linux-kernel - [PATCH RFC 3/3] rcu: Remove redundant check for irq state during unlock

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

Message-ID: <20250705203918.4149863-3-joelagnelf@nvidia.com>
Date: Sat,  5 Jul 2025 16:39:17 -0400
From: Joel Fernandes <joelagnelf@...dia.com>
To: linux-kernel@...r.kernel.org,
	"Paul E. McKenney" <paulmck@...nel.org>,
	Frederic Weisbecker <frederic@...nel.org>,
	Neeraj Upadhyay <neeraj.upadhyay@...nel.org>,
	Joel Fernandes <joelagnelf@...dia.com>,
	Josh Triplett <josh@...htriplett.org>,
	Boqun Feng <boqun.feng@...il.com>,
	Uladzislau Rezki <urezki@...il.com>,
	Steven Rostedt <rostedt@...dmis.org>,
	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
	Lai Jiangshan <jiangshanlai@...il.com>,
	Zqiang <qiang.zhang@...ux.dev>,
	Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
	Clark Williams <clrkwllms@...nel.org>
Cc: rcu@...r.kernel.org,
	linux-rt-devel@...ts.linux.dev
Subject: [PATCH RFC 3/3] rcu: Remove redundant check for irq state during unlock

The check for irqs_were_disabled is redundant in
rcu_unlock_needs_exp_handling() as the caller already checks for this.
This includes the boost case as well. Just remove the redundant check.

This is a first win for the refactor of the needs_exp (formerly
expboost) condition into a new rcu_unlock_needs_exp_handling() function,
as the conditions became more easier to read.

Signed-off-by: Joel Fernandes <joelagnelf@...dia.com>
---
 kernel/rcu/tree_plugin.h | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
index 8504d95bb35b..112973ecebb8 100644
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -659,14 +659,12 @@ static void rcu_preempt_deferred_qs_handler(struct irq_work *iwp)
  * @t: The task being checked
  * @rdp: The per-CPU RCU data
  * @rnp: The RCU node for this CPU
- * @irqs_were_disabled: Whether interrupts were disabled before rcu_read_unlock()
  *
  * Returns true if expedited processing of the rcu_read_unlock() is needed.
  */
 static bool rcu_unlock_needs_exp_handling(struct task_struct *t,
 				      struct rcu_data *rdp,
-				      struct rcu_node *rnp,
-				      bool irqs_were_disabled)
+				      struct rcu_node *rnp)
 {
 	/*
 	 * Check if this task is blocking an expedited grace period.
@@ -692,7 +690,7 @@ static bool rcu_unlock_needs_exp_handling(struct task_struct *t,
 	 * disturbing the system more. Check if either:
 	 * - This CPU has not yet reported a quiescent state, or
 	 * - This task was preempted within an RCU critical section
-	 * In either case, requird expedited handling for strict GP mode.
+	 * In either case, require expedited handling for strict GP mode.
 	 */
 	if (IS_ENABLED(CONFIG_RCU_STRICT_GRACE_PERIOD) &&
 	    ((rdp->grpmask & READ_ONCE(rnp->qsmask)) || t->rcu_blocked_node))
@@ -700,14 +698,14 @@ static bool rcu_unlock_needs_exp_handling(struct task_struct *t,
 
 	/*
 	 * RCU priority boosting case: If a task is subject to RCU priority
-	 * boosting and exits an RCU read-side critical section with interrupts
-	 * disabled, we need expedited handling to ensure timely deboosting.
-	 * Without this, a low-priority task could incorrectly run at high
-	 * real-time priority for an extended period effecting real-time
-	 * responsiveness. This applies to all CONFIG_RCU_BOOST=y kernels,
-	 * not just PREEMPT_RT.
+	 * boosting and exits an RCU read-side critical section, we need
+	 * expedited handling to ensure timely deboosting. Without this,
+	 * a low-priority task could incorrectly run at high real-time
+	 * priority for an extended period effecting real-time
+	 * responsiveness. This applies to all RCU_BOOST=y kernels,
+	 * not just to PREEMPT_RT.
 	 */
-	if (IS_ENABLED(CONFIG_RCU_BOOST) && irqs_were_disabled && t->rcu_blocked_node)
+	if (IS_ENABLED(CONFIG_RCU_BOOST) && t->rcu_blocked_node)
 		return true;
 
 	return false;
@@ -736,7 +734,7 @@ static void rcu_read_unlock_special(struct task_struct *t)
 		struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
 		struct rcu_node *rnp = rdp->mynode;
 
-		needs_exp = rcu_unlock_needs_exp_handling(t, rdp, rnp, irqs_were_disabled);
+		needs_exp = rcu_unlock_needs_exp_handling(t, rdp, rnp);
 	
 		// Need to defer quiescent state until everything is enabled.
 		if (use_softirq && (in_hardirq() || (needs_exp && !irqs_were_disabled))) {
-- 
2.43.0