Received-SPF: pass (google.com: domain of linux-kernel+bounces-21823-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
From: Zqiang <qiang.zhang1211@gmail.com>
To: paulmck@kernel.org,
	frederic@kernel.org,
	quic_neeraju@quicinc.com,
	joel@joelfernandes.org,
	qiang.zhang1211@gmail.com
Cc: rcu@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH] rcu/nocb: Fix WARN_ON_ONCE() in the rcu_nocb_bypass_lock()
Date: Wed, 10 Jan 2024 16:11:28 +0800
Message-Id: <20240110081128.18683-1-qiang.zhang1211@gmail.com>
Precedence: bulk

For the kernels built with CONFIG_RCU_NOCB_CPU_DEFAULT_ALL=y and
CONFIG_RCU_LAZY=y, here are the following scenarios that will trigger
WARN_ON_ONCE() in the rcu_nocb_bypass_lock() and rcu_nocb_wait_contended().

        CPU2                                               CPU11
kthread                                                   
rcu_nocb_cb_kthread                                       ksys_write
rcu_do_batch                                              vfs_write
rcu_torture_timer_cb                                      proc_sys_write
__kmem_cache_free                                         proc_sys_call_handler
kmemleak_free                                             drop_caches_sysctl_handler
delete_object_full                                        drop_slab
__delete_object                                           shrink_slab
put_object                                                lazy_rcu_shrink_scan
call_rcu                                                  rcu_nocb_flush_bypass
__call_rcu_commn                                            rcu_nocb_bypass_lock
                                                            raw_spin_trylock(&rdp->nocb_bypass_lock) fail
                                                            atomic_inc(&rdp->nocb_lock_contended);
rcu_nocb_wait_contended                                     WARN_ON_ONCE(smp_processor_id() != rdp->cpu);
 WARN_ON_ONCE(atomic_read(&rdp->nocb_lock_contended))                                          |
                            |_ _ _ _ _ _ _ _ _ _same rdp and rdp->cpu != 11_ _ _ _ _ _ _ _ _ __|

This commit therefore use the rcu_nocb_try_flush_bypass() instead of
rcu_nocb_flush_bypass() in lazy_rcu_shrink_scan(), if the nocb_bypass
queue is being flushed, the rcu_nocb_try_flush_bypass will return directly.

Signed-off-by: Zqiang <qiang.zhang1211@gmail.com>
--- 
 kernel/rcu/tree_nocb.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/rcu/tree_nocb.h b/kernel/rcu/tree_nocb.h
index d82f96a66600..9b618842c324 100644
--- a/kernel/rcu/tree_nocb.h
+++ b/kernel/rcu/tree_nocb.h
@@ -1381,7 +1381,7 @@ lazy_rcu_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 			rcu_nocb_unlock_irqrestore(rdp, flags);
 			continue;
 		}
-		WARN_ON_ONCE(!rcu_nocb_flush_bypass(rdp, NULL, jiffies, false));
+		rcu_nocb_try_flush_bypass(rdp, jiffies);
 		rcu_nocb_unlock_irqrestore(rdp, flags);
 		wake_nocb_gp(rdp, false);
 		sc->nr_to_scan -= _count;
-- 
2.17.1