[PATCHv4 2/8] khugepaged: Do not stop collapse if less than half PTEs are referenced

From: Kirill A. Shutemov
Date: Thu Apr 16 2020 - 12:01:58 EST


__collapse_huge_page_swapin() checks the number of referenced PTE to
decide if the memory range is hot enough to justify swapin.

We have few problems with the approach:

- It is way too late: we can do the check much earlier and safe time.
khugepaged_scan_pmd() already knows if we have any pages to swap in
and number of referenced page.

- It stops collapse altogether if there's not enough referenced pages,
not only swappingin.

Fix it by making the right check early. We also can avoid additional
page table scanning if khugepaged_scan_pmd() haven't found any swap
entries.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
Fixes: 0db501f7a34c ("mm, thp: convert from optimistic swapin collapsing to conservative")
Reviewed-by: William Kucharski <william.kucharski@xxxxxxxxxx>
Reviewed-and-Tested-by: Zi Yan <ziy@xxxxxxxxxx>
Acked-by: Yang Shi <yang.shi@xxxxxxxxxxxxxxxxx>
---
mm/khugepaged.c | 27 +++++++++++----------------
1 file changed, 11 insertions(+), 16 deletions(-)

diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 99d77ffb79c2..6e69dc9a9fb1 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -899,11 +899,6 @@ static bool __collapse_huge_page_swapin(struct mm_struct *mm,
.pgoff = linear_page_index(vma, address),
};

- /* we only decide to swapin, if there is enough young ptes */
- if (referenced < HPAGE_PMD_NR/2) {
- trace_mm_collapse_huge_page_swapin(mm, swapped_in, referenced, 0);
- return false;
- }
vmf.pte = pte_offset_map(pmd, address);
for (; vmf.address < address + HPAGE_PMD_NR*PAGE_SIZE;
vmf.pte++, vmf.address += PAGE_SIZE) {
@@ -943,7 +938,7 @@ static bool __collapse_huge_page_swapin(struct mm_struct *mm,
static void collapse_huge_page(struct mm_struct *mm,
unsigned long address,
struct page **hpage,
- int node, int referenced)
+ int node, int referenced, int unmapped)
{
pmd_t *pmd, _pmd;
pte_t *pte;
@@ -1000,7 +995,8 @@ static void collapse_huge_page(struct mm_struct *mm,
* If it fails, we release mmap_sem and jump out_nolock.
* Continuing to collapse causes inconsistency.
*/
- if (!__collapse_huge_page_swapin(mm, vma, address, pmd, referenced)) {
+ if (unmapped && !__collapse_huge_page_swapin(mm, vma, address,
+ pmd, referenced)) {
mem_cgroup_cancel_charge(new_page, memcg, true);
up_read(&mm->mmap_sem);
goto out_nolock;
@@ -1233,22 +1229,21 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
mmu_notifier_test_young(vma->vm_mm, address))
referenced++;
}
- if (writable) {
- if (referenced) {
- result = SCAN_SUCCEED;
- ret = 1;
- } else {
- result = SCAN_LACK_REFERENCED_PAGE;
- }
- } else {
+ if (!writable) {
result = SCAN_PAGE_RO;
+ } else if (!referenced || (unmapped && referenced < HPAGE_PMD_NR/2)) {
+ result = SCAN_LACK_REFERENCED_PAGE;
+ } else {
+ result = SCAN_SUCCEED;
+ ret = 1;
}
out_unmap:
pte_unmap_unlock(pte, ptl);
if (ret) {
node = khugepaged_find_target_node();
/* collapse_huge_page will return with the mmap_sem released */
- collapse_huge_page(mm, address, hpage, node, referenced);
+ collapse_huge_page(mm, address, hpage, node,
+ referenced, unmapped);
}
out:
trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
--
2.26.1