Cache insensitive cleanup for ext4/f2fs
Subsystems (3)
v12 cover Cache insensitive cleanup for ext4/f2fs
Hello,
I am trying to respin the series here :
https://www.spinics.net/lists/linux-ext4/msg85081.html
I resent some of the v9 patches and got some reviews from Gabriel, I did changes as requested and here is v12.
Changes in v12:
- revert to v10 comparison with propagating the error code from utf comparison
Changes in v11:
- revert to the original v9 implementation for the comparison helper.
Changes in v10:
- reworked a bit the comparison helper to improve performance by first performing the exact lookup.
* Original commit letter
The case-insensitive implementations in f2fs and ext4 have quite a bit of duplicated code. This series simplifies the ext4 version, with the goal of extracting ext4_ci_compare into a helper library that can be used by both filesystems. It also reduces the clutter from many codeguards for CONFIG_UNICODE; as requested by Linus, they are part of the codeflow now.
While there, I noticed we can leverage the utf8 functions to detect encoded names that are corrupted in the filesystem. Therefore, it also adds an ext4 error on that scenario, to mark the filesystem as corrupted.
This series survived passes of xfstests -g quick.
Gabriel Krisman Bertazi (8):
ext4: Simplify the handling of cached insensitive names
f2fs: Simplify the handling of cached insensitive names
libfs: Introduce case-insensitive string comparison helper
ext4: Reuse generic_ci_match for ci comparisons
f2fs: Reuse generic_ci_match for ci comparisons
ext4: Log error when lookup of encoded dentry fails
ext4: Move CONFIG_UNICODE defguards into the code flow
f2fs: Move CONFIG_UNICODE defguards into the code flow
fs/ext4/crypto.c | 19 ++----- fs/ext4/ext4.h | 35 +++++++----- fs/ext4/namei.c | 129 ++++++++++++++++----------------------------- fs/ext4/super.c | 4 +- fs/f2fs/dir.c | 105 +++++++++++------------------------- fs/f2fs/f2fs.h | 17 +++++- fs/f2fs/namei.c | 10 ++-- fs/f2fs/recovery.c | 5 +- fs/f2fs/super.c | 8 +-- fs/libfs.c | 85 +++++++++++++++++++++++++++++ include/linux/fs.h | 4 ++ 11 files changed, 216 insertions(+), 205 deletions(-)
Signature · 1 line
2.34.1
PATCH v12 1/8 ext4: Simplify the handling of cached insensitive names
From: Gabriel Krisman Bertazi <[email protected]>
Keeping it as qstr avoids the unnecessary conversion in ext4_match
--- fs/ext4/ext4.h | 2 +- fs/ext4/namei.c | 23 +++++++++++------------ 2 files changed, 12 insertions(+), 13 deletions(-)
fs/ext4/ext4.h +1 -1
@@ -2508,7 +2508,7 @@ struct ext4_filename {struct fscrypt_str crypto_buf;#endif #if IS_ENABLED (CONFIG_UNICODE ) -struct fscrypt_str cf_name; +struct qstr cf_name;#endif };
fs/ext4/namei.c +11 -12
@@ -1445,7 +1445,8 @@ static int ext4_ci_compare(const struct inode *parent, const struct qstr *name,int ext4_fname_setup_ci_filename (struct inode *dir,const struct qstr *iname,struct ext4_filename *name) { -struct fscrypt_str *cf_name = &name->cf_name; +struct qstr *cf_name = &name->cf_name; +unsigned char *buf;struct dx_hash_info *hinfo = &name->hinfo;int len;
@@ -1455,18 +1456,18 @@ int ext4_fname_setup_ci_filename(struct inode *dir, const struct qstr *iname,return 0 ; } - cf_name->name =kmalloc (EXT4_NAME_LEN ,GFP_NOFS ); -if (!cf_name->name) + buf =kmalloc (EXT4_NAME_LEN ,GFP_NOFS ); +if (!buf)return -ENOMEM ; - len =utf8_casefold (dir->i_sb->s_encoding, - iname, cf_name->name, -EXT4_NAME_LEN ); + len =utf8_casefold (dir->i_sb->s_encoding, iname, buf,EXT4_NAME_LEN );if (len <=0 ) { -kfree (cf_name->name); - cf_name->name =NULL ; +kfree (buf); + buf =NULL ; } + cf_name->name = buf; cf_name->len = (unsigned ) len; +if (!IS_ENCRYPTED (dir))return 0 ;
@@ -1503,8 +1504,6 @@ static bool ext4_match(struct inode *parent, if (IS_CASEFOLDED(parent) && (!IS_ENCRYPTED(parent) || fscrypt_has_encryption_key(parent))) { if (fname->cf_name.name) { - struct qstr cf = {.name = fname->cf_name.name, - .len = fname->cf_name.len}; if (IS_ENCRYPTED(parent)) { if (fname->hinfo.hash != EXT4_DIRENT_HASH(de) || fname->hinfo.minor_hash !=
@@ -1513,8 +1512,8 @@ static bool ext4_match(struct inode *parent,return false; } } -return !ext4_ci_compare (parent, &cf, de->name, - de->name_len, true); +return !ext4_ci_compare (parent, &fname->cf_name, + de->name, de->name_len, true); }return !ext4_ci_compare (parent, fname->usr_fname, de->name, de->name_len, false);
Signature · 1 line
2.34.1
PATCH v12 2/8 f2fs: Simplify the handling of cached insensitive names
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
Keeping it as qstr avoids the unnecessary conversion in f2fs_match
--- fs/f2fs/dir.c | 53 ++++++++++++++++++++++++++-------------------- fs/f2fs/f2fs.h | 17 ++++++++++++++- fs/f2fs/recovery.c | 5 +---- 3 files changed, 47 insertions(+), 28 deletions(-)
fs/f2fs/dir.c +30 -23
@@ -42,35 +42,49 @@ static unsigned int bucket_blocks(unsigned int level)return 4 ; } +#if IS_ENABLED (CONFIG_UNICODE )/* If @dir is casefolded, initialize @fname->cf_name from @fname->usr_fname. */ int f2fs_init_casefolded_name (const struct inode *dir,struct f2fs_filename *fname) { -#if IS_ENABLED (CONFIG_UNICODE )struct super_block *sb = dir->i_sb; +unsigned char *buf; +int len;if (IS_CASEFOLDED (dir) && !is_dot_dotdot (fname->usr_fname->name, fname->usr_fname->len)) { - fname->cf_name.name =f2fs_kmem_cache_alloc (f2fs_cf_name_slab, -GFP_NOFS , false,F2FS_SB (sb)); -if (!fname->cf_name.name) + buf =f2fs_kmem_cache_alloc (f2fs_cf_name_slab, +GFP_NOFS , false,F2FS_SB (sb)); +if (!buf)return -ENOMEM ; - fname->cf_name.len =utf8_casefold (sb->s_encoding, - fname->usr_fname, - fname->cf_name.name, -F2FS_NAME_LEN ); -if ((int )fname->cf_name.len <=0 ) { -kmem_cache_free (f2fs_cf_name_slab, fname->cf_name.name); - fname->cf_name.name =NULL ; + + len =utf8_casefold (sb->s_encoding, fname->usr_fname, + buf,F2FS_NAME_LEN ); +if (len <=0 ) { +kmem_cache_free (f2fs_cf_name_slab, buf);if (sb_has_strict_encoding (sb))return -EINVAL ;/* fall back to treating name as opaque byte sequence */ +return 0 ; } + fname->cf_name.name = buf; + fname->cf_name.len = len; } -#endif +return 0 ; } +void f2fs_free_casefolded_name (struct f2fs_filename *fname) +{ +unsigned char *buf = (unsigned char *)fname->cf_name.name; + +if (buf) { +kmem_cache_free (f2fs_cf_name_slab, buf); + fname->cf_name.name =NULL ; + } +} +#endif /* CONFIG_UNICODE */ +static int __f2fs_setup_filename (const struct inode *dir,const struct fscrypt_name *crypt_name,struct f2fs_filename *fname)
@@ -142,12 +156,7 @@ void f2fs_free_filename(struct f2fs_filename *fname)kfree (fname->crypto_buf.name); fname->crypto_buf.name =NULL ;#endif -#if IS_ENABLED (CONFIG_UNICODE ) -if (fname->cf_name.name) { -kmem_cache_free (f2fs_cf_name_slab, fname->cf_name.name); - fname->cf_name.name =NULL ; - } -#endif +f2fs_free_casefolded_name (fname); } static unsigned long dir_block_index(unsigned int level,
@@ -235,11 +244,9 @@ static inline int f2fs_match_name(const struct inode *dir,struct fscrypt_name f;#if IS_ENABLED (CONFIG_UNICODE ) -if (fname->cf_name.name) { -struct qstr cf =FSTR_TO_QSTR (&fname->cf_name); - -return f2fs_match_ci_name (dir, &cf, de_name, de_name_len); - } +if (fname->cf_name.name) +return f2fs_match_ci_name (dir, &fname->cf_name, + de_name, de_name_len);#endif f.usr_fname = fname->usr_fname; f.disk_name = fname->disk_name;
fs/f2fs/f2fs.h +16 -1
@@ -530,7 +530,7 @@ struct f2fs_filename {* internal operation where usr_fname is also NULL. In all these cases * we fall back to treating the name as an opaque byte sequence. */ -struct fscrypt_str cf_name; +struct qstr cf_name;#endif };
@@ -3533,8 +3533,23 @@ int f2fs_get_tmpfile(struct mnt_idmap *idmap, struct inode *dir,/* * dir.c */ +unsigned char f2fs_get_de_type (struct f2fs_dir_entry *de); +#if IS_ENABLED (CONFIG_UNICODE )int f2fs_init_casefolded_name (const struct inode *dir,struct f2fs_filename *fname); +void f2fs_free_casefolded_name (struct f2fs_filename *fname); +#else +static inline int f2fs_init_casefolded_name (const struct inode *dir, +struct f2fs_filename *fname) +{ +return 0 ; +} + +static inline void f2fs_free_casefolded_name (struct f2fs_filename *fname) +{ +} +#endif /* CONFIG_UNICODE */ +int f2fs_setup_filename (struct inode *dir,const struct qstr *iname,int lookup,struct f2fs_filename *fname); int f2fs_prepare_lookup(struct inode *dir, struct dentry *dentry,
fs/f2fs/recovery.c +1 -4
@@ -153,11 +153,8 @@ static int init_recovered_filename(const struct inode *dir,if (err)return err;f2fs_hash_filename (dir, fname); -#if IS_ENABLED (CONFIG_UNICODE )/* Case-sensitive match is fine for recovery */ -kmem_cache_free (f2fs_cf_name_slab, fname->cf_name.name); - fname->cf_name.name =NULL ; -#endif +f2fs_free_casefolded_name (fname); }else {f2fs_hash_filename (dir, fname); }
Signature · 1 line
2.34.1
Re: [PATCH v12 2/8] f2fs: Simplify the handling of cached insensitive names
Quoted 65 lines · Eugen Hristev
From: Gabriel Krisman Bertazi <[email protected]>
Keeping it as qstr avoids the unnecessary conversion in
f2fs_matchReviewed-by Eric Biggers <[email protected]>Signed-off-by Gabriel Krisman Bertazi <[email protected]>[[email protected]: port to 6.8-rc3]Signed-off-by Eugen Hristev <[email protected]>--- fs/f2fs/dir.c | 53 ++++++++++++++++++++++++++-------------------- fs/f2fs/f2fs.h | 17 ++++++++++++++- fs/f2fs/recovery.c | 5 +---- 3 files changed, 47 insertions(+), 28 deletions(-)fs/f2fs/dir.c +15
-12@@ -42,35 +42,49 @@ static unsigned int bucket_blocks(unsigned int level)return 4 ; } +#if IS_ENABLED (CONFIG_UNICODE )/* If @dir is casefolded, initialize @fname->cf_name from @fname->usr_fname. */ int f2fs_init_casefolded_name (const struct inode *dir,struct f2fs_filename *fname) { -#if IS_ENABLED (CONFIG_UNICODE )struct super_block *sb = dir->i_sb; +unsigned char *buf; +int len;if (IS_CASEFOLDED (dir) && !is_dot_dotdot (fname->usr_fname->name, fname->usr_fname->len)) { - fname->cf_name.name =f2fs_kmem_cache_alloc (f2fs_cf_name_slab, -GFP_NOFS , false,F2FS_SB (sb)); -if (!fname->cf_name.name) + buf =f2fs_kmem_cache_alloc (f2fs_cf_name_slab, +GFP_NOFS , false,F2FS_SB (sb)); +if (!buf)return -ENOMEM ; - fname->cf_name.len =utf8_casefold (sb->s_encoding, - fname->usr_fname, - fname->cf_name.name, -F2FS_NAME_LEN ); -if ((int )fname->cf_name.len <=0 ) { -kmem_cache_free (f2fs_cf_name_slab, fname->cf_name.name); - fname->cf_name.name =NULL ; + + len =utf8_casefold (sb->s_encoding, fname->usr_fname, + buf,F2FS_NAME_LEN ); +if (len <=0 ) { +kmem_cache_free (f2fs_cf_name_slab, buf);if (sb_has_strict_encoding (sb))return -EINVAL ;/* fall back to treating name as opaque byte sequence */ +return 0 ; } + fname->cf_name.name = buf; + fname->cf_name.len = len; } -#endif +return 0 ; }
fs/f2fs/dir.c static unsigned int bucket_blocks(unsigned int level) · line 85
Quoted 10 lines
fs/f2fs/dir.c +9
-0+void f2fs_free_casefolded_name (struct f2fs_filename *fname) +{ +unsigned char *buf = (unsigned char *)fname->cf_name.name; + +if (buf) { +kmem_cache_free (f2fs_cf_name_slab, buf); + fname->cf_name.name =NULL ; + } +}
If we use kfree here, we can drop the buf !=NULL check.
Quoted 48 lines
fs/f2fs/dir.c +6
-11+#endif /* CONFIG_UNICODE */ +static int __f2fs_setup_filename (const struct inode *dir,const struct fscrypt_name *crypt_name,struct f2fs_filename *fname)@@ -142,12 +156,7 @@ void f2fs_free_filename(struct f2fs_filename *fname)kfree (fname->crypto_buf.name); fname->crypto_buf.name =NULL ;#endif -#if IS_ENABLED (CONFIG_UNICODE ) -if (fname->cf_name.name) { -kmem_cache_free (f2fs_cf_name_slab, fname->cf_name.name); - fname->cf_name.name =NULL ; - } -#endif +f2fs_free_casefolded_name (fname); } static unsigned long dir_block_index(unsigned int level,@@ -235,11 +244,9 @@ static inline int f2fs_match_name(const struct inode *dir,struct fscrypt_name f;#if IS_ENABLED (CONFIG_UNICODE ) -if (fname->cf_name.name) { -struct qstr cf =FSTR_TO_QSTR (&fname->cf_name); - -return f2fs_match_ci_name (dir, &cf, de_name, de_name_len); - } +if (fname->cf_name.name) +return f2fs_match_ci_name (dir, &fname->cf_name, + de_name, de_name_len);#endif f.usr_fname = fname->usr_fname; f.disk_name = fname->disk_name;fs/f2fs/f2fs.h +1
-1@@ -530,7 +530,7 @@ struct f2fs_filename {* internal operation where usr_fname is also NULL. In all these cases * we fall back to treating the name as an opaque byte sequence. */ -struct fscrypt_str cf_name; +struct qstr cf_name;#endif };
fs/f2fs/f2fs.h int f2fs_get_tmpfile(struct mnt_idmap *idmap, struct inode *dir, · line 3536
Quoted 4 lines
fs/f2fs/f2fs.h +1
-0@@ -3533,8 +3533,23 @@ int f2fs_get_tmpfile(struct mnt_idmap *idmap, struct inode *dir,/* * dir.c */ +unsigned char f2fs_get_de_type (struct f2fs_dir_entry *de);
This is not part of the original patch and doesn't make sense here. It seems to be included by a bad rebase?
Signature · 1 line
Gabriel Krisman Bertazi
PATCH v12 3/8 libfs: Introduce case-insensitive string comparison helper
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
generic_ci_match can be used by case-insensitive filesystems to compare strings under lookup with dirents in a case-insensitive way. This function is currently reimplemented by each filesystem supporting casefolding, so this reduces code duplication in filesystem-specific code.
--- fs/libfs.c | 85 ++++++++++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 4 +++ 2 files changed, 89 insertions(+)
fs/libfs.c +85 -0
@@ -1773,6 +1773,91 @@ static const struct dentry_operations generic_ci_dentry_ops = { .d_hash = generic_ci_d_hash, .d_compare = generic_ci_d_compare, }; + +/** +* generic_ci_match() - Match a name (case-insensitively) with a dirent. +* This is a filesystem helper for comparison with directory entries. +* generic_ci_d_compare should be used in VFS' ->d_compare instead. +* +* @parent: Inode of the parent of the dirent under comparison +* @name: name under lookup. +* @folded_name: Optional pre-folded name under lookup +* @de_name: Dirent name. +* @de_name_len: dirent name length. +* +* Test whether a case-insensitive directory entry matches the filename +* being searched. If @folded_name is provided, it is used instead of +* recalculating the casefold of @name. +* +* Return: > 0 if the directory entry matches, 0 if it doesn't match, or +* < 0 on error. +*/ +int generic_ci_match (const struct inode *parent, +const struct qstr *name, +const struct qstr *folded_name, +const u8 *de_name,u32 de_name_len) +{ +const struct super_block *sb = parent->i_sb; +const struct unicode_map *um = sb->s_encoding; +struct fscrypt_str decrypted_name =FSTR_INIT (NULL , de_name_len); +struct qstr dirent =QSTR_INIT (de_name, de_name_len); +int res, match =0 ; + +if (IS_ENCRYPTED (parent)) { +const struct fscrypt_str encrypted_name = +FSTR_INIT ((u8 *) de_name, de_name_len); + +if (WARN_ON_ONCE (!fscrypt_has_encryption_key (parent))) +return -EINVAL ; + + decrypted_name.name =kmalloc (de_name_len,GFP_KERNEL ); +if (!decrypted_name.name) +return -ENOMEM ; + res =fscrypt_fname_disk_to_usr (parent,0 ,0 , &encrypted_name, + &decrypted_name); +if (res <0 ) + goto out; + dirent.name = decrypted_name.name; + dirent.len = decrypted_name.len; + } + +/* +* Attempt a case-sensitive match first. It is cheaper and +* should cover most lookups, including all the sane +* applications that expect a case-sensitive filesystem. +* +* This comparison is safe under RCU because the caller +* guarantees the consistency between str and len. See +* __d_lookup_rcu_op_compare() for details. +*/ +if (folded_name->name) { +if (dirent.len == folded_name->len && + !memcmp (folded_name->name, dirent.name, dirent.len)) { + match =1 ; + goto out; + } + res =utf8_strncasecmp_folded (um, folded_name, &dirent); + }else { +if (dirent.len == name->len && + !memcmp (name->name, dirent.name, dirent.len) && + (!sb_has_strict_encoding (sb) || !utf8_validate (um, name))) { + match =1 ; + goto out; + } + res =utf8_strncasecmp (um, name, &dirent); + } + +out: +kfree (decrypted_name.name); +if (match)/* matched by direct comparison */ +return 1 ; +else if (!res)/* matched by utf8 comparison */ +return 1 ; +else if (res <0 )/* error on utf8 comparison */ +return res; +return 0 ;/* no match */ +} +EXPORT_SYMBOL (generic_ci_match);#endif #ifdef CONFIG_FS_ENCRYPTION
include/linux/fs.h +4 -0
@@ -3296,6 +3296,10 @@ extern int generic_file_fsync(struct file *, loff_t, loff_t, int);extern int generic_check_addressable (unsigned ,u64 );extern void generic_set_encrypted_ci_d_ops (struct dentry *dentry); +extern int generic_ci_match (const struct inode *parent, +const struct qstr *name, +const struct qstr *folded_name, +const u8 *de_name,u32 de_name_len);static inline bool sb_has_encoding (const struct super_block *sb) {
Signature · 1 line
2.34.1
Re: [PATCH v12 3/8] libfs: Introduce case-insensitive string comparison helper
Quoted 7 lines · Eugen Hristev
From: Gabriel Krisman Bertazi <[email protected]>
generic_ci_matchcan be used by case-insensitive filesystems to compare strings under lookup with dirents in a case-insensitive way. This function is currently reimplemented by each filesystem supporting casefolding, so this reduces code duplication in filesystem-specific code.
Just a note that this conflicts with the other patchset to generic helpers that I just applied. The conflict is trivial, If you could base the next iteration on top of my for-next, it would be helpful.
Quoted 63 lines
Signed-off-by Gabriel Krisman Bertazi <[email protected]>[[email protected]: rework to first test the exact match]Signed-off-by Eugen Hristev <[email protected]>--- fs/libfs.c | 85 ++++++++++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 4 +++ 2 files changed, 89 insertions(+)fs/libfs.c +46
-0@@ -1773,6 +1773,91 @@ static const struct dentry_operations generic_ci_dentry_ops = { .d_hash = generic_ci_d_hash, .d_compare = generic_ci_d_compare, }; + +/** +* generic_ci_match() - Match a name (case-insensitively) with a dirent. +* This is a filesystem helper for comparison with directory entries. +* generic_ci_d_compare should be used in VFS' ->d_compare instead. +* +* @parent: Inode of the parent of the dirent under comparison +* @name: name under lookup. +* @folded_name: Optional pre-folded name under lookup +* @de_name: Dirent name. +* @de_name_len: dirent name length. +* +* Test whether a case-insensitive directory entry matches the filename +* being searched. If @folded_name is provided, it is used instead of +* recalculating the casefold of @name. +* +* Return: > 0 if the directory entry matches, 0 if it doesn't match, or +* < 0 on error. +*/ +int generic_ci_match (const struct inode *parent, +const struct qstr *name, +const struct qstr *folded_name, +const u8 *de_name,u32 de_name_len) +{ +const struct super_block *sb = parent->i_sb; +const struct unicode_map *um = sb->s_encoding; +struct fscrypt_str decrypted_name =FSTR_INIT (NULL , de_name_len); +struct qstr dirent =QSTR_INIT (de_name, de_name_len); +int res, match =0 ; + +if (IS_ENCRYPTED (parent)) { +const struct fscrypt_str encrypted_name = +FSTR_INIT ((u8 *) de_name, de_name_len); + +if (WARN_ON_ONCE (!fscrypt_has_encryption_key (parent))) +return -EINVAL ; + + decrypted_name.name =kmalloc (de_name_len,GFP_KERNEL ); +if (!decrypted_name.name) +return -ENOMEM ; + res =fscrypt_fname_disk_to_usr (parent,0 ,0 , &encrypted_name, + &decrypted_name); +if (res <0 ) + goto out; + dirent.name = decrypted_name.name; + dirent.len = decrypted_name.len;
fs/libfs.c static const struct dentry_operations generic_ci_dentry_ops = { · line 1831
Quoted 10 lines
fs/libfs.c +9
-0+ } + +/* +* Attempt a case-sensitive match first. It is cheaper and +* should cover most lookups, including all the sane +* applications that expect a case-sensitive filesystem. +* This comparison is safe under RCU because the caller +* guarantees the consistency between str and len. See +* __d_lookup_rcu_op_compare() for details.
As I mentioned in the previous review, there's no RCU here. This comment makes no sense here.
Quoted 19 lines
fs/libfs.c +19
-0+*/ +if (folded_name->name) { +if (dirent.len == folded_name->len && + !memcmp (folded_name->name, dirent.name, dirent.len)) { + match =1 ; + goto out; + } + res =utf8_strncasecmp_folded (um, folded_name, &dirent); + }else { +if (dirent.len == name->len && + !memcmp (name->name, dirent.name, dirent.len) && + (!sb_has_strict_encoding (sb) || !utf8_validate (um, name))) { + match =1 ; + goto out; + } + res =utf8_strncasecmp (um, name, &dirent); + } + +out:
fs/libfs.c static const struct dentry_operations generic_ci_dentry_ops = {
Quoted 10 lines
fs/libfs.c +9
-0+kfree (decrypted_name.name); +if (match)/* matched by direct comparison */ +return 1 ; +else if (!res)/* matched by utf8 comparison */ +return 1 ; +else if (res <0 )/* error on utf8 comparison */ +return res; +return 0 ;/* no match */ +}
It can be simplified to
if (res < 0)
return res;
return (match || !res);
Quoted 17 lines
fs/libfs.c +0
-0#ifdef CONFIG_FS_ENCRYPTION include/linux/fs.h +4
-0@@ -3296,6 +3296,10 @@ extern int generic_file_fsync(struct file *, loff_t, loff_t, int);extern int generic_check_addressable (unsigned ,u64 );extern void generic_set_encrypted_ci_d_ops (struct dentry *dentry); +extern int generic_ci_match (const struct inode *parent, +const struct qstr *name, +const struct qstr *folded_name, +const u8 *de_name,u32 de_name_len);static inline bool sb_has_encoding (const struct super_block *sb) {
Signature · 1 line
Gabriel Krisman Bertazi
PATCH v12 4/8 ext4: Reuse generic_ci_match for ci comparisons
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
Instead of reimplementing ext4_match_ci, use the new libfs helper.
It also adds a comment explaining why fname->cf_name.name must be checked prior to the encryption hash optimization, because that tripped me before.
--- fs/ext4/namei.c | 91 +++++++++++++++---------------------------------- 1 file changed, 27 insertions(+), 64 deletions(-)
fs/ext4/namei.c +27 -64
@@ -1390,58 +1390,6 @@ static void dx_insert_block(struct dx_frame *frame, u32 hash, ext4_lblk_t block) }#if IS_ENABLED (CONFIG_UNICODE ) -/* -* Test whether a case-insensitive directory entry matches the filename -* being searched for. If quick is set, assume the name being looked up -* is already in the casefolded form. -* -* Returns: 0 if the directory entry matches, more than 0 if it -* doesn't match or less than zero on error. -*/ -static int ext4_ci_compare (const struct inode *parent,const struct qstr *name, -u8 *de_name,size_t de_name_len,bool quick) -{ -const struct super_block *sb = parent->i_sb; -const struct unicode_map *um = sb->s_encoding; -struct fscrypt_str decrypted_name =FSTR_INIT (NULL , de_name_len); -struct qstr entry =QSTR_INIT (de_name, de_name_len); -int ret; - -if (IS_ENCRYPTED (parent)) { -const struct fscrypt_str encrypted_name = -FSTR_INIT (de_name, de_name_len); - - decrypted_name.name =kmalloc (de_name_len,GFP_KERNEL ); -if (!decrypted_name.name) -return -ENOMEM ; - ret =fscrypt_fname_disk_to_usr (parent,0 ,0 , &encrypted_name, - &decrypted_name); -if (ret <0 ) - goto out; - entry.name = decrypted_name.name; - entry.len = decrypted_name.len; - } - -if (quick) - ret =utf8_strncasecmp_folded (um, name, &entry); -else - ret =utf8_strncasecmp (um, name, &entry); -if (ret <0 ) { -/* Handle invalid character sequence as either an error -* or as an opaque byte sequence. -*/ -if (sb_has_strict_encoding (sb)) - ret = -EINVAL ; -else if (name->len != entry.len) - ret =1 ; -else - ret = !!memcmp (name->name, entry.name, entry.len); - } -out: -kfree (decrypted_name.name); -return ret; -} -int ext4_fname_setup_ci_filename (struct inode *dir,const struct qstr *iname,struct ext4_filename *name) {
@@ -1503,20 +1451,35 @@ static bool ext4_match(struct inode *parent,#if IS_ENABLED (CONFIG_UNICODE )if (IS_CASEFOLDED (parent) && (!IS_ENCRYPTED (parent) ||fscrypt_has_encryption_key (parent))) { -if (fname->cf_name.name) { -if (IS_ENCRYPTED (parent)) { -if (fname->hinfo.hash !=EXT4_DIRENT_HASH (de) || - fname->hinfo.minor_hash != -EXT4_DIRENT_MINOR_HASH (de)) { +int ret; -return false; - } - } -return !ext4_ci_compare (parent, &fname->cf_name, - de->name, de->name_len, true); +/* +* Just checking IS_ENCRYPTED(parent) below is not +* sufficient to decide whether one can use the hash for +* skipping the string comparison, because the key might +* have been added right after +* ext4_fname_setup_ci_filename(). In this case, a hash +* mismatch will be a false negative. Therefore, make +* sure cf_name was properly initialized before +* considering the calculated hash. +*/ +if (IS_ENCRYPTED (parent) && fname->cf_name.name && + (fname->hinfo.hash !=EXT4_DIRENT_HASH (de) || + fname->hinfo.minor_hash !=EXT4_DIRENT_MINOR_HASH (de))) +return false; + + ret =generic_ci_match (parent, fname->usr_fname, + &fname->cf_name, de->name, + de->name_len); +if (ret <0 ) { +/* +* Treat comparison errors as not a match. The +* only case where it happens is on a disk +* corruption or ENOMEM. +*/ +return false; } -return !ext4_ci_compare (parent, fname->usr_fname, de->name, - de->name_len, false); +return ret; }#endif
Signature · 1 line
2.34.1
PATCH v12 5/8 f2fs: Reuse generic_ci_match for ci comparisons
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
Now that ci_match is part of libfs, make f2fs reuse it instead of having a different implementation.
--- fs/f2fs/dir.c | 58 ++++----------------------------------------------- 1 file changed, 4 insertions(+), 54 deletions(-)
fs/f2fs/dir.c +4 -54
@@ -185,58 +185,6 @@ static struct f2fs_dir_entry *find_in_block(struct inode *dir,return f2fs_find_target_dentry (&d, fname, max_slots); } -#if IS_ENABLED (CONFIG_UNICODE ) -/* -* Test whether a case-insensitive directory entry matches the filename -* being searched for. -* -* Returns 1 for a match, 0 for no match, and -errno on an error. -*/ -static int f2fs_match_ci_name (const struct inode *dir,const struct qstr *name, -const u8 *de_name,u32 de_name_len) -{ -const struct super_block *sb = dir->i_sb; -const struct unicode_map *um = sb->s_encoding; -struct fscrypt_str decrypted_name =FSTR_INIT (NULL , de_name_len); -struct qstr entry =QSTR_INIT (de_name, de_name_len); -int res; - -if (IS_ENCRYPTED (dir)) { -const struct fscrypt_str encrypted_name = -FSTR_INIT ((u8 *)de_name, de_name_len); - -if (WARN_ON_ONCE (!fscrypt_has_encryption_key (dir))) -return -EINVAL ; - - decrypted_name.name =kmalloc (de_name_len,GFP_KERNEL ); -if (!decrypted_name.name) -return -ENOMEM ; - res =fscrypt_fname_disk_to_usr (dir,0 ,0 , &encrypted_name, - &decrypted_name); -if (res <0 ) - goto out; - entry.name = decrypted_name.name; - entry.len = decrypted_name.len; - } - - res =utf8_strncasecmp_folded (um, name, &entry); -/* -* In strict mode, ignore invalid names. In non-strict mode, -* fall back to treating them as opaque byte sequences. -*/ -if (res <0 && !sb_has_strict_encoding (sb)) { - res = name->len == entry.len && -memcmp (name->name, entry.name, name->len) ==0 ; - }else { -/* utf8_strncasecmp_folded returns 0 on match */ - res = (res ==0 ); - } -out: -kfree (decrypted_name.name); -return res; -} -#endif /* CONFIG_UNICODE */ -static inline int f2fs_match_name (const struct inode *dir,const struct f2fs_filename *fname,const u8 *de_name,u32 de_name_len)
@@ -245,8 +193,10 @@ static inline int f2fs_match_name(const struct inode *dir,#if IS_ENABLED (CONFIG_UNICODE )if (fname->cf_name.name) -return f2fs_match_ci_name (dir, &fname->cf_name, - de_name, de_name_len); +return generic_ci_match (dir, fname->usr_fname, + &fname->cf_name, + de_name, de_name_len); +#endif f.usr_fname = fname->usr_fname; f.disk_name = fname->disk_name;
Signature · 1 line
2.34.1
PATCH v12 6/8 ext4: Log error when lookup of encoded dentry fails
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
If the volume is in strict mode, ext4_ci_compare can report a broken encoding name. This will not trigger on a bad lookup, which is caught earlier, only if the actual disk name is bad.
--- fs/ext4/namei.c | 3 +++ 1 file changed, 3 insertions(+)
fs/ext4/namei.c +3 -0
@@ -1477,6 +1477,9 @@ static bool ext4_match(struct inode *parent,* only case where it happens is on a disk * corruption or ENOMEM. */ +if (ret == -EINVAL ) +EXT4_ERROR_INODE (parent, +"Directory contains filename that is invalid UTF-8" );return false; }return ret;
Signature · 1 line
2.34.1
Re: [PATCH v12 6/8] ext4: Log error when lookup of encoded dentry fails
Quoted 18 lines · Eugen Hristev
From: Gabriel Krisman Bertazi <[email protected]>
If the volume is in strict mode,
ext4_ci_comparecan report a broken encoding name. This will not trigger on a bad lookup, which is caught earlier, only if the actual disk name is bad.Reviewed-by Eric Biggers <[email protected]>Signed-off-by Gabriel Krisman Bertazi <[email protected]>Signed-off-by Eugen Hristev <[email protected]>--- fs/ext4/namei.c | 3 +++ 1 file changed, 3 insertions(+)fs/ext4/namei.c +0
-0
fs/ext4/namei.c static bool ext4_match(struct inode *parent, · line 1485
Quoted 9 lines
fs/ext4/namei.c +3
-0@@ -1477,6 +1477,9 @@ static bool ext4_match(struct inode *parent,* only case where it happens is on a disk * corruption or ENOMEM. */ +if (ret == -EINVAL ) +EXT4_ERROR_INODE (parent, +"Directory contains filename that is invalid UTF-8" );return false; }return ret;
Can you add a patch doing the same for f2fs?
Signature · 1 line
Gabriel Krisman Bertazi
PATCH v12 7/8 ext4: Move CONFIG_UNICODE defguards into the code flow
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
Instead of a bunch of ifdefs, make the unicode built checks part of the code flow where possible, as requested by Torvalds.
--- fs/ext4/crypto.c | 19 +++---------------- fs/ext4/ext4.h | 33 +++++++++++++++++++++------------ fs/ext4/namei.c | 14 +++++--------- fs/ext4/super.c | 4 +--- 4 files changed, 30 insertions(+), 40 deletions(-)
fs/ext4/crypto.c +3 -16
@@ -31,12 +31,7 @@ int ext4_fname_setup_filename(struct inode *dir, const struct qstr *iname,ext4_fname_from_fscrypt_name (fname, &name); -#if IS_ENABLED (CONFIG_UNICODE ) - err =ext4_fname_setup_ci_filename (dir, iname, fname); -if (err) -ext4_fname_free_filename (fname); -#endif -return err; +return ext4_fname_setup_ci_filename (dir, iname, fname); } int ext4_fname_prepare_lookup(struct inode *dir, struct dentry *dentry,
@@ -51,12 +46,7 @@ int ext4_fname_prepare_lookup(struct inode *dir, struct dentry *dentry,ext4_fname_from_fscrypt_name (fname, &name); -#if IS_ENABLED (CONFIG_UNICODE ) - err =ext4_fname_setup_ci_filename (dir, &dentry->d_name, fname); -if (err) -ext4_fname_free_filename (fname); -#endif -return err; +return ext4_fname_setup_ci_filename (dir, &dentry->d_name, fname); }void ext4_fname_free_filename (struct ext4_filename *fname)
@@ -70,10 +60,7 @@ void ext4_fname_free_filename(struct ext4_filename *fname) fname->usr_fname =NULL ; fname->disk_name.name =NULL ; -#if IS_ENABLED (CONFIG_UNICODE ) -kfree (fname->cf_name.name); - fname->cf_name.name =NULL ; -#endif +ext4_fname_free_ci_filename (fname); }static bool uuid_is_zero (__u8 u[16 ])
fs/ext4/ext4.h +21 -12
@@ -2742,8 +2742,25 @@ ext4_fsblk_t ext4_inode_to_goal_block(struct inode *);#if IS_ENABLED (CONFIG_UNICODE )extern int ext4_fname_setup_ci_filename (struct inode *dir, -const struct qstr *iname, -struct ext4_filename *fname); +const struct qstr *iname, +struct ext4_filename *fname); + +static inline void ext4_fname_free_ci_filename (struct ext4_filename *fname) +{ +kfree (fname->cf_name.name); + fname->cf_name.name =NULL ; +} +#else +static inline int ext4_fname_setup_ci_filename (struct inode *dir, +const struct qstr *iname, +struct ext4_filename *fname) +{ +return 0 ; +} + +static inline void ext4_fname_free_ci_filename (struct ext4_filename *fname) +{ +}#endif /* ext4 encryption related stuff goes here crypto.c */
@@ -2766,16 +2783,11 @@ static inline int ext4_fname_setup_filename(struct inode *dir,int lookup, struct ext4_filename *fname) { - int err = 0; fname->usr_fname = iname; fname->disk_name.name = (unsigned char *) iname->name; fname->disk_name.len = iname->len; -#if IS_ENABLED (CONFIG_UNICODE ) - err =ext4_fname_setup_ci_filename (dir, iname, fname); -#endif - -return err; +return ext4_fname_setup_ci_filename (dir, iname, fname); } static inline int ext4_fname_prepare_lookup(struct inode *dir,
@@ -2787,10 +2799,7 @@ static inline int ext4_fname_prepare_lookup(struct inode *dir,static inline void ext4_fname_free_filename (struct ext4_filename *fname) { -#if IS_ENABLED (CONFIG_UNICODE ) -kfree (fname->cf_name.name); - fname->cf_name.name =NULL ; -#endif +ext4_fname_free_ci_filename (fname); } static inline int ext4_ioctl_get_encryption_pwsalt(struct file *filp,
fs/ext4/namei.c +5 -9
@@ -1835,8 +1835,7 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, unsi } } -#if IS_ENABLED (CONFIG_UNICODE ) -if (!inode &&IS_CASEFOLDED (dir)) { +if (IS_ENABLED (CONFIG_UNICODE ) && !inode &&IS_CASEFOLDED (dir)) {/* Eventually we want to call d_add_ci(dentry, NULL) * for negative dentries in the encoding case as * well. For now, prevent the negative dentry
@@ -1844,7 +1843,7 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, unsi*/ return NULL ; } -#endif +return d_splice_alias (inode, dentry); }
@@ -3174,16 +3173,14 @@ static int ext4_rmdir(struct inode *dir, struct dentry *dentry)ext4_fc_track_unlink (handle, dentry); retval =ext4_mark_inode_dirty (handle, dir); -#if IS_ENABLED (CONFIG_UNICODE )/* VFS negative dentries are incompatible with Encoding and * Case-insensitiveness. Eventually we'll want avoid * invalidating the dentries here, alongside with returning the * negative dentries at ext4_lookup(), when it is better * supported by the VFS for the CI case. */ -if (IS_CASEFOLDED (dir)) +if (IS_ENABLED (CONFIG_UNICODE ) &&IS_CASEFOLDED (dir))d_invalidate (dentry); -#endif end_rmdir:brelse (bh);
@@ -3285,16 +3282,15 @@ static int ext4_unlink(struct inode *dir, struct dentry *dentry) goto out_trace; retval =__ext4_unlink (dir, &dentry->d_name,d_inode (dentry), dentry); -#if IS_ENABLED (CONFIG_UNICODE ) +/* VFS negative dentries are incompatible with Encoding and * Case-insensitiveness. Eventually we'll want avoid * invalidating the dentries here, alongside with returning the * negative dentries at ext4_lookup(), when it is better * supported by the VFS for the CI case. */ -if (IS_CASEFOLDED (dir)) +if (IS_ENABLED (CONFIG_UNICODE ) &&IS_CASEFOLDED (dir))d_invalidate (dentry); -#endif out_trace:trace_ext4_unlink_exit (dentry, retval);
fs/ext4/super.c +1 -3
@@ -3609,14 +3609,12 @@ int ext4_feature_set_ok(struct super_block *sb, int readonly)return 0 ; } -#if !IS_ENABLED (CONFIG_UNICODE ) -if (ext4_has_feature_casefold (sb)) { +if (!IS_ENABLED (CONFIG_UNICODE ) &&ext4_has_feature_casefold (sb)) {ext4_msg (sb,KERN_ERR ,"Filesystem with casefold feature cannot be " "mounted without CONFIG_UNICODE" );return 0 ; } -#endif if (readonly)return 1 ;
Signature · 1 line
2.34.1
PATCH v12 8/8 f2fs: Move CONFIG_UNICODE defguards into the code flow
In reply to Eugen HristevFrom: Gabriel Krisman Bertazi <[email protected]>
Instead of a bunch of ifdefs, make the unicode built checks part of the code flow where possible, as requested by Torvalds.
--- fs/f2fs/namei.c | 10 ++++------ fs/f2fs/super.c | 8 ++++---- 2 files changed, 8 insertions(+), 10 deletions(-)
fs/f2fs/namei.c +4 -6
@@ -577,8 +577,7 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry, goto out_iput; } out_splice: -#if IS_ENABLED (CONFIG_UNICODE ) -if (!inode &&IS_CASEFOLDED (dir)) { +if (IS_ENABLED (CONFIG_UNICODE ) && !inode &&IS_CASEFOLDED (dir)) {/* Eventually we want to call d_add_ci(dentry, NULL) * for negative dentries in the encoding case as * well. For now, prevent the negative dentry
@@ -587,7 +586,7 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry,trace_f2fs_lookup_end (dir, dentry, ino, err);return NULL ; } -#endif + new =d_splice_alias (inode, dentry);trace_f2fs_lookup_end (dir, !IS_ERR_OR_NULL (new) ? new : dentry, ino,IS_ERR (new) ?PTR_ERR (new) : err);
@@ -640,16 +639,15 @@ static int f2fs_unlink(struct inode *dir, struct dentry *dentry)f2fs_delete_entry (de, page, dir, inode);f2fs_unlock_op (sbi); -#if IS_ENABLED (CONFIG_UNICODE )/* VFS negative dentries are incompatible with Encoding and * Case-insensitiveness. Eventually we'll want avoid * invalidating the dentries here, alongside with returning the * negative dentries at f2fs_lookup(), when it is better * supported by the VFS for the CI case. */ -if (IS_CASEFOLDED (dir)) +if (IS_ENABLED (CONFIG_UNICODE ) &&IS_CASEFOLDED (dir))d_invalidate (dentry); -#endif +if (IS_DIRSYNC (dir))f2fs_sync_fs (sbi->sb,1 ); fail:
fs/f2fs/super.c +4 -4
@@ -312,7 +312,7 @@ struct kmem_cache *f2fs_cf_name_slab;static int __init f2fs_create_casefold_cache (void ) { f2fs_cf_name_slab =f2fs_kmem_cache_create ("f2fs_casefolded_name" , -F2FS_NAME_LEN ); +F2FS_NAME_LEN );return f2fs_cf_name_slab ?0 : -ENOMEM ; }
@@ -1360,13 +1360,13 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount) return -EINVAL; }#endif -#if !IS_ENABLED (CONFIG_UNICODE ) -if (f2fs_sb_has_casefold (sbi)) { + +if (!IS_ENABLED (CONFIG_UNICODE ) &&f2fs_sb_has_casefold (sbi)) {f2fs_err (sbi,"Filesystem with casefold feature cannot be mounted without CONFIG_UNICODE" );return -EINVAL ; } -#endif +/* * The BLKZONED feature indicates that the drive was formatted with * zone alignment optimization. This is optional for host-aware
Signature · 1 line
2.34.1
Quoted 4 lines · Eugen Hristev
Hello,
I am trying to respin the series here :
https://www.spinics.net/lists/linux-ext4/msg85081.html
This has a reviewed-by tag from Eric, but since its been years and we've been going through more changes now, I'd ask you to drop the r-b until Eric has had a chance to review it and give a new tag.
Thanks,
Quoted 53 lines
I resent some of the v9 patches and got some reviews from Gabriel, I did changes as requested and here is v12.
Changes in v12:
- revert to v10 comparison with propagating the error code from utf comparisonChanges in v11:
- revert to the original v9 implementation for the comparison helper.Changes in v10:
- reworked a bit the comparison helper to improve performance by first performing the exact lookup.* Original commit letter
The case-insensitive implementations in f2fs and ext4 have quite a bit of duplicated code. This series simplifies the ext4 version, with the goal of extracting
ext4_ci_compareinto a helper library that can be used by both filesystems. It also reduces the clutter from many codeguards forCONFIG_UNICODE; as requested by Linus, they are part of the codeflow now.While there, I noticed we can leverage the utf8 functions to detect encoded names that are corrupted in the filesystem. Therefore, it also adds an ext4 error on that scenario, to mark the filesystem as corrupted.
This series survived passes of xfstests -g quick.
Gabriel Krisman Bertazi (8):
ext4: Simplify the handling of cached insensitive names
f2fs: Simplify the handling of cached insensitive names
libfs: Introduce case-insensitive string comparison helper
ext4: Reusegeneric_ci_matchfor ci comparisons
f2fs: Reusegeneric_ci_matchfor ci comparisons
ext4: Log error when lookup of encoded dentry fails
ext4: MoveCONFIG_UNICODEdefguards into the code flow
f2fs: MoveCONFIG_UNICODEdefguards into the code flowfs/ext4/crypto.c | 19 ++----- fs/ext4/ext4.h | 35 +++++++----- fs/ext4/namei.c | 129 ++++++++++++++++----------------------------- fs/ext4/super.c | 4 +- fs/f2fs/dir.c | 105 +++++++++++------------------------- fs/f2fs/f2fs.h | 17 +++++- fs/f2fs/namei.c | 10 ++-- fs/f2fs/recovery.c | 5 +- fs/f2fs/super.c | 8 +-- fs/libfs.c | 85 +++++++++++++++++++++++++++++ include/linux/fs.h | 4 ++ 11 files changed, 216 insertions(+), 205 deletions(-)
Signature · 1 line
Gabriel Krisman Bertazi