syzbot


possible deadlock in ext4_file_write_iter

Status: auto-closed as invalid on 2020/10/01 00:41
Reported-by: syzbot+ccc15f2cc5d9484439c7@syzkaller.appspotmail.com
First crash: 799d, last: 799d
similar bugs (4):
Kernel Title Repro Cause bisect Fix bisect Count Last Reported Patched Status
linux-4.14 possible deadlock in ext4_file_write_iter 3 816d 868d 0/1 auto-closed as invalid on 2020/09/14 06:49
linux-4.14 possible deadlock in ext4_file_write_iter (2) syz inconclusive 13 10d 656d 0/1 upstream: reported syz repro on 2020/10/24 08:36
linux-4.19 possible deadlock in ext4_file_write_iter (2) 1 377d 377d 0/1 auto-closed as invalid on 2021/11/26 16:39
linux-4.19 possible deadlock in ext4_file_write_iter (3) syz error 8 4d05h 229d 0/1 upstream: reported syz repro on 2021/12/24 22:56

Sample crash report:
======================================================
WARNING: possible circular locking dependency detected
4.19.125-syzkaller #0 Not tainted
------------------------------------------------------
syz-executor.3/1887 is trying to acquire lock:
00000000c1f2fa45 ((wq_completion)"dio/%s"sb->s_id){+.+.}, at: flush_workqueue+0xf7/0x1440 kernel/workqueue.c:2660
FAT-fs (loop2): error, invalid access to FAT (entry 0x00000020)

but task is already holding lock:
00000000891b42bf (&sb->s_type->i_mutex_key#9){++++}, at: inode_trylock include/linux/fs.h:768 [inline]
00000000891b42bf (&sb->s_type->i_mutex_key#9){++++}, at: ext4_file_write_iter+0x21f/0xf60 fs/ext4/file.c:238

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #2 (&sb->s_type->i_mutex_key#9){++++}:
       inode_lock include/linux/fs.h:748 [inline]
       __generic_file_fsync+0xb0/0x1f0 fs/libfs.c:987
       ext4_sync_file+0xa28/0x1430 fs/ext4/fsync.c:120
       vfs_fsync_range+0x138/0x220 fs/sync.c:197
       generic_write_sync include/linux/fs.h:2747 [inline]
       dio_complete+0x47c/0x9e0 fs/direct-io.c:329
       process_one_work+0x892/0x1580 kernel/workqueue.c:2155
FAT-fs (loop2): Filesystem has been set read-only
       worker_thread+0x96/0xe20 kernel/workqueue.c:2298
       kthread+0x30b/0x410 kernel/kthread.c:246
       ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:415

-> #1 ((work_completion)(&dio->complete_work)){+.+.}:
       worker_thread+0x96/0xe20 kernel/workqueue.c:2298
FAT-fs (loop2): error, invalid access to FAT (entry 0x00000020)
       kthread+0x30b/0x410 kernel/kthread.c:246
       ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:415

-> #0 ((wq_completion)"dio/%s"sb->s_id){+.+.}:
       flush_workqueue+0x126/0x1440 kernel/workqueue.c:2663
       drain_workqueue+0x1a7/0x460 kernel/workqueue.c:2828
       destroy_workqueue+0x75/0x6f0 kernel/workqueue.c:4172
       sb_init_dio_done_wq+0x72/0x90 fs/direct-io.c:634
       do_blockdev_direct_IO fs/direct-io.c:1284 [inline]
       __blockdev_direct_IO+0x6445/0xfe24 fs/direct-io.c:1418
       ext4_direct_IO_write fs/ext4/inode.c:3769 [inline]
       ext4_direct_IO+0xa8b/0x1ab0 fs/ext4/inode.c:3902
       generic_file_direct_write+0x207/0x490 mm/filemap.c:3065
       __generic_file_write_iter+0x2d4/0x610 mm/filemap.c:3244
       ext4_file_write_iter+0x2fe/0xf60 fs/ext4/file.c:272
FAT-fs (loop5): error, invalid access to FAT (entry 0x00000020)
       call_write_iter include/linux/fs.h:1821 [inline]
       aio_write+0x346/0x530 fs/aio.c:1573
       __io_submit_one fs/aio.c:1850 [inline]
       io_submit_one+0x1005/0x20c0 fs/aio.c:1901
       __do_sys_io_submit fs/aio.c:1945 [inline]
       __se_sys_io_submit+0x142/0x4a0 fs/aio.c:1916
       do_syscall_64+0xf9/0x620 arch/x86/entry/common.c:293
       entry_SYSCALL_64_after_hwframe+0x49/0xbe
attempt to access beyond end of device
loop2: rw=34817, want=157, limit=116

other info that might help us debug this:

FAT-fs (loop5): Filesystem has been set read-only
Chain exists of:
  (wq_completion)"dio/%s"sb->s_id --> (work_completion)(&dio->complete_work) --> &sb->s_type->i_mutex_key#9

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&sb->s_type->i_mutex_key#9);
                               lock((work_completion)(&dio->complete_work));
                               lock(&sb->s_type->i_mutex_key#9);
FAT-fs (loop5): error, invalid access to FAT (entry 0x00000020)
  lock((wq_completion)"dio/%s"sb->s_id);

 *** DEADLOCK ***

1 lock held by syz-executor.3/1887:
 #0: 00000000891b42bf (&sb->s_type->i_mutex_key#9){++++}, at: inode_trylock include/linux/fs.h:768 [inline]
 #0: 00000000891b42bf (&sb->s_type->i_mutex_key#9){++++}, at: ext4_file_write_iter+0x21f/0xf60 fs/ext4/file.c:238

stack backtrace:
CPU: 0 PID: 1887 Comm: syz-executor.3 Not tainted 4.19.125-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x1fc/0x2fe lib/dump_stack.c:118
 print_circular_bug.isra.0.cold+0x2dc/0x425 kernel/locking/lockdep.c:1221
 check_prev_add kernel/locking/lockdep.c:1865 [inline]
 check_prevs_add kernel/locking/lockdep.c:1978 [inline]
 validate_chain kernel/locking/lockdep.c:2419 [inline]
 __lock_acquire+0x3145/0x4380 kernel/locking/lockdep.c:3415
 lock_acquire+0x170/0x3c0 kernel/locking/lockdep.c:3907
 flush_workqueue+0x126/0x1440 kernel/workqueue.c:2663
attempt to access beyond end of device
 drain_workqueue+0x1a7/0x460 kernel/workqueue.c:2828
 destroy_workqueue+0x75/0x6f0 kernel/workqueue.c:4172
loop5: rw=34817, want=157, limit=116
 sb_init_dio_done_wq+0x72/0x90 fs/direct-io.c:634
 do_blockdev_direct_IO fs/direct-io.c:1284 [inline]
 __blockdev_direct_IO+0x6445/0xfe24 fs/direct-io.c:1418
FAT-fs (loop4): error, invalid access to FAT (entry 0x00000020)
 ext4_direct_IO_write fs/ext4/inode.c:3769 [inline]
 ext4_direct_IO+0xa8b/0x1ab0 fs/ext4/inode.c:3902
 generic_file_direct_write+0x207/0x490 mm/filemap.c:3065
FAT-fs (loop4): Filesystem has been set read-only
 __generic_file_write_iter+0x2d4/0x610 mm/filemap.c:3244
 ext4_file_write_iter+0x2fe/0xf60 fs/ext4/file.c:272
 call_write_iter include/linux/fs.h:1821 [inline]
 aio_write+0x346/0x530 fs/aio.c:1573
FAT-fs (loop4): error, invalid access to FAT (entry 0x00000020)
 __io_submit_one fs/aio.c:1850 [inline]
 io_submit_one+0x1005/0x20c0 fs/aio.c:1901
 __do_sys_io_submit fs/aio.c:1945 [inline]
 __se_sys_io_submit+0x142/0x4a0 fs/aio.c:1916
 do_syscall_64+0xf9/0x620 arch/x86/entry/common.c:293
 entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x45ca69
Code: 0d b7 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 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 0f 83 db b6 fb ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007f2ae55ebc78 EFLAGS: 00000246 ORIG_RAX: 00000000000000d1
RAX: ffffffffffffffda RBX: 00000000004e0d40 RCX: 000000000045ca69
RDX: 0000000020000540 RSI: 0000000000000006 RDI: 00007f2ae55a8000
RBP: 000000000078bf00 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00000000ffffffff
R13: 00000000000001fb R14: 00000000004c4486 R15: 00007f2ae55ec6d4
FAT-fs (loop2): error, invalid access to FAT (entry 0x00000020)
FAT-fs (loop2): Filesystem has been set read-only
FAT-fs (loop2): error, invalid access to FAT (entry 0x00000020)
attempt to access beyond end of device
loop2: rw=34817, want=157, limit=116
9pnet: p9_errstr2errno: server reported unknown error αΘ’Vε
Can not set IPV6_FL_F_REFLECT if flowlabel_consistency sysctl is enable
Can not set IPV6_FL_F_REFLECT if flowlabel_consistency sysctl is enable
Can not set IPV6_FL_F_REFLECT if flowlabel_consistency sysctl is enable
Can not set IPV6_FL_F_REFLECT if flowlabel_consistency sysctl is enable
9pnet: p9_errstr2errno: server reported unknown error αΘ’Vε

Crashes (1):
Manager Time Kernel Commit Syzkaller Config Log Report Syz repro C repro VM info Title
ci2-linux-4-19 2020/06/03 00:40 linux-4.19.y 2d16cf4817bc f3ba1b5b .config log report