[v6.6] possible deadlock in ext4_evict_inode

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: 0a805b6ea8cd Linux 6.6.116
git tree: linux-6.6.y
console output: https://syzkaller.appspot.com/x/log.txt?x=176c332f980000
kernel config: https://syzkaller.appspot.com/x/.config?x=12606d4b8832c7e4
dashboard link: https://syzkaller.appspot.com/bug?extid=5135ee32614ae9e41ee0
compiler: Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8

Unfortunately, I don't have any reproducer for this issue yet.

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/13874446b4b2/disk-0a805b6e.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/69a5e59e1ade/vmlinux-0a805b6e.xz
kernel image: https://storage.googleapis.com/syzbot-assets/322c0711f976/bzImage-0a805b6e.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+5135ee...@syzkaller.appspotmail.com

ext4 filesystem being mounted at /365/file0aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa supports timestamps until 2038-01-19 (0x7fffffff)
======================================================
WARNING: possible circular locking dependency detected
syzkaller #0 Not tainted
------------------------------------------------------
syz.6.2751/16898 is trying to acquire lock:
ffff88805bbc4608 (sb_internal){.+.+}-{0:0}, at: __sb_start_write include/linux/fs.h:1633 [inline]
ffff88805bbc4608 (sb_internal){.+.+}-{0:0}, at: sb_start_intwrite include/linux/fs.h:1755 [inline]
ffff88805bbc4608 (sb_internal){.+.+}-{0:0}, at: ext4_evict_inode+0x2b9/0xea0 fs/ext4/inode.c:213

but task is already holding lock:
ffff88807b1b8bd8 (&sbi->s_writepages_rwsem){++++}-{0:0}, at: ext4_writepages_down_write fs/ext4/ext4.h:1779 [inline]
ffff88807b1b8bd8 (&sbi->s_writepages_rwsem){++++}-{0:0}, at: ext4_ext_migrate+0x2f1/0xff0 fs/ext4/migrate.c:437

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (&sbi->s_writepages_rwsem){++++}-{0:0}:
percpu_down_read+0x44/0x1a0 include/linux/percpu-rwsem.h:51
ext4_writepages_down_read fs/ext4/ext4.h:1767 [inline]
ext4_writepages+0x170/0x2f0 fs/ext4/inode.c:2806
do_writepages+0x3a2/0x600 mm/page-writeback.c:2575
__writeback_single_inode+0x153/0xee0 fs/fs-writeback.c:1635
writeback_single_inode+0x211/0x720 fs/fs-writeback.c:1756
write_inode_now+0x161/0x1e0 fs/fs-writeback.c:2802
iput_final fs/inode.c:1807 [inline]
iput+0x5b2/0x920 fs/inode.c:1846
ext4_xattr_block_set+0x273a/0x32a0 fs/ext4/xattr.c:2211
ext4_xattr_move_to_block fs/ext4/xattr.c:2679 [inline]
ext4_xattr_make_inode_space fs/ext4/xattr.c:2754 [inline]
ext4_expand_extra_isize_ea+0x10ea/0x19e0 fs/ext4/xattr.c:2846
__ext4_expand_extra_isize+0x306/0x400 fs/ext4/inode.c:5967
ext4_try_to_expand_extra_isize fs/ext4/inode.c:6010 [inline]
__ext4_mark_inode_dirty+0x45d/0x6e0 fs/ext4/inode.c:6088
ext4_evict_inode+0x7ed/0xea0 fs/ext4/inode.c:252
evict+0x486/0x870 fs/inode.c:705
ext4_orphan_cleanup+0xbd4/0x1400 fs/ext4/orphan.c:470
__ext4_fill_super fs/ext4/super.c:5601 [inline]
ext4_fill_super+0x5de4/0x66c0 fs/ext4/super.c:5724
get_tree_bdev+0x3e4/0x510 fs/super.c:1591
vfs_get_tree+0x8c/0x280 fs/super.c:1764
do_new_mount+0x24b/0xa40 fs/namespace.c:3386
do_mount fs/namespace.c:3726 [inline]
__do_sys_mount fs/namespace.c:3935 [inline]
__se_sys_mount+0x2da/0x3c0 fs/namespace.c:3912
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x55/0xb0 arch/x86/entry/common.c:81
entry_SYSCALL_64_after_hwframe+0x68/0xd2

-> #0 (sb_internal){.+.+}-{0:0}:
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+0x2ddb/0x7c80 kernel/locking/lockdep.c:5137
lock_acquire+0x197/0x410 kernel/locking/lockdep.c:5754
percpu_down_read+0x44/0x1a0 include/linux/percpu-rwsem.h:51
__sb_start_write include/linux/fs.h:1633 [inline]
sb_start_intwrite include/linux/fs.h:1755 [inline]
ext4_evict_inode+0x2b9/0xea0 fs/ext4/inode.c:213
evict+0x486/0x870 fs/inode.c:705
ext4_ext_migrate+0xcfb/0xff0 fs/ext4/migrate.c:587
__ext4_ioctl fs/ext4/ioctl.c:1409 [inline]
ext4_ioctl+0x1c4b/0x3820 fs/ext4/ioctl.c:1627
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:871 [inline]
__se_sys_ioctl+0xfd/0x170 fs/ioctl.c:857
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x55/0xb0 arch/x86/entry/common.c:81
entry_SYSCALL_64_after_hwframe+0x68/0xd2

other info that might help us debug this:

Possible unsafe locking scenario:

CPU0 CPU1
---- ----
lock(&sbi->s_writepages_rwsem);
lock(sb_internal);
lock(&sbi->s_writepages_rwsem);
rlock(sb_internal);

*** DEADLOCK ***

3 locks held by syz.6.2751/16898:
#0: ffff88805bbc4418 (sb_writers#4){.+.+}-{0:0}, at: mnt_want_write_file+0x63/0x200 fs/namespace.c:456
#1: ffff88805de1de10 (&sb->s_type->i_mutex_key#9){++++}-{3:3}, at: inode_lock include/linux/fs.h:804 [inline]
#1: ffff88805de1de10 (&sb->s_type->i_mutex_key#9){++++}-{3:3}, at: __ext4_ioctl fs/ext4/ioctl.c:1408 [inline]
#1: ffff88805de1de10 (&sb->s_type->i_mutex_key#9){++++}-{3:3}, at: ext4_ioctl+0x1c43/0x3820 fs/ext4/ioctl.c:1627
#2: ffff88807b1b8bd8 (&sbi->s_writepages_rwsem){++++}-{0:0}, at: ext4_writepages_down_write fs/ext4/ext4.h:1779 [inline]
#2: ffff88807b1b8bd8 (&sbi->s_writepages_rwsem){++++}-{0:0}, at: ext4_ext_migrate+0x2f1/0xff0 fs/ext4/migrate.c:437

stack backtrace:
CPU: 0 PID: 16898 Comm: syz.6.2751 Not tainted syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/02/2025
Call Trace:
<TASK>
dump_stack_lvl+0x16c/0x230 lib/dump_stack.c:106
check_noncircular+0x2bd/0x3c0 kernel/locking/lockdep.c:2187
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+0x2ddb/0x7c80 kernel/locking/lockdep.c:5137
lock_acquire+0x197/0x410 kernel/locking/lockdep.c:5754
percpu_down_read+0x44/0x1a0 include/linux/percpu-rwsem.h:51
__sb_start_write include/linux/fs.h:1633 [inline]
sb_start_intwrite include/linux/fs.h:1755 [inline]
ext4_evict_inode+0x2b9/0xea0 fs/ext4/inode.c:213
evict+0x486/0x870 fs/inode.c:705
ext4_ext_migrate+0xcfb/0xff0 fs/ext4/migrate.c:587
__ext4_ioctl fs/ext4/ioctl.c:1409 [inline]
ext4_ioctl+0x1c4b/0x3820 fs/ext4/ioctl.c:1627
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:871 [inline]
__se_sys_ioctl+0xfd/0x170 fs/ioctl.c:857
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x55/0xb0 arch/x86/entry/common.c:81
entry_SYSCALL_64_after_hwframe+0x68/0xd2
RIP: 0033:0x7fb44198f6c9
Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 a8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fb44285d038 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 00007fb441be5fa0 RCX: 00007fb44198f6c9
RDX: 0000000000000000 RSI: 0000000000006609 RDI: 0000000000000004
RBP: 00007fb441a11f91 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007fb441be6038 R14: 00007fb441be5fa0 R15: 00007fff48ee15b8
</TASK>
EXT4-fs (loop6): unmounting filesystem 76b65be2-f6da-4727-8c75-0525a5b65a09.


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

If the report is already addressed, let syzbot know by replying with:
#syz fix: exact-commit-title

If you want to overwrite report's subsystems, reply with:
#syz set subsystems: new-subsystem
(See the list of subsystem names on the web dashboard)

If the report is a duplicate of another one, reply with:
#syz dup: exact-subject-of-another-report

If you want to undo deduplication, reply with:
#syz undup