From b3e6bcb94590dea45396b9481e47b809b1be4afa Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 23 May 2023 23:49:48 -0400 Subject: [PATCH 1/6] ext4: add EA_INODE checking to ext4_iget() Add a new flag, EXT4_IGET_EA_INODE which indicates whether the inode is expected to have the EA_INODE flag or not. If the flag is not set/clear as expected, then fail the iget() operation and mark the file system as corrupted. This commit also makes the ext4_iget() always perform the is_bad_inode() check even when the inode is already inode cache. This allows us to remove the is_bad_inode() check from the callers of ext4_iget() in the ea_inode code. Reported-by: syzbot+cbb68193bdb95af4340a@syzkaller.appspotmail.com Reported-by: syzbot+62120febbd1ee3c3c860@syzkaller.appspotmail.com Reported-by: syzbot+edce54daffee36421b4c@syzkaller.appspotmail.com Cc: stable@kernel.org Signed-off-by: Theodore Ts'o Link: https://lore.kernel.org/r/20230524034951.779531-2-tytso@mit.edu Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 3 ++- fs/ext4/inode.c | 31 ++++++++++++++++++++++++++----- fs/ext4/xattr.c | 36 +++++++----------------------------- 3 files changed, 35 insertions(+), 35 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 6948d673bba2..9525c52b78dc 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2901,7 +2901,8 @@ typedef enum { EXT4_IGET_NORMAL = 0, EXT4_IGET_SPECIAL = 0x0001, /* OK to iget a system inode */ EXT4_IGET_HANDLE = 0x0002, /* Inode # is from a handle */ - EXT4_IGET_BAD = 0x0004 /* Allow to iget a bad inode */ + EXT4_IGET_BAD = 0x0004, /* Allow to iget a bad inode */ + EXT4_IGET_EA_INODE = 0x0008 /* Inode should contain an EA value */ } ext4_iget_flags; extern struct inode *__ext4_iget(struct super_block *sb, unsigned long ino, diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index ce5f21b6c2b3..258f3cbed347 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -4641,6 +4641,21 @@ static inline void ext4_inode_set_iversion_queried(struct inode *inode, u64 val) inode_set_iversion_queried(inode, val); } +static const char *check_igot_inode(struct inode *inode, ext4_iget_flags flags) + +{ + if (flags & EXT4_IGET_EA_INODE) { + if (!(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL)) + return "missing EA_INODE flag"; + } else { + if ((EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL)) + return "unexpected EA_INODE flag"; + } + if (is_bad_inode(inode) && !(flags & EXT4_IGET_BAD)) + return "unexpected bad inode w/o EXT4_IGET_BAD"; + return NULL; +} + struct inode *__ext4_iget(struct super_block *sb, unsigned long ino, ext4_iget_flags flags, const char *function, unsigned int line) @@ -4650,6 +4665,7 @@ struct inode *__ext4_iget(struct super_block *sb, unsigned long ino, struct ext4_inode_info *ei; struct ext4_super_block *es = EXT4_SB(sb)->s_es; struct inode *inode; + const char *err_str; journal_t *journal = EXT4_SB(sb)->s_journal; long ret; loff_t size; @@ -4677,8 +4693,14 @@ struct inode *__ext4_iget(struct super_block *sb, unsigned long ino, inode = iget_locked(sb, ino); if (!inode) return ERR_PTR(-ENOMEM); - if (!(inode->i_state & I_NEW)) + if (!(inode->i_state & I_NEW)) { + if ((err_str = check_igot_inode(inode, flags)) != NULL) { + ext4_error_inode(inode, function, line, 0, err_str); + iput(inode); + return ERR_PTR(-EFSCORRUPTED); + } return inode; + } ei = EXT4_I(inode); iloc.bh = NULL; @@ -4944,10 +4966,9 @@ struct inode *__ext4_iget(struct super_block *sb, unsigned long ino, if (IS_CASEFOLDED(inode) && !ext4_has_feature_casefold(inode->i_sb)) ext4_error_inode(inode, function, line, 0, "casefold flag without casefold feature"); - if (is_bad_inode(inode) && !(flags & EXT4_IGET_BAD)) { - ext4_error_inode(inode, function, line, 0, - "bad inode without EXT4_IGET_BAD flag"); - ret = -EUCLEAN; + if ((err_str = check_igot_inode(inode, flags)) != NULL) { + ext4_error_inode(inode, function, line, 0, err_str); + ret = -EFSCORRUPTED; goto bad_inode; } diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c index dfc2e223bd10..a27208129a80 100644 --- a/fs/ext4/xattr.c +++ b/fs/ext4/xattr.c @@ -433,7 +433,7 @@ static int ext4_xattr_inode_iget(struct inode *parent, unsigned long ea_ino, return -EFSCORRUPTED; } - inode = ext4_iget(parent->i_sb, ea_ino, EXT4_IGET_NORMAL); + inode = ext4_iget(parent->i_sb, ea_ino, EXT4_IGET_EA_INODE); if (IS_ERR(inode)) { err = PTR_ERR(inode); ext4_error(parent->i_sb, @@ -441,23 +441,6 @@ static int ext4_xattr_inode_iget(struct inode *parent, unsigned long ea_ino, err); return err; } - - if (is_bad_inode(inode)) { - ext4_error(parent->i_sb, - "error while reading EA inode %lu is_bad_inode", - ea_ino); - err = -EIO; - goto error; - } - - if (!(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL)) { - ext4_error(parent->i_sb, - "EA inode %lu does not have EXT4_EA_INODE_FL flag", - ea_ino); - err = -EINVAL; - goto error; - } - ext4_xattr_inode_set_class(inode); /* @@ -478,9 +461,6 @@ static int ext4_xattr_inode_iget(struct inode *parent, unsigned long ea_ino, *ea_inode = inode; return 0; -error: - iput(inode); - return err; } /* Remove entry from mbcache when EA inode is getting evicted */ @@ -1556,11 +1536,10 @@ ext4_xattr_inode_cache_find(struct inode *inode, const void *value, while (ce) { ea_inode = ext4_iget(inode->i_sb, ce->e_value, - EXT4_IGET_NORMAL); - if (!IS_ERR(ea_inode) && - !is_bad_inode(ea_inode) && - (EXT4_I(ea_inode)->i_flags & EXT4_EA_INODE_FL) && - i_size_read(ea_inode) == value_len && + EXT4_IGET_EA_INODE); + if (IS_ERR(ea_inode)) + goto next_entry; + if (i_size_read(ea_inode) == value_len && !ext4_xattr_inode_read(ea_inode, ea_data, value_len) && !ext4_xattr_inode_verify_hashes(ea_inode, NULL, ea_data, value_len) && @@ -1570,9 +1549,8 @@ ext4_xattr_inode_cache_find(struct inode *inode, const void *value, kvfree(ea_data); return ea_inode; } - - if (!IS_ERR(ea_inode)) - iput(ea_inode); + iput(ea_inode); + next_entry: ce = mb_cache_entry_find_next(ea_inode_cache, ce); } kvfree(ea_data); From b928dfdcb27d8fa59917b794cfba53052a2f050f Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 23 May 2023 23:49:49 -0400 Subject: [PATCH 2/6] ext4: set lockdep subclass for the ea_inode in ext4_xattr_inode_cache_find() If the ea_inode has been pushed out of the inode cache while there is still a reference in the mb_cache, the lockdep subclass will not be set on the inode, which can lead to some lockdep false positives. Fixes: 33d201e0277b ("ext4: fix lockdep warning about recursive inode locking") Cc: stable@kernel.org Reported-by: syzbot+d4b971e744b1f5439336@syzkaller.appspotmail.com Signed-off-by: Theodore Ts'o Link: https://lore.kernel.org/r/20230524034951.779531-3-tytso@mit.edu Signed-off-by: Theodore Ts'o --- fs/ext4/xattr.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c index a27208129a80..ff7ab63c5b4f 100644 --- a/fs/ext4/xattr.c +++ b/fs/ext4/xattr.c @@ -1539,6 +1539,7 @@ ext4_xattr_inode_cache_find(struct inode *inode, const void *value, EXT4_IGET_EA_INODE); if (IS_ERR(ea_inode)) goto next_entry; + ext4_xattr_inode_set_class(ea_inode); if (i_size_read(ea_inode) == value_len && !ext4_xattr_inode_read(ea_inode, ea_data, value_len) && !ext4_xattr_inode_verify_hashes(ea_inode, NULL, ea_data, From 2bc7e7c1a3bc9bd0cbf0f71006f6fe7ef24a00c2 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 23 May 2023 23:49:50 -0400 Subject: [PATCH 3/6] ext4: disallow ea_inodes with extended attributes An ea_inode stores the value of an extended attribute; it can not have extended attributes itself, or this will cause recursive nightmares. Add a check in ext4_iget() to make sure this is the case. Cc: stable@kernel.org Reported-by: syzbot+e44749b6ba4d0434cd47@syzkaller.appspotmail.com Signed-off-by: Theodore Ts'o Link: https://lore.kernel.org/r/20230524034951.779531-4-tytso@mit.edu Signed-off-by: Theodore Ts'o --- fs/ext4/inode.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 258f3cbed347..02de439bf1f0 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -4647,6 +4647,9 @@ static const char *check_igot_inode(struct inode *inode, ext4_iget_flags flags) if (flags & EXT4_IGET_EA_INODE) { if (!(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL)) return "missing EA_INODE flag"; + if (ext4_test_inode_state(inode, EXT4_STATE_XATTR) || + EXT4_I(inode)->i_file_acl) + return "ea_inode with extended attributes"; } else { if ((EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL)) return "unexpected EA_INODE flag"; From aff3bea95388299eec63440389b4545c8041b357 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Tue, 23 May 2023 23:49:51 -0400 Subject: [PATCH 4/6] ext4: add lockdep annotations for i_data_sem for ea_inode's Treat i_data_sem for ea_inodes as being in their own lockdep class to avoid lockdep complaints about ext4_setattr's use of inode_lock() on normal inodes potentially causing lock ordering with i_data_sem on ea_inodes in ext4_xattr_inode_write(). However, ea_inodes will be operated on by ext4_setattr(), so this isn't a problem. Cc: stable@kernel.org Link: https://syzkaller.appspot.com/bug?extid=298c5d8fb4a128bc27b0 Reported-by: syzbot+298c5d8fb4a128bc27b0@syzkaller.appspotmail.com Signed-off-by: Theodore Ts'o Link: https://lore.kernel.org/r/20230524034951.779531-5-tytso@mit.edu Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 2 ++ fs/ext4/xattr.c | 4 ++++ 2 files changed, 6 insertions(+) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 9525c52b78dc..8104a21b001a 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -918,11 +918,13 @@ do { \ * where the second inode has larger inode number * than the first * I_DATA_SEM_QUOTA - Used for quota inodes only + * I_DATA_SEM_EA - Used for ea_inodes only */ enum { I_DATA_SEM_NORMAL = 0, I_DATA_SEM_OTHER, I_DATA_SEM_QUOTA, + I_DATA_SEM_EA }; diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c index ff7ab63c5b4f..13d7f17a9c8c 100644 --- a/fs/ext4/xattr.c +++ b/fs/ext4/xattr.c @@ -121,7 +121,11 @@ ext4_expand_inode_array(struct ext4_xattr_inode_array **ea_inode_array, #ifdef CONFIG_LOCKDEP void ext4_xattr_inode_set_class(struct inode *ea_inode) { + struct ext4_inode_info *ei = EXT4_I(ea_inode); + lockdep_set_subclass(&ea_inode->i_rwsem, 1); + (void) ei; /* shut up clang warning if !CONFIG_LOCKDEP */ + lockdep_set_subclass(&ei->i_data_sem, I_DATA_SEM_EA); } #endif From 1077b2d53ef53629c14106aecf633bebd286c04c Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 24 May 2023 12:44:53 +0200 Subject: [PATCH 5/6] ext4: fix fsync for non-directories Commit e360c6ed7274 ("ext4: Drop special handling of journalled data from ext4_sync_file()") simplified ext4_sync_file() by dropping special handling of journalled data mode as it was not needed anymore. However that branch was also used for directories and symlinks and since the fastcommit code does not track metadata changes to non-regular files, the change has caused e.g. fsync(2) on directories to not commit transaction as it should. Fix the problem by adding handling for non-regular files. Fixes: e360c6ed7274 ("ext4: Drop special handling of journalled data from ext4_sync_file()") Reported-by: Eric Whitney Link: https://lore.kernel.org/all/ZFqO3xVnmhL7zv1x@debian-BULLSEYE-live-builder-AMD64 Signed-off-by: Jan Kara Tested-by: Eric Whitney Link: https://lore.kernel.org/r/20230524104453.8734-1-jack@suse.cz Signed-off-by: Theodore Ts'o --- fs/ext4/fsync.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c index f65fdb27ce14..2a143209aa0c 100644 --- a/fs/ext4/fsync.c +++ b/fs/ext4/fsync.c @@ -108,6 +108,13 @@ static int ext4_fsync_journal(struct inode *inode, bool datasync, journal_t *journal = EXT4_SB(inode->i_sb)->s_journal; tid_t commit_tid = datasync ? ei->i_datasync_tid : ei->i_sync_tid; + /* + * Fastcommit does not really support fsync on directories or other + * special files. Force a full commit. + */ + if (!S_ISREG(inode->i_mode)) + return ext4_force_commit(inode->i_sb); + if (journal->j_flags & JBD2_BARRIER && !jbd2_trans_will_send_data_barrier(journal, commit_tid)) *needs_barrier = true; From eb1f822c76beeaa76ab8b6737ab9dc9f9798408c Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Fri, 26 May 2023 23:57:29 -0400 Subject: [PATCH 6/6] ext4: enable the lazy init thread when remounting read/write In commit a44be64bbecb ("ext4: don't clear SB_RDONLY when remounting r/w until quota is re-enabled") we defer clearing tyhe SB_RDONLY flag in struct super. However, we didn't defer when we checked sb_rdonly() to determine the lazy itable init thread should be enabled, with the next result that the lazy inode table initialization would not be properly started. This can cause generic/231 to fail in ext4's nojournal mode. Fix this by moving when we decide to start or stop the lazy itable init thread to after we clear the SB_RDONLY flag when we are remounting the file system read/write. Fixes a44be64bbecb ("ext4: don't clear SB_RDONLY when remounting r/w until...") Signed-off-by: Theodore Ts'o Link: https://lore.kernel.org/r/20230527035729.1001605-1-tytso@mit.edu Signed-off-by: Theodore Ts'o --- fs/ext4/super.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 9680fe753e59..56a5d1c469fc 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -6588,18 +6588,6 @@ static int __ext4_remount(struct fs_context *fc, struct super_block *sb) } } - /* - * Reinitialize lazy itable initialization thread based on - * current settings - */ - if (sb_rdonly(sb) || !test_opt(sb, INIT_INODE_TABLE)) - ext4_unregister_li_request(sb); - else { - ext4_group_t first_not_zeroed; - first_not_zeroed = ext4_has_uninit_itable(sb); - ext4_register_li_request(sb, first_not_zeroed); - } - /* * Handle creation of system zone data early because it can fail. * Releasing of existing data is done when we are sure remount will @@ -6637,6 +6625,18 @@ static int __ext4_remount(struct fs_context *fc, struct super_block *sb) if (enable_rw) sb->s_flags &= ~SB_RDONLY; + /* + * Reinitialize lazy itable initialization thread based on + * current settings + */ + if (sb_rdonly(sb) || !test_opt(sb, INIT_INODE_TABLE)) + ext4_unregister_li_request(sb); + else { + ext4_group_t first_not_zeroed; + first_not_zeroed = ext4_has_uninit_itable(sb); + ext4_register_li_request(sb, first_not_zeroed); + } + if (!ext4_has_feature_mmp(sb) || sb_rdonly(sb)) ext4_stop_mmpd(sbi);