mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Oscar Salvador <osalvador@suse.de>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: David Hildenbrand <david@kernel.org>,
	Michal Hocko <mhocko@suse.com>,
	Muchun Song <muchun.song@linux.dev>,
	Vlastimil Babka <vbabka@kernel.org>,
	Lorenzo Stoakes <lorenzo.stoakes@oracle.com>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	Oscar Salvador <osalvador@suse.de>
Subject: [RFC PATCH v3 8/8] mm: Make /proc/pid/clear_refs use the new generic pagewalk API
Date: Mon, 25 May 2026 18:55:28 +0200	[thread overview]
Message-ID: <20260525165528.184397-9-osalvador@suse.de> (raw)
In-Reply-To: <20260525165528.184397-1-osalvador@suse.de>

Have /proc/pid/clear_refs make use of the new generic API, and remove
the code which was using the old one.

Signed-off-by: Oscar Salvador <osalvador@suse.de>
---
 fs/proc/task_mmu.c | 133 ++++++++++++++++++++++++---------------------
 1 file changed, 70 insertions(+), 63 deletions(-)

diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 20ffb26692cc..5f09a5b26b61 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -1571,72 +1571,54 @@ static inline void clear_soft_dirty_pmd(struct vm_area_struct *vma,
 }
 #endif
 
-static int clear_refs_pte_range(pmd_t *pmd, unsigned long addr,
-				unsigned long end, struct mm_walk *walk)
+static void clear_refs_pmd_range(struct pt_range_walk *ptw,
+				 enum clear_refs_types type)
 {
-	struct clear_refs_private *cp = walk->private;
-	struct vm_area_struct *vma = walk->vma;
-	pte_t *pte, ptent;
-	spinlock_t *ptl;
-	struct folio *folio;
+	struct vm_area_struct *vma = ptw->vma;
+	unsigned long addr = ptw->curr_addr;
 
-	ptl = pmd_trans_huge_lock(pmd, vma);
-	if (ptl) {
-		if (cp->type == CLEAR_REFS_SOFT_DIRTY) {
-			clear_soft_dirty_pmd(vma, addr, pmd);
-			goto out;
-		}
-
-		if (!pmd_present(*pmd))
-			goto out;
+	if (type == CLEAR_REFS_SOFT_DIRTY) {
+		clear_soft_dirty_pmd(vma, addr, ptw->pmdp);
+		return;
+	}
 
-		folio = pmd_folio(*pmd);
+	if (!pmd_present(ptw->pmd) || !ptw->folio)
+		return;
 
-		/* Clear accessed and referenced bits. */
-		pmdp_test_and_clear_young(vma, addr, pmd);
-		folio_test_clear_young(folio);
-		folio_clear_referenced(folio);
-out:
-		spin_unlock(ptl);
-		return 0;
-	}
+	pmdp_test_and_clear_young(vma, addr, ptw->pmdp);
+	folio_test_clear_young(ptw->folio);
+	folio_clear_referenced(ptw->folio);
+}
 
-	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
-	if (!pte) {
-		walk->action = ACTION_AGAIN;
-		return 0;
-	}
-	for (; addr != end; pte++, addr += PAGE_SIZE) {
-		ptent = ptep_get(pte);
+static void clear_refs_pte_range(struct pt_range_walk *ptw,
+				enum clear_refs_types type)
+{
+	struct vm_area_struct *vma = ptw->vma;
+	unsigned long addr = ptw->curr_addr;
+	unsigned long end = pmd_addr_end(addr, vma->vm_end);
+	pte_t *ptep = ptw->ptep;
 
-		if (cp->type == CLEAR_REFS_SOFT_DIRTY) {
-			clear_soft_dirty(vma, addr, pte);
+	for (; addr != end; ptep++, addr += PAGE_SIZE) {
+		if (type == CLEAR_REFS_SOFT_DIRTY) {
+			clear_soft_dirty(vma, addr, ptep);
 			continue;
 		}
 
-		if (!pte_present(ptent))
-			continue;
-
-		folio = vm_normal_folio(vma, addr, ptent);
-		if (!folio)
+		if (!ptw->present || !ptw->folio)
 			continue;
 
 		/* Clear accessed and referenced bits. */
-		ptep_test_and_clear_young(vma, addr, pte);
-		folio_test_clear_young(folio);
-		folio_clear_referenced(folio);
+		ptep_test_and_clear_young(vma, addr, ptep);
+		folio_test_clear_young(ptw->folio);
+		folio_clear_referenced(ptw->folio);
 	}
-	pte_unmap_unlock(pte - 1, ptl);
-	cond_resched();
-	return 0;
+
+	ptw->next_addr = end;
 }
 
-static int clear_refs_test_walk(unsigned long start, unsigned long end,
-				struct mm_walk *walk)
+static int clear_refs_test_vma(struct vm_area_struct *vma,
+			       enum clear_refs_types type)
 {
-	struct clear_refs_private *cp = walk->private;
-	struct vm_area_struct *vma = walk->vma;
-
 	if (vma->vm_flags & VM_PFNMAP)
 		return 1;
 
@@ -1646,19 +1628,13 @@ static int clear_refs_test_walk(unsigned long start, unsigned long end,
 	 * Writing 3 to /proc/pid/clear_refs only affects file mapped pages.
 	 * Writing 4 to /proc/pid/clear_refs affects all pages.
 	 */
-	if (cp->type == CLEAR_REFS_ANON && vma->vm_file)
+	if (type == CLEAR_REFS_ANON && vma->vm_file)
 		return 1;
-	if (cp->type == CLEAR_REFS_MAPPED && !vma->vm_file)
+	if (type == CLEAR_REFS_MAPPED && !vma->vm_file)
 		return 1;
 	return 0;
 }
 
-static const struct mm_walk_ops clear_refs_walk_ops = {
-	.pmd_entry		= clear_refs_pte_range,
-	.test_walk		= clear_refs_test_walk,
-	.walk_lock		= PGWALK_WRLOCK,
-};
-
 static ssize_t clear_refs_write(struct file *file, const char __user *buf,
 				size_t count, loff_t *ppos)
 {
@@ -1688,9 +1664,6 @@ static ssize_t clear_refs_write(struct file *file, const char __user *buf,
 	if (mm) {
 		VMA_ITERATOR(vmi, mm, 0);
 		struct mmu_notifier_range range;
-		struct clear_refs_private cp = {
-			.type = type,
-		};
 
 		if (mmap_write_lock_killable(mm)) {
 			count = -EINTR;
@@ -1712,13 +1685,47 @@ static ssize_t clear_refs_write(struct file *file, const char __user *buf,
 				vm_flags_clear(vma, VM_SOFTDIRTY);
 				vma_set_page_prot(vma);
 			}
-
+			vma_iter_init(&vmi, mm, 0);
 			inc_tlb_flush_pending(mm);
 			mmu_notifier_range_init(&range, MMU_NOTIFY_SOFT_DIRTY,
 						0, mm, 0, -1UL);
 			mmu_notifier_invalidate_range_start(&range);
 		}
-		walk_page_range(mm, 0, -1, &clear_refs_walk_ops, &cp);
+
+		for_each_vma(vmi, vma) {
+			struct pt_range_walk ptw = {
+				.mm = mm,
+			};
+			enum pt_range_walk_type pt_type;
+			pt_type_flags_t flags = PT_TYPE_ALL;
+			int ret;
+
+			/* We ignore hugetlb vmas */
+			if (is_vm_hugetlb_page(vma))
+				continue;
+
+			ret = clear_refs_test_vma(vma, type);
+			if (ret > 0)
+				continue;
+			else if (ret < 0)
+				break;
+
+			pt_type = pt_range_walk_start(&ptw, vma, vma->vm_start, vma->vm_end, flags);
+			while (pt_type != PTW_DONE) {
+
+				/*
+				 * Since we ignore hugetlb vmas, we just care for PMD
+				 * or PTE mapped pages.
+				 */
+				if (ptw.level == PTW_PMD_LEVEL)
+					clear_refs_pmd_range(&ptw, type);
+				else if (ptw.level == PTW_PTE_LEVEL)
+					clear_refs_pte_range(&ptw, type);
+
+				pt_type = pt_range_walk_next(&ptw, vma, vma->vm_start, vma->vm_end, flags);
+			}
+			pt_range_walk_done(&ptw);
+		}
 		if (type == CLEAR_REFS_SOFT_DIRTY) {
 			mmu_notifier_invalidate_range_end(&range);
 			flush_tlb_mm(mm);
-- 
2.53.0


  parent reply	other threads:[~2026-05-25 16:56 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-25 16:55 [RFC PATCH v3 0/8] Implement a " Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 1/8] mm: Add softleaf_from_pud Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 2/8] mm: Add {pmd,pud}_huge_lock helper Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 3/8] mm: Implement folio_pmd_batch Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 4/8] mm: Implement pt_range_walk Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 5/8] mm: Make /proc/pid/smaps use the new generic pagewalk API Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 6/8] mm: Make /proc/pid/numa_maps " Oscar Salvador
2026-05-25 16:55 ` [RFC PATCH v3 7/8] mm: Make /proc/pid/pagemap " Oscar Salvador
2026-05-25 16:55 ` Oscar Salvador [this message]
2026-05-26  5:49 ` [syzbot ci] Re: Implement a " syzbot ci
2026-06-11 12:01   ` Oscar Salvador (SUSE)
2026-06-11 13:22     ` syzbot ci
2026-06-11 13:52   ` Oscar Salvador (SUSE)
2026-06-11 15:08     ` syzbot ci
2026-06-11 12:07 ` [RFC PATCH v3 0/8] " Lorenzo Stoakes
2026-06-11 12:29   ` Oscar Salvador (SUSE)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260525165528.184397-9-osalvador@suse.de \
    --to=osalvador@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=david@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=lorenzo.stoakes@oracle.com \
    --cc=mhocko@suse.com \
    --cc=muchun.song@linux.dev \
    --cc=vbabka@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®