[PATCH] mm: shmem: don't set large-order range for internal anonymous shmem mapping

[PATCH] mm: shmem: don't set large-order range for internal anonymous shmem mapping

[Baolin Wang]

Anonymous shmem large order allocations are dynamically controlled via the
global THP sysfs knob (/sys/kernel/mm/transparent_hugepage/shmem_enabled)
and the per-size mTHP knobs (/sys/kernel/mm/transparent_hugepage/hugepages-<size>kB/shmem_enabled).

Therefore, anonymous shmem uses shmem_allowable_huge_orders() to check
which large orders are allowed, rather than relying on mapping_max_folio_order().
Moreover, mapping_max_folio_order() is intended to control large order
allocations only for tmpfs mounts. Clarify this by not setting a large-order
range for internal anonymous shmem mappings, to avoid confusion, as discussed
in the previous thread[1].

[1] https://lore.kernel.org/all/ec927492-4577-4192-8fad-85eb1bb43121@linux.alibaba.com/
Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
---
 mm/shmem.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/mm/shmem.c b/mm/shmem.c
index 4ecefe02881d..a60fe067969c 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -3088,8 +3088,17 @@ static struct inode *__shmem_get_inode(struct mnt_idmap *idmap,
 	if (sbinfo->noswap)
 		mapping_set_unevictable(inode->i_mapping);
 
-	/* Don't consider 'deny' for emergencies and 'force' for testing */
-	if (sbinfo->huge)
+	/*
+	 * Only set the large order range for tmpfs mounts. The large order
+	 * selection for the internal anonymous shmem mount is configured
+	 * dynamically via the 'shmem_enabled' interfaces, so there is no
+	 * need to set a large order range for the internal anonymous shmem
+	 * mapping.
+	 *
+	 * Note: Don't consider 'deny' for emergencies and 'force' for
+	 * testing.
+	 */
+	if (sbinfo->huge && !(sb->s_flags & SB_KERNMOUNT))
 		mapping_set_large_folios(inode->i_mapping);
 
 	switch (mode & S_IFMT) {
-- 
2.47.3

Re: [PATCH] mm: shmem: don't set large-order range for internal anonymous shmem mapping

[Lance Yang]

On Tue, Apr 07, 2026 at 02:07:27PM +0800, Baolin Wang wrote:
>Anonymous shmem large order allocations are dynamically controlled via the
>global THP sysfs knob (/sys/kernel/mm/transparent_hugepage/shmem_enabled)
>and the per-size mTHP knobs (/sys/kernel/mm/transparent_hugepage/hugepages-<size>kB/shmem_enabled).
>
>Therefore, anonymous shmem uses shmem_allowable_huge_orders() to check
>which large orders are allowed, rather than relying on mapping_max_folio_order().
>Moreover, mapping_max_folio_order() is intended to control large order
>allocations only for tmpfs mounts. Clarify this by not setting a large-order
>range for internal anonymous shmem mappings, to avoid confusion, as discussed
>in the previous thread[1].
>
>[1] https://lore.kernel.org/all/ec927492-4577-4192-8fad-85eb1bb43121@linux.alibaba.com/
>Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
>---
> mm/shmem.c | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
>diff --git a/mm/shmem.c b/mm/shmem.c
>index 4ecefe02881d..a60fe067969c 100644
>--- a/mm/shmem.c
>+++ b/mm/shmem.c
>@@ -3088,8 +3088,17 @@ static struct inode *__shmem_get_inode(struct mnt_idmap *idmap,
> 	if (sbinfo->noswap)
> 		mapping_set_unevictable(inode->i_mapping);
> 
>-	/* Don't consider 'deny' for emergencies and 'force' for testing */
>-	if (sbinfo->huge)
>+	/*
>+	 * Only set the large order range for tmpfs mounts. The large order
>+	 * selection for the internal anonymous shmem mount is configured
>+	 * dynamically via the 'shmem_enabled' interfaces, so there is no
>+	 * need to set a large order range for the internal anonymous shmem
>+	 * mapping.
>+	 *
>+	 * Note: Don't consider 'deny' for emergencies and 'force' for
>+	 * testing.
>+	 */
>+	if (sbinfo->huge && !(sb->s_flags & SB_KERNMOUNT))

FWIW, SB_KERNMOUNT is broader than "internal anonymous shmem" and covers
all shm_mnt users too.

So maybe "internal shmem mount" would be a better description of what
this code is actually checking.

> 		mapping_set_large_folios(inode->i_mapping);
> 
> 	switch (mode & S_IFMT) {

Cheers,
Lance

Re: [PATCH] mm: shmem: don't set large-order range for internal anonymous shmem mapping

[Baolin Wang]

On 4/7/26 2:49 PM, Lance Yang wrote:
> 
> On Tue, Apr 07, 2026 at 02:07:27PM +0800, Baolin Wang wrote:
>> Anonymous shmem large order allocations are dynamically controlled via the
>> global THP sysfs knob (/sys/kernel/mm/transparent_hugepage/shmem_enabled)
>> and the per-size mTHP knobs (/sys/kernel/mm/transparent_hugepage/hugepages-<size>kB/shmem_enabled).
>>
>> Therefore, anonymous shmem uses shmem_allowable_huge_orders() to check
>> which large orders are allowed, rather than relying on mapping_max_folio_order().
>> Moreover, mapping_max_folio_order() is intended to control large order
>> allocations only for tmpfs mounts. Clarify this by not setting a large-order
>> range for internal anonymous shmem mappings, to avoid confusion, as discussed
>> in the previous thread[1].
>>
>> [1] https://lore.kernel.org/all/ec927492-4577-4192-8fad-85eb1bb43121@linux.alibaba.com/
>> Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
>> ---
>> mm/shmem.c | 13 +++++++++++--
>> 1 file changed, 11 insertions(+), 2 deletions(-)
>>
>> diff --git a/mm/shmem.c b/mm/shmem.c
>> index 4ecefe02881d..a60fe067969c 100644
>> --- a/mm/shmem.c
>> +++ b/mm/shmem.c
>> @@ -3088,8 +3088,17 @@ static struct inode *__shmem_get_inode(struct mnt_idmap *idmap,
>> 	if (sbinfo->noswap)
>> 		mapping_set_unevictable(inode->i_mapping);
>>
>> -	/* Don't consider 'deny' for emergencies and 'force' for testing */
>> -	if (sbinfo->huge)
>> +	/*
>> +	 * Only set the large order range for tmpfs mounts. The large order
>> +	 * selection for the internal anonymous shmem mount is configured
>> +	 * dynamically via the 'shmem_enabled' interfaces, so there is no
>> +	 * need to set a large order range for the internal anonymous shmem
>> +	 * mapping.
>> +	 *
>> +	 * Note: Don't consider 'deny' for emergencies and 'force' for
>> +	 * testing.
>> +	 */
>> +	if (sbinfo->huge && !(sb->s_flags & SB_KERNMOUNT))
> 
> FWIW, SB_KERNMOUNT is broader than "internal anonymous shmem" and covers
> all shm_mnt users too.
> 
> So maybe "internal shmem mount" would be a better description of what
> this code is actually checking.

Right, good point. Will fix. Thanks.