Skip to contentSkip to conversation

Support large folios for tmpfs

PATCH Superseded Baolin Wang 4 participants 23 messages Nov 12, 2024 – Nov 22, 2024

Subsystems (6)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Baolin Wang

author reviewer
#0 lore raw

v2 cover Support large folios for tmpfs

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays with other file systems supporting any sized large folios, and extending anonymous to support mTHP, we should not restrict tmpfs to allocating only PMD-sized huge folios, making it more special. Instead, we should allow tmpfs can allocate any sized large folios.

Considering that tmpfs already has the 'huge=' option to control the huge folios allocation, we can extend the 'huge=' option to allow any sized huge folios. The semantics of the 'huge=' mount option are:

huge=never: no any sized huge folios
huge=always: any sized huge folios
huge=within_size: like 'always' but respect the i_size huge=advise: like 'always' if requested with fadvise()/madvise()

Note: for tmpfs mmap() faults, due to the lack of a write size hint, still allocate the PMD-sized huge folios if huge=always/within_size/advise is set.

Moreover, the 'deny' and 'force' testing options controlled by '/sys/kernel/mm/transparent_hugepage/shmem_enabled', still retain the same semantics. The 'deny' can disable any sized large folios for tmpfs, while the 'force' can enable PMD sized large folios for tmpfs.

Any comments and suggestions are appreciated. Thanks.

Changes from v1:
- Add reviewed tag from Barry and David. Thanks.
- Fix building warnings reported by kernel test robot.
- Add a new patch to control the default huge policy for tmpfs.

Changes from RFC v3:
- Drop the huge=write_size option.
- Allow any sized huge folios for 'hgue' option.
- Update the documentation, per David.

Changes from RFC v2:
- Drop mTHP interfaces to control huge page allocation, per Matthew.
- Add a new helper to calculate the order, suggested by Matthew.
- Add a new huge=write_size option to allocate large folios based on the write size.
- Add a new patch to update the documentation.

Changes from RFC v1:
- Drop patch 1.
- Use 'write_end' to calculate the length in shmem_allowable_huge_orders().
- Update shmem_mapping_size_order() per Daniel.

Baolin Wang (4):
mm: factor out the order calculation into a new helper
mm: shmem: change shmem_huge_global_enabled() to return huge order
bitmap
mm: shmem: add large folio support for tmpfs
mm: shmem: add a kernel command line to change the default huge policy
for tmpfs

David Hildenbrand (1):
docs: tmpfs: update the huge folios policy for tmpfs and shmem

 .../admin-guide/kernel-parameters.txt         |   7 +
 Documentation/admin-guide/mm/transhuge.rst    |  64 ++++++--
 include/linux/pagemap.h                       |  16 +-
 mm/shmem.c                                    | 148 ++++++++++++++----
 4 files changed, 183 insertions(+), 52 deletions(-)
Signature · 1 line
2.39.3
author reviewer
#1 lore raw

PATCH v2 1/5 mm: factor out the order calculation into a new helper

Factor out the order calculation into a new helper, which can be reused by shmem in the following patch.

---
 include/linux/pagemap.h | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
include/linux/pagemap.h +13 -3
index bcf0865a38ae..d796c8a33647 100644
@@ -727,6 +727,16 @@ typedef unsigned int __bitwise fgf_t;
 
 #define FGP_WRITEBEGIN		(FGP_LOCK | FGP_WRITE | FGP_CREAT | FGP_STABLE)
 
+static inline unsigned int filemap_get_order(size_t size)
+{
+	unsigned int shift = ilog2(size);
+
+	if (shift <= PAGE_SHIFT)
+		return 0;
+
+	return shift - PAGE_SHIFT;
+}
+
 /**
  * fgf_set_order - Encode a length in the fgf_t flags.
  * @size: The suggested size of the folio to create.
@@ -740,11 +750,11 @@ typedef unsigned int __bitwise fgf_t;
  */
 static inline fgf_t fgf_set_order(size_t size)
 {
-	unsigned int shift = ilog2(size);
+	unsigned int order = filemap_get_order(size);
 
-	if (shift <= PAGE_SHIFT)
+	if (!order)
 		return 0;
-	return (__force fgf_t)((shift - PAGE_SHIFT) << 26);
+	return (__force fgf_t)(order << 26);
 }
 
 void *filemap_get_entry(struct address_space *mapping, pgoff_t index);
Signature · 1 line
2.39.3
author reviewer
#5 lore raw

PATCH v2 3/5 mm: shmem: add large folio support for tmpfs

In reply to Baolin Wang

Add large folio support for tmpfs write and fallocate paths matching the same high order preference mechanism used in the iomap buffered IO path as used in __filemap_get_folio().

Add shmem_mapping_size_orders() to get a hint for the orders of the folio based on the file size which takes care of the mapping requirements.

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays with other file systems supporting any sized large folios, and extending anonymous to support mTHP, we should not restrict tmpfs to allocating only PMD-sized huge folios, making it more special. Instead, we should allow tmpfs can allocate any sized large folios.

Considering that tmpfs already has the 'huge=' option to control the huge folios allocation, we can extend the 'huge=' option to allow any sized huge folios. The semantics of the 'huge=' mount option are:

huge=never: no any sized huge folios
huge=always: any sized huge folios
huge=within_size: like 'always' but respect the i_size huge=advise: like 'always' if requested with fadvise()/madvise()

Note: for tmpfs mmap() faults, due to the lack of a write size hint, still allocate the PMD-sized huge folios if huge=always/within_size/advise is set.

Moreover, the 'deny' and 'force' testing options controlled by '/sys/kernel/mm/transparent_hugepage/shmem_enabled', still retain the same semantics. The 'deny' can disable any sized large folios for tmpfs, while the 'force' can enable PMD sized large folios for tmpfs.

---
 mm/shmem.c | 91 +++++++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 77 insertions(+), 14 deletions(-)
mm/shmem.c +77 -14
index 86b2e417dc6f..a3203cf8860f 100644
@@ -549,10 +549,50 @@ static bool shmem_confirm_swap(struct address_space *mapping,
 
 static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
 
+/**
+ * shmem_mapping_size_orders - Get allowable folio orders for the given file size.
+ * @mapping: Target address_space.
+ * @index: The page index.
+ * @write_end: end of a write, could extend inode size.
+ *
+ * This returns huge orders for folios (when supported) based on the file size
+ * which the mapping currently allows at the given index. The index is relevant
+ * due to alignment considerations the mapping might have. The returned order
+ * may be less than the size passed.
+ *
+ * Return: The orders.
+ */
+static inline unsigned int
+shmem_mapping_size_orders(struct address_space *mapping, pgoff_t index, loff_t write_end)
+{
+	unsigned int order;
+	size_t size;
+
+	if (!mapping_large_folio_support(mapping) || !write_end)
+		return 0;
+
+	/* Calculate the write size based on the write_end */
+	size = write_end - (index << PAGE_SHIFT);
+	order = filemap_get_order(size);
+	if (!order)
+		return 0;
+
+	/* If we're not aligned, allocate a smaller folio */
+	if (index & ((1UL << order) - 1))
+		order = __ffs(index);
+
+	order = min_t(size_t, order, MAX_PAGECACHE_ORDER);
+	return order > 0 ? BIT(order + 1) - 1 : 0;
+}
+
 static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index,
 					      loff_t write_end, bool shmem_huge_force,
+					      struct vm_area_struct *vma,
 					      unsigned long vm_flags)
 {
+	unsigned long within_size_orders;
+	unsigned int order;
+	pgoff_t aligned_index;
 	loff_t i_size;
 
 	if (HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER)
@@ -564,15 +604,41 @@ static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index
 	if (shmem_huge_force || shmem_huge == SHMEM_HUGE_FORCE)
 		return BIT(HPAGE_PMD_ORDER);
 
+	/*
+	 * The huge order allocation for anon shmem is controlled through
+	 * the mTHP interface, so we still use PMD-sized huge order to
+	 * check whether global control is enabled.
+	 *
+	 * For tmpfs mmap()'s huge order, we still use PMD-sized order to
+	 * allocate huge pages due to lack of a write size hint.
+	 *
+	 * Otherwise, tmpfs will allow getting a highest order hint based on
+	 * the size of write and fallocate paths, then will try each allowable
+	 * huge orders.
+	 */
 	switch (SHMEM_SB(inode->i_sb)->huge) {
 	case SHMEM_HUGE_ALWAYS:
-		return BIT(HPAGE_PMD_ORDER);
-	case SHMEM_HUGE_WITHIN_SIZE:
-		index = round_up(index + 1, HPAGE_PMD_NR);
-		i_size = max(write_end, i_size_read(inode));
-		i_size = round_up(i_size, PAGE_SIZE);
-		if (i_size >> PAGE_SHIFT >= index)
+		if (vma)
 			return BIT(HPAGE_PMD_ORDER);
+
+		return shmem_mapping_size_orders(inode->i_mapping, index, write_end);
+	case SHMEM_HUGE_WITHIN_SIZE:
+		if (vma)
+			within_size_orders = BIT(HPAGE_PMD_ORDER);
+		else
+			within_size_orders = shmem_mapping_size_orders(inode->i_mapping,
+								       index, write_end);
+
+		order = highest_order(within_size_orders);
+		while (within_size_orders) {
+			aligned_index = round_up(index + 1, 1 << order);
+			i_size = max(write_end, i_size_read(inode));
+			i_size = round_up(i_size, PAGE_SIZE);
+			if (i_size >> PAGE_SHIFT >= aligned_index)
+				return within_size_orders;
+
+			order = next_order(&within_size_orders, order);
+		}
 		fallthrough;
 	case SHMEM_HUGE_ADVISE:
 		if (vm_flags & VM_HUGEPAGE)
@@ -776,6 +842,7 @@ static unsigned long shmem_unused_huge_shrink(struct shmem_sb_info *sbinfo,
 
 static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index,
 					      loff_t write_end, bool shmem_huge_force,
+					      struct vm_area_struct *vma,
 					      unsigned long vm_flags)
 {
 	return 0;
@@ -1173,7 +1240,7 @@ static int shmem_getattr(struct mnt_idmap *idmap,
 	generic_fillattr(idmap, request_mask, inode, stat);
 	inode_unlock_shared(inode);
 
-	if (shmem_huge_global_enabled(inode, 0, 0, false, 0))
+	if (shmem_huge_global_enabled(inode, 0, 0, false, NULL, 0))
 		stat->blksize = HPAGE_PMD_SIZE;
 
 	if (request_mask & STATX_BTIME) {
@@ -1690,14 +1757,10 @@ unsigned long shmem_allowable_huge_orders(struct inode *inode,
 		return 0;
 
 	global_orders = shmem_huge_global_enabled(inode, index, write_end,
-						  shmem_huge_force, vm_flags);
-	if (!vma || !vma_is_anon_shmem(vma)) {
-		/*
-		 * For tmpfs, we now only support PMD sized THP if huge page
-		 * is enabled, otherwise fallback to order 0.
-		 */
+						  shmem_huge_force, vma, vm_flags);
+	/* Tmpfs huge pages allocation */
+	if (!vma || !vma_is_anon_shmem(vma))
 		return global_orders;
-	}
 
 	/*
 	 * Following the 'deny' semantics of the top level, force the huge
Signature · 1 line
2.39.3
author reviewer
#11 lore raw

PATCH v2 4/5 mm: shmem: add a kernel command line to change the default huge policy for tmpfs

In reply to Baolin Wang

Now the tmpfs can allow to allocate any sized large folios, and the default huge policy is still 'never'. Thus adding a new command line to change the default huge policy will be helpful to use the large folios for tmpfs, which is similar to the 'transparent_hugepage_shmem' cmdline for shmem.

---
 .../admin-guide/kernel-parameters.txt         |  7 ++++++
 Documentation/admin-guide/mm/transhuge.rst    |  6 +++++
 mm/shmem.c                                    | 23 ++++++++++++++++++-
 3 files changed, 35 insertions(+), 1 deletion(-)
Documentation/admin-guide/kernel-parameters.txt +7 -0
index b48d744d99b0..007e6cfada3e 100644
@@ -6943,6 +6943,13 @@
 			See Documentation/admin-guide/mm/transhuge.rst
 			for more details.
 
+	transparent_hugepage_tmpfs= [KNL]
+			Format: [always|within_size|advise|never]
+			Can be used to control the default hugepage allocation policy
+			for the tmpfs mount.
+			See Documentation/admin-guide/mm/transhuge.rst
+			for more details.
+
 	trusted.source=	[KEYS]
 			Format: <string>
 			This parameter identifies the trust source as a backend
Documentation/admin-guide/mm/transhuge.rst +6 -0
index 5034915f4e8e..9ae775eaacbe 100644
@@ -332,6 +332,12 @@ allocation policy for the internal shmem mount by using the kernel parameter
 seven valid policies for shmem (``always``, ``within_size``, ``advise``,
 ``never``, ``deny``, and ``force``).
 
+Similarly to ``transparent_hugepage_shmem``, you can control the default
+hugepage allocation policy for the tmpfs mount by using the kernel parameter
+``transparent_hugepage_tmpfs=<policy>``, where ``<policy>`` is one of the
+four valid policies for tmpfs (``always``, ``within_size``, ``advise``,
+``never``). The tmpfs mount default policy is ``never``.
+
 In the same manner as ``thp_anon`` controls each supported anonymous THP
 size, ``thp_shmem`` controls each supported shmem THP size. ``thp_shmem``
 has the same format as ``thp_anon``, but also supports the policy
mm/shmem.c +22 -1
index a3203cf8860f..021760e91cea 100644
@@ -548,6 +548,7 @@ static bool shmem_confirm_swap(struct address_space *mapping,
 /* ifdef here to avoid bloating shmem.o when not necessary */
 
 static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
+static int tmpfs_huge __read_mostly = SHMEM_HUGE_NEVER;
 
 /**
  * shmem_mapping_size_orders - Get allowable folio orders for the given file size.
@@ -4780,7 +4781,12 @@ static int shmem_fill_super(struct super_block *sb, struct fs_context *fc)
 	sbinfo->gid = ctx->gid;
 	sbinfo->full_inums = ctx->full_inums;
 	sbinfo->mode = ctx->mode;
-	sbinfo->huge = ctx->huge;
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+	if (ctx->seen & SHMEM_SEEN_HUGE)
+		sbinfo->huge = ctx->huge;
+	else
+		sbinfo->huge = tmpfs_huge;
+#endif
 	sbinfo->mpol = ctx->mpol;
 	ctx->mpol = NULL;
 
@@ -5259,6 +5265,21 @@ static int __init setup_transparent_hugepage_shmem(char *str)
 }
 __setup("transparent_hugepage_shmem=", setup_transparent_hugepage_shmem);
 
+static int __init setup_transparent_hugepage_tmpfs(char *str)
+{
+	int huge;
+
+	huge = shmem_parse_huge(str);
+	if (huge < 0) {
+		pr_warn("transparent_hugepage_tmpfs= cannot parse, ignored\n");
+		return huge;
+	}
+
+	tmpfs_huge = huge;
+	return 1;
+}
+__setup("transparent_hugepage_tmpfs=", setup_transparent_hugepage_tmpfs);
+
 static char str_dup[PAGE_SIZE] __initdata;
 static int __init setup_thp_shmem(char *str)
 {
Signature · 1 line
2.39.3
author reviewer
#15 lore raw

PATCH v2 5/5 docs: tmpfs: update the huge folios policy for tmpfs and shmem

In reply to Baolin Wang

From: David Hildenbrand <[email protected]>

Update the huge folios policy for tmpfs and shmem.

---
 Documentation/admin-guide/mm/transhuge.rst | 58 +++++++++++++++-------
 1 file changed, 41 insertions(+), 17 deletions(-)
Documentation/admin-guide/mm/transhuge.rst +41 -17
index 9ae775eaacbe..ba6edff728ed 100644
@@ -358,8 +358,21 @@ default to ``never``.
 Hugepages in tmpfs/shmem
 ========================
 
-You can control hugepage allocation policy in tmpfs with mount option
-``huge=``. It can have following values:
+Traditionally, tmpfs only supported a single huge page size ("PMD"). Today,
+it also supports smaller sizes just like anonymous memory, often referred
+to as "multi-size THP" (mTHP). Huge pages of any size are commonly
+represented in the kernel as "large folios".
+
+While there is fine control over the huge page sizes to use for the internal
+shmem mount (see below), ordinary tmpfs mounts will make use of all available
+huge page sizes without any control over the exact sizes, behaving more like
+other file systems.
+
+tmpfs mounts
+------------
+
+The THP allocation policy for tmpfs mounts can be adjusted using the mount
+option: ``huge=``. It can have following values:
 
 always
     Attempt to allocate huge pages every time we need a new page;
@@ -374,19 +387,19 @@ within_size
 advise
     Only allocate huge pages if requested with fadvise()/madvise();
 
-The default policy is ``never``.
+Remember, that the kernel may use huge pages of all available sizes, and
+that no fine control as for the internal tmpfs mount is available.
+
+The default policy in the past was ``never``, but it can now be adjusted
+using the kernel parameter ``transparent_hugepage_tmpfs=<policy>``.
 
 ``mount -o remount,huge= /mountpoint`` works fine after mount: remounting
 ``huge=never`` will not attempt to break up huge pages at all, just stop more
 from being allocated.
 
-There's also sysfs knob to control hugepage allocation policy for internal
-shmem mount: /sys/kernel/mm/transparent_hugepage/shmem_enabled. The mount
-is used for SysV SHM, memfds, shared anonymous mmaps (of /dev/zero or
-MAP_ANONYMOUS), GPU drivers' DRM objects, Ashmem.
-
-In addition to policies listed above, shmem_enabled allows two further
-values:
+In addition to policies listed above, the sysfs knob
+/sys/kernel/mm/transparent_hugepage/shmem_enabled will affect the
+allocation policy of tmpfs mounts, when set to the following values:
 
 deny
     For use in emergencies, to force the huge option off from
@@ -394,13 +407,24 @@ deny
 force
     Force the huge option on for all - very useful for testing;
 
-Shmem can also use "multi-size THP" (mTHP) by adding a new sysfs knob to
-control mTHP allocation:
-'/sys/kernel/mm/transparent_hugepage/hugepages-<size>kB/shmem_enabled',
-and its value for each mTHP is essentially consistent with the global
-setting.  An 'inherit' option is added to ensure compatibility with these
-global settings.  Conversely, the options 'force' and 'deny' are dropped,
-which are rather testing artifacts from the old ages.
+shmem / internal tmpfs
+----------------------
+The mount internal tmpfs mount is used for SysV SHM, memfds, shared anonymous
+mmaps (of /dev/zero or MAP_ANONYMOUS), GPU drivers' DRM  objects, Ashmem.
+
+To control the THP allocation policy for this internal tmpfs mount, the
+sysfs knob /sys/kernel/mm/transparent_hugepage/shmem_enabled and the knobs
+per THP size in
+'/sys/kernel/mm/transparent_hugepage/hugepages-<size>kB/shmem_enabled'
+can be used.
+
+The global knob has the same semantics as the ``huge=`` mount options
+for tmpfs mounts, except that the different huge page sizes can be controlled
+individually, and will only use the setting of the global knob when the
+per-size knob is set to 'inherit'.
+
+The options 'force' and 'deny' are dropped for the individual sizes, which
+are rather testing artifacts from the old ages.
 
 always
     Attempt to allocate <size> huge pages every time we need a new page;
Signature · 1 line
2.39.3
author reviewer
#3 lore raw

PATCH v2 2/5 mm: shmem: change shmem_huge_global_enabled() to return huge order bitmap

In reply to Baolin Wang

Change the shmem_huge_global_enabled() to return the suitable huge order bitmap, and return 0 if huge pages are not allowed. This is a preparation for supporting various huge orders allocation of tmpfs in the following patches.

No functional changes.

---
 mm/shmem.c | 40 ++++++++++++++++++++--------------------
 1 file changed, 20 insertions(+), 20 deletions(-)
mm/shmem.c +20 -20
index 579e58cb3262..86b2e417dc6f 100644
@@ -549,37 +549,37 @@ static bool shmem_confirm_swap(struct address_space *mapping,
 
 static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
 
-static bool shmem_huge_global_enabled(struct inode *inode, pgoff_t index,
-				      loff_t write_end, bool shmem_huge_force,
-				      unsigned long vm_flags)
+static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index,
+					      loff_t write_end, bool shmem_huge_force,
+					      unsigned long vm_flags)
 {
 	loff_t i_size;
 
 	if (HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER)
-		return false;
+		return 0;
 	if (!S_ISREG(inode->i_mode))
-		return false;
+		return 0;
 	if (shmem_huge == SHMEM_HUGE_DENY)
-		return false;
+		return 0;
 	if (shmem_huge_force || shmem_huge == SHMEM_HUGE_FORCE)
-		return true;
+		return BIT(HPAGE_PMD_ORDER);
 
 	switch (SHMEM_SB(inode->i_sb)->huge) {
 	case SHMEM_HUGE_ALWAYS:
-		return true;
+		return BIT(HPAGE_PMD_ORDER);
 	case SHMEM_HUGE_WITHIN_SIZE:
 		index = round_up(index + 1, HPAGE_PMD_NR);
 		i_size = max(write_end, i_size_read(inode));
 		i_size = round_up(i_size, PAGE_SIZE);
 		if (i_size >> PAGE_SHIFT >= index)
-			return true;
+			return BIT(HPAGE_PMD_ORDER);
 		fallthrough;
 	case SHMEM_HUGE_ADVISE:
 		if (vm_flags & VM_HUGEPAGE)
-			return true;
+			return BIT(HPAGE_PMD_ORDER);
 		fallthrough;
 	default:
-		return false;
+		return 0;
 	}
 }
 
@@ -774,11 +774,11 @@ static unsigned long shmem_unused_huge_shrink(struct shmem_sb_info *sbinfo,
 	return 0;
 }
 
-static bool shmem_huge_global_enabled(struct inode *inode, pgoff_t index,
-				      loff_t write_end, bool shmem_huge_force,
-				      unsigned long vm_flags)
+static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index,
+					      loff_t write_end, bool shmem_huge_force,
+					      unsigned long vm_flags)
 {
-	return false;
+	return 0;
 }
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
@@ -1682,21 +1682,21 @@ unsigned long shmem_allowable_huge_orders(struct inode *inode,
 	unsigned long mask = READ_ONCE(huge_shmem_orders_always);
 	unsigned long within_size_orders = READ_ONCE(huge_shmem_orders_within_size);
 	unsigned long vm_flags = vma ? vma->vm_flags : 0;
-	bool global_huge;
+	unsigned int global_orders;
 	loff_t i_size;
 	int order;
 
 	if (thp_disabled_by_hw() || (vma && vma_thp_disabled(vma, vm_flags)))
 		return 0;
 
-	global_huge = shmem_huge_global_enabled(inode, index, write_end,
-						shmem_huge_force, vm_flags);
+	global_orders = shmem_huge_global_enabled(inode, index, write_end,
+						  shmem_huge_force, vm_flags);
 	if (!vma || !vma_is_anon_shmem(vma)) {
 		/*
 		 * For tmpfs, we now only support PMD sized THP if huge page
 		 * is enabled, otherwise fallback to order 0.
 		 */
-		return global_huge ? BIT(HPAGE_PMD_ORDER) : 0;
+		return global_orders;
 	}
 
 	/*
@@ -1729,7 +1729,7 @@ unsigned long shmem_allowable_huge_orders(struct inode *inode,
 	if (vm_flags & VM_HUGEPAGE)
 		mask |= READ_ONCE(huge_shmem_orders_madvise);
 
-	if (global_huge)
+	if (global_orders > 0)
 		mask |= READ_ONCE(huge_shmem_orders_inherit);
 
 	return THP_ORDERS_ALL_FILE_DEFAULT & mask;
Signature · 1 line
2.39.3
Acked-by on v2 2/5
#4 lore raw

Re: [PATCH v2 2/5] mm: shmem: change shmem_huge_global_enabled() to return huge order bitmap

Quoted 6 lines · Baolin Wang

Change the shmem_huge_global_enabled() to return the suitable huge order bitmap, and return 0 if huge pages are not allowed. This is a preparation for supporting various huge orders allocation of tmpfs in the following patches.

No functional changes.

Signature · 3 lines
Cheers,

David / dhildenb
#6 lore raw

Re: [PATCH v2 3/5] mm: shmem: add large folio support for tmpfs

In reply to Baolin Wang
Quoted 82 lines · Baolin Wang

Add large folio support for tmpfs write and fallocate paths matching the same high order preference mechanism used in the iomap buffered IO path as used in __filemap_get_folio().

Add shmem_mapping_size_orders() to get a hint for the orders of the folio based on the file size which takes care of the mapping requirements.

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays with other file systems supporting any sized large folios, and extending anonymous to support mTHP, we should not restrict tmpfs to allocating only PMD-sized huge folios, making it more special. Instead, we should allow tmpfs can allocate any sized large folios.

Considering that tmpfs already has the 'huge=' option to control the huge folios allocation, we can extend the 'huge=' option to allow any sized huge folios. The semantics of the 'huge=' mount option are:

huge=never: no any sized huge folios
huge=always: any sized huge folios
huge=within_size: like 'always' but respect the i_size huge=advise: like 'always' if requested with fadvise()/madvise()

Note: for tmpfs mmap() faults, due to the lack of a write size hint, still allocate the PMD-sized huge folios if huge=always/within_size/advise is set.

Moreover, the 'deny' and 'force' testing options controlled by '/sys/kernel/mm/transparent_hugepage/shmem_enabled', still retain the same semantics. The 'deny' can disable any sized large folios for tmpfs, while the 'force' can enable PMD sized large folios for tmpfs.

---
  mm/shmem.c | 91 +++++++++++++++++++++++++++++++++++++++++++++---------
  1 file changed, 77 insertions(+), 14 deletions(-)
mm/shmem.c +36 -0
index 86b2e417dc6f..a3203cf8860f 100644
@@ -549,10 +549,50 @@ static bool shmem_confirm_swap(struct address_space *mapping,
 
  static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
 
+/**
+ * shmem_mapping_size_orders - Get allowable folio orders for the given file size.
+ * @mapping: Target address_space.
+ * @index: The page index.
+ * @write_end: end of a write, could extend inode size.
+ *
+ * This returns huge orders for folios (when supported) based on the file size
+ * which the mapping currently allows at the given index. The index is relevant
+ * due to alignment considerations the mapping might have. The returned order
+ * may be less than the size passed.
+ *
+ * Return: The orders.
+ */
+static inline unsigned int
+shmem_mapping_size_orders(struct address_space *mapping, pgoff_t index, loff_t write_end)
+{
+	unsigned int order;
+	size_t size;
+
+	if (!mapping_large_folio_support(mapping) || !write_end)
+		return 0;
+
+	/* Calculate the write size based on the write_end */
+	size = write_end - (index << PAGE_SHIFT);
+	order = filemap_get_order(size);
+	if (!order)
+		return 0;
+
+	/* If we're not aligned, allocate a smaller folio */
+	if (index & ((1UL << order) - 1))
+		order = __ffs(index);
+
+	order = min_t(size_t, order, MAX_PAGECACHE_ORDER);
+	return order > 0 ? BIT(order + 1) - 1 : 0;
+}
+
  static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index,

mm/shmem.c static bool shmem_confirm_swap(struct address_space *mapping, · line 598

Quoted 10 lines
mm/shmem.c +4 -0
  					      loff_t write_end, bool shmem_huge_force,
+					      struct vm_area_struct *vma,
  					      unsigned long vm_flags)
  {
+	unsigned long within_size_orders;
+	unsigned int order;
+	pgoff_t aligned_index;
  	loff_t i_size;
 
  	if (HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER)

We can allow all orders up to MAX_PAGECACHE_ORDER, shmem_mapping_size_orders() handles it properly.

So maybe we should drop this condition and use instead below where we have

return BIT(HPAGE_PMD_ORDER);

instead something like.

return HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ? 0 : BIT(HPAGE_PMD_ORDER);

Ideally, factoring it out somehow

int maybe_pmd_order = HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ? 0 : BIT(HPAGE_PMD_ORDER);

...

return maybe_pmd_order;

mm/shmem.c static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index · line 605

Quoted 3 lines
mm/shmem.c +0 -0
@@ -564,15 +604,41 @@ static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index
  	if (shmem_huge_force || shmem_huge == SHMEM_HUGE_FORCE)
  		return BIT(HPAGE_PMD_ORDER);

Why not force-enable all orders (of course, respecting MAX_PAGECACHE_ORDER and possibly VMA)?

Quoted 35 lines
mm/shmem.c +25 -6
+	/*
+	 * The huge order allocation for anon shmem is controlled through
+	 * the mTHP interface, so we still use PMD-sized huge order to
+	 * check whether global control is enabled.
+	 *
+	 * For tmpfs mmap()'s huge order, we still use PMD-sized order to
+	 * allocate huge pages due to lack of a write size hint.
+	 *
+	 * Otherwise, tmpfs will allow getting a highest order hint based on
+	 * the size of write and fallocate paths, then will try each allowable
+	 * huge orders.
+	 */
  	switch (SHMEM_SB(inode->i_sb)->huge) {
  	case SHMEM_HUGE_ALWAYS:
-		return BIT(HPAGE_PMD_ORDER);
-	case SHMEM_HUGE_WITHIN_SIZE:
-		index = round_up(index + 1, HPAGE_PMD_NR);
-		i_size = max(write_end, i_size_read(inode));
-		i_size = round_up(i_size, PAGE_SIZE);
-		if (i_size >> PAGE_SHIFT >= index)
+		if (vma)
  			return BIT(HPAGE_PMD_ORDER);
+
+		return shmem_mapping_size_orders(inode->i_mapping, index, write_end);
+	case SHMEM_HUGE_WITHIN_SIZE:
+		if (vma)
+			within_size_orders = BIT(HPAGE_PMD_ORDER);
+		else
+			within_size_orders = shmem_mapping_size_orders(inode->i_mapping,
+								       index, write_end);
+
+		order = highest_order(within_size_orders);
+		while (within_size_orders) {
+			aligned_index = round_up(index + 1, 1 << order);

mm/shmem.c static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index · line 644

Quoted 10 lines
mm/shmem.c +7 -0
+			i_size = max(write_end, i_size_read(inode));
+			i_size = round_up(i_size, PAGE_SIZE);
+			if (i_size >> PAGE_SHIFT >= aligned_index)
+				return within_size_orders;
+
+			order = next_order(&within_size_orders, order);
+		}
  		fallthrough;
  	case SHMEM_HUGE_ADVISE:
  		if (vm_flags & VM_HUGEPAGE)

I think the point here is that "write" -> no VMA -> vm_flags == 0 -> no code changes needed :)

Signature · 3 lines
Cheers,

David / dhildenb
#7 lore raw

Re: [PATCH v2 3/5] mm: shmem: add large folio support for tmpfs

Quoted 162 lines · David Hildenbrand
Quoted 82 lines · Baolin Wang

Add large folio support for tmpfs write and fallocate paths matching the same high order preference mechanism used in the iomap buffered IO path as used in __filemap_get_folio().

Add shmem_mapping_size_orders() to get a hint for the orders of the folio based on the file size which takes care of the mapping requirements.

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays with other file systems supporting any sized large folios, and extending anonymous to support mTHP, we should not restrict tmpfs to allocating only PMD-sized huge folios, making it more special. Instead, we should allow tmpfs can allocate any sized large folios.

Considering that tmpfs already has the 'huge=' option to control the huge folios allocation, we can extend the 'huge=' option to allow any sized huge folios. The semantics of the 'huge=' mount option are:

huge=never: no any sized huge folios
huge=always: any sized huge folios
huge=within_size: like 'always' but respect the i_size huge=advise: like 'always' if requested with fadvise()/madvise()

Note: for tmpfs mmap() faults, due to the lack of a write size hint, still allocate the PMD-sized huge folios if huge=always/within_size/advise is set.

Moreover, the 'deny' and 'force' testing options controlled by '/sys/kernel/mm/transparent_hugepage/shmem_enabled', still retain the same semantics. The 'deny' can disable any sized large folios for tmpfs, while the 'force' can enable PMD sized large folios for tmpfs.

---
   mm/shmem.c | 91 +++++++++++++++++++++++++++++++++++++++++++++---------
   1 file changed, 77 insertions(+), 14 deletions(-)
mm/shmem.c +36 -0
index 86b2e417dc6f..a3203cf8860f 100644
@@ -549,10 +549,50 @@ static bool shmem_confirm_swap(struct address_space *mapping,
 
   static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
 
+/**
+ * shmem_mapping_size_orders - Get allowable folio orders for the given file size.
+ * @mapping: Target address_space.
+ * @index: The page index.
+ * @write_end: end of a write, could extend inode size.
+ *
+ * This returns huge orders for folios (when supported) based on the file size
+ * which the mapping currently allows at the given index. The index is relevant
+ * due to alignment considerations the mapping might have. The returned order
+ * may be less than the size passed.
+ *
+ * Return: The orders.
+ */
+static inline unsigned int
+shmem_mapping_size_orders(struct address_space *mapping, pgoff_t index, loff_t write_end)
+{
+	unsigned int order;
+	size_t size;
+
+	if (!mapping_large_folio_support(mapping) || !write_end)
+		return 0;
+
+	/* Calculate the write size based on the write_end */
+	size = write_end - (index << PAGE_SHIFT);
+	order = filemap_get_order(size);
+	if (!order)
+		return 0;
+
+	/* If we're not aligned, allocate a smaller folio */
+	if (index & ((1UL << order) - 1))
+		order = __ffs(index);
+
+	order = min_t(size_t, order, MAX_PAGECACHE_ORDER);
+	return order > 0 ? BIT(order + 1) - 1 : 0;
+}
+
   static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index,

mm/shmem.c static bool shmem_confirm_swap(struct address_space *mapping, · line 598

Quoted 10 lines
mm/shmem.c +4 -0
   					      loff_t write_end, bool shmem_huge_force,
+					      struct vm_area_struct *vma,
   					      unsigned long vm_flags)
   {
+	unsigned long within_size_orders;
+	unsigned int order;
+	pgoff_t aligned_index;
   	loff_t i_size;
 
   	if (HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER)

We can allow all orders up to MAX_PAGECACHE_ORDER, shmem_mapping_size_orders() handles it properly.

So maybe we should drop this condition and use instead below where we have

return BIT(HPAGE_PMD_ORDER);

instead something like.

return HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ? 0 : BIT(HPAGE_PMD_ORDER);

Ideally, factoring it out somehow

int maybe_pmd_order = HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ? 0 :
BIT(HPAGE_PMD_ORDER);

...

return maybe_pmd_order;

mm/shmem.c static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index · line 605

Quoted 3 lines
mm/shmem.c +0 -0
@@ -564,15 +604,41 @@ static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index
   	if (shmem_huge_force || shmem_huge == SHMEM_HUGE_FORCE)
   		return BIT(HPAGE_PMD_ORDER);

Why not force-enable all orders (of course, respecting MAX_PAGECACHE_ORDER and possibly VMA)?

Quoted 40 lines
mm/shmem.c +30 -6
+	/*
+	 * The huge order allocation for anon shmem is controlled through
+	 * the mTHP interface, so we still use PMD-sized huge order to
+	 * check whether global control is enabled.
+	 *
+	 * For tmpfs mmap()'s huge order, we still use PMD-sized order to
+	 * allocate huge pages due to lack of a write size hint.
+	 *
+	 * Otherwise, tmpfs will allow getting a highest order hint based on
+	 * the size of write and fallocate paths, then will try each allowable
+	 * huge orders.
+	 */
   	switch (SHMEM_SB(inode->i_sb)->huge) {
   	case SHMEM_HUGE_ALWAYS:
-		return BIT(HPAGE_PMD_ORDER);
-	case SHMEM_HUGE_WITHIN_SIZE:
-		index = round_up(index + 1, HPAGE_PMD_NR);
-		i_size = max(write_end, i_size_read(inode));
-		i_size = round_up(i_size, PAGE_SIZE);
-		if (i_size >> PAGE_SHIFT >= index)
+		if (vma)
   			return BIT(HPAGE_PMD_ORDER);
+
+		return shmem_mapping_size_orders(inode->i_mapping, index, write_end);
+	case SHMEM_HUGE_WITHIN_SIZE:
+		if (vma)
+			within_size_orders = BIT(HPAGE_PMD_ORDER);
+		else
+			within_size_orders = shmem_mapping_size_orders(inode->i_mapping,
+								       index, write_end);
+
+		order = highest_order(within_size_orders);
+		while (within_size_orders) {
+			aligned_index = round_up(index + 1, 1 << order);
+			i_size = max(write_end, i_size_read(inode));
+			i_size = round_up(i_size, PAGE_SIZE);
+			if (i_size >> PAGE_SHIFT >= aligned_index)
+				return within_size_orders;
+
Quoted 8 lines

mm/shmem.c static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index · line 644

Quoted 5 lines
mm/shmem.c +2 -0
+			order = next_order(&within_size_orders, order);
+		}
   		fallthrough;
   	case SHMEM_HUGE_ADVISE:
   		if (vm_flags & VM_HUGEPAGE)

I think the point here is that "write" -> no VMA -> vm_flags == 0 -> no code changes needed :)

... and now I wonder about documented "fadvise", because this here is only concerned with madvise?

Signature · 3 lines
Cheers,

David / dhildenb
author reviewer
#8 lore raw

Re: [PATCH v2 3/5] mm: shmem: add large folio support for tmpfs

In reply to David Hildenbrand
Quoted 107 lines · David Hildenbrand
Quoted 87 lines · Baolin Wang

Add large folio support for tmpfs write and fallocate paths matching the same high order preference mechanism used in the iomap buffered IO path as used in __filemap_get_folio().

Add shmem_mapping_size_orders() to get a hint for the orders of the folio based on the file size which takes care of the mapping requirements.

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays with other file systems supporting any sized large folios, and extending anonymous to support mTHP, we should not restrict tmpfs to allocating only PMD-sized huge folios, making it more special. Instead, we should allow tmpfs can allocate any sized large folios.

Considering that tmpfs already has the 'huge=' option to control the huge folios allocation, we can extend the 'huge=' option to allow any sized huge folios. The semantics of the 'huge=' mount option are:

huge=never: no any sized huge folios
huge=always: any sized huge folios
huge=within_size: like 'always' but respect the i_size huge=advise: like 'always' if requested with fadvise()/madvise()

Note: for tmpfs mmap() faults, due to the lack of a write size hint, still allocate the PMD-sized huge folios if huge=always/within_size/advise is set.

Moreover, the 'deny' and 'force' testing options controlled by '/sys/kernel/mm/transparent_hugepage/shmem_enabled', still retain the same semantics. The 'deny' can disable any sized large folios for tmpfs, while the 'force' can enable PMD sized large folios for tmpfs.

---
  mm/shmem.c | 91 +++++++++++++++++++++++++++++++++++++++++++++---------

  1 file changed, 77 insertions(+), 14 deletions(-)
mm/shmem.c +2 -0
index 86b2e417dc6f..a3203cf8860f 100644
@@ -549,10 +549,50 @@ static bool shmem_confirm_swap(struct 
 address_space *mapping,
   static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
+/**
+ * shmem_mapping_size_orders - Get allowable folio orders for the 

given file size.
+ * @mapping: Target address_space.
+ * @index: The page index.
+ * @write_end: end of a write, could extend inode size.
+ *
+ * This returns huge orders for folios (when supported) based on the file size
+ * which the mapping currently allows at the given index. The index is relevant
+ * due to alignment considerations the mapping might have. The returned order
+ * may be less than the size passed.
+ *
+ * Return: The orders.
+ */

mm/shmem.c +2 -0
+static inline unsigned int
+shmem_mapping_size_orders(struct address_space *mapping, pgoff_t 

index, loff_t write_end)

mm/shmem.c +21 -0
+{
+    unsigned int order;
+    size_t size;
+
+    if (!mapping_large_folio_support(mapping) || !write_end)
+        return 0;
+
+    /* Calculate the write size based on the write_end */
+    size = write_end - (index << PAGE_SHIFT);
+    order = filemap_get_order(size);
+    if (!order)
+        return 0;
+
+    /* If we're not aligned, allocate a smaller folio */
+    if (index & ((1UL << order) - 1))
+        order = __ffs(index);
+
+    order = min_t(size_t, order, MAX_PAGECACHE_ORDER);
+    return order > 0 ? BIT(order + 1) - 1 : 0;
+}
+

  static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index, loff_t write_end, bool shmem_huge_force,

Quoted 8 lines

+                          struct vm_area_struct *vma, unsigned long vm_flags) {
+    unsigned long within_size_orders;
+    unsigned int order;
+    pgoff_t aligned_index;
      loff_t i_size;
      if (HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER)

We can allow all orders up to MAX_PAGECACHE_ORDER, shmem_mapping_size_orders() handles it properly.

So maybe we should drop this condition and use instead below where we have

return BIT(HPAGE_PMD_ORDER);

instead something like.

return HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ? 0 : BIT(HPAGE_PMD_ORDER);

Quoted 8 lines

Ideally, factoring it out somehow

int maybe_pmd_order = HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ? 0 : BIT(HPAGE_PMD_ORDER);

...

return maybe_pmd_order;

Good point. Will do.

Quoted 7 lines
Quoted 5 lines
mm/shmem.c +0 -0
@@ -564,15 +604,41 @@ static unsigned int 

shmem_huge_global_enabled(struct inode *inode, pgoff_t index if (shmem_huge_force || shmem_huge == SHMEM_HUGE_FORCE) return BIT(HPAGE_PMD_ORDER);

Why not force-enable all orders (of course, respecting MAX_PAGECACHE_ORDER and possibly VMA)?

The ‘force’ option will affect the tmpfs mmap()'s huge allocation, which I intend to handle in a separate patch as we discussed. Additionally, for the huge page allocation of tmpfs mmap(), I am also considering the readahead approach for the pagecache.

Quoted 41 lines
Quoted 41 lines

+    /*
+     * The huge order allocation for anon shmem is controlled through +     * the mTHP interface, so we still use PMD-sized huge order to +     * check whether global control is enabled.
+     *
+     * For tmpfs mmap()'s huge order, we still use PMD-sized order to +     * allocate huge pages due to lack of a write size hint. +     *
+     * Otherwise, tmpfs will allow getting a highest order hint based on +     * the size of write and fallocate paths, then will try each allowable
+     * huge orders.
+     */
      switch (SHMEM_SB(inode->i_sb)->huge) {
      case SHMEM_HUGE_ALWAYS:
-        return BIT(HPAGE_PMD_ORDER);
-    case SHMEM_HUGE_WITHIN_SIZE:
-        index = round_up(index + 1, HPAGE_PMD_NR);
-        i_size = max(write_end, i_size_read(inode));
-        i_size = round_up(i_size, PAGE_SIZE);
-        if (i_size >> PAGE_SHIFT >= index)
+        if (vma)
              return BIT(HPAGE_PMD_ORDER);
+
+        return shmem_mapping_size_orders(inode->i_mapping, index, write_end);
+    case SHMEM_HUGE_WITHIN_SIZE:
+        if (vma)
+            within_size_orders = BIT(HPAGE_PMD_ORDER);
+        else
+            within_size_orders = shmem_mapping_size_orders(inode->i_mapping,
+                                       index, write_end);
+
+        order = highest_order(within_size_orders);
+        while (within_size_orders) {
+            aligned_index = round_up(index + 1, 1 << order); +            i_size = max(write_end, i_size_read(inode));
+            i_size = round_up(i_size, PAGE_SIZE);
+            if (i_size >> PAGE_SHIFT >= aligned_index)
+                return within_size_orders;

Quoted 8 lines
Quoted 6 lines

+
+            order = next_order(&within_size_orders, order); +        }
          fallthrough;
      case SHMEM_HUGE_ADVISE:
          if (vm_flags & VM_HUGEPAGE)

I think the point here is that "write" -> no VMA -> vm_flags == 0 -> no code changes needed :)

Yes. Currently the fadvise() have no HUGEPAGE handling, so I will drop the 'fadvise' in the doc.

author reviewer
#10 lore raw

PATCH 1/1 mm: shmem: add large folio support for tmpfs fix

In reply to Baolin Wang

As David suggested: "We can allow all orders up to MAX_PAGECACHE_ORDER, since shmem_mapping_size_orders() handles it properly", therefore we can drop the 'MAX_PAGECACHE_ORDER' condition.

---
 mm/shmem.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
mm/shmem.c +6 -6
index a3203cf8860f..d54b24d65193 100644
@@ -590,19 +590,19 @@ static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index
 					      struct vm_area_struct *vma,
 					      unsigned long vm_flags)
 {
+	unsigned int maybe_pmd_order = HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER ?
+		0 : BIT(HPAGE_PMD_ORDER);
 	unsigned long within_size_orders;
 	unsigned int order;
 	pgoff_t aligned_index;
 	loff_t i_size;
 
-	if (HPAGE_PMD_ORDER > MAX_PAGECACHE_ORDER)
-		return 0;
 	if (!S_ISREG(inode->i_mode))
 		return 0;
 	if (shmem_huge == SHMEM_HUGE_DENY)
 		return 0;
 	if (shmem_huge_force || shmem_huge == SHMEM_HUGE_FORCE)
-		return BIT(HPAGE_PMD_ORDER);
+		return maybe_pmd_order;
 
 	/*
 	 * The huge order allocation for anon shmem is controlled through
@@ -619,12 +619,12 @@ static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index
 	switch (SHMEM_SB(inode->i_sb)->huge) {
 	case SHMEM_HUGE_ALWAYS:
 		if (vma)
-			return BIT(HPAGE_PMD_ORDER);
+			return maybe_pmd_order;
 
 		return shmem_mapping_size_orders(inode->i_mapping, index, write_end);
 	case SHMEM_HUGE_WITHIN_SIZE:
 		if (vma)
-			within_size_orders = BIT(HPAGE_PMD_ORDER);
+			within_size_orders = maybe_pmd_order;
 		else
 			within_size_orders = shmem_mapping_size_orders(inode->i_mapping,
 								       index, write_end);
@@ -642,7 +642,7 @@ static unsigned int shmem_huge_global_enabled(struct inode *inode, pgoff_t index
 		fallthrough;
 	case SHMEM_HUGE_ADVISE:
 		if (vm_flags & VM_HUGEPAGE)
-			return BIT(HPAGE_PMD_ORDER);
+			return maybe_pmd_order;
 		fallthrough;
 	default:
 		return 0;
Signature · 1 line
2.39.3
author reviewer
#16 lore raw

PATCH 1/1 docs: tmpfs: update the huge folios policy for tmpfs and shmem fix

In reply to Baolin Wang

Drop 'fadvise()' from the doc, since fadvise() has no HUGEPAGE advise currently.

---
 Documentation/admin-guide/mm/transhuge.rst | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
Documentation/admin-guide/mm/transhuge.rst +4 -4
index ba6edff728ed..333958ef0d5f 100644
@@ -382,10 +382,10 @@ never
 
 within_size
     Only allocate huge page if it will be fully within i_size.
-    Also respect fadvise()/madvise() hints;
+    Also respect madvise() hints;
 
 advise
-    Only allocate huge pages if requested with fadvise()/madvise();
+    Only allocate huge pages if requested with madvise();
 
 Remember, that the kernel may use huge pages of all available sizes, and
 that no fine control as for the internal tmpfs mount is available.
@@ -438,10 +438,10 @@ never
 
 within_size
     Only allocate <size> huge page if it will be fully within i_size.
-    Also respect fadvise()/madvise() hints;
+    Also respect madvise() hints;
 
 advise
-    Only allocate <size> huge pages if requested with fadvise()/madvise();
+    Only allocate <size> huge pages if requested with madvise();
 
 Need of application restart
 ===========================
Signature · 1 line
2.39.3
#19 lore raw
In reply to Baolin Wang
Quoted 1 line · Baolin Wang

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays

Nitpick:
We are mixing here folios/page, PMD-size huge. For anyone not aware of Memory Folios conversion in the kernel I think this makes it confusing. Tmpfs has never supported folios so, this is not true. Can we rephrase it?

Below you are also mixing terms huge/large folios etc. Can we be consistent? I'd stick with folios (for order-0), and large folios (! order-0). I'd use huge term only when referring to PMD-size pages.

Quoted 3 lines

with other file systems supporting any sized large folios, and extending anonymous to support mTHP, we should not restrict tmpfs to allocating only PMD-sized huge folios, making it more special. Instead, we should allow

Again here.

Quoted 4 lines

tmpfs can allocate any sized large folios.

Considering that tmpfs already has the 'huge=' option to control the huge folios allocation, we can extend the 'huge=' option to allow any sized huge

'huge=' has never controlled folios.

Quoted 55 lines

folios. The semantics of the 'huge=' mount option are:

huge=never: no any sized huge folios
huge=always: any sized huge folios
huge=within_size: like 'always' but respect the i_size huge=advise: like 'always' if requested with fadvise()/madvise()

Note: for tmpfs mmap() faults, due to the lack of a write size hint, still allocate the PMD-sized huge folios if huge=always/within_size/advise is set.

Moreover, the 'deny' and 'force' testing options controlled by '/sys/kernel/mm/transparent_hugepage/shmem_enabled', still retain the same semantics. The 'deny' can disable any sized large folios for tmpfs, while the 'force' can enable PMD sized large folios for tmpfs.

Any comments and suggestions are appreciated. Thanks.

Changes from v1:
- Add reviewed tag from Barry and David. Thanks.
- Fix building warnings reported by kernel test robot.
- Add a new patch to control the default huge policy for tmpfs.

Changes from RFC v3:
- Drop the huge=write_size option.
- Allow any sized huge folios for 'hgue' option.
- Update the documentation, per David.

Changes from RFC v2:
- Drop mTHP interfaces to control huge page allocation, per Matthew.
- Add a new helper to calculate the order, suggested by Matthew.
- Add a new huge=write_size option to allocate large folios based on the write size.
- Add a new patch to update the documentation.

Changes from RFC v1:
- Drop patch 1.
- Use 'write_end' to calculate the length in shmem_allowable_huge_orders().
- Update shmem_mapping_size_order() per Daniel.

Baolin Wang (4):
mm: factor out the order calculation into a new helper
mm: shmem: change shmem_huge_global_enabled() to return huge order
bitmap
mm: shmem: add large folio support for tmpfs
mm: shmem: add a kernel command line to change the default huge policy
for tmpfs

David Hildenbrand (1):
docs: tmpfs: update the huge folios policy for tmpfs and shmem

 .../admin-guide/kernel-parameters.txt         |   7 +
 Documentation/admin-guide/mm/transhuge.rst    |  64 ++++++--
 include/linux/pagemap.h                       |  16 +-
 mm/shmem.c                                    | 148 ++++++++++++++----
 4 files changed, 183 insertions(+), 52 deletions(-)
#20 lore raw
Quoted 8 lines · Daniel Gomez
Quoted 1 line · Baolin Wang

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays

Nitpick:
We are mixing here folios/page, PMD-size huge. For anyone not aware of Memory Folios conversion in the kernel I think this makes it confusing. Tmpfs has never supported folios so, this is not true. Can we rephrase it?

We had the exact same discussion when we added mTHP support to anonymous memory.

I suggest you read:

https://lkml.kernel.org/r/[email protected]

Folios are an implementation detail on how we manage metadata. Nobody in user space should even have to be aware of how we manage metadata for larger chunks of memory ("huge pages") in the kernel.

Signature · 3 lines
Cheers,

David / dhildenb
#9 lore raw

Re: [PATCH v2 3/5] mm: shmem: add large folio support for tmpfs

In reply to Baolin Wang
Quoted 10 lines
Quoted 5 lines
Quoted 2 lines

          return BIT(HPAGE_PMD_ORDER);

Why not force-enable all orders (of course, respecting MAX_PAGECACHE_ORDER and possibly VMA)?

The ‘force’ option will affect the tmpfs mmap()'s huge allocation, which I intend to handle in a separate patch as we discussed. Additionally, for the huge page allocation of tmpfs mmap(), I am also considering the readahead approach for the pagecache.

Okay, we can change this later. Likely force/deny are a blast from the past either way.

[...]

Quoted 18 lines
Quoted 9 lines
Quoted 9 lines
mm/shmem.c +9 -0
+
+        order = highest_order(within_size_orders);
+        while (within_size_orders) {
+            aligned_index = round_up(index + 1, 1 << order);
+            i_size = max(write_end, i_size_read(inode));
+            i_size = round_up(i_size, PAGE_SIZE);
+            if (i_size >> PAGE_SHIFT >= aligned_index)
+                return within_size_orders;
+
Quoted 8 lines
Quoted 5 lines

+            order = next_order(&within_size_orders, order); +        }
          fallthrough;
      case SHMEM_HUGE_ADVISE:
          if (vm_flags & VM_HUGEPAGE)

I think the point here is that "write" -> no VMA -> vm_flags == 0 -> no code changes needed :)

Quoted 2 lines

Yes. Currently the fadvise() have no HUGEPAGE handling, so I will drop the 'fadvise' in the doc.

Interesting that we documented it :)

Signature · 3 lines
Cheers,

David / dhildenb
Reviewed-by on v2 1/5
#2 lore raw

Re: [PATCH v2 1/5] mm: factor out the order calculation into a new helper

In reply to Baolin Wang
Quoted 7 lines · Baolin Wang

Factor out the order calculation into a new helper, which can be reused by shmem in the following patch.

Quoted 40 lines
---
 include/linux/pagemap.h | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
include/linux/pagemap.h +13 -3
index bcf0865a38ae..d796c8a33647 100644
@@ -727,6 +727,16 @@ typedef unsigned int __bitwise fgf_t;
 
 #define FGP_WRITEBEGIN		(FGP_LOCK | FGP_WRITE | FGP_CREAT | FGP_STABLE)
 
+static inline unsigned int filemap_get_order(size_t size)
+{
+	unsigned int shift = ilog2(size);
+
+	if (shift <= PAGE_SHIFT)
+		return 0;
+
+	return shift - PAGE_SHIFT;
+}
+
 /**
  * fgf_set_order - Encode a length in the fgf_t flags.
  * @size: The suggested size of the folio to create.
@@ -740,11 +750,11 @@ typedef unsigned int __bitwise fgf_t;
  */
 static inline fgf_t fgf_set_order(size_t size)
 {
-	unsigned int shift = ilog2(size);
+	unsigned int order = filemap_get_order(size);
 
-	if (shift <= PAGE_SHIFT)
+	if (!order)
 		return 0;
-	return (__force fgf_t)((shift - PAGE_SHIFT) << 26);
+	return (__force fgf_t)(order << 26);
 }
 
 void *filemap_get_entry(struct address_space *mapping, pgoff_t index);
#12 lore raw

Re: [PATCH v2 4/5] mm: shmem: add a kernel command line to change the default huge policy for tmpfs

In reply to Baolin Wang
Quoted 4 lines · Baolin Wang

Now the tmpfs can allow to allocate any sized large folios, and the default huge policy is still 'never'. Thus adding a new command line to change the default huge policy will be helpful to use the large folios for tmpfs, which is similar to the 'transparent_hugepage_shmem' cmdline for shmem.

I think it would be good to include a summary of why tmpfs is not enabling large folios by default as the other fs. David has been pretty good at repeating the reasons over and over and it would be very valuable to have them included here.

Quoted 91 lines
---
 .../admin-guide/kernel-parameters.txt         |  7 ++++++
 Documentation/admin-guide/mm/transhuge.rst    |  6 +++++
 mm/shmem.c                                    | 23 ++++++++++++++++++-
 3 files changed, 35 insertions(+), 1 deletion(-)
Documentation/admin-guide/kernel-parameters.txt +7 -0
index b48d744d99b0..007e6cfada3e 100644
@@ -6943,6 +6943,13 @@
 			See Documentation/admin-guide/mm/transhuge.rst
 			for more details.
 
+	transparent_hugepage_tmpfs= [KNL]
+			Format: [always|within_size|advise|never]
+			Can be used to control the default hugepage allocation policy
+			for the tmpfs mount.
+			See Documentation/admin-guide/mm/transhuge.rst
+			for more details.
+
 	trusted.source=	[KEYS]
 			Format: <string>
 			This parameter identifies the trust source as a backend
Documentation/admin-guide/mm/transhuge.rst +6 -0
index 5034915f4e8e..9ae775eaacbe 100644
@@ -332,6 +332,12 @@ allocation policy for the internal shmem mount by using the kernel parameter
 seven valid policies for shmem (``always``, ``within_size``, ``advise``,
 ``never``, ``deny``, and ``force``).
 
+Similarly to ``transparent_hugepage_shmem``, you can control the default
+hugepage allocation policy for the tmpfs mount by using the kernel parameter
+``transparent_hugepage_tmpfs=<policy>``, where ``<policy>`` is one of the
+four valid policies for tmpfs (``always``, ``within_size``, ``advise``,
+``never``). The tmpfs mount default policy is ``never``.
+
 In the same manner as ``thp_anon`` controls each supported anonymous THP
 size, ``thp_shmem`` controls each supported shmem THP size. ``thp_shmem``
 has the same format as ``thp_anon``, but also supports the policy
mm/shmem.c +22 -1
index a3203cf8860f..021760e91cea 100644
@@ -548,6 +548,7 @@ static bool shmem_confirm_swap(struct address_space *mapping,
 /* ifdef here to avoid bloating shmem.o when not necessary */
 
 static int shmem_huge __read_mostly = SHMEM_HUGE_NEVER;
+static int tmpfs_huge __read_mostly = SHMEM_HUGE_NEVER;
 
 /**
  * shmem_mapping_size_orders - Get allowable folio orders for the given file size.
@@ -4780,7 +4781,12 @@ static int shmem_fill_super(struct super_block *sb, struct fs_context *fc)
 	sbinfo->gid = ctx->gid;
 	sbinfo->full_inums = ctx->full_inums;
 	sbinfo->mode = ctx->mode;
-	sbinfo->huge = ctx->huge;
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+	if (ctx->seen & SHMEM_SEEN_HUGE)
+		sbinfo->huge = ctx->huge;
+	else
+		sbinfo->huge = tmpfs_huge;
+#endif
 	sbinfo->mpol = ctx->mpol;
 	ctx->mpol = NULL;
 
@@ -5259,6 +5265,21 @@ static int __init setup_transparent_hugepage_shmem(char *str)
 }
 __setup("transparent_hugepage_shmem=", setup_transparent_hugepage_shmem);
 
+static int __init setup_transparent_hugepage_tmpfs(char *str)
+{
+	int huge;
+
+	huge = shmem_parse_huge(str);
+	if (huge < 0) {
+		pr_warn("transparent_hugepage_tmpfs= cannot parse, ignored\n");
+		return huge;
+	}
+
+	tmpfs_huge = huge;
+	return 1;
+}
+__setup("transparent_hugepage_tmpfs=", setup_transparent_hugepage_tmpfs);
+
 static char str_dup[PAGE_SIZE] __initdata;
 static int __init setup_thp_shmem(char *str)
 {
#13 lore raw

Re: [PATCH v2 4/5] mm: shmem: add a kernel command line to change the default huge policy for tmpfs

Quoted 11 lines · Daniel Gomez
Quoted 4 lines · Baolin Wang

Now the tmpfs can allow to allocate any sized large folios, and the default huge policy is still 'never'. Thus adding a new command line to change the default huge policy will be helpful to use the large folios for tmpfs, which is similar to the 'transparent_hugepage_shmem' cmdline for shmem.

I think it would be good to include a summary of why tmpfs is not enabling large folios by default as the other fs. David has been pretty good at repeating the reasons over and over and it would be very valuable to have them included here.

Yes. We also discussed in v4 the idea of having a Kconfig option to just change the default policy to "always". We could mention that here as well.

Signature · 3 lines
Cheers,

David / dhildenb
#21 lore raw
In reply to David Hildenbrand
Quoted 13 lines · David Hildenbrand
Quoted 8 lines · Daniel Gomez
Quoted 1 line · Baolin Wang

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays

Nitpick:
We are mixing here folios/page, PMD-size huge. For anyone not aware of Memory Folios conversion in the kernel I think this makes it confusing. Tmpfs has never supported folios so, this is not true. Can we rephrase it?

We had the exact same discussion when we added mTHP support to anonymous memory.

Quoted 7 lines

I suggest you read:

https://lkml.kernel.org/r/[email protected]

Folios are an implementation detail on how we manage metadata. Nobody in user space should even have to be aware of how we manage metadata for larger chunks of memory ("huge pages") in the kernel.

I read it and I can't find where the use of "PMD-size huge folios" could be a valid term. Tmpfs has never supported "folios", so I think using "PMD-size huge pages" is more appropiate.

#22 lore raw
Quoted 18 lines · Daniel Gomez
Quoted 17 lines · David Hildenbrand
Quoted 8 lines · Daniel Gomez
Quoted 1 line · Baolin Wang

Traditionally, tmpfs only supported PMD-sized huge folios. However nowadays

Nitpick:
We are mixing here folios/page, PMD-size huge. For anyone not aware of Memory Folios conversion in the kernel I think this makes it confusing. Tmpfs has never supported folios so, this is not true. Can we rephrase it?

We had the exact same discussion when we added mTHP support to anonymous memory.

I suggest you read:

https://lkml.kernel.org/r/[email protected]

Quoted 7 lines
Quoted 3 lines

Folios are an implementation detail on how we manage metadata. Nobody in user space should even have to be aware of how we manage metadata for larger chunks of memory ("huge pages") in the kernel.

I read it and I can't find where the use of "PMD-size huge folios" could be a valid term. Tmpfs has never supported "folios", so I think using "PMD-size huge pages" is more appropiate.

Oh sorry, I completely agree. Yes, we should use that.

Signature · 3 lines
Cheers,

David / dhildenb
author reviewer
#14 lore raw

Re: [PATCH v2 4/5] mm: shmem: add a kernel command line to change the default huge policy for tmpfs

In reply to David Hildenbrand
Quoted 8 lines · David Hildenbrand
Quoted 7 lines · Daniel Gomez
Quoted 4 lines · Baolin Wang

Now the tmpfs can allow to allocate any sized large folios, and the default huge policy is still 'never'. Thus adding a new command line to change the default huge policy will be helpful to use the large folios for tmpfs, which is similar to the 'transparent_hugepage_shmem' cmdline for shmem.

Quoted 4 lines
Quoted 4 lines

I think it would be good to include a summary of why tmpfs is not enabling large folios by default as the other fs. David has been pretty good at repeating the reasons over and over and it would be very valuable to have them included here.

OK. I'd like to directly quote David's previous comments. So hope Andew can help include the updated commit message:

=====
Now the tmpfs can allow to allocate any sized large folios, and the default huge policy is still prefered to be 'never'. Due to tmpfs not behaving like other file systems in some cases as previously explained by David[1]:

"
I think I raised this in the past, but tmpfs/shmem is just like any other file system .. except it sometimes really isn't and behaves much more like (swappable) anonymous memory. (or mlocked files)

There are many systems out there that run without swap enabled, or with extremely minimal swap (IIRC until recently kubernetes was completely incompatible with swapping). Swap can even be disabled today for shmem using a mount option.

That's a big difference to all other file systems where you are guaranteed to have backend storage where you can simply evict under memory pressure (might temporarily fail, of course).

I *think* that's the reason why we have the "huge=" parameter that also controls the THP allocations during page faults (IOW possible memory over-allocation). Maybe also because it was a new feature, and we only had a single THP size.
"

Thus adding a new command line to change the default huge policy will be helpful to use the large folios for tmpfs, which is similar to the 'transparent_hugepage_shmem' cmdline for shmem.

[1] https://lore.kernel.org/all/[email protected]/

Quoted 1 line

Yes. We also discussed in v4 the idea of having a Kconfig option to just change the default policy to "always". We could mention that here as well.

Reviewed-by on v1 1/1
#17 lore raw

Re: [PATCH] docs: tmpfs: update the huge folios policy for tmpfs and shmem fix

In reply to Baolin Wang
Quoted 5 lines · Baolin Wang

Drop 'fadvise()' from the doc, since fadvise() has no HUGEPAGE advise currently.

I couldn’t find any mention of HUGEPAGE in fadvise() either.

FADV_NORMAL
FADV_RANDOM
FADV_SEQUENTIAL
FADV_WILLNEED
FADV_DONTNEED
FADV_NOREUSE

Quoted 37 lines
---
 Documentation/admin-guide/mm/transhuge.rst | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
Documentation/admin-guide/mm/transhuge.rst +4 -5
index ba6edff728ed..333958ef0d5f 100644
@@ -382,10 +382,10 @@ never
 
 within_size
     Only allocate huge page if it will be fully within i_size.
-    Also respect fadvise()/madvise() hints;
+    Also respect madvise() hints;
 
 advise
-    Only allocate huge pages if requested with fadvise()/madvise();
+    Only allocate huge pages if requested with madvise();
 
 Remember, that the kernel may use huge pages of all available sizes, and
 that no fine control as for the internal tmpfs mount is available.
@@ -438,10 +438,10 @@ never
 
 within_size
     Only allocate <size> huge page if it will be fully within i_size.
-    Also respect fadvise()/madvise() hints;
+    Also respect madvise() hints;
 
 advise
-    Only allocate <size> huge pages if requested with fadvise()/madvise();
+    Only allocate <size> huge pages if requested with madvise();
 
 Need of application restart
 ===========================
--

2.39.3

#18 lore raw

Re: [PATCH] docs: tmpfs: update the huge folios policy for tmpfs and shmem fix

Quoted 10 lines · Barry Song
Quoted 5 lines · Baolin Wang

Drop 'fadvise()' from the doc, since fadvise() has no HUGEPAGE advise currently.

Quoted 8 lines

I couldn’t find any mention of HUGEPAGE in fadvise() either.

FADV_NORMAL
FADV_RANDOM
FADV_SEQUENTIAL
FADV_WILLNEED
FADV_DONTNEED
FADV_NOREUSE

Probably it was forward-looking, and that change never happened.

Signature · 3 lines
Cheers,

David / dhildenb