1. Patchew
  2. linux
  3. View series

[PATCH v8 0/6] mm: Fix vmemmap optimization accounting and initialization

Muchun Song posted 6 patches 1 month, 2 weeks ago 
arch/arm64/mm/mmu.c            |  5 +--
arch/loongarch/mm/init.c       |  5 +--
arch/powerpc/mm/mem.c          |  5 +--
arch/riscv/mm/init.c           |  5 +--
arch/s390/mm/init.c            |  5 +--
arch/x86/mm/init_64.c          |  5 +--
include/linux/memory_hotplug.h |  8 +++--
mm/memory_hotplug.c            | 13 ++++----
mm/memremap.c                  |  4 +--
mm/mm_init.c                   | 47 ++++++++++++++++-----------
mm/sparse-vmemmap.c            | 58 ++++++++++++++++++++++++++--------
11 files changed, 105 insertions(+), 55 deletions(-)
[PATCH v8 0/6] mm: Fix vmemmap optimization accounting and initialization
Posted by Muchun Song 1 month, 2 weeks ago 
The series fixes several bugs in vmemmap optimization, mainly around
incorrect page accounting and memmap initialization in DAX and memory
hotplug paths. It also fixes pageblock migratetype initialization and
struct page initialization for ZONE_DEVICE compound pages.

Patches 1-4 fix vmemmap accounting issues. Patch 1 fixes an accounting
underflow in the section activation failure path by moving vmemmap page
accounting into the lower-level allocation and freeing helpers. Patch 2
fixes incorrect altmap passing in the memory hotplug error path. Patch 3
passes pgmap through memory deactivation paths so the teardown side can
determine whether vmemmap optimization was in effect. Patch 4 uses that
information to account the optimized DAX vmemmap size correctly.

Patches 5-6 fix initialization issues in mm/mm_init. One makes sure all
pageblocks in ZONE_DEVICE compound pages get their migratetype
initialized. The other fixes a case where DAX memory hotplug reuses an
unoptimized early-section memmap while compound_nr_pages() still assumes
vmemmap optimization, leaving tail struct pages uninitialized.

Changes in v8:
- In patch 4, move VM_WARN_ON_ONCE(nr_pages > PAGES_PER_SECTION) to the
  top of section_nr_vmemmap_pages().
- In patch 4, add Acked-by from David Hildenbrand.

Muchun Song (6):
  mm/sparse-vmemmap: Fix vmemmap accounting underflow
  mm/memory_hotplug: Fix incorrect altmap passing in error path
  mm/sparse-vmemmap: Pass @pgmap argument to memory deactivation paths
  mm/sparse-vmemmap: Fix DAX vmemmap accounting with optimization
  mm/mm_init: Fix pageblock migratetype for ZONE_DEVICE compound pages
  mm/mm_init: Fix uninitialized struct pages for ZONE_DEVICE

 arch/arm64/mm/mmu.c            |  5 +--
 arch/loongarch/mm/init.c       |  5 +--
 arch/powerpc/mm/mem.c          |  5 +--
 arch/riscv/mm/init.c           |  5 +--
 arch/s390/mm/init.c            |  5 +--
 arch/x86/mm/init_64.c          |  5 +--
 include/linux/memory_hotplug.h |  8 +++--
 mm/memory_hotplug.c            | 13 ++++----
 mm/memremap.c                  |  4 +--
 mm/mm_init.c                   | 47 ++++++++++++++++-----------
 mm/sparse-vmemmap.c            | 58 ++++++++++++++++++++++++++--------
 11 files changed, 105 insertions(+), 55 deletions(-)


base-commit: 39704f00f747aba3144289870b5fd8ac230a9aaf
-- 
2.20.1
Re: [PATCH v8 0/6] mm: Fix vmemmap optimization accounting and initialization
Posted by Liam R. Howlett 1 month, 2 weeks ago 
On 26/04/28 04:18PM, Muchun Song wrote:
> The series fixes several bugs in vmemmap optimization, mainly around
> incorrect page accounting and memmap initialization in DAX and memory
> hotplug paths. It also fixes pageblock migratetype initialization and
> struct page initialization for ZONE_DEVICE compound pages.
> 
> Patches 1-4 fix vmemmap accounting issues. Patch 1 fixes an accounting
> underflow in the section activation failure path by moving vmemmap page
> accounting into the lower-level allocation and freeing helpers. Patch 2
> fixes incorrect altmap passing in the memory hotplug error path. Patch 3
> passes pgmap through memory deactivation paths so the teardown side can
> determine whether vmemmap optimization was in effect. Patch 4 uses that
> information to account the optimized DAX vmemmap size correctly.
> 
> Patches 5-6 fix initialization issues in mm/mm_init. One makes sure all
> pageblocks in ZONE_DEVICE compound pages get their migratetype
> initialized. The other fixes a case where DAX memory hotplug reuses an
> unoptimized early-section memmap while compound_nr_pages() still assumes
> vmemmap optimization, leaving tail struct pages uninitialized.
> 
> Changes in v8:
> - In patch 4, move VM_WARN_ON_ONCE(nr_pages > PAGES_PER_SECTION) to the
>   top of section_nr_vmemmap_pages().
> - In patch 4, add Acked-by from David Hildenbrand.

Thanks!

Acked-by: Liam R. Howlett <liam@infradead.org>

> 
> Muchun Song (6):
>   mm/sparse-vmemmap: Fix vmemmap accounting underflow
>   mm/memory_hotplug: Fix incorrect altmap passing in error path
>   mm/sparse-vmemmap: Pass @pgmap argument to memory deactivation paths
>   mm/sparse-vmemmap: Fix DAX vmemmap accounting with optimization
>   mm/mm_init: Fix pageblock migratetype for ZONE_DEVICE compound pages
>   mm/mm_init: Fix uninitialized struct pages for ZONE_DEVICE
> 
>  arch/arm64/mm/mmu.c            |  5 +--
>  arch/loongarch/mm/init.c       |  5 +--
>  arch/powerpc/mm/mem.c          |  5 +--
>  arch/riscv/mm/init.c           |  5 +--
>  arch/s390/mm/init.c            |  5 +--
>  arch/x86/mm/init_64.c          |  5 +--
>  include/linux/memory_hotplug.h |  8 +++--
>  mm/memory_hotplug.c            | 13 ++++----
>  mm/memremap.c                  |  4 +--
>  mm/mm_init.c                   | 47 ++++++++++++++++-----------
>  mm/sparse-vmemmap.c            | 58 ++++++++++++++++++++++++++--------
>  11 files changed, 105 insertions(+), 55 deletions(-)
> 
> 
> base-commit: 39704f00f747aba3144289870b5fd8ac230a9aaf
> -- 
> 2.20.1
> 
>
Re: [PATCH v8 0/6] mm: Fix vmemmap optimization accounting and initialization
Posted by Andrew Morton 1 month, 2 weeks ago 
On Tue, 28 Apr 2026 16:18:49 +0800 Muchun Song <songmuchun@bytedance.com> wrote:

> The series fixes several bugs in vmemmap optimization, mainly around
> incorrect page accounting and memmap initialization in DAX and memory
> hotplug paths. It also fixes pageblock migratetype initialization and
> struct page initialization for ZONE_DEVICE compound pages.
> 
> Patches 1-4 fix vmemmap accounting issues. Patch 1 fixes an accounting
> underflow in the section activation failure path by moving vmemmap page
> accounting into the lower-level allocation and freeing helpers. Patch 2
> fixes incorrect altmap passing in the memory hotplug error path. Patch 3
> passes pgmap through memory deactivation paths so the teardown side can
> determine whether vmemmap optimization was in effect. Patch 4 uses that
> information to account the optimized DAX vmemmap size correctly.
> 
> Patches 5-6 fix initialization issues in mm/mm_init. One makes sure all
> pageblocks in ZONE_DEVICE compound pages get their migratetype
> initialized. The other fixes a case where DAX memory hotplug reuses an
> unoptimized early-section memmap while compound_nr_pages() still assumes
> vmemmap optimization, leaving tail struct pages uninitialized.

Thanks, I updated mm-new to this version.

> Changes in v8:
> - In patch 4, move VM_WARN_ON_ONCE(nr_pages > PAGES_PER_SECTION) to the
>   top of section_nr_vmemmap_pages().
> - In patch 4, add Acked-by from David Hildenbrand.

Here's how v8 altered mm.git:


--- a/mm/sparse-vmemmap.c~b
+++ a/mm/sparse-vmemmap.c
@@ -654,6 +654,7 @@ static int __meminit section_nr_vmemmap_
 	const unsigned long pages_per_compound = 1UL << order;
 
 	VM_WARN_ON_ONCE(!IS_ALIGNED(pfn | nr_pages, PAGES_PER_SUBSECTION));
+	VM_WARN_ON_ONCE(nr_pages > PAGES_PER_SECTION);
 
 	if (!vmemmap_can_optimize(altmap, pgmap))
 		return DIV_ROUND_UP(nr_pages * sizeof(struct page), PAGE_SIZE);
@@ -664,7 +665,6 @@ static int __meminit section_nr_vmemmap_
 	}
 
 	VM_WARN_ON_ONCE(!IS_ALIGNED(pfn | nr_pages, PAGES_PER_SECTION));
-	VM_WARN_ON_ONCE(nr_pages > PAGES_PER_SECTION);
 
 	if (IS_ALIGNED(pfn, pages_per_compound))
 		return VMEMMAP_RESERVE_NR;
_

Patchew 2.1-devel-ea86937

© 2016 - 2026 Red Hat, Inc.