[to-be-updated] mm-vma-fix-self-merge-check-in-copy_vma.patch removed from -mm tree

Andrew Morton <[email protected]>
Newsgroups org.kernel.vger.mm-commits
Message-ID <[email protected]>
The quilt patch titled
     Subject: mm/vma: fix self-merge check in copy_vma()
has been removed from the -mm tree.  Its filename was
     mm-vma-fix-self-merge-check-in-copy_vma.patch

This patch was dropped because an updated version will be issued

------------------------------------------------------
From: "Lorenzo Stoakes (ARM)" <[email protected]>
Subject: mm/vma: fix self-merge check in copy_vma()
Date: Thu, 06 Aug 2026 21:21:32 +0100

The existing logic is very confusing so improve things.  Firstly rename
the confusing faulted_in_anon_vma variable to can_self_merge and update
this when the page offset is updated.

What is being checked for is a 'self-merge' - that is between the VMA
being remapped and its prior VMA (remember that this is copy_vma() - if a
non-MREMAP_DONTUNMAP remap the original VMA is only removed afterwards).

This can happen if the VMA is moved immediately adjacent to itself, either
before or after it:

		|----------------|----------------|
		|		 |                |
		v		 |		  v
	|...............||---------------||...............|
	|      new      ||      old      ||      new      |
	|...............||---------------||---------------|

In these cases the old VMA is simply expanded to cover the new range.

It is also possible for the move to both self-merge and merge with a prior
VMA if it is placed between a preceding VMA and its old self:

				|---------------|
				|		|
				v		|
	|---------------||...............||---------------|
	|      prev     ||     new       ||     old       |
	|---------------||...............||---------------|

In this case, the old VMA is removed and 'prev' is expanded and replaces
it.

Since copy_vma_and_data() which calls copy_vma() intends to reference the
old VMA after the merge, it must have this pointer updated.

This kind of self-merge is not possible with a succeeding merge, as the
merge always prefers to expand the preceding VMA if possible.

copy_vma() accounts for this by explicitly checking to see if a self-merge
occurred and updating the vmap pointer if so.  However it incorrect did so
even for a subsequent merge (this is simply a noop so it had no impact).

So change this to only check for the case which matters - a backwards
merge - and rearrange the parameters to make it clearer we're doing that -
i.e.  check new_vma->vm_start < old_vma_start (having already renamed
vma_start to old_vma_start to make it clear this is the previous VMA).

Also update the existing wall-of-text comment to be a lot clearer.

While we're here, replace the VM_BUG_ON_VMA() with a VM_WARN_ON_ONCE_VMA()
and update the VMA userland tests accordingly.

No functional change intended.

Link: https://lore.kernel.org/[email protected]
Signed-off-by: Lorenzo Stoakes (ARM) <[email protected]>
Acked-by: David Hildenbrand (Arm) <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Alexander Deucher <[email protected]>
Cc: Alexander Gordeev <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Alistair Popple <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Arnd Bergmann <[email protected]>
Cc: Baolin Wang <[email protected]>
Cc: Baoquan He <[email protected]>
Cc: Barry Song <[email protected]>
Cc: Boris Brezillon <[email protected]>
Cc: Byungchul Park <[email protected]>
Cc: Chengming Zhou <[email protected]>
Cc: Chris Li <[email protected]>
Cc: Christan König <[email protected]>
Cc: Christian Borntraeger <[email protected]>
Cc: Claudio Imbrenda <[email protected]>
Cc: Dave Airlie <[email protected]>
Cc: Dev Jain <[email protected]>
Cc: Gerald Schaefer <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Gregory Price (Meta) <[email protected]>
Cc: Harry Yoo <[email protected]>
Cc: Heiko Carstens <[email protected]>
Cc: Huang Ray <[email protected]>
Cc: "Huang, Ying" <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jan Kara <[email protected]>
Cc: Jann Horn <[email protected]>
Cc: Janosch Frank <[email protected]>
Cc: Jason Gunthorpe <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: John Hubbard <[email protected]>
Cc: Joshua Hahn <[email protected]>
Cc: Kairui Song <[email protected]>
Cc: Kees Cook <[email protected]>
Cc: Kemeng Shi <[email protected]>
Cc: Lance Yang <[email protected]>
Cc: Liam R. Howlett <[email protected]>
Cc: Liviu Dudau <[email protected]>
Cc: Maarten Lankhorst <[email protected]>
Cc: Marc Rutland <[email protected]>
Cc: "Masami Hiramatsu (Google)" <[email protected]>
Cc: Matthew Auld <[email protected]>
Cc: Matthew Brost <[email protected]>
Cc: Matthew Wilcox (Oracle) <[email protected]>
Cc: Maxime Ripard <[email protected]>
Cc: Miaohe Lin <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Mike Rapoport <[email protected]>
Cc: Muchun Song <[email protected]>
Cc: Namhyung kim <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Nhat Pham <[email protected]>
Cc: Nico Pache <[email protected]>
Cc: Oleg Nesterov <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Pedro Falcato <[email protected]>
Cc: Peter Xu <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Rakie Kim <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Rodrigo Vivi <[email protected]>
Cc: Ryan Roberts <[email protected]>
Cc: Steven Price <[email protected]>
Cc: Suren Baghdasaryan <[email protected]>
Cc: Sven Schnelle <[email protected]>
Cc: <[email protected]>
Cc: Thomas Hellström <[email protected]>
Cc: Thomas Zimemrmann <[email protected]>
Cc: Vasily Gorbik <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Xu Xin <[email protected]>
Cc: Zi Yan <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
---

 mm/vma.c                         |   35 +++++++++++++----------------
 tools/testing/vma/vma_internal.h |    1 
 2 files changed, 17 insertions(+), 19 deletions(-)

--- a/mm/vma.c~mm-vma-fix-self-merge-check-in-copy_vma
+++ a/mm/vma.c
@@ -1911,10 +1911,10 @@ struct vm_area_struct *copy_vma(struct v
 	bool *need_rmap_locks)
 {
 	struct vm_area_struct *vma = *vmap;
-	unsigned long vma_start = vma->vm_start;
+	unsigned long old_vma_start = vma->vm_start;
 	struct mm_struct *mm = vma->vm_mm;
 	struct vm_area_struct *new_vma;
-	bool faulted_in_anon_vma = true;
+	bool can_self_merge = false;
 	VMA_ITERATOR(vmi, mm, addr);
 	VMG_VMA_STATE(vmg, &vmi, NULL, vma, addr, addr + len);
 
@@ -1924,7 +1924,7 @@ struct vm_area_struct *copy_vma(struct v
 	 */
 	if (unlikely(vma_is_anonymous(vma) && !vma->anon_vma)) {
 		pgoff = addr >> PAGE_SHIFT;
-		faulted_in_anon_vma = false;
+		can_self_merge = true;
 	}
 
 	/*
@@ -1944,24 +1944,21 @@ struct vm_area_struct *copy_vma(struct v
 	new_vma = vma_merge_copied_range(&vmg);
 
 	if (new_vma) {
-		/*
-		 * Source vma may have been merged into new_vma
-		 */
-		if (unlikely(vma_start >= new_vma->vm_start &&
-			     vma_start < new_vma->vm_end)) {
+		/* Self-merged and VMA replaced. */
+		if (unlikely(new_vma->vm_start < old_vma_start &&
+			     new_vma->vm_end > old_vma_start)) {
 			/*
-			 * The only way we can get a vma_merge with
-			 * self during an mremap is if the vma hasn't
-			 * been faulted in yet and we were allowed to
-			 * reset the dst vma->vm_pgoff to the
-			 * destination address of the mremap to allow
-			 * the merge to happen. mremap must change the
-			 * vm_pgoff linearity between src and dst vmas
-			 * (in turn preventing a vma_merge) to be
-			 * safe. It is only safe to keep the vm_pgoff
-			 * linear if there are no pages mapped yet.
+			 * The only way a VMA can both self-merge and be
+			 * replaced is if the remap places the new VMA
+			 * immediately prior to its old self ('next') and
+			 * immediately after another VMA ('prev') causing the
+			 * next to be removed and prev to be expanded to cover
+			 * the entire range.
+			 *
+			 * This should only be possible if the page offset was
+			 * updated, i.e. the VMA is unfaulted.
 			 */
-			VM_BUG_ON_VMA(faulted_in_anon_vma, new_vma);
+			VM_WARN_ON_ONCE_VMA(!can_self_merge, new_vma);
 			*vmap = vma = new_vma;
 		}
 		*need_rmap_locks =
--- a/tools/testing/vma/vma_internal.h~mm-vma-fix-self-merge-check-in-copy_vma
+++ a/tools/testing/vma/vma_internal.h
@@ -53,6 +53,7 @@ typedef __bitwise unsigned int vm_fault_
 
 #define VM_WARN_ON(_expr) (WARN_ON(_expr))
 #define VM_WARN_ON_ONCE(_expr) (WARN_ON_ONCE(_expr))
+#define VM_WARN_ON_ONCE_VMA(_expr, _vma) (WARN_ON_ONCE(_expr))
 #define VM_WARN_ON_VMG(_expr, _vmg) (WARN_ON(_expr))
 #define VM_BUG_ON(_expr) (BUG_ON(_expr))
 #define VM_BUG_ON_VMA(_expr, _vma) (BUG_ON(_expr))
_

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

tools-testing-vma-add-tests-for-copy_vma-self-merge.patch
mm-propagate-vma-anonymous-page-offset-on-map-remap-split-merge.patch
mm-rmap-track-whether-the-page-vma-mapped-pgoff-is-anonymous.patch
mm-clean-up-vma_address_end.patch
mm-huge_memory-update-remove_migration_pmd-to-accept-a-folio.patch
mm-migrate-calculate-large-folio-page-index-using-pfn.patch
mm-rmap-use-anon-pgoff-to-track-map_private-file-backed-anon-folios.patch
tools-testing-vma-expand-vma-merge-tests-to-assert-anon-pgoff.patch
tools-testing-selftests-mm-test-anonymous-page-offset-merge-behaviour.patch
mm-vma-only-permit-map_private-dev-zero-to-be-mapped-anonymous.patch
mm-vma-make-map_private-mapped-dev-zero-mappings-truly-anonymous.patch
tools-testing-vma-add-test-to-assert-map_private-dev-zero-is-anon.patch
tools-testing-selftests-mm-add-map_private-dev-zero-merge-tests.patch
mm-add-some-missing-includes-to-mm-local-headers.patch
maintainers-add-drivers-char-memc-to-mm-misc-memory-mapping-sections.patch
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.