[syzbot] possible deadlock in pipe_lock (5)

[syzbot] possible deadlock in pipe_lock (5)

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit:    8cae8cd89f05 seq_file: disallow extremely large seq buffer..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=1083e8cc300000
kernel config:  https://syzkaller.appspot.com/x/.config?x=7273c75708b55890
dashboard link: https://syzkaller.appspot.com/bug?extid=579885d1a9a833336209
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=163905f2300000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=165bd0ea300000

The issue was bisected to:

commit 82a763e61e2b601309d696d4fa514c77d64ee1be
Author: Miklos Szeredi <mszeredi@redhat.com>
Date:   Mon Dec 14 14:26:14 2020 +0000

    ovl: simplify file splice

bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=13440c82300000
final oops:     https://syzkaller.appspot.com/x/report.txt?x=10c40c82300000
console output: https://syzkaller.appspot.com/x/log.txt?x=17440c82300000

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+579885d1a9a833336209@syzkaller.appspotmail.com
Fixes: 82a763e61e2b ("ovl: simplify file splice")

======================================================
WARNING: possible circular locking dependency detected
5.14.0-rc2-syzkaller #0 Not tainted
------------------------------------------------------
syz-executor005/9173 is trying to acquire lock:
ffff888027e98468 (&pipe->mutex/1){+.+.}-{3:3}, at: pipe_lock_nested fs/pipe.c:66 [inline]
ffff888027e98468 (&pipe->mutex/1){+.+.}-{3:3}, at: pipe_lock+0x5a/0x70 fs/pipe.c:74

but task is already holding lock:
ffff888147cfc460 (sb_writers#5){.+.+}-{0:0}, at: __do_splice+0x134/0x250 fs/splice.c:1144

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #2 (sb_writers#5){.+.+}-{0:0}:
       percpu_down_read include/linux/percpu-rwsem.h:51 [inline]
       __sb_start_write include/linux/fs.h:1763 [inline]
       sb_start_write include/linux/fs.h:1833 [inline]
       file_start_write include/linux/fs.h:3040 [inline]
       ovl_write_iter+0x1015/0x17f0 fs/overlayfs/file.c:357
       call_write_iter include/linux/fs.h:2114 [inline]
       do_iter_readv_writev+0x46f/0x740 fs/read_write.c:740
       do_iter_write+0x188/0x670 fs/read_write.c:866
       vfs_iter_write+0x70/0xa0 fs/read_write.c:907
       iter_file_splice_write+0x723/0xc70 fs/splice.c:689
       do_splice_from fs/splice.c:767 [inline]
       do_splice+0xb7e/0x1960 fs/splice.c:1079
       __do_splice+0x134/0x250 fs/splice.c:1144
       __do_sys_splice fs/splice.c:1350 [inline]
       __se_sys_splice fs/splice.c:1332 [inline]
       __x64_sys_splice+0x198/0x250 fs/splice.c:1332
       do_syscall_x64 arch/x86/entry/common.c:50 [inline]
       do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
       entry_SYSCALL_64_after_hwframe+0x44/0xae

-> #1 (&ovl_i_mutex_key[depth]){+.+.}-{3:3}:
       down_write+0x92/0x150 kernel/locking/rwsem.c:1406
       inode_lock include/linux/fs.h:774 [inline]
       ovl_write_iter+0x17d/0x17f0 fs/overlayfs/file.c:341
       call_write_iter include/linux/fs.h:2114 [inline]
       do_iter_readv_writev+0x46f/0x740 fs/read_write.c:740
       do_iter_write+0x188/0x670 fs/read_write.c:866
       vfs_iter_write+0x70/0xa0 fs/read_write.c:907
       iter_file_splice_write+0x723/0xc70 fs/splice.c:689
       do_splice_from fs/splice.c:767 [inline]
       do_splice+0xb7e/0x1960 fs/splice.c:1079
       __do_splice+0x134/0x250 fs/splice.c:1144
       __do_sys_splice fs/splice.c:1350 [inline]
       __se_sys_splice fs/splice.c:1332 [inline]
       __x64_sys_splice+0x198/0x250 fs/splice.c:1332
       do_syscall_x64 arch/x86/entry/common.c:50 [inline]
       do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
       entry_SYSCALL_64_after_hwframe+0x44/0xae

-> #0 (&pipe->mutex/1){+.+.}-{3:3}:
       check_prev_add kernel/locking/lockdep.c:3051 [inline]
       check_prevs_add kernel/locking/lockdep.c:3174 [inline]
       validate_chain kernel/locking/lockdep.c:3789 [inline]
       __lock_acquire+0x2a07/0x54a0 kernel/locking/lockdep.c:5015
       lock_acquire kernel/locking/lockdep.c:5625 [inline]
       lock_acquire+0x1ab/0x510 kernel/locking/lockdep.c:5590
       __mutex_lock_common kernel/locking/mutex.c:959 [inline]
       __mutex_lock+0x12a/0x10a0 kernel/locking/mutex.c:1104
       pipe_lock_nested fs/pipe.c:66 [inline]
       pipe_lock+0x5a/0x70 fs/pipe.c:74
       iter_file_splice_write+0x183/0xc70 fs/splice.c:635
       do_splice_from fs/splice.c:767 [inline]
       do_splice+0xb7e/0x1960 fs/splice.c:1079
       __do_splice+0x134/0x250 fs/splice.c:1144
       __do_sys_splice fs/splice.c:1350 [inline]
       __se_sys_splice fs/splice.c:1332 [inline]
       __x64_sys_splice+0x198/0x250 fs/splice.c:1332
       do_syscall_x64 arch/x86/entry/common.c:50 [inline]
       do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
       entry_SYSCALL_64_after_hwframe+0x44/0xae

other info that might help us debug this:

Chain exists of:
  &pipe->mutex/1 --> &ovl_i_mutex_key[depth] --> sb_writers#5

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(sb_writers#5);
                               lock(&ovl_i_mutex_key[depth]);
                               lock(sb_writers#5);
  lock(&pipe->mutex/1);

 *** DEADLOCK ***

1 lock held by syz-executor005/9173:
 #0: ffff888147cfc460 (sb_writers#5){.+.+}-{0:0}, at: __do_splice+0x134/0x250 fs/splice.c:1144

stack backtrace:
CPU: 0 PID: 9173 Comm: syz-executor005 Not tainted 5.14.0-rc2-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:105
 check_noncircular+0x25f/0x2e0 kernel/locking/lockdep.c:2131
 check_prev_add kernel/locking/lockdep.c:3051 [inline]
 check_prevs_add kernel/locking/lockdep.c:3174 [inline]
 validate_chain kernel/locking/lockdep.c:3789 [inline]
 __lock_acquire+0x2a07/0x54a0 kernel/locking/lockdep.c:5015
 lock_acquire kernel/locking/lockdep.c:5625 [inline]
 lock_acquire+0x1ab/0x510 kernel/locking/lockdep.c:5590
 __mutex_lock_common kernel/locking/mutex.c:959 [inline]
 __mutex_lock+0x12a/0x10a0 kernel/locking/mutex.c:1104
 pipe_lock_nested fs/pipe.c:66 [inline]
 pipe_lock+0x5a/0x70 fs/pipe.c:74
 iter_file_splice_write+0x183/0xc70 fs/splice.c:635
 do_splice_from fs/splice.c:767 [inline]
 do_splice+0xb7e/0x1960 fs/splice.c:1079
 __do_splice+0x134/0x250 fs/splice.c:1144
 __do_sys_splice fs/splice.c:1350 [inline]
 __se_sys_splice fs/splice.c:1332 [inline]
 __x64_sys_splice+0x198/0x250 fs/splice.c:1332
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x446219
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 41 15 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:00007f1d5746e2e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000113
RAX: ffffffffffffffda RBX: 00000000004cb4e0 RCX: 0000000000446219
RDX: 0000000000000006 RSI: 0000000000000000 RDI: 0000000000000004
RBP: 000000000049b06c R08: 000000000004ffdc R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 3d72696472657070
R13: 0079616c7265766f R14: 69662f7375622f2e R15: 00000000004cb4e8


---
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 syzkaller@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
For information about bisection process see: https://goo.gl/tpsmEJ#bisection
syzbot can test patches for this issue, for details see:
https://goo.gl/tpsmEJ#testing-patches

Re: [syzbot] possible deadlock in pipe_lock (5)

[Miklos Szeredi]

[-- Attachment #1: Type: text/plain, Size: 1185 bytes --]

On Sun, Jul 25, 2021 at 3:31 AM Hillf Danton <hdanton@sina.com> wrote:
>
> On Sat, 24 Jul 2021 12:07:20 -0700
> >syzbot found the following issue on:
> >
> >HEAD commit:    8cae8cd89f05 seq_file: disallow extremely large seq buffer..
> >git tree:       upstream
> >console output: https://syzkaller.appspot.com/x/log.txt?x=1083e8cc300000
> >kernel config:  https://syzkaller.appspot.com/x/.config?x=7273c75708b55890
> >dashboard link: https://syzkaller.appspot.com/bug?extid=579885d1a9a833336209
> >syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=163905f2300000
> >C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=165bd0ea300000
> >
> >The issue was bisected to:
> >
> >commit 82a763e61e2b601309d696d4fa514c77d64ee1be
> >Author: Miklos Szeredi <mszeredi@redhat.com>
> >Date:   Mon Dec 14 14:26:14 2020 +0000
> >
> >    ovl: simplify file splice
>
>
> If this commit is innocent then is it false positive lockdep warning again,
> given another report [1]?

Appears to be legit.

Attached partial revert + sync with ovl_write_iter() should fix it
(fingers crossed).

#syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master

Thanks,
Miklos

[-- Attachment #2: ovl-fix-deadlock-in-splice-write.patch --]
[-- Type: text/x-patch, Size: 1861 bytes --]

diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
index 4d53d3b7e5fe..d081faa55e83 100644
--- a/fs/overlayfs/file.c
+++ b/fs/overlayfs/file.c
@@ -392,6 +392,51 @@ static ssize_t ovl_write_iter(struct kiocb *iocb, struct iov_iter *iter)
 	return ret;
 }
 
+/*
+ * Calling iter_file_splice_write() directly from overlay's f_op may deadlock
+ * due to lock order inversion between pipe->mutex in iter_file_splice_write()
+ * and file_start_write(real.file) in ovl_write_iter().
+ *
+ * So do everything ovl_write_iter() does and call iter_file_splice_write() on
+ * the real file.
+ */
+static ssize_t ovl_splice_write(struct pipe_inode_info *pipe, struct file *out,
+				loff_t *ppos, size_t len, unsigned int flags)
+{
+	struct fd real;
+	const struct cred *old_cred;
+	struct inode *inode = file_inode(out);
+	struct inode *realinode = ovl_inode_real(inode);
+	ssize_t ret;
+
+	inode_lock(inode);
+	/* Update mode */
+	ovl_copyattr(realinode, inode);
+	ret = file_remove_privs(out);
+	if (ret)
+		goto out_unlock;
+
+	ret = ovl_real_fdget(out, &real);
+	if (ret)
+		goto out_unlock;
+
+	old_cred = ovl_override_creds(inode->i_sb);
+	file_start_write(real.file);
+
+	ret = iter_file_splice_write(pipe, real.file, ppos, len, flags);
+
+	file_end_write(real.file);
+	/* Update size */
+	ovl_copyattr(realinode, inode);
+	revert_creds(old_cred);
+	fdput(real);
+
+out_unlock:
+	inode_unlock(inode);
+
+	return ret;
+}
+
 static int ovl_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 {
 	struct fd real;
@@ -603,7 +648,7 @@ const struct file_operations ovl_file_operations = {
 	.fadvise	= ovl_fadvise,
 	.flush		= ovl_flush,
 	.splice_read    = generic_file_splice_read,
-	.splice_write   = iter_file_splice_write,
+	.splice_write   = ovl_splice_write,
 
 	.copy_file_range	= ovl_copy_file_range,
 	.remap_file_range	= ovl_remap_file_range,

Re: [syzbot] possible deadlock in pipe_lock (5)

[syzbot]

Hello,

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

Reported-and-tested-by: syzbot+579885d1a9a833336209@syzkaller.appspotmail.com

Tested on:

commit:         7d549995 Merge tag 'for-linus' of git://git.kernel.org..
git tree:       upstream
kernel config:  https://syzkaller.appspot.com/x/.config?x=40eef000d7648480
dashboard link: https://syzkaller.appspot.com/bug?extid=579885d1a9a833336209
compiler:       gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.1
patch:          https://syzkaller.appspot.com/x/patch.diff?x=104bb746300000

Note: testing is done by a robot and is best-effort only.