linux-kernel - [PATCH (repost)] sched/core: defer printk() while rq lock is held

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <94c1499a-2325-4bbf-b7bc-04a1e9090488@I-love.SAKURA.ne.jp>
Date: Fri, 19 Jul 2024 22:17:01 +0900
From: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
To: syzbot <syzbot+18cfb7f63482af8641df@...kaller.appspotmail.com>,
        syzkaller-bugs@...glegroups.com, Ingo Molnar <mingo@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Juri Lelli <juri.lelli@...hat.com>,
        Vincent Guittot <vincent.guittot@...aro.org>
Cc: tglx@...utronix.de, linux-kernel@...r.kernel.org,
        linux-trace-kernel@...r.kernel.org, mhiramat@...nel.org,
        netdev@...r.kernel.org, andrii@...nel.org, ast@...nel.org,
        bpf@...r.kernel.org, daniel@...earbox.net,
        Steven Rostedt <rostedt@...dmis.org>
Subject: [PATCH (repost)] sched/core: defer printk() while rq lock is held

syzbot is reporting circular locking dependency inside __bpf_prog_run()
when trace_sched_switch() hook is called from __schedule(), for fault
injection calls printk() despite rq lock is already held.

Since any debugging functionality such as lockdep, fault injection,
KASAN/KCSAN/KMSAN etc. might call printk(), guard the whole section
between raw_spin_rq_{lock,lock_nested,trylock}() and raw_spin_rq_unlock()
using printk_deferred_{enter,exit}().

Reported-by: syzbot <syzbot+18cfb7f63482af8641df@...kaller.appspotmail.com>
Closes: https://syzkaller.appspot.com/bug?extid=18cfb7f63482af8641df
Signed-off-by: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
---
This is a repost of https://lkml.kernel.org/r/b55e5f24-01ad-4a3d-94dc-e8a6bc15ac42@I-love.SAKURA.ne.jp .
Scheduler developers, can you agree with addressing this problem at locations where rq lock is held?

 kernel/sched/core.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index bcf2c4cc0522..134f5196b9c4 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -559,6 +559,7 @@ void raw_spin_rq_lock_nested(struct rq *rq, int subclass)
 		raw_spin_lock_nested(&rq->__lock, subclass);
 		/* preempt_count *MUST* be > 1 */
 		preempt_enable_no_resched();
+		printk_deferred_enter();
 		return;
 	}
 
@@ -568,6 +569,7 @@ void raw_spin_rq_lock_nested(struct rq *rq, int subclass)
 		if (likely(lock == __rq_lockp(rq))) {
 			/* preempt_count *MUST* be > 1 */
 			preempt_enable_no_resched();
+			printk_deferred_enter();
 			return;
 		}
 		raw_spin_unlock(lock);
@@ -584,6 +586,8 @@ bool raw_spin_rq_trylock(struct rq *rq)
 	if (sched_core_disabled()) {
 		ret = raw_spin_trylock(&rq->__lock);
 		preempt_enable();
+		if (ret)
+			printk_deferred_enter();
 		return ret;
 	}
 
@@ -592,6 +596,8 @@ bool raw_spin_rq_trylock(struct rq *rq)
 		ret = raw_spin_trylock(lock);
 		if (!ret || (likely(lock == __rq_lockp(rq)))) {
 			preempt_enable();
+			if (ret)
+				printk_deferred_enter();
 			return ret;
 		}
 		raw_spin_unlock(lock);
@@ -600,6 +606,7 @@ bool raw_spin_rq_trylock(struct rq *rq)
 
 void raw_spin_rq_unlock(struct rq *rq)
 {
+	printk_deferred_exit();
 	raw_spin_unlock(rq_lockp(rq));
 }
 
-- 
2.43.0