From: Zi Yan <[email protected]>
PageTransHuge returns true for both thp and hugetlb, so thp stats was
counting both thp and hugetlb migrations. Exclude hugetlb migration by
setting is_thp variable right.
Fixes: 1a5bae25e3cf ("mm/vmstat: add events for THP migration without split")
Signed-off-by: Zi Yan <[email protected]>
Cc: Daniel Jordan <[email protected]>
Cc: Anshuman Khandual <[email protected]>
---
mm/migrate.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mm/migrate.c b/mm/migrate.c
index 941b89383cf3..d1ad964165e5 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1445,7 +1445,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
* Capture required information that might get lost
* during migration.
*/
- is_thp = PageTransHuge(page);
+ is_thp = PageTransHuge(page) && !PageHuge(page);
nr_subpages = thp_nr_pages(page);
cond_resched();
--
2.28.0
On Thu, Sep 17, 2020 at 04:27:29PM -0400, Zi Yan wrote:
> From: Zi Yan <[email protected]>
>
> PageTransHuge returns true for both thp and hugetlb, so thp stats was
> counting both thp and hugetlb migrations. Exclude hugetlb migration by
> setting is_thp variable right.
Yeah, shoot.
> Fixes: 1a5bae25e3cf ("mm/vmstat: add events for THP migration without split")
> Signed-off-by: Zi Yan <[email protected]>
Reviewed-by: Daniel Jordan <[email protected]>
If you wanted, you could also do this.
diff --git a/mm/migrate.c b/mm/migrate.c
index d1ad964165e5..6bc9559afc70 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1471,7 +1471,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
* we encounter them after the rest of the list
* is processed.
*/
- if (PageTransHuge(page) && !PageHuge(page)) {
+ if (is_thp) {
lock_page(page);
rc = split_huge_page_to_list(page, from);
unlock_page(page);
@@ -1480,8 +1480,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
nr_thp_split++;
goto retry;
}
- }
- if (is_thp) {
+
nr_thp_failed++;
nr_failed += nr_subpages;
goto out;
On 17 Sep 2020, at 16:59, Daniel Jordan wrote:
> On Thu, Sep 17, 2020 at 04:27:29PM -0400, Zi Yan wrote:
>> From: Zi Yan <[email protected]>
>>
>> PageTransHuge returns true for both thp and hugetlb, so thp stats was
>> counting both thp and hugetlb migrations. Exclude hugetlb migration by
>> setting is_thp variable right.
>
> Yeah, shoot.
>
>> Fixes: 1a5bae25e3cf ("mm/vmstat: add events for THP migration without split")
>> Signed-off-by: Zi Yan <[email protected]>
>
> Reviewed-by: Daniel Jordan <[email protected]>
Thanks.
>
> If you wanted, you could also do this.
>
> diff --git a/mm/migrate.c b/mm/migrate.c
> index d1ad964165e5..6bc9559afc70 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -1471,7 +1471,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> * we encounter them after the rest of the list
> * is processed.
> */
> - if (PageTransHuge(page) && !PageHuge(page)) {
> + if (is_thp) {
> lock_page(page);
> rc = split_huge_page_to_list(page, from);
> unlock_page(page);
> @@ -1480,8 +1480,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> nr_thp_split++;
> goto retry;
> }
> - }
> - if (is_thp) {
> +
> nr_thp_failed++;
> nr_failed += nr_subpages;
> goto out;
Yeah, makes sense to me. Let me send V2 to include this.
—
Best Regards,
Yan Zi