From 6dbce04d8417ae706596366e16841d77c454ba52 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 16 Nov 2020 13:10:12 +0100 Subject: rcu: Allow rcu_irq_enter_check_tick() from NMI MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Eugenio managed to tickle #PF from NMI context which resulted in hitting a WARN in RCU through irqentry_enter() -> __rcu_irq_enter_check_tick(). However, this situation is perfectly sane and does not warrant an WARN. The #PF will (necessarily) be atomic and not require messing with the tick state, so early return is correct. This commit therefore removes the WARN. Fixes: aaf2bc50df1f ("rcu: Abstract out rcu_irq_enter_check_tick() from rcu_nmi_enter()") Reported-by: "Eugenio Pérez" Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Andy Lutomirski Signed-off-by: Paul E. McKenney --- kernel/rcu/tree.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/rcu/tree.c') diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index 06895ef85d69..93e1808ac3fc 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -928,8 +928,8 @@ void __rcu_irq_enter_check_tick(void) { struct rcu_data *rdp = this_cpu_ptr(&rcu_data); - // Enabling the tick is unsafe in NMI handlers. - if (WARN_ON_ONCE(in_nmi())) + // If we're here from NMI there's nothing to do. + if (in_nmi()) return; RCU_LOCKDEP_WARN(rcu_dynticks_curr_cpu_in_eqs(), -- cgit v1.2.1