[PATCH] ext4: fix deadlock in ext4_xattr_inode_iget

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



[Syzbot reported]
WARNING: possible circular locking dependency detected
6.8.0-syzkaller-08951-gfe46a7dd189e #0 Not tainted
------------------------------------------------------
syz-executor545/5275 is trying to acquire lock:
ffff888077730400 (&ea_inode->i_rwsem#8/1){+.+.}-{3:3}, at: inode_lock include/linux/fs.h:793 [inline]
ffff888077730400 (&ea_inode->i_rwsem#8/1){+.+.}-{3:3}, at: ext4_xattr_inode_iget+0x173/0x440 fs/ext4/xattr.c:461

but task is already holding lock:
ffff888077730c88 (&ei->i_data_sem/3){++++}-{3:3}, at: ext4_setattr+0x1ba0/0x29d0 fs/ext4/inode.c:5417

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (&ei->i_data_sem/3){++++}-{3:3}:
       down_write+0x3a/0x50 kernel/locking/rwsem.c:1579
       ext4_update_i_disksize fs/ext4/ext4.h:3383 [inline]
       ext4_xattr_inode_write fs/ext4/xattr.c:1446 [inline]
       ext4_xattr_inode_lookup_create fs/ext4/xattr.c:1594 [inline]
       ext4_xattr_set_entry+0x3a14/0x3cf0 fs/ext4/xattr.c:1719
       ext4_xattr_ibody_set+0x126/0x380 fs/ext4/xattr.c:2287
       ext4_xattr_set_handle+0x98d/0x1480 fs/ext4/xattr.c:2444
       ext4_xattr_set+0x149/0x380 fs/ext4/xattr.c:2558
       __vfs_setxattr+0x176/0x1e0 fs/xattr.c:200
       __vfs_setxattr_noperm+0x127/0x5e0 fs/xattr.c:234
       __vfs_setxattr_locked+0x182/0x260 fs/xattr.c:295
       vfs_setxattr+0x146/0x350 fs/xattr.c:321
       do_setxattr+0x146/0x170 fs/xattr.c:629
       setxattr+0x15d/0x180 fs/xattr.c:652
       path_setxattr+0x179/0x1e0 fs/xattr.c:671
       __do_sys_lsetxattr fs/xattr.c:694 [inline]
       __se_sys_lsetxattr fs/xattr.c:690 [inline]
       __x64_sys_lsetxattr+0xc1/0x160 fs/xattr.c:690
       do_syscall_x64 arch/x86/entry/common.c:52 [inline]
       do_syscall_64+0xd5/0x260 arch/x86/entry/common.c:83
       entry_SYSCALL_64_after_hwframe+0x6d/0x75

-> #0 (&ea_inode->i_rwsem#8/1){+.+.}-{3:3}:
       check_prev_add kernel/locking/lockdep.c:3134 [inline]
       check_prevs_add kernel/locking/lockdep.c:3253 [inline]
       validate_chain kernel/locking/lockdep.c:3869 [inline]
       __lock_acquire+0x2478/0x3b30 kernel/locking/lockdep.c:5137
       lock_acquire kernel/locking/lockdep.c:5754 [inline]
       lock_acquire+0x1b1/0x540 kernel/locking/lockdep.c:5719
       down_write+0x3a/0x50 kernel/locking/rwsem.c:1579
       inode_lock include/linux/fs.h:793 [inline]
       ext4_xattr_inode_iget+0x173/0x440 fs/ext4/xattr.c:461
       ext4_xattr_inode_get+0x16c/0x870 fs/ext4/xattr.c:535
       ext4_xattr_move_to_block fs/ext4/xattr.c:2640 [inline]
       ext4_xattr_make_inode_space fs/ext4/xattr.c:2742 [inline]
       ext4_expand_extra_isize_ea+0x1367/0x1ae0 fs/ext4/xattr.c:2834
       __ext4_expand_extra_isize+0x346/0x480 fs/ext4/inode.c:5789
       ext4_try_to_expand_extra_isize fs/ext4/inode.c:5832 [inline]
       __ext4_mark_inode_dirty+0x55a/0x860 fs/ext4/inode.c:5910
       ext4_setattr+0x1c14/0x29d0 fs/ext4/inode.c:5420
       notify_change+0x745/0x11c0 fs/attr.c:497
       do_truncate+0x15c/0x220 fs/open.c:65
       handle_truncate fs/namei.c:3300 [inline]
       do_open fs/namei.c:3646 [inline]
       path_openat+0x24b9/0x2990 fs/namei.c:3799
       do_filp_open+0x1dc/0x430 fs/namei.c:3826
       do_sys_openat2+0x17a/0x1e0 fs/open.c:1406
       do_sys_open fs/open.c:1421 [inline]
       __do_sys_openat fs/open.c:1437 [inline]
       __se_sys_openat fs/open.c:1432 [inline]
       __x64_sys_openat+0x175/0x210 fs/open.c:1432
       do_syscall_x64 arch/x86/entry/common.c:52 [inline]
       do_syscall_64+0xd5/0x260 arch/x86/entry/common.c:83
       entry_SYSCALL_64_after_hwframe+0x6d/0x75

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&ei->i_data_sem/3);
                               lock(&ea_inode->i_rwsem#8/1);
                               lock(&ei->i_data_sem/3);
  lock(&ea_inode->i_rwsem#8/1);

 *** DEADLOCK ***
[Fix]
According to mark inode dirty context, it does not need to be protected by lock
i_data_sem, and if it is protected by i_data_sem, a deadlock will occur.

Reported-by: syzbot+ee72b9a7aad1e5a77c5c@xxxxxxxxxxxxxxxxxxxxxxxxx
Signed-off-by: Edward Adam Davis <eadavis@xxxxxx>
---
 fs/ext4/inode.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 537803250ca9..d2cbe3dddfab 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -5417,6 +5417,7 @@ int ext4_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 			down_write(&EXT4_I(inode)->i_data_sem);
 			old_disksize = EXT4_I(inode)->i_disksize;
 			EXT4_I(inode)->i_disksize = attr->ia_size;
+			up_write(&EXT4_I(inode)->i_data_sem);
 			rc = ext4_mark_inode_dirty(handle, inode);
 			if (!error)
 				error = rc;
@@ -5425,6 +5426,7 @@ int ext4_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
 			 * with i_disksize to avoid races with writeback code
 			 * running ext4_wb_update_i_disksize().
 			 */
+			down_write(&EXT4_I(inode)->i_data_sem);
 			if (!error)
 				i_size_write(inode, attr->ia_size);
 			else
-- 
2.43.0





[Index of Archives]     [Reiser Filesystem Development]     [Ceph FS]     [Kernel Newbies]     [Security]     [Netfilter]     [Bugtraq]     [Linux FS]     [Yosemite National Park]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Samba]     [Device Mapper]     [Linux Media]

  Powered by Linux