The quilt patch titled
Subject: mm: avoid zeroing user movable page twice with init_on_alloc=1
has been removed from the -mm tree. Its filename was
mm-avoid-zeroing-user-movable-page-twice-with-init_on_alloc=1.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: Zi Yan <ziy@xxxxxxxxxx>
Subject: mm: avoid zeroing user movable page twice with init_on_alloc=1
Date: Fri, 11 Oct 2024 11:03:04 -0400
Commit 6471384af2a6 ("mm: security: introduce init_on_alloc=1 and
init_on_free=1 boot options") forces allocated page to be zeroed in
post_alloc_hook() when init_on_alloc=1.
For order-0 folios, if arch does not define
vma_alloc_zeroed_movable_folio(), the default implementation again zeros
the page return from the buddy allocator. So the page is zeroed twice.
Fix it by passing __GFP_ZERO instead to avoid double page zeroing. At the
moment, s390,arm64,x86,alpha,m68k are not impacted since they define their
own vma_alloc_zeroed_movable_folio().
For >0 order folios (mTHP and PMD THP), folio_zero_user() is called to
zero the folio again. Fix it by calling folio_zero_user() only if
init_on_alloc is set. All arch are impacted.
Add alloc_zeroed() helper to encapsulate the init_on_alloc check.
[ziy@xxxxxxxxxx: comment fixes, per David]
Link: https://lkml.kernel.org/r/97DB52E1-C594-49B5-9736-89AC302FAB01@xxxxxxxxxx
Link: https://lkml.kernel.org/r/20241011150304.709590-1-ziy@xxxxxxxxxx
Signed-off-by: Zi Yan <ziy@xxxxxxxxxx>
Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
Acked-by: David Hildenbrand <david@xxxxxxxxxx>
Cc: Alexander Potapenko <glider@xxxxxxxxxx>
Cc: "Huang, Ying" <ying.huang@xxxxxxxxx>
Cc: John Hubbard <jhubbard@xxxxxxxxxx>
Cc: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Kefeng Wang <wangkefeng.wang@xxxxxxxxxx>
Cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Cc: Miaohe Lin <linmiaohe@xxxxxxxxxx>
Cc: Ryan Roberts <ryan.roberts@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/highmem.h | 8 +-------
mm/huge_memory.c | 8 +++++++-
mm/internal.h | 6 ++++++
mm/memory.c | 10 +++++++++-
4 files changed, 23 insertions(+), 9 deletions(-)
--- a/include/linux/highmem.h~mm-avoid-zeroing-user-movable-page-twice-with-init_on_alloc=1
+++ a/include/linux/highmem.h
@@ -224,13 +224,7 @@ static inline
struct folio *vma_alloc_zeroed_movable_folio(struct vm_area_struct *vma,
unsigned long vaddr)
{
- struct folio *folio;
-
- folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0, vma, vaddr);
- if (folio)
- clear_user_highpage(&folio->page, vaddr);
-
- return folio;
+ return vma_alloc_folio(GFP_HIGHUSER_MOVABLE | __GFP_ZERO, 0, vma, vaddr);
}
#endif
--- a/mm/huge_memory.c~mm-avoid-zeroing-user-movable-page-twice-with-init_on_alloc=1
+++ a/mm/huge_memory.c
@@ -1165,7 +1165,13 @@ static struct folio *vma_alloc_anon_foli
}
folio_throttle_swaprate(folio, gfp);
- folio_zero_user(folio, addr);
+ /*
+ * When a folio is not zeroed during allocation (__GFP_ZERO not used),
+ * folio_zero_user() is used to make sure that the page corresponding
+ * to the faulting address will be hot in the cache after zeroing.
+ */
+ if (!alloc_zeroed())
+ folio_zero_user(folio, addr);
/*
* The memory barrier inside __folio_mark_uptodate makes sure that
* folio_zero_user writes become visible before the set_pmd_at()
--- a/mm/internal.h~mm-avoid-zeroing-user-movable-page-twice-with-init_on_alloc=1
+++ a/mm/internal.h
@@ -1276,6 +1276,12 @@ void touch_pud(struct vm_area_struct *vm
void touch_pmd(struct vm_area_struct *vma, unsigned long addr,
pmd_t *pmd, bool write);
+static inline bool alloc_zeroed(void)
+{
+ return static_branch_maybe(CONFIG_INIT_ON_ALLOC_DEFAULT_ON,
+ &init_on_alloc);
+}
+
enum {
/* mark page accessed */
FOLL_TOUCH = 1 << 16,
--- a/mm/memory.c~mm-avoid-zeroing-user-movable-page-twice-with-init_on_alloc=1
+++ a/mm/memory.c
@@ -4719,7 +4719,15 @@ static struct folio *alloc_anon_folio(st
goto next;
}
folio_throttle_swaprate(folio, gfp);
- folio_zero_user(folio, vmf->address);
+ /*
+ * When a folio is not zeroed during allocation
+ * (__GFP_ZERO not used), folio_zero_user() is used
+ * to make sure that the page corresponding to the
+ * faulting address will be hot in the cache after
+ * zeroing.
+ */
+ if (!alloc_zeroed())
+ folio_zero_user(folio, vmf->address);
return folio;
}
next:
_
Patches currently in -mm which might be from ziy@xxxxxxxxxx are
