Re: [syzbot] [bcachefs?] possible deadlock in bch2_symlink

Re: [syzbot] [bcachefs?] possible deadlock in bch2_symlink

[Alan Huang]

+cc overlayfs

> On Feb 8, 2025, at 17:14, syzbot <syzbot+7836a68852a10ec3d790@syzkaller.appspotmail.com> wrote:
> 
> syzbot has found a reproducer for the following issue on:
> 
> HEAD commit:    7ee983c850b4 Merge tag 'drm-fixes-2025-02-08' of https://g..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=17375ca4580000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=1909f2f0d8e641ce
> dashboard link: https://syzkaller.appspot.com/bug?extid=7836a68852a10ec3d790
> compiler:       Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=14e0cca4580000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=155361b0580000
> 
> Downloadable assets:
> disk image (non-bootable): https://storage.googleapis.com/syzbot-assets/7feb34a89c2a/non_bootable_disk-7ee983c8.raw.xz
> vmlinux: https://storage.googleapis.com/syzbot-assets/f2f78699fc41/vmlinux-7ee983c8.xz
> kernel image: https://storage.googleapis.com/syzbot-assets/ca55e6e8dd01/bzImage-7ee983c8.xz
> mounted in repro: https://storage.googleapis.com/syzbot-assets/aa79c539b21d/mount_0.gz
> 
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+7836a68852a10ec3d790@syzkaller.appspotmail.com
> 
> bcachefs (loop0): reading snapshots table
> bcachefs (loop0): reading snapshots done
> bcachefs (loop0): done starting filesystem
> ============================================
> WARNING: possible recursive locking detected
> 6.14.0-rc1-syzkaller-00181-g7ee983c850b4 #0 Not tainted
> --------------------------------------------
> syz-executor294/5305 is trying to acquire lock:
> ffff888044775078 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: inode_lock include/linux/fs.h:877 [inline]
> ffff888044775078 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: bch2_symlink+0x176/0x310 fs/bcachefs/fs.c:839
> 
> but task is already holding lock:
> ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: inode_lock include/linux/fs.h:877 [inline]
> ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_copy_up_workdir fs/overlayfs/copy_up.c:782 [inline]
> ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_do_copy_up fs/overlayfs/copy_up.c:1001 [inline]
> ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_copy_up_one fs/overlayfs/copy_up.c:1202 [inline]
> ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_copy_up_flags+0x19cb/0x47c0 fs/overlayfs/copy_up.c:1257
> 
> other info that might help us debug this:
> Possible unsafe locking scenario:
> 
>       CPU0
>       ----
>  lock(&sb->s_type->i_mutex_key#16);
>  lock(&sb->s_type->i_mutex_key#16);
> 
> *** DEADLOCK ***
> 
> May be due to missing lock nesting notation
> 
> 7 locks held by syz-executor294/5305:
> #0: ffff88801df94420 (sb_writers#10){.+.+}-{0:0}, at: mnt_want_write+0x3f/0x90 fs/namespace.c:547
> #1: ffff88804470de50 (&ovl_i_mutex_dir_key[depth]/1){+.+.}-{4:4}, at: inode_lock_nested include/linux/fs.h:912 [inline]
> #1: ffff88804470de50 (&ovl_i_mutex_dir_key[depth]/1){+.+.}-{4:4}, at: lock_rename fs/namei.c:3217 [inline]
> #1: ffff88804470de50 (&ovl_i_mutex_dir_key[depth]/1){+.+.}-{4:4}, at: do_renameat2+0x62c/0x13f0 fs/namei.c:5161
> #2: ffff88804470e418 (&ovl_i_mutex_key[depth]){+.+.}-{4:4}, at: inode_lock include/linux/fs.h:877 [inline]
> #2: ffff88804470e418 (&ovl_i_mutex_key[depth]){+.+.}-{4:4}, at: lock_two_nondirectories+0xe1/0x170 fs/inode.c:1281
> #3: ffff88804470e9e0 (&ovl_i_mutex_key[depth]/4){+.+.}-{4:4}, at: vfs_rename+0x6a2/0xf00 fs/namei.c:5040
> #4: ffff88804470e7d0 (&ovl_i_lock_key[depth]){+.+.}-{4:4}, at: ovl_inode_lock_interruptible fs/overlayfs/overlayfs.h:650 [inline]
> #4: ffff88804470e7d0 (&ovl_i_lock_key[depth]){+.+.}-{4:4}, at: ovl_copy_up_start+0x53/0x310 fs/overlayfs/util.c:727
> #5: ffff8880477b0420 (sb_writers#9){.+.+}-{0:0}, at: ovl_copy_up_workdir fs/overlayfs/copy_up.c:781 [inline]
> #5: ffff8880477b0420 (sb_writers#9){.+.+}-{0:0}, at: ovl_do_copy_up fs/overlayfs/copy_up.c:1001 [inline]
> #5: ffff8880477b0420 (sb_writers#9){.+.+}-{0:0}, at: ovl_copy_up_one fs/overlayfs/copy_up.c:1202 [inline]
> #5: ffff8880477b0420 (sb_writers#9){.+.+}-{0:0}, at: ovl_copy_up_flags+0x19b4/0x47c0 fs/overlayfs/copy_up.c:1257
> #6: ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: inode_lock include/linux/fs.h:877 [inline]
> #6: ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_copy_up_workdir fs/overlayfs/copy_up.c:782 [inline]
> #6: ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_do_copy_up fs/overlayfs/copy_up.c:1001 [inline]
> #6: ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_copy_up_one fs/overlayfs/copy_up.c:1202 [inline]
> #6: ffff888044774148 (&sb->s_type->i_mutex_key#16){++++}-{4:4}, at: ovl_copy_up_flags+0x19cb/0x47c0 fs/overlayfs/copy_up.c:1257
> 
> stack backtrace:
> CPU: 0 UID: 0 PID: 5305 Comm: syz-executor294 Not tainted 6.14.0-rc1-syzkaller-00181-g7ee983c850b4 #0
> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
> Call Trace:
> <TASK>
> __dump_stack lib/dump_stack.c:94 [inline]
> dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120
> print_deadlock_bug+0x483/0x620 kernel/locking/lockdep.c:3039
> check_deadlock kernel/locking/lockdep.c:3091 [inline]
> validate_chain+0x15e2/0x5920 kernel/locking/lockdep.c:3893
> __lock_acquire+0x1397/0x2100 kernel/locking/lockdep.c:5228
> lock_acquire+0x1ed/0x550 kernel/locking/lockdep.c:5851
> down_write+0x99/0x220 kernel/locking/rwsem.c:1577
> inode_lock include/linux/fs.h:877 [inline]
> bch2_symlink+0x176/0x310 fs/bcachefs/fs.c:839
> vfs_symlink+0x137/0x2e0 fs/namei.c:4671
> ovl_do_symlink+0x85/0xd0 fs/overlayfs/overlayfs.h:267
> ovl_create_real+0x346/0x550 fs/overlayfs/dir.c:206
> ovl_copy_up_workdir fs/overlayfs/copy_up.c:783 [inline]
> ovl_do_copy_up fs/overlayfs/copy_up.c:1001 [inline]
> ovl_copy_up_one fs/overlayfs/copy_up.c:1202 [inline]
> ovl_copy_up_flags+0x19fe/0x47c0 fs/overlayfs/copy_up.c:1257
> ovl_rename+0x62a/0x1760 fs/overlayfs/dir.c:1150
> vfs_rename+0xbdb/0xf00 fs/namei.c:5069
> do_renameat2+0xd94/0x13f0 fs/namei.c:5226
> __do_sys_rename fs/namei.c:5273 [inline]
> __se_sys_rename fs/namei.c:5271 [inline]
> __x64_sys_rename+0x82/0x90 fs/namei.c:5271
> do_syscall_x64 arch/x86/entry/common.c:52 [inline]
> do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83
> entry_SYSCALL_64_after_hwframe+0x77/0x7f
> RIP: 0033:0x7f0d8dddad19
> Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 f1 17 00 00 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 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007ffd9dbea2d8 EFLAGS: 00000246 ORIG_RAX: 0000000000000052
> RAX: ffffffffffffffda RBX: 0031656c69662f2e RCX: 00007f0d8dddad19
> RDX: 00007f0d8dddad19 RSI: 0000400000000840 RDI: 0000400000000800
> RBP: 0031656c69662f2e R08: 000055558454d4c0 R09: 000055558454d4c0
> R10: 000055558454d4c0 R11: 0000000000000246 R12: 00007ffd9dbea300
> R13: 00007ffd9dbea528 R14: 431bde82d7b634db R15: 00007f0d8de2303b
> </TASK>
> syz-executor294 (5305) used greatest stack depth: 10768 bytes left
> 
> 
> ---
> If you want syzbot to run the reproducer, reply with:
> #syz test: git://repo/address.git branch-or-commit-hash
> If you attach or paste a git patch, syzbot will apply it before testing.
> 

Re: [syzbot] [bcachefs?] possible deadlock in bch2_symlink

[Al Viro]

On Mon, Aug 04, 2025 at 11:02:54AM +0800, Alan Huang wrote:
> +cc overlayfs

Sigh...

1) ovl_copy_up_workdir() should lock wdir with I_MUTEX_PARENT, same
as filename_create().

2) what the hell is bch2_symlink() need to lock the new inode for?
page_symlink() doesn't need that; are there any bcachefs-specific reasons
to bother with that?

Re: [syzbot] [bcachefs?] possible deadlock in bch2_symlink

[Kent Overstreet]

On Mon, Aug 04, 2025 at 04:23:12AM +0100, Al Viro wrote:
> On Mon, Aug 04, 2025 at 11:02:54AM +0800, Alan Huang wrote:
> > +cc overlayfs
> 
> Sigh...
> 
> 2) what the hell is bch2_symlink() need to lock the new inode for?
> page_symlink() doesn't need that; are there any bcachefs-specific reasons
> to bother with that?

No, there isn't. I'll look it over and kill that tomorrow...

Re: [syzbot] [bcachefs?] possible deadlock in bch2_symlink

[Amir Goldstein]

On Mon, Aug 4, 2025 at 5:23 AM Al Viro <viro@zeniv.linux.org.uk> wrote:
>
> On Mon, Aug 04, 2025 at 11:02:54AM +0800, Alan Huang wrote:
> > +cc overlayfs
>
> Sigh...
>
> 1) ovl_copy_up_workdir() should lock wdir with I_MUTEX_PARENT, same
> as filename_create().

#syz test: https://github.com/amir73il/linux ovl-fixes

Re: [syzbot] [bcachefs?] possible deadlock in bch2_symlink

[syzbot]

Hello,

syzbot has tested the proposed patch and the reproducer did not trigger any issue:

Reported-by: syzbot+7836a68852a10ec3d790@syzkaller.appspotmail.com
Tested-by: syzbot+7836a68852a10ec3d790@syzkaller.appspotmail.com

Tested on:

commit:         9e7e7e36 ovl: use I_MUTEX_PARENT when locking parent i..
git tree:       https://github.com/amir73il/linux ovl-fixes
console output: https://syzkaller.appspot.com/x/log.txt?x=11d09ba2580000
kernel config:  https://syzkaller.appspot.com/x/.config?x=41498776bd40da1
dashboard link: https://syzkaller.appspot.com/bug?extid=7836a68852a10ec3d790
compiler:       Debian clang version 20.1.7 (++20250616065708+6146a88f6049-1~exp1~20250616065826.132), Debian LLD 20.1.7

Note: no patches were applied.
Note: testing is done by a robot and is best-effort only.