mm: prepare page_referenced() and page_idle to new THP refcounting

Both page_referenced() and page_idle_clear_pte_refs_one() assume that
THP can only be mapped with PMD, so there's no reason to look on PTEs
for PageTransHuge() pages.  That's no true anymore: THP can be mapped
with PTEs too.

The patch removes PageTransHuge() test from the functions and opencode
page table check.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Vladimir Davydov <vdavydov@parallels.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 82 insertions, 35 deletions

diff --git a/mm/rmap.c b/mm/rmap.c
index 31d8866..6127c00 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -814,58 +814,105 @@ static int page_referenced_one(struct page *page, struct vm_area_struct *vma,
 	spinlock_t *ptl;
 	int referenced = 0;
 	struct page_referenced_arg *pra = arg;
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
 
-	if (unlikely(PageTransHuge(page))) {
-		pmd_t *pmd;
-
-		/*
-		 * rmap might return false positives; we must filter
-		 * these out using page_check_address_pmd().
-		 */
-		pmd = page_check_address_pmd(page, mm, address, &ptl);
-		if (!pmd)
+	if (unlikely(PageHuge(page))) {
+		/* when pud is not present, pte will be NULL */
+		pte = huge_pte_offset(mm, address);
+		if (!pte)
 			return SWAP_AGAIN;
 
-		if (vma->vm_flags & VM_LOCKED) {
+		ptl = huge_pte_lockptr(page_hstate(page), mm, pte);
+		goto check_pte;
+	}
+
+	pgd = pgd_offset(mm, address);
+	if (!pgd_present(*pgd))
+		return SWAP_AGAIN;
+	pud = pud_offset(pgd, address);
+	if (!pud_present(*pud))
+		return SWAP_AGAIN;
+	pmd = pmd_offset(pud, address);
+
+	if (pmd_trans_huge(*pmd)) {
+		int ret = SWAP_AGAIN;
+
+		ptl = pmd_lock(mm, pmd);
+		if (!pmd_present(*pmd))
+			goto unlock_pmd;
+		if (unlikely(!pmd_trans_huge(*pmd))) {
 			spin_unlock(ptl);
+			goto map_pte;
+		}
+
+		if (pmd_page(*pmd) != page)
+			goto unlock_pmd;
+
+		if (vma->vm_flags & VM_LOCKED) {
 			pra->vm_flags |= VM_LOCKED;
-			return SWAP_FAIL; /* To break the loop */
+			ret = SWAP_FAIL; /* To break the loop */
+			goto unlock_pmd;
 		}
 
 		if (pmdp_clear_flush_young_notify(vma, address, pmd))
 			referenced++;
 		spin_unlock(ptl);
+		goto found;
+unlock_pmd:
+		spin_unlock(ptl);
+		return ret;
 	} else {
-		pte_t *pte;
+		pmd_t pmde = *pmd;
 
-		/*
-		 * rmap might return false positives; we must filter
-		 * these out using page_check_address().
-		 */
-		pte = page_check_address(page, mm, address, &ptl, 0);
-		if (!pte)
+		barrier();
+		if (!pmd_present(pmde) || pmd_trans_huge(pmde))
 			return SWAP_AGAIN;
+	}
+map_pte:
+	pte = pte_offset_map(pmd, address);
+	if (!pte_present(*pte)) {
+		pte_unmap(pte);
+		return SWAP_AGAIN;
+	}
 
-		if (vma->vm_flags & VM_LOCKED) {
-			pte_unmap_unlock(pte, ptl);
-			pra->vm_flags |= VM_LOCKED;
-			return SWAP_FAIL; /* To break the loop */
-		}
+	ptl = pte_lockptr(mm, pmd);
+check_pte:
+	spin_lock(ptl);
 
-		if (ptep_clear_flush_young_notify(vma, address, pte)) {
-			/*
-			 * Don't treat a reference through a sequentially read
-			 * mapping as such.  If the page has been used in
-			 * another mapping, we will catch it; if this other
-			 * mapping is already gone, the unmap path will have
-			 * set PG_referenced or activated the page.
-			 */
-			if (likely(!(vma->vm_flags & VM_SEQ_READ)))
-				referenced++;
-		}
+	if (!pte_present(*pte)) {
+		pte_unmap_unlock(pte, ptl);
+		return SWAP_AGAIN;
+	}
+
+	/* THP can be referenced by any subpage */
+	if (pte_pfn(*pte) - page_to_pfn(page) >= hpage_nr_pages(page)) {
+		pte_unmap_unlock(pte, ptl);
+		return SWAP_AGAIN;
+	}
+
+	if (vma->vm_flags & VM_LOCKED) {
 		pte_unmap_unlock(pte, ptl);
+		pra->vm_flags |= VM_LOCKED;
+		return SWAP_FAIL; /* To break the loop */
 	}
 
+	if (ptep_clear_flush_young_notify(vma, address, pte)) {
+		/*
+		 * Don't treat a reference through a sequentially read
+		 * mapping as such.  If the page has been used in
+		 * another mapping, we will catch it; if this other
+		 * mapping is already gone, the unmap path will have
+		 * set PG_referenced or activated the page.
+		 */
+		if (likely(!(vma->vm_flags & VM_SEQ_READ)))
+			referenced++;
+	}
+	pte_unmap_unlock(pte, ptl);
+
+found:
 	if (referenced)
 		clear_page_idle(page);
 	if (test_and_clear_page_young(page))
@@ -912,7 +959,7 @@ int page_referenced(struct page *page,
 	int ret;
 	int we_locked = 0;
 	struct page_referenced_arg pra = {
-		.mapcount = page_mapcount(page),
+		.mapcount = total_mapcount(page),
 		.memcg = memcg,
 	};
 	struct rmap_walk_control rwc = {