Re: [syzbot] [bluetooth?] INFO: task hung in hci_conn_failed

From: Hillf Danton <hdanton@sina.com>
To: Boqun Feng <boqun.feng@gmail.com>
Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
	syzbot <syzbot+a984066a63e9c1e62662@syzkaller.appspotmail.com>,
	eadavis@qq.com, Peter Zijlstra <peterz@infradead.org>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com
Subject: Re: [syzbot] [bluetooth?] INFO: task hung in hci_conn_failed
Date: Tue, 30 Jan 2024 19:30:37 +0800	[thread overview]
Message-ID: <20240130113037.1390-1-hdanton@sina.com> (raw)
In-Reply-To: <ZbgJqc06rEZ_FqRw@boqun-archlinux>

On Mon, 29 Jan 2024 12:25:13 -0800 Boqun Feng <boqun.feng@gmail.com>
> 
> Looks to me that debug_show_all_locks() doesn't hold the lockdep lock,
> so it's really a best effort race read of all tasks on lock hold
> information. Maybe Hillf wants the following? (Completely untested, it
> will stop your whole system and print lock holding information).

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

--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -6611,6 +6611,8 @@ void debug_show_all_locks(void)
 	}
 	pr_warn("\nShowing all locks held in the system:\n");
 
+	local_irq_disable();
+	lockdep_lock();
 	rcu_read_lock();
 	for_each_process_thread(g, p) {
 		if (!p->lockdep_depth)
@@ -6620,6 +6622,8 @@ void debug_show_all_locks(void)
 		touch_all_softlockup_watchdogs();
 	}
 	rcu_read_unlock();
+	lockdep_unlock();
+	local_irq_enable();
 
 	pr_warn("\n");
 	pr_warn("=============================================\n\n");
--
