[RFC 0/4] add static huge zero folio support

[RFC 0/4] add static huge zero folio support

[Pankaj Raghav (Samsung)]

From: Pankaj Raghav <p.raghav@samsung.com>

NOTE: I am resending as an RFC again based on Lorenzo's feedback. The
old series can be found here [1].

There are many places in the kernel where we need to zeroout larger
chunks but the maximum segment we can zeroout at a time by ZERO_PAGE
is limited by PAGE_SIZE.

This concern was raised during the review of adding Large Block Size support
to XFS[2][3].

This is especially annoying in block devices and filesystems where we
attach multiple ZERO_PAGEs to the bio in different bvecs. With multipage
bvec support in block layer, it is much more efficient to send out
larger zero pages as a part of a single bvec.

Some examples of places in the kernel where this could be useful:
- blkdev_issue_zero_pages()
- iomap_dio_zero()
- vmalloc.c:zero_iter()
- rxperf_process_call()
- fscrypt_zeroout_range_inline_crypt()
- bch2_checksum_update()
...

Usually huge_zero_folio is allocated on demand, and it will be
deallocated by the shrinker if there are no users of it left. At the moment,
huge_zero_folio infrastructure refcount is tied to the process lifetime
that created it. This might not work for bio layer as the completions
can be async and the process that created the huge_zero_folio might no
longer be alive. And, one of the main point that came during discussion
is to have something bigger than zero page as a drop-in replacement.

Add a config option STATIC_HUGE_ZERO_FOLIO that will always allocate
the huge_zero_folio, and it will never drop the reference. This makes
using the huge_zero_folio without having to pass any mm struct and does
not tie the lifetime of the zero folio to anything, making it a drop-in
replacement for ZERO_PAGE.

I have converted blkdev_issue_zero_pages() as an example as a part of
this series. I also noticed close to 4% performance improvement just by
replacing ZERO_PAGE with static huge_zero_folio.

I will send patches to individual subsystems using the huge_zero_folio
once this gets upstreamed.

Looking forward to some feedback.

[1] https://lore.kernel.org/linux-mm/20250707142319.319642-1-kernel@pankajraghav.com/
[2] https://lore.kernel.org/linux-xfs/20231027051847.GA7885@lst.de/
[3] https://lore.kernel.org/linux-xfs/ZitIK5OnR7ZNY0IG@infradead.org/

Changes since last series[1]:
- Instead of allocating a new page through memblock, use the same
  infrastructure as huge_zero_folio but raise the reference and never
  drop it. (David)
- And some minor cleanups based on Lorenzo's feedback.

Pankaj Raghav (4):
  mm: rename huge_zero_page_shrinker to huge_zero_folio_shrinker
  mm: add static huge zero folio
  mm: add largest_zero_folio() routine
  block: use largest_zero_folio in __blkdev_issue_zero_pages()

 arch/x86/Kconfig        |  1 +
 block/blk-lib.c         | 15 ++++++------
 include/linux/huge_mm.h | 24 +++++++++++++++++++
 mm/Kconfig              | 12 ++++++++++
 mm/huge_memory.c        | 52 +++++++++++++++++++++++++++++++----------
 5 files changed, 85 insertions(+), 19 deletions(-)


base-commit: 1b0686bd18c1aa9d7f01943829faa5befe6ab3d1
-- 
2.49.0

[RFC 1/4] mm: rename huge_zero_page_shrinker to huge_zero_folio_shrinker

[Pankaj Raghav (Samsung)]

From: Pankaj Raghav <p.raghav@samsung.com>

As we already moved from exposing huge_zero_page to huge_zero_folio,
change the name of the shrinker to reflect that.

No functional changes.

Suggested-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 mm/huge_memory.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 2b4ea5a2ce7d..5d8365d1d3e9 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -266,15 +266,15 @@ void mm_put_huge_zero_folio(struct mm_struct *mm)
 		put_huge_zero_page();
 }
 
-static unsigned long shrink_huge_zero_page_count(struct shrinker *shrink,
-					struct shrink_control *sc)
+static unsigned long shrink_huge_zero_folio_count(struct shrinker *shrink,
+						  struct shrink_control *sc)
 {
 	/* we can free zero page only if last reference remains */
 	return atomic_read(&huge_zero_refcount) == 1 ? HPAGE_PMD_NR : 0;
 }
 
-static unsigned long shrink_huge_zero_page_scan(struct shrinker *shrink,
-				       struct shrink_control *sc)
+static unsigned long shrink_huge_zero_folio_scan(struct shrinker *shrink,
+						 struct shrink_control *sc)
 {
 	if (atomic_cmpxchg(&huge_zero_refcount, 1, 0) == 1) {
 		struct folio *zero_folio = xchg(&huge_zero_folio, NULL);
@@ -287,7 +287,7 @@ static unsigned long shrink_huge_zero_page_scan(struct shrinker *shrink,
 	return 0;
 }
 
-static struct shrinker *huge_zero_page_shrinker;
+static struct shrinker *huge_zero_folio_shrinker;
 
 #ifdef CONFIG_SYSFS
 static ssize_t enabled_show(struct kobject *kobj,
@@ -849,8 +849,8 @@ static inline void hugepage_exit_sysfs(struct kobject *hugepage_kobj)
 
 static int __init thp_shrinker_init(void)
 {
-	huge_zero_page_shrinker = shrinker_alloc(0, "thp-zero");
-	if (!huge_zero_page_shrinker)
+	huge_zero_folio_shrinker = shrinker_alloc(0, "thp-zero");
+	if (!huge_zero_folio_shrinker)
 		return -ENOMEM;
 
 	deferred_split_shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE |
@@ -858,13 +858,13 @@ static int __init thp_shrinker_init(void)
 						 SHRINKER_NONSLAB,
 						 "thp-deferred_split");
 	if (!deferred_split_shrinker) {
-		shrinker_free(huge_zero_page_shrinker);
+		shrinker_free(huge_zero_folio_shrinker);
 		return -ENOMEM;
 	}
 
-	huge_zero_page_shrinker->count_objects = shrink_huge_zero_page_count;
-	huge_zero_page_shrinker->scan_objects = shrink_huge_zero_page_scan;
-	shrinker_register(huge_zero_page_shrinker);
+	huge_zero_folio_shrinker->count_objects = shrink_huge_zero_folio_count;
+	huge_zero_folio_shrinker->scan_objects = shrink_huge_zero_folio_scan;
+	shrinker_register(huge_zero_folio_shrinker);
 
 	deferred_split_shrinker->count_objects = deferred_split_count;
 	deferred_split_shrinker->scan_objects = deferred_split_scan;
@@ -875,7 +875,7 @@ static int __init thp_shrinker_init(void)
 
 static void __init thp_shrinker_exit(void)
 {
-	shrinker_free(huge_zero_page_shrinker);
+	shrinker_free(huge_zero_folio_shrinker);
 	shrinker_free(deferred_split_shrinker);
 }
 
-- 
2.49.0

[RFC 2/4] mm: add static huge zero folio

[Pankaj Raghav (Samsung)]

From: Pankaj Raghav <p.raghav@samsung.com>

There are many places in the kernel where we need to zeroout larger
chunks but the maximum segment we can zeroout at a time by ZERO_PAGE
is limited by PAGE_SIZE.

This is especially annoying in block devices and filesystems where we
attach multiple ZERO_PAGEs to the bio in different bvecs. With multipage
bvec support in block layer, it is much more efficient to send out
larger zero pages as a part of single bvec.

This concern was raised during the review of adding LBS support to
XFS[1][2].

Usually huge_zero_folio is allocated on demand, and it will be
deallocated by the shrinker if there are no users of it left. At moment,
huge_zero_folio infrastructure refcount is tied to the process lifetime
that created it. This might not work for bio layer as the completions
can be async and the process that created the huge_zero_folio might no
longer be alive. And, one of the main point that came during discussion
is to have something bigger than zero page as a drop-in replacement.

Add a config option STATIC_HUGE_ZERO_FOLIO that will always allocate
the huge_zero_folio, and it will never drop the reference. This makes
using the huge_zero_folio without having to pass any mm struct and does
not tie the lifetime of the zero folio to anything, making it a drop-in
replacement for ZERO_PAGE.

If STATIC_PMD_ZERO_PAGE config option is enabled, then
mm_get_huge_zero_folio() will simply return this page instead of
dynamically allocating a new PMD page.

This option can waste memory in small systems or systems with 64k base
page size. So make it an opt-in and also add an option from individual
architecture so that we don't enable this feature for larger base page
size systems.

[1] https://lore.kernel.org/linux-xfs/20231027051847.GA7885@lst.de/
[2] https://lore.kernel.org/linux-xfs/ZitIK5OnR7ZNY0IG@infradead.org/

Co-Developed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 arch/x86/Kconfig        |  1 +
 include/linux/huge_mm.h | 16 ++++++++++++++++
 mm/Kconfig              | 12 ++++++++++++
 mm/huge_memory.c        | 28 ++++++++++++++++++++++++++++
 4 files changed, 57 insertions(+)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 0ce86e14ab5e..8e2aa1887309 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -153,6 +153,7 @@ config X86
 	select ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP	if X86_64
 	select ARCH_WANT_HUGETLB_VMEMMAP_PREINIT if X86_64
 	select ARCH_WANTS_THP_SWAP		if X86_64
+	select ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO if X86_64
 	select ARCH_HAS_PARANOID_L1D_FLUSH
 	select ARCH_WANT_IRQS_OFF_ACTIVATE_MM
 	select BUILDTIME_TABLE_SORT
diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 7748489fde1b..0ddd9c78f9f4 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -476,6 +476,7 @@ vm_fault_t do_huge_pmd_numa_page(struct vm_fault *vmf);
 
 extern struct folio *huge_zero_folio;
 extern unsigned long huge_zero_pfn;
+extern atomic_t huge_zero_folio_is_static;
 
 static inline bool is_huge_zero_folio(const struct folio *folio)
 {
@@ -494,6 +495,16 @@ static inline bool is_huge_zero_pmd(pmd_t pmd)
 
 struct folio *mm_get_huge_zero_folio(struct mm_struct *mm);
 void mm_put_huge_zero_folio(struct mm_struct *mm);
+struct folio *__get_static_huge_zero_folio(void);
+
+static inline struct folio *get_static_huge_zero_folio(void)
+{
+       if (!IS_ENABLED(CONFIG_STATIC_HUGE_ZERO_FOLIO))
+               return NULL;
+       if (likely(atomic_read(&huge_zero_folio_is_static)))
+               return huge_zero_folio;
+       return __get_static_huge_zero_folio();
+}
 
 static inline bool thp_migration_supported(void)
 {
@@ -685,6 +696,11 @@ static inline int change_huge_pud(struct mmu_gather *tlb,
 {
 	return 0;
 }
+
+static inline struct folio *get_static_huge_zero_folio(void)
+{
+       return NULL;
+}
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
 static inline int split_folio_to_list_to_order(struct folio *folio,
diff --git a/mm/Kconfig b/mm/Kconfig
index 0287e8d94aea..14721171846f 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -835,6 +835,18 @@ config ARCH_WANT_GENERAL_HUGETLB
 config ARCH_WANTS_THP_SWAP
 	def_bool n
 
+config ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO
+	def_bool n
+
+config STATIC_HUGE_ZERO_FOLIO
+	bool "Allocate a PMD sized folio for zeroing"
+	depends on ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO
+	help
+	  Typically huge_zero_folio, which is a PMD page of zeroes, is allocated
+	  on demand and deallocated when not in use. This option will
+	  allocate huge_zero_folio but it will never free it.
+	  Not suitable for memory constrained systems.
+
 config MM_ID
 	def_bool n
 
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 5d8365d1d3e9..6c890a1482f3 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -75,6 +75,8 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
 static bool split_underused_thp = true;
 
 static atomic_t huge_zero_refcount;
+static atomic_t huge_zero_static_fail_count __read_mostly;
+atomic_t huge_zero_folio_is_static __read_mostly;
 struct folio *huge_zero_folio __read_mostly;
 unsigned long huge_zero_pfn __read_mostly = ~0UL;
 unsigned long huge_anon_orders_always __read_mostly;
@@ -266,6 +268,32 @@ void mm_put_huge_zero_folio(struct mm_struct *mm)
 		put_huge_zero_page();
 }
 
+#ifdef CONFIG_STATIC_HUGE_ZERO_FOLIO
+struct folio *__get_static_huge_zero_folio(void)
+{
+       /*
+        * If we failed to allocate a huge zero folio multiple times,
+        * just refrain from trying.
+        */
+       if (atomic_read(&huge_zero_static_fail_count) > 2)
+               return NULL;
+
+       /*
+        * Our raised reference will prevent the shrinker from ever having
+        * success.
+        */
+       if (!get_huge_zero_page()) {
+              atomic_inc(&huge_zero_static_fail_count);
+               return NULL;
+       }
+
+       if (atomic_cmpxchg(&huge_zero_folio_is_static, 0, 1) != 0)
+               put_huge_zero_page();
+
+       return huge_zero_folio;
+}
+#endif /* CONFIG_STATIC_HUGE_ZERO_FOLIO */
+
 static unsigned long shrink_huge_zero_folio_count(struct shrinker *shrink,
 						  struct shrink_control *sc)
 {
-- 
2.49.0

[RFC 3/4] mm: add largest_zero_folio() routine

[Pankaj Raghav (Samsung)]

From: Pankaj Raghav <p.raghav@samsung.com>

Add largest_zero_folio() routine so that huge_zero_folio can be
used directly when CONFIG_STATIC_HUGE_ZERO_FOLIO is enabled. This will
return ZERO_PAGE folio if CONFIG_STATIC_HUGE_ZERO_FOLIO is disabled or
if we failed to allocate a huge_zero_folio.

Co-Developed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 include/linux/huge_mm.h | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 0ddd9c78f9f4..a34c5427aaf6 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -714,4 +714,12 @@ static inline int split_folio_to_order(struct folio *folio, int new_order)
 	return split_folio_to_list_to_order(folio, NULL, new_order);
 }
 
+static inline struct folio *largest_zero_folio(void)
+{
+       struct folio *folio = get_static_huge_zero_folio();
+
+       if (folio)
+               return folio;
+       return page_folio(ZERO_PAGE(0));
+}
 #endif /* _LINUX_HUGE_MM_H */
-- 
2.49.0

[RFC 4/4] block: use largest_zero_folio in __blkdev_issue_zero_pages()

[Pankaj Raghav (Samsung)]

From: Pankaj Raghav <p.raghav@samsung.com>

Use largest_zero_folio() in __blkdev_issue_zero_pages().
On systems with CONFIG_STATIC_HUGE_ZERO_FOLIO enabled, we will end up
sending larger bvecs instead of multiple small ones.

Noticed a 4% increase in performance on a commercial NVMe SSD which does
not support OP_WRITE_ZEROES. The device's MDTS was 128K. The performance
gains might be bigger if the device supports bigger MDTS.

Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 block/blk-lib.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/block/blk-lib.c b/block/blk-lib.c
index 4c9f20a689f7..3030a772d3aa 100644
--- a/block/blk-lib.c
+++ b/block/blk-lib.c
@@ -196,6 +196,8 @@ static void __blkdev_issue_zero_pages(struct block_device *bdev,
 		sector_t sector, sector_t nr_sects, gfp_t gfp_mask,
 		struct bio **biop, unsigned int flags)
 {
+	struct folio *zero_folio = largest_zero_folio();
+
 	while (nr_sects) {
 		unsigned int nr_vecs = __blkdev_sectors_to_bio_pages(nr_sects);
 		struct bio *bio;
@@ -208,15 +210,14 @@ static void __blkdev_issue_zero_pages(struct block_device *bdev,
 			break;
 
 		do {
-			unsigned int len, added;
+			unsigned int len;
 
-			len = min_t(sector_t,
-				PAGE_SIZE, nr_sects << SECTOR_SHIFT);
-			added = bio_add_page(bio, ZERO_PAGE(0), len, 0);
-			if (added < len)
+			len = min_t(sector_t, folio_size(zero_folio),
+				    nr_sects << SECTOR_SHIFT);
+			if (!bio_add_folio(bio, zero_folio, len, 0))
 				break;
-			nr_sects -= added >> SECTOR_SHIFT;
-			sector += added >> SECTOR_SHIFT;
+			nr_sects -= len >> SECTOR_SHIFT;
+			sector += len >> SECTOR_SHIFT;
 		} while (nr_sects);
 
 		*biop = bio_chain_and_submit(*biop, bio);
-- 
2.49.0

Re: [RFC 1/4] mm: rename huge_zero_page_shrinker to huge_zero_folio_shrinker

[David Hildenbrand]

On 22.07.25 11:42, Pankaj Raghav (Samsung) wrote:
> From: Pankaj Raghav <p.raghav@samsung.com>
> 
> As we already moved from exposing huge_zero_page to huge_zero_folio,
> change the name of the shrinker to reflect that.
> 
> No functional changes.
> 
> Suggested-by: David Hildenbrand <david@redhat.com>
> Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
> ---

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Cheers,

David / dhildenb

Re: [RFC 0/4] add static huge zero folio support

[David Hildenbrand]

On 22.07.25 11:42, Pankaj Raghav (Samsung) wrote:
> From: Pankaj Raghav <p.raghav@samsung.com>
> 
> NOTE: I am resending as an RFC again based on Lorenzo's feedback. The
> old series can be found here [1].
> 
> There are many places in the kernel where we need to zeroout larger
> chunks but the maximum segment we can zeroout at a time by ZERO_PAGE
> is limited by PAGE_SIZE.
> 
> This concern was raised during the review of adding Large Block Size support
> to XFS[2][3].
> 
> This is especially annoying in block devices and filesystems where we
> attach multiple ZERO_PAGEs to the bio in different bvecs. With multipage
> bvec support in block layer, it is much more efficient to send out
> larger zero pages as a part of a single bvec.
> 
> Some examples of places in the kernel where this could be useful:
> - blkdev_issue_zero_pages()
> - iomap_dio_zero()
> - vmalloc.c:zero_iter()
> - rxperf_process_call()
> - fscrypt_zeroout_range_inline_crypt()
> - bch2_checksum_update()
> ...
> 
> Usually huge_zero_folio is allocated on demand, and it will be
> deallocated by the shrinker if there are no users of it left. At the moment,
> huge_zero_folio infrastructure refcount is tied to the process lifetime
> that created it. This might not work for bio layer as the completions
> can be async and the process that created the huge_zero_folio might no
> longer be alive. And, one of the main point that came during discussion
> is to have something bigger than zero page as a drop-in replacement.
> 
> Add a config option STATIC_HUGE_ZERO_FOLIO that will always allocate
> the huge_zero_folio, and it will never drop the reference. This makes
> using the huge_zero_folio without having to pass any mm struct and does
> not tie the lifetime of the zero folio to anything, making it a drop-in
> replacement for ZERO_PAGE.
> 
> I have converted blkdev_issue_zero_pages() as an example as a part of
> this series. I also noticed close to 4% performance improvement just by
> replacing ZERO_PAGE with static huge_zero_folio.
> 
> I will send patches to individual subsystems using the huge_zero_folio
> once this gets upstreamed.
> 
> Looking forward to some feedback.

Please run scripts/checkpatch.pl on your patches.

There are quite some warning for patch #2 and #3, in particular, around 
using spaces vs. tabs.

-- 
Cheers,

David / dhildenb

Re: [RFC 2/4] mm: add static huge zero folio

[David Hildenbrand]

On 22.07.25 11:42, Pankaj Raghav (Samsung) wrote:
> From: Pankaj Raghav <p.raghav@samsung.com>
> 
> There are many places in the kernel where we need to zeroout larger
> chunks but the maximum segment we can zeroout at a time by ZERO_PAGE
> is limited by PAGE_SIZE.
> 
> This is especially annoying in block devices and filesystems where we
> attach multiple ZERO_PAGEs to the bio in different bvecs. With multipage
> bvec support in block layer, it is much more efficient to send out
> larger zero pages as a part of single bvec.
> 
> This concern was raised during the review of adding LBS support to
> XFS[1][2].
> 
> Usually huge_zero_folio is allocated on demand, and it will be
> deallocated by the shrinker if there are no users of it left. At moment,
> huge_zero_folio infrastructure refcount is tied to the process lifetime
> that created it. This might not work for bio layer as the completions
> can be async and the process that created the huge_zero_folio might no
> longer be alive. And, one of the main point that came during discussion
> is to have something bigger than zero page as a drop-in replacement.
> 
> Add a config option STATIC_HUGE_ZERO_FOLIO that will always allocate
> the huge_zero_folio, and it will never drop the reference. This makes
> using the huge_zero_folio without having to pass any mm struct and does
> not tie the lifetime of the zero folio to anything, making it a drop-in
> replacement for ZERO_PAGE.
> 
> If STATIC_PMD_ZERO_PAGE config option is enabled, then
> mm_get_huge_zero_folio() will simply return this page instead of
> dynamically allocating a new PMD page.
> 
> This option can waste memory in small systems or systems with 64k base
> page size. So make it an opt-in and also add an option from individual
> architecture so that we don't enable this feature for larger base page
> size systems.
> 
> [1] https://lore.kernel.org/linux-xfs/20231027051847.GA7885@lst.de/
> [2] https://lore.kernel.org/linux-xfs/ZitIK5OnR7ZNY0IG@infradead.org/
> 
> Co-Developed-by: David Hildenbrand <david@redhat.com>

"Co-developed-by:"

And must be followed by

Signed-of-by: David Hildenbrand <david@redhat.com>

As mentioned to the cover letter: spaces vs. tabs.

> Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
> ---
>   arch/x86/Kconfig        |  1 +
>   include/linux/huge_mm.h | 16 ++++++++++++++++
>   mm/Kconfig              | 12 ++++++++++++
>   mm/huge_memory.c        | 28 ++++++++++++++++++++++++++++
>   4 files changed, 57 insertions(+)
> 
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index 0ce86e14ab5e..8e2aa1887309 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -153,6 +153,7 @@ config X86
>   	select ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP	if X86_64
>   	select ARCH_WANT_HUGETLB_VMEMMAP_PREINIT if X86_64
>   	select ARCH_WANTS_THP_SWAP		if X86_64
> +	select ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO if X86_64
>   	select ARCH_HAS_PARANOID_L1D_FLUSH
>   	select ARCH_WANT_IRQS_OFF_ACTIVATE_MM
>   	select BUILDTIME_TABLE_SORT
> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> index 7748489fde1b..0ddd9c78f9f4 100644
> --- a/include/linux/huge_mm.h
> +++ b/include/linux/huge_mm.h
> @@ -476,6 +476,7 @@ vm_fault_t do_huge_pmd_numa_page(struct vm_fault *vmf);
>   
>   extern struct folio *huge_zero_folio;
>   extern unsigned long huge_zero_pfn;
> +extern atomic_t huge_zero_folio_is_static;
>   
>   static inline bool is_huge_zero_folio(const struct folio *folio)
>   {
> @@ -494,6 +495,16 @@ static inline bool is_huge_zero_pmd(pmd_t pmd)
>   
>   struct folio *mm_get_huge_zero_folio(struct mm_struct *mm);
>   void mm_put_huge_zero_folio(struct mm_struct *mm);
> +struct folio *__get_static_huge_zero_folio(void);
> +
> +static inline struct folio *get_static_huge_zero_folio(void)
> +{
> +       if (!IS_ENABLED(CONFIG_STATIC_HUGE_ZERO_FOLIO))
> +               return NULL;
> +       if (likely(atomic_read(&huge_zero_folio_is_static)))
> +               return huge_zero_folio;
 > +       return __get_static_huge_zero_folio();> +}
>   
>   static inline bool thp_migration_supported(void)
>   {
> @@ -685,6 +696,11 @@ static inline int change_huge_pud(struct mmu_gather *tlb,
>   {
>   	return 0;
>   }
> +
> +static inline struct folio *get_static_huge_zero_folio(void)
> +{
> +       return NULL;
> +}
>   #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
>   
>   static inline int split_folio_to_list_to_order(struct folio *folio,
> diff --git a/mm/Kconfig b/mm/Kconfig
> index 0287e8d94aea..14721171846f 100644
> --- a/mm/Kconfig
> +++ b/mm/Kconfig
> @@ -835,6 +835,18 @@ config ARCH_WANT_GENERAL_HUGETLB
>   config ARCH_WANTS_THP_SWAP
>   	def_bool n
>   
> +config ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO
> +	def_bool n
> +
> +config STATIC_HUGE_ZERO_FOLIO
> +	bool "Allocate a PMD sized folio for zeroing"
> +	depends on ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO
> +	help
> +	  Typically huge_zero_folio, which is a PMD page of zeroes, is allocated
> +	  on demand and deallocated when not in use. This option will
> +	  allocate huge_zero_folio but it will never free it.
> +	  Not suitable for memory constrained systems.

Maybe something like

"
Without this config enabled, the huge zero folio is allocated on demand 
and freed under memory pressure once no longer in use. To detect 
remaining users reliably, references to the huge zero folio must be 
tracked precisely, so it is commonly only available for mapping it into 
user page tables.

With this config enabled, the huge zero folio can also be used for other 
purposes that do not implement precise reference counting: it is still 
allocated on demand, but never freed, allowing for more wide-spread use,
for example, when performing I/O similar to the traditional shared 
zeropage."

Not suitable for memory constrained systems.
"

Should we make it clear that this is currently limited to THP configs?

depends on TRANSPARENT_HUGEPAGE

> +
>   config MM_ID
>   	def_bool n
>   
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 5d8365d1d3e9..6c890a1482f3 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -75,6 +75,8 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
>   static bool split_underused_thp = true;
>   
>   static atomic_t huge_zero_refcount;
> +static atomic_t huge_zero_static_fail_count __read_mostly;
> +atomic_t huge_zero_folio_is_static __read_mostly;
>   struct folio *huge_zero_folio __read_mostly;
>   unsigned long huge_zero_pfn __read_mostly = ~0UL;
>   unsigned long huge_anon_orders_always __read_mostly;
> @@ -266,6 +268,32 @@ void mm_put_huge_zero_folio(struct mm_struct *mm)
>   		put_huge_zero_page();
>   }
>   
> +#ifdef CONFIG_STATIC_HUGE_ZERO_FOLIO
> +struct folio *__get_static_huge_zero_folio(void)

Do we want to play safe and have a

if (unlikely(!slab_is_available()))
	return NULL;

> +{
> +       /*
> +        * If we failed to allocate a huge zero folio multiple times,
> +        * just refrain from trying.
> +        */

Hmmm, I wonder if we want to retry "some time later" again. Meaning, 
we'd base it on the jiffies, maybe?

See print_bad_pte() for an example.

> +       if (atomic_read(&huge_zero_static_fail_count) > 2)
> +               return NULL;
> +

We could make some smart decision regarding totalram_pages() and just 
disable it. A bit tricky, we can do that as a follow-up.

> +       /*
> +        * Our raised reference will prevent the shrinker from ever having
> +        * success.
> +        */
> +       if (!get_huge_zero_page()) {
> +              atomic_inc(&huge_zero_static_fail_count);
> +               return NULL;
> +       }
> +
> +       if (atomic_cmpxchg(&huge_zero_folio_is_static, 0, 1) != 0)
> +               put_huge_zero_page();
> +
> +       return huge_zero_folio;
> +}
> +#endif /* CONFIG_STATIC_HUGE_ZERO_FOLIO */
> +
>   static unsigned long shrink_huge_zero_folio_count(struct shrinker *shrink,
>   						  struct shrink_control *sc)
>   {


-- 
Cheers,

David / dhildenb

Re: [RFC 0/4] add static huge zero folio support

[Pankaj Raghav (Samsung)]

> > 
> > I will send patches to individual subsystems using the huge_zero_folio
> > once this gets upstreamed.
> > 
> > Looking forward to some feedback.
> 
> Please run scripts/checkpatch.pl on your patches.
> 
> There are quite some warning for patch #2 and #3, in particular, around
> using spaces vs. tabs.

Ah, you are right. I usually run it as a post-commit hook but I forgot
to add it when I changed my workflow.

Thanks for pointing it out. I also got a unused variable warning for huge_zero_static_fail_count
as we don't use it when CONFIG_STATIC_HUGE_ZERO_FOLIO is disabled.

I will change them in the new version.

--
Pankaj

Re: [RFC 2/4] mm: add static huge zero folio

[Pankaj Raghav (Samsung)]

On Wed, Jul 23, 2025 at 11:06:05AM +0200, David Hildenbrand wrote:
> On 22.07.25 11:42, Pankaj Raghav (Samsung) wrote:
> > From: Pankaj Raghav <p.raghav@samsung.com>
> > 
> > There are many places in the kernel where we need to zeroout larger
> > chunks but the maximum segment we can zeroout at a time by ZERO_PAGE
> > is limited by PAGE_SIZE.
> > 
> > This is especially annoying in block devices and filesystems where we
> > attach multiple ZERO_PAGEs to the bio in different bvecs. With multipage
> > bvec support in block layer, it is much more efficient to send out
> > larger zero pages as a part of single bvec.
> > 
> > This concern was raised during the review of adding LBS support to
> > XFS[1][2].
> > 
> > Usually huge_zero_folio is allocated on demand, and it will be
> > deallocated by the shrinker if there are no users of it left. At moment,
> > huge_zero_folio infrastructure refcount is tied to the process lifetime
> > that created it. This might not work for bio layer as the completions
> > can be async and the process that created the huge_zero_folio might no
> > longer be alive. And, one of the main point that came during discussion
> > is to have something bigger than zero page as a drop-in replacement.
> > 
> > Add a config option STATIC_HUGE_ZERO_FOLIO that will always allocate
> > the huge_zero_folio, and it will never drop the reference. This makes
> > using the huge_zero_folio without having to pass any mm struct and does
> > not tie the lifetime of the zero folio to anything, making it a drop-in
> > replacement for ZERO_PAGE.
> > 
> > If STATIC_PMD_ZERO_PAGE config option is enabled, then
> > mm_get_huge_zero_folio() will simply return this page instead of
> > dynamically allocating a new PMD page.
> > 
> > This option can waste memory in small systems or systems with 64k base
> > page size. So make it an opt-in and also add an option from individual
> > architecture so that we don't enable this feature for larger base page
> > size systems.
> > 
> > [1] https://lore.kernel.org/linux-xfs/20231027051847.GA7885@lst.de/
> > [2] https://lore.kernel.org/linux-xfs/ZitIK5OnR7ZNY0IG@infradead.org/
> > 
> > Co-Developed-by: David Hildenbrand <david@redhat.com>
> 
> "Co-developed-by:"
> 
> And must be followed by
> 
> Signed-of-by: David Hildenbrand <david@redhat.com>

Sounds good. Actually, I didn't want to add your sign-off without
your consent. But I will add it to the patch :)

> 
> As mentioned to the cover letter: spaces vs. tabs.
> 
> > Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
> > ---
> >   arch/x86/Kconfig        |  1 +
> >   include/linux/huge_mm.h | 16 ++++++++++++++++
> >   mm/Kconfig              | 12 ++++++++++++
> >   mm/huge_memory.c        | 28 ++++++++++++++++++++++++++++
> >   4 files changed, 57 insertions(+)
> > 
> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > index 0ce86e14ab5e..8e2aa1887309 100644
> > --- a/arch/x86/Kconfig
> > +++ b/arch/x86/Kconfig
> > @@ -153,6 +153,7 @@ config X86
> >   	select ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP	if X86_64
> >   	select ARCH_WANT_HUGETLB_VMEMMAP_PREINIT if X86_64
> >   	select ARCH_WANTS_THP_SWAP		if X86_64
> > +	select ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO if X86_64
> >   	select ARCH_HAS_PARANOID_L1D_FLUSH
> >   	select ARCH_WANT_IRQS_OFF_ACTIVATE_MM
> >   	select BUILDTIME_TABLE_SORT
> > diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> > index 7748489fde1b..0ddd9c78f9f4 100644
> > --- a/include/linux/huge_mm.h
> > +++ b/include/linux/huge_mm.h
> > @@ -476,6 +476,7 @@ vm_fault_t do_huge_pmd_numa_page(struct vm_fault *vmf);
> >   extern struct folio *huge_zero_folio;
> >   extern unsigned long huge_zero_pfn;
> > +extern atomic_t huge_zero_folio_is_static;
> >   static inline bool is_huge_zero_folio(const struct folio *folio)
> >   {
> > @@ -494,6 +495,16 @@ static inline bool is_huge_zero_pmd(pmd_t pmd)
> >   struct folio *mm_get_huge_zero_folio(struct mm_struct *mm);
> >   void mm_put_huge_zero_folio(struct mm_struct *mm);
> > +struct folio *__get_static_huge_zero_folio(void);
> > +
> > +static inline struct folio *get_static_huge_zero_folio(void)
> > +{
> > +       if (!IS_ENABLED(CONFIG_STATIC_HUGE_ZERO_FOLIO))
> > +               return NULL;
> > +       if (likely(atomic_read(&huge_zero_folio_is_static)))
> > +               return huge_zero_folio;
> > +       return __get_static_huge_zero_folio();> +}
> >   static inline bool thp_migration_supported(void)
> >   {
> > @@ -685,6 +696,11 @@ static inline int change_huge_pud(struct mmu_gather *tlb,
> >   {
> >   	return 0;
> >   }
> > +
> > +static inline struct folio *get_static_huge_zero_folio(void)
> > +{
> > +       return NULL;
> > +}
> >   #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
> >   static inline int split_folio_to_list_to_order(struct folio *folio,
> > diff --git a/mm/Kconfig b/mm/Kconfig
> > index 0287e8d94aea..14721171846f 100644
> > --- a/mm/Kconfig
> > +++ b/mm/Kconfig
> > @@ -835,6 +835,18 @@ config ARCH_WANT_GENERAL_HUGETLB
> >   config ARCH_WANTS_THP_SWAP
> >   	def_bool n
> > +config ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO
> > +	def_bool n
> > +
> > +config STATIC_HUGE_ZERO_FOLIO
> > +	bool "Allocate a PMD sized folio for zeroing"
> > +	depends on ARCH_WANTS_STATIC_HUGE_ZERO_FOLIO
> > +	help
> > +	  Typically huge_zero_folio, which is a PMD page of zeroes, is allocated
> > +	  on demand and deallocated when not in use. This option will
> > +	  allocate huge_zero_folio but it will never free it.
> > +	  Not suitable for memory constrained systems.
> 
> Maybe something like
> 
> "
> Without this config enabled, the huge zero folio is allocated on demand and
> freed under memory pressure once no longer in use. To detect remaining users
> reliably, references to the huge zero folio must be tracked precisely, so it
> is commonly only available for mapping it into user page tables.
> 
> With this config enabled, the huge zero folio can also be used for other
> purposes that do not implement precise reference counting: it is still
> allocated on demand, but never freed, allowing for more wide-spread use,
> for example, when performing I/O similar to the traditional shared
> zeropage."
> 
> Not suitable for memory constrained systems.
> "

Sounds much better! I will add it.

> 
> Should we make it clear that this is currently limited to THP configs?
> 
> depends on TRANSPARENT_HUGEPAGE

You are right. As we use the existing infrastructure, we do become
dependent on THP.
> 
> > +
> >   config MM_ID
> >   	def_bool n
> > diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> > index 5d8365d1d3e9..6c890a1482f3 100644
> > --- a/mm/huge_memory.c
> > +++ b/mm/huge_memory.c
> > @@ -75,6 +75,8 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,
> >   static bool split_underused_thp = true;
> >   static atomic_t huge_zero_refcount;
> > +static atomic_t huge_zero_static_fail_count __read_mostly;
> > +atomic_t huge_zero_folio_is_static __read_mostly;
> >   struct folio *huge_zero_folio __read_mostly;
> >   unsigned long huge_zero_pfn __read_mostly = ~0UL;
> >   unsigned long huge_anon_orders_always __read_mostly;
> > @@ -266,6 +268,32 @@ void mm_put_huge_zero_folio(struct mm_struct *mm)
> >   		put_huge_zero_page();
> >   }
> > +#ifdef CONFIG_STATIC_HUGE_ZERO_FOLIO
> > +struct folio *__get_static_huge_zero_folio(void)
> 
> Do we want to play safe and have a
> 
> if (unlikely(!slab_is_available()))
> 	return NULL;
> 
Yes, sounds good.

> > +{
> > +       /*
> > +        * If we failed to allocate a huge zero folio multiple times,
> > +        * just refrain from trying.
> > +        */
> 
> Hmmm, I wonder if we want to retry "some time later" again. Meaning, we'd
> base it on the jiffies, maybe?
> 
> See print_bad_pte() for an example.

That is a good idea. I was thinking somethign like that while I was
making the changes. This seems more logical.

> 
> > +       if (atomic_read(&huge_zero_static_fail_count) > 2)
> > +               return NULL;
> > +
> 
> We could make some smart decision regarding totalram_pages() and just
> disable it. A bit tricky, we can do that as a follow-up.
> 

oooh. Yeah, I will add it in my todos to make this as a follow up :)


Thanks for all your comments David! Can I send it next series as a normal patch
series instead of an RFC? It looks like this series is shaping up nicely.

--
Pankaj