Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:8 as permitted sender) client-ip=2620:137:e000::3:8;
From:   "Vishal Moola (Oracle)" <vishal.moola@gmail.com>
To:     linux-mm@kvack.org
Cc:     linux-kernel@vger.kernel.org, akpm@linux-foundation.org,
        shy828301@gmail.com,
        "Vishal Moola (Oracle)" <vishal.moola@gmail.com>
Subject: [PATCH 5/5] mm/khugepaged: Convert collapse_pte_mapped_thp() to use folios
Date:   Mon, 16 Oct 2023 13:05:10 -0700
Message-Id: <20231016200510.7387-6-vishal.moola@gmail.com>
In-Reply-To: <20231016200510.7387-1-vishal.moola@gmail.com>
References: <20231016200510.7387-1-vishal.moola@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

This removes 2 calls to compound_head() and helps convert khugepaged to
use folios throughout.

Previously, if the address passed to collapse_pte_mapped_thp()
corresponded to a tail page, the scan would fail immediately. Using
filemap_lock_folio() we can get the corresponding folio back and try to
operate on the folio instead.

Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
---
 mm/khugepaged.c | 45 ++++++++++++++++++++-------------------------
 1 file changed, 20 insertions(+), 25 deletions(-)

diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 70bba8ddea13..99c437979848 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -1471,7 +1471,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 	bool notified = false;
 	unsigned long haddr = addr & HPAGE_PMD_MASK;
 	struct vm_area_struct *vma = vma_lookup(mm, haddr);
-	struct page *hpage;
+	struct folio *folio;
 	pte_t *start_pte, *pte;
 	pmd_t *pmd, pgt_pmd;
 	spinlock_t *pml = NULL, *ptl;
@@ -1504,19 +1504,14 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 	if (userfaultfd_wp(vma))
 		return SCAN_PTE_UFFD_WP;
 
-	hpage = find_lock_page(vma->vm_file->f_mapping,
+	folio = filemap_lock_folio(vma->vm_file->f_mapping,
 			       linear_page_index(vma, haddr));
-	if (!hpage)
+	if (!folio)
 		return SCAN_PAGE_NULL;
 
-	if (!PageHead(hpage)) {
-		result = SCAN_FAIL;
-		goto drop_hpage;
-	}
-
-	if (compound_order(hpage) != HPAGE_PMD_ORDER) {
+	if (folio_order(folio) != HPAGE_PMD_ORDER) {
 		result = SCAN_PAGE_COMPOUND;
-		goto drop_hpage;
+		goto drop_folio;
 	}
 
 	result = find_pmd_or_thp_or_none(mm, haddr, &pmd);
@@ -1530,13 +1525,13 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 		 */
 		goto maybe_install_pmd;
 	default:
-		goto drop_hpage;
+		goto drop_folio;
 	}
 
 	result = SCAN_FAIL;
 	start_pte = pte_offset_map_lock(mm, pmd, haddr, &ptl);
 	if (!start_pte)		/* mmap_lock + page lock should prevent this */
-		goto drop_hpage;
+		goto drop_folio;
 
 	/* step 1: check all mapped PTEs are to the right huge page */
 	for (i = 0, addr = haddr, pte = start_pte;
@@ -1561,7 +1556,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 		 * Note that uprobe, debugger, or MAP_PRIVATE may change the
 		 * page table, but the new page will not be a subpage of hpage.
 		 */
-		if (hpage + i != page)
+		if (folio_page(folio, i) != page)
 			goto abort;
 	}
 
@@ -1576,7 +1571,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 	 * page_table_lock) ptl nests inside pml. The less time we hold pml,
 	 * the better; but userfaultfd's mfill_atomic_pte() on a private VMA
 	 * inserts a valid as-if-COWed PTE without even looking up page cache.
-	 * So page lock of hpage does not protect from it, so we must not drop
+	 * So page lock of folio does not protect from it, so we must not drop
 	 * ptl before pgt_pmd is removed, so uffd private needs pml taken now.
 	 */
 	if (userfaultfd_armed(vma) && !(vma->vm_flags & VM_SHARED))
@@ -1600,7 +1595,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 			continue;
 		/*
 		 * We dropped ptl after the first scan, to do the mmu_notifier:
-		 * page lock stops more PTEs of the hpage being faulted in, but
+		 * page lock stops more PTEs of the folio being faulted in, but
 		 * does not stop write faults COWing anon copies from existing
 		 * PTEs; and does not stop those being swapped out or migrated.
 		 */
@@ -1609,7 +1604,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 			goto abort;
 		}
 		page = vm_normal_page(vma, addr, ptent);
-		if (hpage + i != page)
+		if (folio_page(folio, i) != page)
 			goto abort;
 
 		/*
@@ -1628,8 +1623,8 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 
 	/* step 3: set proper refcount and mm_counters. */
 	if (nr_ptes) {
-		page_ref_sub(hpage, nr_ptes);
-		add_mm_counter(mm, mm_counter_file(hpage), -nr_ptes);
+		folio_ref_sub(folio, nr_ptes);
+		add_mm_counter(mm, mm_counter_file(&folio->page), -nr_ptes);
 	}
 
 	/* step 4: remove empty page table */
@@ -1653,14 +1648,14 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 maybe_install_pmd:
 	/* step 5: install pmd entry */
 	result = install_pmd
-			? set_huge_pmd(vma, haddr, pmd, hpage)
+			? set_huge_pmd(vma, haddr, pmd, &folio->page)
 			: SCAN_SUCCEED;
-	goto drop_hpage;
+	goto drop_folio;
 abort:
 	if (nr_ptes) {
 		flush_tlb_mm(mm);
-		page_ref_sub(hpage, nr_ptes);
-		add_mm_counter(mm, mm_counter_file(hpage), -nr_ptes);
+		folio_ref_sub(folio, nr_ptes);
+		add_mm_counter(mm, mm_counter_file(&folio->page), -nr_ptes);
 	}
 	if (start_pte)
 		pte_unmap_unlock(start_pte, ptl);
@@ -1668,9 +1663,9 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
 		spin_unlock(pml);
 	if (notified)
 		mmu_notifier_invalidate_range_end(&range);
-drop_hpage:
-	unlock_page(hpage);
-	put_page(hpage);
+drop_folio:
+	folio_unlock(folio);
+	folio_put(folio);
 	return result;
 }
 
-- 
2.40.1