[PATCH 0/3] f2fs: support encrypted inline data

[PATCH 0/3] f2fs: support encrypted inline data

[LiaoYuanhong-vivo]

From: Liao Yuanhong <liaoyuanhong@vivo.com>

F2FS currently avoids inline data for encrypted regular files.  This is
because inline data is stored in the inode block, outside the regular
bio-based data path where fscrypt and blk-crypto normally operate.
As a result, devices that enable blk-crypto for encrypted file contents
cannot use F2FS inline data for encrypted regular files, which wastes
space for small files.

This series adds support for keeping small encrypted regular-file
contents as inline data.  The f2fs side defines a new on-disk feature,
encrypted_inline_data, under which inline payloads of encrypted regular
files are interpreted as ciphertext.  The payload is encrypted before
being stored in the inode block and decrypted back into page-cache
plaintext on read.

The fscrypt side prepares a software contents-key transform even when
normal file contents use blk-crypto, so filesystems can encrypt
filesystem-managed data regions that do not go through bio submission.
The new fscrypt helper operates on fscrypt data units and leaves the
filesystem responsible for deciding which filesystem-managed byte ranges
need this treatment.

The software crypto operation is limited to the inline payload.  Since
these files are small enough to remain inline, the expected read/write
performance difference between hardware and software crypto is small,
while the space saving from keeping the data inline is significant.

The feature is guarded by CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA and by the
F2FS encrypted_inline_data on-disk feature bit.  Filesystems with this
feature set are rejected if the kernel lacks the config option.

Hardware-wrapped keys are not supported by this initial version. I would
like to discuss whether this feature should remain disabled for
hardware-wrapped keys, or whether there is an acceptable way to support the
combination in the future.

The f2fs-tools support for formatting filesystems with this feature will be
submitted separately.

Basic testing passed.  Encrypted small files can be kept as inline data,
and read/write verification succeeded.

Liao Yuanhong (3):
  fscrypt: prepare software keys for filesystem-managed data units
  f2fs: support encrypted inline data
  Documentation: f2fs: document encrypted inline data

 Documentation/ABI/testing/sysfs-fs-f2fs |   5 +-
 Documentation/filesystems/f2fs.rst      |  27 ++++++
 fs/crypto/crypto.c                      |  63 +++++++++++++
 fs/crypto/fscrypt_private.h             |   3 +-
 fs/crypto/keysetup.c                    |  59 +++++++++---
 fs/f2fs/Kconfig                         |  14 +++
 fs/f2fs/data.c                          |   8 +-
 fs/f2fs/f2fs.h                          |  37 +++++++-
 fs/f2fs/file.c                          |  24 ++++-
 fs/f2fs/inline.c                        | 119 +++++++++++++++++++++---
 fs/f2fs/super.c                         |  12 +++
 fs/f2fs/sysfs.c                         |   8 ++
 include/linux/fscrypt.h                 |  28 ++++++
 13 files changed, 370 insertions(+), 37 deletions(-)

-- 
2.34.1

[PATCH 1/3] fscrypt: prepare software keys for filesystem-managed data units

[LiaoYuanhong-vivo]

From: Liao Yuanhong <liaoyuanhong@vivo.com>

Some filesystems store small data regions inside filesystem metadata rather
than submitting them through the normal bio path.  F2FS inline data is one
such case.  When an encrypted file uses blk-crypto, these regions still
need software fscrypt handling because no data bio is submitted for them.

Add fscrypt_encrypt_data_unit_inplace() and
fscrypt_decrypt_data_unit_inplace().  They use the same data-unit crypto
path as fscrypt_encrypt_block_inplace() and
fscrypt_decrypt_block_inplace(), but take a data-unit index instead of a
filesystem logical block number.

Also add fscrypt_inode_supports_data_unit_inplace() so filesystems can
check whether an inode has a software transform available for this path.

Factor the software skcipher setup into fscrypt_prepare_software_key().
The existing fscrypt_prepare_key() path now reuses it, and inline-crypto
key setup can use it in addition to preparing the blk-crypto key.  This
lets an inline-crypto inode keep its normal blk-crypto contents path while
also having a software transform for filesystem-managed data units.

Signed-off-by: Liao Yuanhong <liaoyuanhong@vivo.com>

---
 fs/crypto/crypto.c          | 63 +++++++++++++++++++++++++++++++++++++
 fs/crypto/fscrypt_private.h |  3 +-
 fs/crypto/keysetup.c        | 59 +++++++++++++++++++++++++---------
 include/linux/fscrypt.h     | 28 +++++++++++++++++
 4 files changed, 138 insertions(+), 15 deletions(-)

diff --git a/fs/crypto/crypto.c b/fs/crypto/crypto.c
index 570a2231c945..c43acbc8b4ea 100644
--- a/fs/crypto/crypto.c
+++ b/fs/crypto/crypto.c
@@ -208,6 +208,44 @@ struct page *fscrypt_encrypt_pagecache_blocks(struct folio *folio,
 }
 EXPORT_SYMBOL(fscrypt_encrypt_pagecache_blocks);
 
+/**
+ * fscrypt_encrypt_data_unit_inplace() - Encrypt a data unit in-place
+ * @inode: The inode to which this data unit belongs
+ * @page:  The page containing the data unit to encrypt
+ * @len:   Size of data unit to encrypt.  This must be a multiple of
+ *	   FSCRYPT_CONTENTS_ALIGNMENT.
+ * @offs:  Byte offset within @page at which the data unit begins
+ * @index: Fscrypt data unit index within the file
+ *
+ * Return: 0 on success; -errno on failure
+ */
+int fscrypt_encrypt_data_unit_inplace(const struct inode *inode,
+				      struct page *page, unsigned int len,
+				      unsigned int offs, u64 index)
+{
+	const struct fscrypt_inode_info *ci = fscrypt_get_inode_info_raw(inode);
+
+	if (!fscrypt_inode_supports_data_unit_inplace(inode))
+		return -EOPNOTSUPP;
+
+	return fscrypt_crypt_data_unit(ci, FS_ENCRYPT, index, page, page, len,
+				       offs);
+}
+EXPORT_SYMBOL(fscrypt_encrypt_data_unit_inplace);
+
+bool fscrypt_inode_supports_data_unit_inplace(const struct inode *inode)
+{
+	const struct fscrypt_inode_info *ci = fscrypt_get_inode_info_raw(inode);
+
+	if (!IS_ENABLED(CONFIG_FS_ENCRYPTION_INLINE_CRYPT))
+		return false;
+	if (!ci)
+		return false;
+	/* pairs with smp_store_release() in fscrypt_prepare_software_key() */
+	return smp_load_acquire(&ci->ci_enc_key.tfm);
+}
+EXPORT_SYMBOL(fscrypt_inode_supports_data_unit_inplace);
+
 /**
  * fscrypt_encrypt_block_inplace() - Encrypt a filesystem block in-place
  * @inode:     The inode to which this block belongs
@@ -282,6 +320,31 @@ int fscrypt_decrypt_pagecache_blocks(struct folio *folio, size_t len,
 }
 EXPORT_SYMBOL(fscrypt_decrypt_pagecache_blocks);
 
+/**
+ * fscrypt_decrypt_data_unit_inplace() - Decrypt a data unit in-place
+ * @inode: The inode to which this data unit belongs
+ * @page:  The page containing the data unit to decrypt
+ * @len:   Size of data unit to decrypt.  This must be a multiple of
+ *	   FSCRYPT_CONTENTS_ALIGNMENT.
+ * @offs:  Byte offset within @page at which the data unit begins
+ * @index: Fscrypt data unit index within the file
+ *
+ * Return: 0 on success; -errno on failure
+ */
+int fscrypt_decrypt_data_unit_inplace(const struct inode *inode,
+				      struct page *page, unsigned int len,
+				      unsigned int offs, u64 index)
+{
+	const struct fscrypt_inode_info *ci = fscrypt_get_inode_info_raw(inode);
+
+	if (!fscrypt_inode_supports_data_unit_inplace(inode))
+		return -EOPNOTSUPP;
+
+	return fscrypt_crypt_data_unit(ci, FS_DECRYPT, index, page, page, len,
+				       offs);
+}
+EXPORT_SYMBOL(fscrypt_decrypt_data_unit_inplace);
+
 /**
  * fscrypt_decrypt_block_inplace() - Decrypt a filesystem block in-place
  * @inode:     The inode to which this block belongs
diff --git a/fs/crypto/fscrypt_private.h b/fs/crypto/fscrypt_private.h
index 8d3c278a7591..b5c0b881fd4b 100644
--- a/fs/crypto/fscrypt_private.h
+++ b/fs/crypto/fscrypt_private.h
@@ -236,7 +236,8 @@ struct fscrypt_symlink_data {
  * @tfm: crypto API transform object
  * @blk_key: key for blk-crypto
  *
- * Normally only one of the fields will be non-NULL.
+ * Most users need only one prepared form.  Inline-crypto users that also need
+ * filesystem-layer software crypto for non-bio data regions may prepare both.
  */
 struct fscrypt_prepared_key {
 	struct crypto_sync_skcipher *tfm;
diff --git a/fs/crypto/keysetup.c b/fs/crypto/keysetup.c
index ce327bfdada4..716911b97d8e 100644
--- a/fs/crypto/keysetup.c
+++ b/fs/crypto/keysetup.c
@@ -144,6 +144,23 @@ fscrypt_allocate_skcipher(struct fscrypt_mode *mode, const u8 *raw_key,
 	return ERR_PTR(err);
 }
 
+static int fscrypt_prepare_software_key(struct fscrypt_prepared_key *prep_key,
+					const u8 *raw_key,
+					const struct fscrypt_inode_info *ci)
+{
+	struct crypto_sync_skcipher *tfm;
+
+	/* pairs with smp_store_release() below */
+	if (smp_load_acquire(&prep_key->tfm))
+		return 0;
+	tfm = fscrypt_allocate_skcipher(ci->ci_mode, raw_key, ci->ci_inode);
+	if (IS_ERR(tfm))
+		return PTR_ERR(tfm);
+	/* pairs with smp_load_acquire() above */
+	smp_store_release(&prep_key->tfm, tfm);
+	return 0;
+}
+
 /*
  * Prepare the crypto transform object or blk-crypto key in @prep_key, given the
  * raw key, encryption mode (@ci->ci_mode), flag indicating which encryption
@@ -153,24 +170,12 @@ fscrypt_allocate_skcipher(struct fscrypt_mode *mode, const u8 *raw_key,
 int fscrypt_prepare_key(struct fscrypt_prepared_key *prep_key,
 			const u8 *raw_key, const struct fscrypt_inode_info *ci)
 {
-	struct crypto_sync_skcipher *tfm;
-
 	if (fscrypt_using_inline_encryption(ci))
 		return fscrypt_prepare_inline_crypt_key(prep_key, raw_key,
 							ci->ci_mode->keysize,
 							false, ci);
 
-	tfm = fscrypt_allocate_skcipher(ci->ci_mode, raw_key, ci->ci_inode);
-	if (IS_ERR(tfm))
-		return PTR_ERR(tfm);
-	/*
-	 * Pairs with the smp_load_acquire() in fscrypt_is_key_prepared().
-	 * I.e., here we publish ->tfm with a RELEASE barrier so that
-	 * concurrent tasks can ACQUIRE it.  Note that this concurrency is only
-	 * possible for per-mode keys, not for per-file keys.
-	 */
-	smp_store_release(&prep_key->tfm, tfm);
-	return 0;
+	return fscrypt_prepare_software_key(prep_key, raw_key, ci);
 }
 
 /* Destroy a crypto transform object and/or blk-crypto key. */
@@ -190,6 +195,20 @@ int fscrypt_set_per_file_enc_key(struct fscrypt_inode_info *ci,
 	return fscrypt_prepare_key(&ci->ci_enc_key, raw_key, ci);
 }
 
+static int
+fscrypt_prepare_inline_crypt_and_software_key(struct fscrypt_prepared_key *prep_key,
+					      const u8 *raw_key,
+					      const struct fscrypt_inode_info *ci)
+{
+	int err;
+
+	err = fscrypt_prepare_software_key(prep_key, raw_key, ci);
+	if (err)
+		return err;
+	return fscrypt_prepare_inline_crypt_key(prep_key, raw_key,
+					       ci->ci_mode->keysize, false, ci);
+}
+
 static int setup_per_mode_enc_key(struct fscrypt_inode_info *ci,
 				  struct fscrypt_master_key *mk,
 				  struct fscrypt_prepared_key *keys,
@@ -255,7 +274,16 @@ static int setup_per_mode_enc_key(struct fscrypt_inode_info *ci,
 	}
 	fscrypt_hkdf_expand(&mk->mk_secret.hkdf, hkdf_context, hkdf_info,
 			    hkdf_infolen, mode_key, mode->keysize);
-	err = fscrypt_prepare_key(prep_key, mode_key, ci);
+	if (!use_hw_wrapped_key && fscrypt_using_inline_encryption(ci)) {
+		/*
+		 * Filesystem-managed regions such as F2FS inline_data need the
+		 * same contents key as a software tfm.
+		 */
+		err = fscrypt_prepare_inline_crypt_and_software_key(prep_key,
+								    mode_key, ci);
+	} else {
+		err = fscrypt_prepare_key(prep_key, mode_key, ci);
+	}
 	memzero_explicit(mode_key, mode->keysize);
 	if (err)
 		goto out_unlock;
@@ -381,6 +409,7 @@ static int fscrypt_setup_v2_file_key(struct fscrypt_inode_info *ci,
 		   FSCRYPT_POLICY_FLAG_IV_INO_LBLK_32) {
 		err = fscrypt_setup_iv_ino_lblk_32_key(ci, mk);
 	} else {
+		struct fscrypt_prepared_key *prep_key = &ci->ci_enc_key;
 		u8 derived_key[FSCRYPT_MAX_RAW_KEY_SIZE];
 
 		fscrypt_hkdf_expand(&mk->mk_secret.hkdf,
@@ -388,6 +417,8 @@ static int fscrypt_setup_v2_file_key(struct fscrypt_inode_info *ci,
 				    ci->ci_nonce, FSCRYPT_FILE_NONCE_SIZE,
 				    derived_key, ci->ci_mode->keysize);
 		err = fscrypt_set_per_file_enc_key(ci, derived_key);
+		if (!err && fscrypt_using_inline_encryption(ci))
+			err = fscrypt_prepare_software_key(prep_key, derived_key, ci);
 		memzero_explicit(derived_key, ci->ci_mode->keysize);
 	}
 	if (err)
diff --git a/include/linux/fscrypt.h b/include/linux/fscrypt.h
index 54712ec61ffb..3762a7526fcc 100644
--- a/include/linux/fscrypt.h
+++ b/include/linux/fscrypt.h
@@ -346,12 +346,20 @@ void fscrypt_enqueue_decrypt_work(struct work_struct *);
 
 struct page *fscrypt_encrypt_pagecache_blocks(struct folio *folio,
 		size_t len, size_t offs, gfp_t gfp_flags);
+
+int fscrypt_encrypt_data_unit_inplace(const struct inode *inode,
+				      struct page *page, unsigned int len,
+				      unsigned int offs, u64 index);
+bool fscrypt_inode_supports_data_unit_inplace(const struct inode *inode);
 int fscrypt_encrypt_block_inplace(const struct inode *inode, struct page *page,
 				  unsigned int len, unsigned int offs,
 				  u64 lblk_num);
 
 int fscrypt_decrypt_pagecache_blocks(struct folio *folio, size_t len,
 				     size_t offs);
+int fscrypt_decrypt_data_unit_inplace(const struct inode *inode,
+				      struct page *page, unsigned int len,
+				      unsigned int offs, u64 index);
 int fscrypt_decrypt_block_inplace(const struct inode *inode, struct page *page,
 				  unsigned int len, unsigned int offs,
 				  u64 lblk_num);
@@ -519,6 +527,19 @@ static inline struct page *fscrypt_encrypt_pagecache_blocks(struct folio *folio,
 	return ERR_PTR(-EOPNOTSUPP);
 }
 
+static inline int fscrypt_encrypt_data_unit_inplace(const struct inode *inode,
+						    struct page *page, unsigned int len,
+						    unsigned int offs, u64 index)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline bool
+fscrypt_inode_supports_data_unit_inplace(const struct inode *inode)
+{
+	return false;
+}
+
 static inline int fscrypt_encrypt_block_inplace(const struct inode *inode,
 						struct page *page,
 						unsigned int len,
@@ -533,6 +554,13 @@ static inline int fscrypt_decrypt_pagecache_blocks(struct folio *folio,
 	return -EOPNOTSUPP;
 }
 
+static inline int fscrypt_decrypt_data_unit_inplace(const struct inode *inode,
+						    struct page *page, unsigned int len,
+						    unsigned int offs, u64 index)
+{
+	return -EOPNOTSUPP;
+}
+
 static inline int fscrypt_decrypt_block_inplace(const struct inode *inode,
 						struct page *page,
 						unsigned int len,
-- 
2.34.1

[PATCH 2/3] f2fs: support encrypted inline data

[LiaoYuanhong-vivo]

From: Liao Yuanhong <liaoyuanhong@vivo.com>

F2FS normally disables inline data for encrypted regular files because
inline data is stored in the inode block and does not pass through the
regular fscrypt data I/O path.  This wastes space for small encrypted
files on filesystems that otherwise use inline_data, including devices
that enable blk-crypto for encrypted file contents.

Add encrypted inline data support for encrypted regular files.  Inline
payloads are decrypted into page-cache plaintext on read and encrypted
before being copied back into the inode block.

The inline read path decrypts the inode payload through a temporary page.
The inline write path encrypts page-cache plaintext before storing it back
into the inode block.

Update the inline-data size checks to use the encrypted inline capacity,
since the stored payload is rounded to the fscrypt contents alignment.  If
an encrypted inline-data file is truncated from a non-zero offset, convert
it to normal data blocks first and then use the normal truncate path.

Only the inline payload is encrypted in software.  These files are small
enough to remain in the inode block, so the performance difference
between hardware and software crypto is expected to be small compared to
the space saved by retaining inline data.

Signed-off-by: Liao Yuanhong <liaoyuanhong@vivo.com>

---
 fs/f2fs/Kconfig  |  14 ++++++
 fs/f2fs/data.c   |   8 ++--
 fs/f2fs/f2fs.h   |  37 ++++++++++++++-
 fs/f2fs/file.c   |  24 +++++++++-
 fs/f2fs/inline.c | 119 +++++++++++++++++++++++++++++++++++++++++------
 fs/f2fs/super.c  |  12 +++++
 fs/f2fs/sysfs.c  |   8 ++++
 7 files changed, 202 insertions(+), 20 deletions(-)

diff --git a/fs/f2fs/Kconfig b/fs/f2fs/Kconfig
index 5916a02fb46d..9e31923b8df8 100644
--- a/fs/f2fs/Kconfig
+++ b/fs/f2fs/Kconfig
@@ -92,6 +92,20 @@ config F2FS_FAULT_INJECTION
 
 	  If unsure, say N.
 
+config F2FS_FS_ENCRYPTED_INLINE_DATA
+	bool "F2FS encrypted inline data support"
+	depends on F2FS_FS && FS_ENCRYPTION_INLINE_CRYPT
+	help
+	  Allow encrypted regular files to keep inline data inside the inode
+	  while encrypting that inode-managed payload in software.
+
+	  This does not change normal data block encryption.  Normal data
+	  blocks continue to use the existing fscrypt path, such as blk-crypto
+	  when inline encryption is enabled.
+
+	  Filesystems carrying the encrypted_inline_data incompat feature
+	  require this option in order to be mounted correctly.
+
 config F2FS_FS_COMPRESSION
 	bool "F2FS compression feature"
 	depends on F2FS_FS
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 657fd5986c73..9371ffa7c96d 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -3700,7 +3700,7 @@ static int prepare_write_begin(struct f2fs_sb_info *sbi,
 
 	/* f2fs_lock_op avoids race between write CP and convert_inline_page */
 	if (f2fs_has_inline_data(inode)) {
-		if (pos + len > MAX_INLINE_DATA(inode))
+		if (pos + len > f2fs_max_inline_data(inode))
 			flag = F2FS_GET_BLOCK_DEFAULT;
 		f2fs_map_lock(sbi, &lc, flag);
 		locked = true;
@@ -3720,8 +3720,10 @@ static int prepare_write_begin(struct f2fs_sb_info *sbi,
 	set_new_dnode(&dn, inode, ifolio, ifolio, 0);
 
 	if (f2fs_has_inline_data(inode)) {
-		if (pos + len <= MAX_INLINE_DATA(inode)) {
-			f2fs_do_read_inline_data(folio, ifolio);
+		if (pos + len <= f2fs_max_inline_data(inode)) {
+			err = f2fs_do_read_inline_data(folio, ifolio);
+			if (err)
+				goto out;
 			set_inode_flag(inode, FI_DATA_EXIST);
 			if (inode->i_nlink)
 				folio_set_f2fs_inline(ifolio);
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 832b2f8beb11..0a2d75baf23e 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -276,6 +276,7 @@ struct f2fs_mount_info {
 #define F2FS_FEATURE_RO				0x00004000
 #define F2FS_FEATURE_DEVICE_ALIAS		0x00008000
 #define F2FS_FEATURE_PACKED_SSA			0x00010000
+#define F2FS_FEATURE_ENCRYPTED_INLINE_DATA	0x00020000
 
 #define __F2FS_HAS_FEATURE(raw_super, mask)				\
 	((raw_super->feature & cpu_to_le32(mask)) != 0)
@@ -4502,7 +4503,7 @@ extern struct kmem_cache *f2fs_inode_entry_slab;
 bool f2fs_may_inline_data(struct inode *inode);
 bool f2fs_sanity_check_inline_data(struct inode *inode, struct folio *ifolio);
 bool f2fs_may_inline_dentry(struct inode *inode);
-void f2fs_do_read_inline_data(struct folio *folio, struct folio *ifolio);
+int f2fs_do_read_inline_data(struct folio *folio, struct folio *ifolio);
 void f2fs_truncate_inline_inode(struct inode *inode, struct folio *ifolio,
 		u64 from);
 int f2fs_read_inline_data(struct inode *inode, struct folio *folio);
@@ -4595,6 +4596,39 @@ static inline bool f2fs_encrypted_file(struct inode *inode)
 	return IS_ENCRYPTED(inode) && S_ISREG(inode->i_mode);
 }
 
+static inline bool f2fs_sb_has_encrypted_inline_data(struct f2fs_sb_info *sbi);
+
+static inline bool f2fs_uses_encrypted_inline_data(struct inode *inode)
+{
+#ifdef CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA
+	/*
+	 * When the filesystem allows encrypted inline data, inline payloads
+	 * in encrypted regular files are interpreted as ciphertext.
+	 */
+	return f2fs_sb_has_encrypted_inline_data(F2FS_I_SB(inode)) &&
+	       f2fs_encrypted_file(inode);
+#else
+	return false;
+#endif
+}
+
+static inline unsigned int f2fs_max_inline_data(struct inode *inode)
+{
+	unsigned int max_bytes = MAX_INLINE_DATA(inode);
+
+	/*
+	 * Encrypted inline data is rounded up to the fscrypt contents
+	 * alignment before being stored back into the inode.  This is an
+	 * on-disk layout constraint, so it must not depend on whether the
+	 * inode's key has been prepared yet.
+	 */
+#ifdef CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA
+	if (f2fs_uses_encrypted_inline_data(inode))
+		max_bytes = round_down(max_bytes, FSCRYPT_CONTENTS_ALIGNMENT);
+#endif
+	return max_bytes;
+}
+
 static inline void f2fs_set_encrypted_inode(struct inode *inode)
 {
 #ifdef CONFIG_FS_ENCRYPTION
@@ -4827,6 +4861,7 @@ F2FS_FEATURE_FUNCS(compression, COMPRESSION);
 F2FS_FEATURE_FUNCS(readonly, RO);
 F2FS_FEATURE_FUNCS(device_alias, DEVICE_ALIAS);
 F2FS_FEATURE_FUNCS(packed_ssa, PACKED_SSA);
+F2FS_FEATURE_FUNCS(encrypted_inline_data, ENCRYPTED_INLINE_DATA);
 
 #ifdef CONFIG_BLK_DEV_ZONED
 static inline bool f2fs_zone_is_seq(struct f2fs_sb_info *sbi, int devi,
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 71385ca4163d..ec243bb9039b 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -825,12 +825,32 @@ int f2fs_do_truncate_blocks(struct inode *inode, u64 from, bool lock)
 	}
 
 	if (f2fs_has_inline_data(inode)) {
+		if (f2fs_uses_encrypted_inline_data(inode) && from) {
+			f2fs_folio_put(ifolio, true);
+			if (lock)
+				f2fs_unlock_op(sbi, &lc);
+
+			err = f2fs_convert_inline_inode(inode);
+
+			if (lock)
+				f2fs_lock_op(sbi, &lc);
+			if (err)
+				goto out;
+
+			ifolio = f2fs_get_inode_folio(sbi, inode->i_ino);
+			if (IS_ERR(ifolio)) {
+				err = PTR_ERR(ifolio);
+				goto out;
+			}
+			goto truncate_blocks;
+		}
 		f2fs_truncate_inline_inode(inode, ifolio, from);
 		f2fs_folio_put(ifolio, true);
 		truncate_page = true;
 		goto out;
 	}
 
+truncate_blocks:
 	set_new_dnode(&dn, inode, ifolio, NULL, 0);
 	err = f2fs_get_dnode_of_data(&dn, free_from, LOOKUP_NODE_RA);
 	if (err) {
@@ -1147,7 +1167,7 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 	if (attr->ia_valid & ATTR_SIZE) {
 		loff_t old_size = i_size_read(inode);
 
-		if (attr->ia_size > MAX_INLINE_DATA(inode)) {
+		if (attr->ia_size > f2fs_max_inline_data(inode)) {
 			/*
 			 * should convert inline inode before i_size_write to
 			 * keep smaller than inline_data size with inline flag.
@@ -5007,7 +5027,7 @@ static int f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *iter,
 
 	if (f2fs_has_inline_data(inode)) {
 		/* If the data will fit inline, don't bother. */
-		if (pos + count <= MAX_INLINE_DATA(inode))
+		if (pos + count <= f2fs_max_inline_data(inode))
 			return 0;
 		ret = f2fs_convert_inline_inode(inode);
 		if (ret)
diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
index 099f72089701..6bcf103584ce 100644
--- a/fs/f2fs/inline.c
+++ b/fs/f2fs/inline.c
@@ -21,7 +21,7 @@ static bool support_inline_data(struct inode *inode)
 		return false;
 	if (!S_ISREG(inode->i_mode) && !S_ISLNK(inode->i_mode))
 		return false;
-	if (i_size_read(inode) > MAX_INLINE_DATA(inode))
+	if (i_size_read(inode) > f2fs_max_inline_data(inode))
 		return false;
 	return true;
 }
@@ -31,6 +31,9 @@ bool f2fs_may_inline_data(struct inode *inode)
 	if (!support_inline_data(inode))
 		return false;
 
+	if (f2fs_uses_encrypted_inline_data(inode))
+		return fscrypt_inode_supports_data_unit_inplace(inode);
+
 	return !f2fs_post_read_required(inode);
 }
 
@@ -65,7 +68,9 @@ bool f2fs_sanity_check_inline_data(struct inode *inode, struct folio *ifolio)
 	 * been synchronized to inmem fields.
 	 */
 	return (S_ISREG(inode->i_mode) &&
-		(file_is_encrypt(inode) || file_is_verity(inode) ||
+		((file_is_encrypt(inode) &&
+		  !f2fs_sb_has_encrypted_inline_data(F2FS_I_SB(inode))) ||
+		 file_is_verity(inode) ||
 		(F2FS_I(inode)->i_flags & F2FS_COMPR_FL)));
 }
 
@@ -80,22 +85,60 @@ bool f2fs_may_inline_dentry(struct inode *inode)
 	return true;
 }
 
-void f2fs_do_read_inline_data(struct folio *folio, struct folio *ifolio)
+int f2fs_do_read_inline_data(struct folio *folio, struct folio *ifolio)
 {
 	struct inode *inode = folio->mapping->host;
+	unsigned int len = min_t(loff_t, i_size_read(inode),
+				 f2fs_max_inline_data(inode));
 
 	if (folio_test_uptodate(folio))
-		return;
+		return 0;
 
 	f2fs_bug_on(F2FS_I_SB(inode), folio->index);
 
-	folio_zero_segment(folio, MAX_INLINE_DATA(inode), folio_size(folio));
+	if (f2fs_uses_encrypted_inline_data(inode)) {
+		struct page *tmp_page;
+		void *kaddr;
+		int err;
+
+		folio_zero_segment(folio, 0, folio_size(folio));
 
-	/* Copy the whole inline data block */
-	memcpy_to_folio(folio, 0, inline_data_addr(inode, ifolio),
-		       MAX_INLINE_DATA(inode));
+		/*
+		 * Decrypt through a temporary page because inline data occupies
+		 * only a byte range inside the inode folio.
+		 */
+		tmp_page = alloc_page(GFP_NOFS | __GFP_ZERO);
+		if (!tmp_page)
+			return -ENOMEM;
+
+		len = round_up(len, FSCRYPT_CONTENTS_ALIGNMENT);
+		if (len) {
+			memcpy_to_page(tmp_page, 0, inline_data_addr(inode, ifolio),
+				       len);
+			err = fscrypt_decrypt_data_unit_inplace(inode, tmp_page,
+								len, 0, 0);
+			if (err) {
+				__free_page(tmp_page);
+				return err;
+			}
+		}
+
+		kaddr = kmap_local_page(tmp_page);
+		memcpy_to_folio(folio, 0, kaddr,
+				min_t(loff_t, i_size_read(inode),
+				      f2fs_max_inline_data(inode)));
+		kunmap_local(kaddr);
+		__free_page(tmp_page);
+	} else {
+		folio_zero_segment(folio, MAX_INLINE_DATA(inode),
+				   folio_size(folio));
+		/* Copy the whole inline data block */
+		memcpy_to_folio(folio, 0, inline_data_addr(inode, ifolio),
+				MAX_INLINE_DATA(inode));
+	}
 	if (!folio_test_uptodate(folio))
 		folio_mark_uptodate(folio);
+	return 0;
 }
 
 void f2fs_truncate_inline_inode(struct inode *inode, struct folio *ifolio,
@@ -119,6 +162,7 @@ void f2fs_truncate_inline_inode(struct inode *inode, struct folio *ifolio,
 int f2fs_read_inline_data(struct inode *inode, struct folio *folio)
 {
 	struct folio *ifolio;
+	int ret = 0;
 
 	ifolio = f2fs_get_inode_folio(F2FS_I_SB(inode), inode->i_ino);
 	if (IS_ERR(ifolio)) {
@@ -134,7 +178,13 @@ int f2fs_read_inline_data(struct inode *inode, struct folio *folio)
 	if (folio->index)
 		folio_zero_segment(folio, 0, folio_size(folio));
 	else
-		f2fs_do_read_inline_data(folio, ifolio);
+		ret = f2fs_do_read_inline_data(folio, ifolio);
+
+	if (!folio->index && ret) {
+		f2fs_folio_put(ifolio, true);
+		folio_unlock(folio);
+		return ret;
+	}
 
 	if (!folio_test_uptodate(folio))
 		folio_mark_uptodate(folio);
@@ -186,7 +236,9 @@ int f2fs_convert_inline_folio(struct dnode_of_data *dn, struct folio *folio)
 
 	f2fs_bug_on(F2FS_F_SB(folio), folio_test_writeback(folio));
 
-	f2fs_do_read_inline_data(folio, dn->inode_folio);
+	err = f2fs_do_read_inline_data(folio, dn->inode_folio);
+	if (err)
+		return err;
 	folio_mark_dirty(folio);
 
 	/* clear dirty state */
@@ -267,6 +319,8 @@ int f2fs_write_inline_data(struct inode *inode, struct folio *folio)
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct folio *ifolio;
+	void *inline_addr;
+	int err = 0;
 
 	ifolio = f2fs_get_inode_folio(sbi, inode->i_ino);
 	if (IS_ERR(ifolio))
@@ -280,8 +334,44 @@ int f2fs_write_inline_data(struct inode *inode, struct folio *folio)
 	f2fs_bug_on(F2FS_I_SB(inode), folio->index);
 
 	f2fs_folio_wait_writeback(ifolio, NODE, true, true);
-	memcpy_from_folio(inline_data_addr(inode, ifolio),
-			 folio, 0, MAX_INLINE_DATA(inode));
+	inline_addr = inline_data_addr(inode, ifolio);
+
+	if (f2fs_uses_encrypted_inline_data(inode)) {
+		struct page *tmp_page;
+		void *kaddr;
+		unsigned int len = min_t(loff_t, i_size_read(inode),
+					 f2fs_max_inline_data(inode));
+
+		tmp_page = alloc_page(GFP_NOFS | __GFP_ZERO);
+		if (!tmp_page) {
+			err = -ENOMEM;
+			goto out;
+		}
+
+		len = round_up(len, FSCRYPT_CONTENTS_ALIGNMENT);
+		if (len) {
+			kaddr = kmap_local_page(tmp_page);
+			memcpy_from_folio(kaddr, folio, 0,
+					  min_t(loff_t, i_size_read(inode),
+						f2fs_max_inline_data(inode)));
+			kunmap_local(kaddr);
+			err = fscrypt_encrypt_data_unit_inplace(inode, tmp_page,
+								len, 0, 0);
+		}
+		if (!err) {
+			memset(inline_addr, 0, MAX_INLINE_DATA(inode));
+			if (len) {
+				kaddr = kmap_local_page(tmp_page);
+				memcpy(inline_addr, kaddr, len);
+				kunmap_local(kaddr);
+			}
+		}
+		__free_page(tmp_page);
+		if (err)
+			goto out;
+	} else {
+		memcpy_from_folio(inline_addr, folio, 0, MAX_INLINE_DATA(inode));
+	}
 	folio_mark_dirty(ifolio);
 
 	f2fs_clear_page_cache_dirty_tag(folio);
@@ -290,8 +380,9 @@ int f2fs_write_inline_data(struct inode *inode, struct folio *folio)
 	set_inode_flag(inode, FI_DATA_EXIST);
 
 	folio_clear_f2fs_inline(ifolio);
+out:
 	f2fs_folio_put(ifolio, true);
-	return 0;
+	return err;
 }
 
 int f2fs_recover_inline_data(struct inode *inode, struct folio *nfolio)
@@ -826,7 +917,7 @@ int f2fs_inline_data_fiemap(struct inode *inode,
 			return PTR_ERR(ifolio);
 		f2fs_folio_wait_writeback(ifolio, NODE, true, true);
 	}
-	ilen = min_t(size_t, MAX_INLINE_DATA(inode), i_size_read(inode));
+	ilen = min_t(size_t, f2fs_max_inline_data(inode), i_size_read(inode));
 	if (start >= ilen)
 		goto out;
 	if (start + len < ilen)
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index c6afdbd6e1cd..9eddcde7939c 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1549,6 +1549,18 @@ static int f2fs_check_opt_consistency(struct fs_context *fc,
 		return -EINVAL;
 	}
 
+	if (f2fs_sb_has_encrypted_inline_data(sbi)) {
+		if (!IS_ENABLED(CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA)) {
+			f2fs_err(sbi,
+				 "encrypted_inline_data requires CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA");
+			return -EINVAL;
+		}
+		if (!f2fs_sb_has_encrypt(sbi)) {
+			f2fs_err(sbi, "encrypted inline_data requires encryption feature");
+			return -EINVAL;
+		}
+	}
+
 	/*
 	 * The BLKZONED feature indicates that the drive was formatted with
 	 * zone alignment optimization. This is optional for host-aware
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index 665687244c93..600eaee75926 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -1399,6 +1399,9 @@ F2FS_FEATURE_RO_ATTR(pin_file);
 F2FS_FEATURE_RO_ATTR(linear_lookup);
 #endif
 F2FS_FEATURE_RO_ATTR(packed_ssa);
+#ifdef CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA
+F2FS_FEATURE_RO_ATTR(encrypted_inline_data);
+#endif
 F2FS_FEATURE_RO_ATTR(fserror);
 
 #define ATTR_LIST(name) (&f2fs_attr_##name.attr)
@@ -1567,6 +1570,9 @@ static struct attribute *f2fs_feat_attrs[] = {
 	BASE_ATTR_LIST(linear_lookup),
 #endif
 	BASE_ATTR_LIST(packed_ssa),
+#ifdef CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA
+	BASE_ATTR_LIST(encrypted_inline_data),
+#endif
 	BASE_ATTR_LIST(fserror),
 	NULL,
 };
@@ -1604,6 +1610,7 @@ F2FS_SB_FEATURE_RO_ATTR(compression, COMPRESSION);
 F2FS_SB_FEATURE_RO_ATTR(readonly, RO);
 F2FS_SB_FEATURE_RO_ATTR(device_alias, DEVICE_ALIAS);
 F2FS_SB_FEATURE_RO_ATTR(packed_ssa, PACKED_SSA);
+F2FS_SB_FEATURE_RO_ATTR(encrypted_inline_data, ENCRYPTED_INLINE_DATA);
 
 static struct attribute *f2fs_sb_feat_attrs[] = {
 	ATTR_LIST(sb_encryption),
@@ -1622,6 +1629,7 @@ static struct attribute *f2fs_sb_feat_attrs[] = {
 	ATTR_LIST(sb_readonly),
 	ATTR_LIST(sb_device_alias),
 	ATTR_LIST(sb_packed_ssa),
+	ATTR_LIST(sb_encrypted_inline_data),
 	NULL,
 };
 ATTRIBUTE_GROUPS(f2fs_sb_feat);
-- 
2.34.1

[PATCH 3/3] Documentation: f2fs: document encrypted inline data

[LiaoYuanhong-vivo]

From: Liao Yuanhong <liaoyuanhong@vivo.com>

Document the F2FS encrypted_inline_data feature, including the on-disk
feature requirement, the CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA dependency,
how inline payloads are encrypted and decrypted, and the truncate
behavior.

Also list encrypted_inline_data in the supported F2FS feature sysfs
documentation.

Signed-off-by: Liao Yuanhong <liaoyuanhong@vivo.com>

---
 Documentation/ABI/testing/sysfs-fs-f2fs |  5 +++--
 Documentation/filesystems/f2fs.rst      | 27 +++++++++++++++++++++++++
 2 files changed, 30 insertions(+), 2 deletions(-)

diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs
index 27d5e88facbe..dad483fb2fc1 100644
--- a/Documentation/ABI/testing/sysfs-fs-f2fs
+++ b/Documentation/ABI/testing/sysfs-fs-f2fs
@@ -258,7 +258,8 @@ Description:	Expand /sys/fs/f2fs/<disk>/features to meet sysfs rule.
 		encryption, block_zoned (aka blkzoned), extra_attr,
 		project_quota (aka projquota), inode_checksum,
 		flexible_inline_xattr, quota_ino, inode_crtime, lost_found,
-		verity, sb_checksum, casefold, readonly, compression.
+		verity, sb_checksum, casefold, readonly, compression,
+		encrypted_inline_data.
 		Note that, pin_file is moved into /sys/fs/f2fs/features/.
 
 What:		/sys/fs/f2fs/features/
@@ -271,7 +272,7 @@ Description:	Shows all enabled kernel features.
 		inode_crtime, lost_found, verity, sb_checksum,
 		casefold, readonly, compression, test_dummy_encryption_v2,
 		atomic_write, pin_file, encrypted_casefold, linear_lookup,
-		fserror.
+		fserror, encrypted_inline_data.
 
 What:		/sys/fs/f2fs/<disk>/inject_rate
 Date:		May 2016
diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst
index 5bc37a1c4e51..1f3e02352dd1 100644
--- a/Documentation/filesystems/f2fs.rst
+++ b/Documentation/filesystems/f2fs.rst
@@ -420,6 +420,33 @@ lookup_mode=%s		 Control the directory lookup behavior for casefolded
 			     ================== ========================================
 ======================== ============================================================
 
+Encrypted inline data
+=====================
+
+F2FS normally disables inline data for encrypted regular files, since inline
+data is stored inside the inode block and does not pass through the regular
+block I/O path.  When a filesystem is formatted with the encrypted_inline_data
+feature, encrypted regular files may keep small file contents in the inode
+block.  The inline payload is encrypted with fscrypt contents-key semantics
+before it is written to the inode, and it is decrypted back to page-cache
+plaintext when it is read.
+
+This feature requires the encrypt feature on disk and kernel support for
+CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA.  It is intended to be used together with
+the inline_data mount option.  When the normal encrypted file contents path uses
+blk-crypto, fscrypt also prepares a software contents-key transform for the
+filesystem-managed inline payload.
+
+Encrypted inline data is stored in fscrypt contents-aligned units.  Therefore,
+the maximum plaintext size that can stay inline may be slightly smaller than the
+ordinary inline data capacity.  If an encrypted inline-data file is truncated
+from a non-zero offset, F2FS first converts the inline payload to normal data
+blocks and then applies the truncate operation.
+
+Recovery copies inline payloads as on-disk bytes.  Encryption and decryption are
+performed only when moving data between the inode inline area and page-cache
+plaintext.
+
 Debugfs Entries
 ===============
 
-- 
2.34.1

Re: [PATCH 0/3] f2fs: support encrypted inline data

[Eric Biggers]

On Wed, May 13, 2026 at 06:04:27PM +0800, LiaoYuanhong-vivo wrote:
> From: Liao Yuanhong <liaoyuanhong@vivo.com>
> 
> F2FS currently avoids inline data for encrypted regular files.  This is
> because inline data is stored in the inode block, outside the regular
> bio-based data path where fscrypt and blk-crypto normally operate.
> As a result, devices that enable blk-crypto for encrypted file contents
> cannot use F2FS inline data for encrypted regular files, which wastes
> space for small files.
> 
> This series adds support for keeping small encrypted regular-file
> contents as inline data.  The f2fs side defines a new on-disk feature,
> encrypted_inline_data, under which inline payloads of encrypted regular
> files are interpreted as ciphertext.  The payload is encrypted before
> being stored in the inode block and decrypted back into page-cache
> plaintext on read.
> 
> The fscrypt side prepares a software contents-key transform even when
> normal file contents use blk-crypto, so filesystems can encrypt
> filesystem-managed data regions that do not go through bio submission.
> The new fscrypt helper operates on fscrypt data units and leaves the
> filesystem responsible for deciding which filesystem-managed byte ranges
> need this treatment.
> 
> The software crypto operation is limited to the inline payload.  Since
> these files are small enough to remain inline, the expected read/write
> performance difference between hardware and software crypto is small,
> while the space saving from keeping the data inline is significant.
> 
> The feature is guarded by CONFIG_F2FS_FS_ENCRYPTED_INLINE_DATA and by the
> F2FS encrypted_inline_data on-disk feature bit.  Filesystems with this
> feature set are rejected if the kernel lacks the config option.
> 
> Hardware-wrapped keys are not supported by this initial version. I would
> like to discuss whether this feature should remain disabled for
> hardware-wrapped keys, or whether there is an acceptable way to support the
> combination in the future.
> 
> The f2fs-tools support for formatting filesystems with this feature will be
> submitted separately.
> 
> Basic testing passed.  Encrypted small files can be kept as inline data,
> and read/write verification succeeded.

Honestly, I'm not convinced this is worth the complexity and the
additional memory use.

First, it works only in the combination: 'f2fs && inlinecrypt &&
!hw_wrapped_keys'.  That really limits how many users would use this.
'f2fs && inlinecrypt' de facto targets it to Android devices rather than
"regular" Linux systems.  But at the same time, the "best practice" on
such devices is to use HW-wrapped keys, which has already been widely
adopted.  So this would be useful only on devices where the SoC doesn't
support HW-wrapped keys.  Its usefulness will go away when support for
HW-wrapped keys is added.

Second, in the per-file key case this makes every file use an additional
1 KiB of memory or so (assuming AES-XTS) to hold the "software key",
just in case the file ever has inline data.  That seems problematic, and
maybe not a great direction to be going in right now, given the ongoing
RAM shortage.

There also seem to be quite a few bugs/issues.  Sashiko found quite a
few
(https://sashiko.dev/#/message/20260513100431.299904-1-liaoyuanhong%40vivo.com).
But just from a quick readthrough, anything that calls
fscrypt_is_key_prepared() seems to be broken now, as that function isn't
aware that both fields of fscrypt_prepared_key can be needed.

I'm also not seeing what differentiates the new
fscrypt_{en,decrypt}_data_unit_inplace() from the existing
fscrypt_{en,decrypt}_block_inplace().  They seem redundant.

There's already a lot of complexity in fscrypt, with the different
settings and the different ways the filesystems do en/decryption.  With
this, plus the concurrent work to add support for extent-based
encryption (for btrfs), it's really quite hard to keep track of
everything.  So I have to wonder if this patchset is really worth it.

So, overall, I think this would need a bit more work.  But also I'm
wondering if it's actually worthwhile.  Do you plan to never enable
HW-wrapped keys, for example?  And you're fine with using more RAM?

- Eric