[syzbot] [hfs?] possible deadlock in hfs_extend_file (3)

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: fe46a7dd189e Merge tag 'sound-6.9-rc1' of git://git.kernel..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=1556f7cb180000
kernel config: https://syzkaller.appspot.com/x/.config?x=aef2a55903e5791c
dashboard link: https://syzkaller.appspot.com/bug?extid=2a62f58f1a4951a549bb
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

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

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/089e25869df5/disk-fe46a7dd.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/423b1787914f/vmlinux-fe46a7dd.xz
kernel image: https://storage.googleapis.com/syzbot-assets/4c043e30c07d/bzImage-fe46a7dd.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+2a62f5...@syzkaller.appspotmail.com

======================================================
WARNING: possible circular locking dependency detected
6.8.0-syzkaller-08951-gfe46a7dd189e #0 Not tainted
------------------------------------------------------
syz-executor.3/5818 is trying to acquire lock:
ffff888060dd1af8 (&HFS_I(tree->inode)->extents_lock){+.+.}-{3:3}, at: hfs_extend_file+0xa2/0xb10 fs/hfs/extent.c:397

but task is already holding lock:
ffff88802c4140b0 (&tree->tree_lock#2/1){+.+.}-{3:3}, at: hfs_find_init+0x183/0x220 fs/hfs/bfind.c:33

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (&tree->tree_lock#2/1){+.+.}-{3:3}:
__mutex_lock_common kernel/locking/mutex.c:608 [inline]
__mutex_lock+0x175/0x9c0 kernel/locking/mutex.c:752
hfs_find_init+0x183/0x220 fs/hfs/bfind.c:33
hfs_ext_read_extent+0x19c/0x9e0 fs/hfs/extent.c:200
hfs_extend_file+0x4e4/0xb10 fs/hfs/extent.c:401
hfs_bmap_reserve+0x29c/0x380 fs/hfs/btree.c:234
hfs_cat_create+0x22b/0x810 fs/hfs/catalog.c:104
hfs_create+0x6b/0xf0 fs/hfs/dir.c:202
lookup_open.isra.0+0x10a1/0x13c0 fs/namei.c:3497
open_last_lookups fs/namei.c:3566 [inline]
path_openat+0x92f/0x2990 fs/namei.c:3796
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+0xd2/0x260 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x6d/0x75

-> #0 (&HFS_I(tree->inode)->extents_lock){+.+.}-{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
__mutex_lock_common kernel/locking/mutex.c:608 [inline]
__mutex_lock+0x175/0x9c0 kernel/locking/mutex.c:752
hfs_extend_file+0xa2/0xb10 fs/hfs/extent.c:397
hfs_bmap_reserve+0x29c/0x380 fs/hfs/btree.c:234
__hfs_ext_write_extent+0x3cf/0x520 fs/hfs/extent.c:121
__hfs_ext_cache_extent fs/hfs/extent.c:174 [inline]
hfs_ext_read_extent+0x809/0x9e0 fs/hfs/extent.c:202
hfs_extend_file+0x4e4/0xb10 fs/hfs/extent.c:401
hfs_get_block+0x17f/0x830 fs/hfs/extent.c:353
__block_write_begin_int+0x4fb/0x16e0 fs/buffer.c:2105
__block_write_begin fs/buffer.c:2154 [inline]
block_write_begin+0xb1/0x4a0 fs/buffer.c:2213
cont_write_begin+0x53d/0x740 fs/buffer.c:2567
hfs_write_begin+0x87/0x150 fs/hfs/inode.c:53
generic_perform_write+0x272/0x620 mm/filemap.c:3930
__generic_file_write_iter+0x1fd/0x240 mm/filemap.c:4025
generic_file_write_iter+0xe7/0x350 mm/filemap.c:4051
call_write_iter include/linux/fs.h:2108 [inline]
new_sync_write fs/read_write.c:497 [inline]
vfs_write+0x6db/0x1100 fs/read_write.c:590
ksys_write+0x12f/0x260 fs/read_write.c:643
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xd2/0x260 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(&tree->tree_lock#2/1);
lock(&HFS_I(tree->inode)->extents_lock);
lock(&tree->tree_lock#2/1);
lock(&HFS_I(tree->inode)->extents_lock);

*** DEADLOCK ***

5 locks held by syz-executor.3/5818:
#0: ffff888011652348 (&f->f_pos_lock){+.+.}-{3:3}, at: __fdget_pos+0xeb/0x180 fs/file.c:1191
#1: ffff88802ce58420 (sb_writers#16){.+.+}-{0:0}, at: ksys_write+0x12f/0x260 fs/read_write.c:643
#2: ffff888060dd6aa8 (&sb->s_type->i_mutex_key#23){+.+.}-{3:3}, at: inode_lock include/linux/fs.h:793 [inline]
#2: ffff888060dd6aa8 (&sb->s_type->i_mutex_key#23){+.+.}-{3:3}, at: generic_file_write_iter+0x92/0x350 mm/filemap.c:4048
#3: ffff888060dd68f8 (&HFS_I(inode)->extents_lock#2){+.+.}-{3:3}, at: hfs_extend_file+0xa2/0xb10 fs/hfs/extent.c:397
#4: ffff88802c4140b0 (&tree->tree_lock#2/1){+.+.}-{3:3}, at: hfs_find_init+0x183/0x220 fs/hfs/bfind.c:33

stack backtrace:
CPU: 0 PID: 5818 Comm: syz-executor.3 Not tainted 6.8.0-syzkaller-08951-gfe46a7dd189e #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 03/27/2024
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:114
check_noncircular+0x31a/0x400 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+0x2478/0x3b30 kernel/locking/lockdep.c:5137
lock_acquire kernel/locking/lockdep.c:5754 [inline]
lock_acquire+0x1b1/0x540 kernel/locking/lockdep.c:5719
__mutex_lock_common kernel/locking/mutex.c:608 [inline]
__mutex_lock+0x175/0x9c0 kernel/locking/mutex.c:752
hfs_extend_file+0xa2/0xb10 fs/hfs/extent.c:397
hfs_bmap_reserve+0x29c/0x380 fs/hfs/btree.c:234
__hfs_ext_write_extent+0x3cf/0x520 fs/hfs/extent.c:121
__hfs_ext_cache_extent fs/hfs/extent.c:174 [inline]
hfs_ext_read_extent+0x809/0x9e0 fs/hfs/extent.c:202
hfs_extend_file+0x4e4/0xb10 fs/hfs/extent.c:401
hfs_get_block+0x17f/0x830 fs/hfs/extent.c:353
__block_write_begin_int+0x4fb/0x16e0 fs/buffer.c:2105
__block_write_begin fs/buffer.c:2154 [inline]
block_write_begin+0xb1/0x4a0 fs/buffer.c:2213
cont_write_begin+0x53d/0x740 fs/buffer.c:2567
hfs_write_begin+0x87/0x150 fs/hfs/inode.c:53
generic_perform_write+0x272/0x620 mm/filemap.c:3930
__generic_file_write_iter+0x1fd/0x240 mm/filemap.c:4025
generic_file_write_iter+0xe7/0x350 mm/filemap.c:4051
call_write_iter include/linux/fs.h:2108 [inline]
new_sync_write fs/read_write.c:497 [inline]
vfs_write+0x6db/0x1100 fs/read_write.c:590
ksys_write+0x12f/0x260 fs/read_write.c:643
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xd2/0x260 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x6d/0x75
RIP: 0033:0x7efc4247de69
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 e1 20 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 b0 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007efc41fff0c8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 00007efc425abf80 RCX: 00007efc4247de69
RDX: 000000000208e24b RSI: 00000000200004c0 RDI: 0000000000000005
RBP: 00007efc424ca47a R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 000000000000000b R14: 00007efc425abf80 R15: 00007ffea65dcd48
</TASK>


---
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

[Jeongjun Park]

======================================================
WARNING: possible circular locking dependency detected

6.9.0-rc4-syzkaller-00274-g3b68086599f8 #0 Not tainted
------------------------------------------------------
kworker/u8:6/1059 is trying to acquire lock:
ffff88805bd4a7f8 (&HFS_I(tree->inode)->extents_lock){+.+.}-{3:3}, at: hfs_extend_file+0xff/0x1450 fs/hfs/extent.c:397

but task is already holding lock:

ffff88806a48c0b0 (&tree->tree_lock#2/1){+.+.}-{3:3}, at: hfs_find_init+0x16e/0x1f0

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (&tree->tree_lock#2/1){+.+.}-{3:3}:

lock_acquire+0x1ed/0x550 kernel/locking/lockdep.c:5754
__mutex_lock_common kernel/locking/mutex.c:608 [inline]
__mutex_lock+0x136/0xd70 kernel/locking/mutex.c:752
hfs_find_init+0x16e/0x1f0
hfs_ext_read_extent fs/hfs/extent.c:200 [inline]
hfs_extend_file+0x31b/0x1450 fs/hfs/extent.c:401
hfs_bmap_reserve+0xd9/0x400 fs/hfs/btree.c:234
hfs_cat_create+0x1e0/0x970 fs/hfs/catalog.c:104
hfs_create+0x66/0xe0 fs/hfs/dir.c:202
lookup_open fs/namei.c:3497 [inline]
open_last_lookups fs/namei.c:3566 [inline]
path_openat+0x1425/0x3240 fs/namei.c:3796
do_filp_open+0x235/0x490 fs/namei.c:3826
do_sys_openat2+0x13e/0x1d0 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+0x247/0x2a0 fs/open.c:1432
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xf5/0x240 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f

-> #0 (&HFS_I(tree->inode)->extents_lock){+.+.}-{3:3}:
check_prev_add kernel/locking/lockdep.c:3134 [inline]
check_prevs_add kernel/locking/lockdep.c:3253 [inline]

validate_chain+0x18cb/0x58e0 kernel/locking/lockdep.c:3869
__lock_acquire+0x1346/0x1fd0 kernel/locking/lockdep.c:5137
lock_acquire+0x1ed/0x550 kernel/locking/lockdep.c:5754
__mutex_lock_common kernel/locking/mutex.c:608 [inline]
__mutex_lock+0x136/0xd70 kernel/locking/mutex.c:752
hfs_extend_file+0xff/0x1450 fs/hfs/extent.c:397
hfs_bmap_reserve+0xd9/0x400 fs/hfs/btree.c:234
__hfs_ext_write_extent+0x22e/0x4f0 fs/hfs/extent.c:121
hfs_ext_write_extent+0x154/0x1d0 fs/hfs/extent.c:144
hfs_write_inode+0xbc/0xec0 fs/hfs/inode.c:427
write_inode fs/fs-writeback.c:1498 [inline]
__writeback_single_inode+0x6b9/0x10b0 fs/fs-writeback.c:1715
writeback_sb_inodes+0x905/0x1260 fs/fs-writeback.c:1941
wb_writeback+0x457/0xce0 fs/fs-writeback.c:2117
wb_do_writeback fs/fs-writeback.c:2264 [inline]
wb_workfn+0x410/0x1090 fs/fs-writeback.c:2304
process_one_work kernel/workqueue.c:3254 [inline]
process_scheduled_works+0xa10/0x17c0 kernel/workqueue.c:3335
worker_thread+0x86d/0xd70 kernel/workqueue.c:3416
kthread+0x2f0/0x390 kernel/kthread.c:388
ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244

other info that might help us debug this:

Possible unsafe locking scenario:

CPU0 CPU1
---- ----
lock(&tree->tree_lock#2/1);
lock(&HFS_I(tree->inode)->extents_lock);
lock(&tree->tree_lock#2/1);
lock(&HFS_I(tree->inode)->extents_lock);

*** DEADLOCK ***

3 locks held by kworker/u8:6/1059:
#0: ffff88801be87148 ((wq_completion)writeback){+.+.}-{0:0}, at: process_one_work kernel/workqueue.c:3229 [inline]
#0: ffff88801be87148 ((wq_completion)writeback){+.+.}-{0:0}, at: process_scheduled_works+0x8e0/0x17c0 kernel/workqueue.c:3335
#1: ffffc90004197d00 ((work_completion)(&(&wb->dwork)->work)){+.+.}-{0:0}, at: process_one_work kernel/workqueue.c:3230 [inline]
#1: ffffc90004197d00 ((work_completion)(&(&wb->dwork)->work)){+.+.}-{0:0}, at: process_scheduled_works+0x91b/0x17c0 kernel/workqueue.c:3335
#2: ffff88806a48c0b0 (&tree->tree_lock#2/1){+.+.}-{3:3}, at: hfs_find_init+0x16e/0x1f0

======================================================

When a file expansion operation occurs in the hfs file system,
unnecessary locking occurs due to recursion. This situation does
not appear to be easy to reproduce, but it is very strange logic
and must be fixed.

Whether this recursion is intended behavior or not, I think it is a
good idea to prevent deadlock by placing mutex_lock() in a higher
function than hfs_extend_file().

Reported-by: syzbot+2a62f5...@syzkaller.appspotmail.com
Fixes: 39f8d472f280 ("hfs: convert extents_lock in a mutex")
Fixes: 1267a07be5eb ("hfs: fix return value of hfs_get_block()")
Fixes: 54640c7502e5 ("hfs: prevent btree data loss on ENOSPC")
Signed-off-by: Jeongjun Park <aha3...@gmail.com>
---
fs/hfs/catalog.c | 4 ++++
fs/hfs/extent.c | 26 +++++++++++++++-----------
2 files changed, 19 insertions(+), 11 deletions(-)

diff --git a/fs/hfs/catalog.c b/fs/hfs/catalog.c
index d63880e7d9d6..cece4333f7a7 100644
--- a/fs/hfs/catalog.c
+++ b/fs/hfs/catalog.c
@@ -101,7 +101,9 @@ int hfs_cat_create(u32 cnid, struct inode *dir, const struct qstr *str, struct i
* Fail early and avoid ENOSPC during the btree operations. We may
* have to split the root node at most once.
*/
+ mutex_lock(&HFS_I(fd.tree->inode)->extents_lock);
err = hfs_bmap_reserve(fd.tree, 2 * fd.tree->depth);
+ mutex_unlock(&HFS_I(fd.tree->inode)->extents_lock);
if (err)
goto err2;

@@ -307,7 +309,9 @@ int hfs_cat_move(u32 cnid, struct inode *src_dir, const struct qstr *src_name,
* Fail early and avoid ENOSPC during the btree operations. We may
* have to split the root node at most once.
*/
+ mutex_lock(&HFS_I(src_fd.tree->inode)->extents_lock);
err = hfs_bmap_reserve(src_fd.tree, 2 * src_fd.tree->depth);
+ mutex_unlock(&HFS_I(src_fd.tree->inode)->extents_lock);
if (err)
goto out;

diff --git a/fs/hfs/extent.c b/fs/hfs/extent.c
index 6d1878b99b30..25de1d48b667 100644
--- a/fs/hfs/extent.c
+++ b/fs/hfs/extent.c
@@ -338,38 +338,41 @@ int hfs_get_block(struct inode *inode, sector_t block,
{
struct super_block *sb;
u16 dblock, ablock;
- int res;
+ int res = 0;

sb = inode->i_sb;
/* Convert inode block to disk allocation block */
ablock = (u32)block / HFS_SB(sb)->fs_div;

+ mutex_lock(&HFS_I(inode)->extents_lock);
if (block >= HFS_I(inode)->fs_blocks) {
if (!create)
- return 0;
- if (block > HFS_I(inode)->fs_blocks)
- return -EIO;
+ goto out;
+ if (block > HFS_I(inode)->fs_blocks){
+ res = -EIO;
+ goto out;
+ }
if (ablock >= HFS_I(inode)->alloc_blocks) {
res = hfs_extend_file(inode);
if (res)
- return res;
+ goto out;
}
} else
create = 0;

if (ablock < HFS_I(inode)->first_blocks) {
dblock = hfs_ext_find_block(HFS_I(inode)->first_extents, ablock);
+ mutex_unlock(&HFS_I(inode)->extents_lock);
goto done;
}

- mutex_lock(&HFS_I(inode)->extents_lock);
res = hfs_ext_read_extent(inode, ablock);
if (!res)
dblock = hfs_ext_find_block(HFS_I(inode)->cached_extents,
ablock - HFS_I(inode)->cached_start);
else {
- mutex_unlock(&HFS_I(inode)->extents_lock);
- return -EIO;
+ res = -EIO;
+ goto out;
}
mutex_unlock(&HFS_I(inode)->extents_lock);

@@ -385,7 +388,10 @@ int hfs_get_block(struct inode *inode, sector_t block,
inode_add_bytes(inode, sb->s_blocksize);
mark_inode_dirty(inode);
}
- return 0;
+ return res;
+out:
+ mutex_unlock(&HFS_I(inode)->extents_lock);
+ return res;
}

int hfs_extend_file(struct inode *inode)
@@ -394,7 +400,6 @@ int hfs_extend_file(struct inode *inode)
u32 start, len, goal;
int res;

- mutex_lock(&HFS_I(inode)->extents_lock);
if (HFS_I(inode)->alloc_blocks == HFS_I(inode)->first_blocks)
goal = hfs_ext_lastblock(HFS_I(inode)->first_extents);
else {
@@ -444,7 +449,6 @@ int hfs_extend_file(struct inode *inode)
goto insert_extent;
}
out:
- mutex_unlock(&HFS_I(inode)->extents_lock);
if (!res) {
HFS_I(inode)->alloc_blocks += len;
mark_inode_dirty(inode);
--
2.34.1