[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20240807163506.434885-3-neeraj.upadhyay@kernel.org>
Date: Wed, 7 Aug 2024 22:05:06 +0530
From: neeraj.upadhyay@...nel.org
To: rcu@...r.kernel.org
Cc: linux-kernel@...r.kernel.org,
paulmck@...nel.org,
joel@...lfernandes.org,
frederic@...nel.org,
boqun.feng@...il.com,
urezki@...il.com,
rostedt@...dmis.org,
mathieu.desnoyers@...icios.com,
jiangshanlai@...il.com,
qiang.zhang1211@...il.com,
neeraj.iitr10@...il.com,
Neeraj Upadhyay <neeraj.upadhyay@...nel.org>
Subject: [PATCH 2/2] context_tracking: Invoke Tasks-RCU enter/exit for NMI context
From: Neeraj Upadhyay <neeraj.upadhyay@...nel.org>
rcu_task_enter() and rcu_task_exit() are not called on NMI
entry and exit. So, Tasks-RCU-Rude grace period wait is required to
ensure that NMI handlers have entered/exited into Tasks-RCU eqs.
For architectures which do not require Tasks-RCU-Rude (as the code
sections where RCU is not watching are marked as noinstr), when
those architectures switch to not using Tasks-RCU-Rude, NMI handlers
task exit to eqs will need to be handled correctly for Tasks-RCU holdout
tasks running on nohz_full CPUs. As it is safe to call these two
functions from NMI context, remove the in_nmi() check to ensure that
Tasks-RCU entry/exit is marked correctly for NMI handlers.
Reported-by: Frederic Weisbecker <frederic@...nel.org>
Suggested-by: Frederic Weisbecker <frederic@...nel.org>
Suggested-by: "Paul E. McKenney" <paulmck@...nel.org>
Signed-off-by: Neeraj Upadhyay <neeraj.upadhyay@...nel.org>
---
kernel/context_tracking.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c
index 152b485a62db..626dd7a173a5 100644
--- a/kernel/context_tracking.c
+++ b/kernel/context_tracking.c
@@ -239,8 +239,7 @@ void noinstr ct_nmi_exit(void)
ct_kernel_exit_state(CT_RCU_WATCHING);
// ... but is no longer watching here.
- if (!in_nmi())
- rcu_task_exit();
+ rcu_task_exit();
}
/**
@@ -273,8 +272,7 @@ void noinstr ct_nmi_enter(void)
*/
if (!rcu_is_watching_curr_cpu()) {
- if (!in_nmi())
- rcu_task_enter();
+ rcu_task_enter();
// RCU is not watching here ...
ct_kernel_enter_state(CT_RCU_WATCHING);
--
2.40.1
Powered by blists - more mailing lists