So we can record the source pages' mlocked status in migrate_folio_unmap(),
> and only drain the lru list when the mlocked status is set in migrate_folio_move().
> In addition, the page was already isolated from lru when migrating, so checking
> the mlocked status is stable by folio_test_mlocked() in migrate_folio_unmap().
>
> After this patch, I can see the hotpot of the lru_add_drain() is gone:
> - 9.41% migrate_pages_batch
> - 6.15% migrate_folio_move
> - 3.64% move_to_new_folio
> + 1.80% migrate_folio_extra
> + 1.70% buffer_migrate_folio
> + 1.41% rmap_walk
> + 0.62% folio_add_lru
> + 3.07% migrate_folio_unmap
>
> Meanwhile, the compaction latency shows some improvements when running
> thpscale:
> base patched
> Amean fault-both-1 1131.22 ( 0.00%) 1112.55 * 1.65%*
> Amean fault-both-3 2489.75 ( 0.00%) 2324.15 * 6.65%*
> Amean fault-both-5 3257.37 ( 0.00%) 3183.18 * 2.28%*
> Amean fault-both-7 4257.99 ( 0.00%) 4079.04 * 4.20%*
> Amean fault-both-12 6614.02 ( 0.00%) 6075.60 * 8.14%*
> Amean fault-both-18 10607.78 ( 0.00%) 8978.86 * 15.36%*
> Amean fault-both-24 14911.65 ( 0.00%) 11619.55 * 22.08%*
> Amean fault-both-30 14954.67 ( 0.00%) 14925.66 * 0.19%*
> Amean fault-both-32 16654.87 ( 0.00%) 15580.31 * 6.45%*
>
> Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
> ---
> Changes from v2:
> - Use BIT() to define macros.
> - Simplify handing page_was_mapped and page_was_mlocked variables with
> introducing old_page_state variable.
>
> Changes from v1:
> - Use separate flags in __migrate_folio_record() to avoid to pack flags
> in each call site per Ying.
> ---
> mm/migrate.c | 48 +++++++++++++++++++++++++++++-------------------
> 1 file changed, 29 insertions(+), 19 deletions(-)
>
LGTM. Thanks. Reviewed-by: Zi Yan <ziy@nvidia.com>
--
Best Regards,
Yan, Zi
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 854 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2023-10-23 13:16 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-10-21 4:33 [PATCH v3] mm: migrate: record the mlocked page status to remove unnecessary lru drain Baolin Wang
2023-10-23 0:54 ` Huang, Ying
2023-10-23 13:16 ` Zi Yan
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox