Skip to contentSkip to conversation

mm-madvise-batch-tlb-flushes-for-madv_dontneed.patch removed from -mm tree

merged mm-stable Andrew Morton 1 participant 1 message May 12, 2025

lore
Share & tools
lore

Andrew Morton

author tree notice Applied
#0 lore raw

The quilt patch titled
Subject: mm/madvise: batch tlb flushes for MADV_DONTNEED[_LOCKED]
has been removed from the -mm tree. Its filename was
mm-madvise-batch-tlb-flushes-for-madv_dontneed.patch

This patch was dropped because it was merged into the mm-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

------------------------------------------------------
From: SeongJae Park <[email protected]>
Subject: mm/madvise: batch tlb flushes for MADV_DONTNEED[_LOCKED]
Date: Wed, 9 Apr 2025 17:00:22 -0700

MADV_DONTNEED[_LOCKED] handling for [process_]madvise() flushes tlb for each vma of each address range. Update the logic to do tlb flushes in a batched way. Initialize an mmu_gather object from do_madvise() and vector_madvise(), which are the entry level functions for [process_]madvise(), respectively. And pass those objects to the function for per-vma work, via madvise_behavior struct. Make the per-vma logic not flushes tlb on their own but just saves the tlb entries to the received mmu_gather object. For this internal logic change, make zap_page_range_single_batched() non-static and use it directly from madvise_dontneed_single_vma(). Finally, the entry level functions flush the tlb entries that gathered for the entire user request, at once.

---

 mm/internal.h |    3 +++
 mm/madvise.c  |   11 ++++++++---
 mm/memory.c   |    4 ++--
 3 files changed, 13 insertions(+), 5 deletions(-)
mm/internal.h renamed from mm/internal.h~mm-madvise-batch-tlb-flushes-for-madv_dontneed +3 -0
@@ -430,6 +430,9 @@ void unmap_page_range(struct mmu_gather
 			     struct vm_area_struct *vma,
 			     unsigned long addr, unsigned long end,
 			     struct zap_details *details);
+void zap_page_range_single_batched(struct mmu_gather *tlb,
+		struct vm_area_struct *vma, unsigned long addr,
+		unsigned long size, struct zap_details *details);
 int folio_unmap_invalidate(struct address_space *mapping, struct folio *folio,
 			   gfp_t gfp);
 
mm/madvise.c renamed from mm/madvise.c~mm-madvise-batch-tlb-flushes-for-madv_dontneed +8 -3
@@ -851,7 +851,8 @@ static int madvise_free_single_vma(struc
  * An interface that causes the system to free clean pages and flush
  * dirty pages is already available as msync(MS_INVALIDATE).
  */
-static long madvise_dontneed_single_vma(struct vm_area_struct *vma,
+static long madvise_dontneed_single_vma(struct madvise_behavior *madv_behavior,
+					struct vm_area_struct *vma,
 					unsigned long start, unsigned long end)
 {
 	struct zap_details details = {
@@ -859,7 +860,8 @@ static long madvise_dontneed_single_vma(
 		.even_cows = true,
 	};
 
-	zap_page_range_single(vma, start, end - start, &details);
+	zap_page_range_single_batched(
+			madv_behavior->tlb, vma, start, end - start, &details);
 	return 0;
 }
 
@@ -950,7 +952,8 @@ static long madvise_dontneed_free(struct
 	}
 
 	if (behavior == MADV_DONTNEED || behavior == MADV_DONTNEED_LOCKED)
-		return madvise_dontneed_single_vma(vma, start, end);
+		return madvise_dontneed_single_vma(
+				madv_behavior, vma, start, end);
 	else if (behavior == MADV_FREE)
 		return madvise_free_single_vma(madv_behavior, vma, start, end);
 	else
@@ -1628,6 +1631,8 @@ static void madvise_unlock(struct mm_str
 static bool madvise_batch_tlb_flush(int behavior)
 {
 	switch (behavior) {
+	case MADV_DONTNEED:
+	case MADV_DONTNEED_LOCKED:
 	case MADV_FREE:
 		return true;
 	default:
mm/memory.c renamed from mm/memory.c~mm-madvise-batch-tlb-flushes-for-madv_dontneed +2 -2
@@ -1998,7 +1998,7 @@ void unmap_vmas(struct mmu_gather *tlb,
 	mmu_notifier_invalidate_range_end(&range);
 }
 
-/*
+/**
  * zap_page_range_single_batched - remove user pages in a given range
  * @tlb: pointer to the caller's struct mmu_gather
  * @vma: vm_area_struct holding the applicable pages
@@ -2009,7 +2009,7 @@ void unmap_vmas(struct mmu_gather *tlb,
  * @tlb shouldn't be NULL.  The range must fit into one VMA.  If @vma is for
  * hugetlb, @tlb is flushed and re-initialized by this function.
  */
-static void zap_page_range_single_batched(struct mmu_gather *tlb,
+void zap_page_range_single_batched(struct mmu_gather *tlb,
 		struct vm_area_struct *vma, unsigned long address,
 		unsigned long size, struct zap_details *details)
 {

_

Patches currently in -mm which might be from [email protected] are

mm-damon-core-introduce-damos-quota-goal-metrics-for-memory-node-utilization.patch mm-damon-sysfs-schemes-implement-file-for-quota-goal-nid-parameter.patch mm-damon-sysfs-schemes-connect-damos_quota_goal-nid-with-core-layer.patch docs-mm-damon-design-document-node_mem_usedfree_bp.patch docs-admin-guide-mm-damon-usage-document-nid-file.patch docs-abi-damon-document-nid-file.patch samples-damon-implement-a-damon-module-for-memory-tiering.patch