Re: [PATCH v4 3/3] mm/vmscan: avoid split lazyfree THP during shrink_folio_list()

From: Baolin Wang
Date: Tue May 07 2024 - 05:34:15 EST




On 2024/5/7 16:26, Lance Yang wrote:
On Tue, May 7, 2024 at 2:32 PM Lance Yang <ioworker0@xxxxxxxxx> wrote:

Hey Baolin,

Thanks a lot for taking time to review!

On Tue, May 7, 2024 at 12:01 PM Baolin Wang
<baolin.wang@xxxxxxxxxxxxxxxxx> wrote:



On 2024/5/1 12:27, Lance Yang wrote:
When the user no longer requires the pages, they would use
madvise(MADV_FREE) to mark the pages as lazy free. Subsequently, they
typically would not re-write to that memory again.

During memory reclaim, if we detect that the large folio and its PMD are
both still marked as clean and there are no unexpected references
(such as GUP), so we can just discard the memory lazily, improving the
efficiency of memory reclamation in this case. On an Intel i5 CPU, reclaiming 1GiB of lazyfree THPs using
mem_cgroup_force_empty() results in the following runtimes in seconds
(shorter is better):

--------------------------------------------
| Old | New | Change |
--------------------------------------------
| 0.683426 | 0.049197 | -92.80% |
--------------------------------------------

Suggested-by: Zi Yan <ziy@xxxxxxxxxx>
Suggested-by: David Hildenbrand <david@xxxxxxxxxx>
Signed-off-by: Lance Yang <ioworker0@xxxxxxxxx>
---
include/linux/huge_mm.h | 9 +++++
mm/huge_memory.c | 73 +++++++++++++++++++++++++++++++++++++++++
mm/rmap.c | 3 ++
3 files changed, 85 insertions(+)

diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 38c4b5537715..017cee864080 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -411,6 +411,8 @@ static inline bool thp_migration_supported(void)

void split_huge_pmd_locked(struct vm_area_struct *vma, unsigned long address,
pmd_t *pmd, bool freeze, struct folio *folio);
+bool unmap_huge_pmd_locked(struct vm_area_struct *vma, unsigned long addr,
+ pmd_t *pmdp, struct folio *folio);

static inline void align_huge_pmd_range(struct vm_area_struct *vma,
unsigned long *start,
@@ -492,6 +494,13 @@ static inline void align_huge_pmd_range(struct vm_area_struct *vma,
unsigned long *start,
unsigned long *end) {}

+static inline bool unmap_huge_pmd_locked(struct vm_area_struct *vma,
+ unsigned long addr, pmd_t *pmdp,
+ struct folio *folio)
+{
+ return false;
+}
+
#define split_huge_pud(__vma, __pmd, __address) \
do { } while (0)

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 145505a1dd05..90fdef847a88 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2690,6 +2690,79 @@ static void unmap_folio(struct folio *folio)
try_to_unmap_flush();
}

+static bool __discard_trans_pmd_locked(struct vm_area_struct *vma,
+ unsigned long addr, pmd_t *pmdp,
+ struct folio *folio)
+{
+ struct mm_struct *mm = vma->vm_mm;
+ int ref_count, map_count;
+ pmd_t orig_pmd = *pmdp;
+ struct mmu_gather tlb;
+ struct page *page;
+
+ if (pmd_dirty(orig_pmd) || folio_test_dirty(folio))
+ return false;
+ if (unlikely(!pmd_present(orig_pmd) || !pmd_trans_huge(orig_pmd)))
+ return false;
+
+ page = pmd_page(orig_pmd);
+ if (unlikely(page_folio(page) != folio))
+ return false;
+
+ tlb_gather_mmu(&tlb, mm);
+ orig_pmd = pmdp_huge_get_and_clear(mm, addr, pmdp);
+ tlb_remove_pmd_tlb_entry(&tlb, pmdp, addr);
+
+ /*
+ * Syncing against concurrent GUP-fast:
+ * - clear PMD; barrier; read refcount
+ * - inc refcount; barrier; read PMD
+ */
+ smp_mb();
+
+ ref_count = folio_ref_count(folio);
+ map_count = folio_mapcount(folio);
+
+ /*
+ * Order reads for folio refcount and dirty flag
+ * (see comments in __remove_mapping()).
+ */
+ smp_rmb();
+
+ /*
+ * If the PMD or folio is redirtied at this point, or if there are
+ * unexpected references, we will give up to discard this folio
+ * and remap it.
+ *
+ * The only folio refs must be one from isolation plus the rmap(s).
+ */
+ if (ref_count != map_count + 1 || folio_test_dirty(folio) ||
+ pmd_dirty(orig_pmd)) {
+ set_pmd_at(mm, addr, pmdp, orig_pmd);
+ return false;
+ }
+
+ folio_remove_rmap_pmd(folio, page, vma);
+ zap_deposited_table(mm, pmdp);
+ add_mm_counter(mm, MM_ANONPAGES, -HPAGE_PMD_NR);
+ folio_put(folio);

IIUC, you missed handling mlock vma, see mlock_drain_local() in
try_to_unmap_one().

Good spot!

I suddenly realized that I overlooked another thing: If we detect that a
PMD-mapped THP is within the range of the VM_LOCKED VMA, we
should check whether the TTU_IGNORE_MLOCK flag is set in
try_to_unmap_one(). If the flag is set, we will remove the PMD mapping
from the folio. Otherwise, the folio should be mlocked, which avoids
splitting the folio and then mlocking each page again.

My previous response above is flawed - sorry :(

If we detect that a PMD-mapped THP is within the range of the
VM_LOCKED VMA.

1) If the TTU_IGNORE_MLOCK flag is set, we will try to remove the
PMD mapping from the folio, as this series has done.

Right.

2) If the flag is not set, the large folio should be mlocked to prevent it
from being picked during memory reclaim? Currently, we just leave it

Yes. From commit 1acbc3f93614 ("mm: handle large folio when large folio in VM_LOCKED VMA range"), large folios of the mlocked VMA will be handled during page reclaim phase.

as is and do not to mlock it, IIUC.

Original code already handle the mlock case after the PMD-mapped THP is split in try_to_unmap_one():
/*
* If the folio is in an mlock()d vma, we must not swap it out.
*/
if (!(flags & TTU_IGNORE_MLOCK) &&
(vma->vm_flags & VM_LOCKED)) {
/* Restore the mlock which got missed */
if (!folio_test_large(folio))
mlock_vma_folio(folio, vma);
page_vma_mapped_walk_done(&pvmw);
ret = false;
break;
}