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: <20190821231906.4224-3-swood@redhat.com>
Date:   Wed, 21 Aug 2019 18:19:05 -0500
From:   Scott Wood <swood@...hat.com>
To:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc:     linux-rt-users@...r.kernel.org, linux-kernel@...r.kernel.org,
        "Paul E . McKenney" <paulmck@...ux.ibm.com>,
        Joel Fernandes <joel@...lfernandes.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Peter Zijlstra <peterz@...radead.org>,
        Juri Lelli <juri.lelli@...hat.com>,
        Clark Williams <williams@...hat.com>,
        Scott Wood <swood@...hat.com>
Subject: [PATCH RT v2 2/3] sched: migrate_enable: Use sleeping_lock to indicate involuntary sleep

Without this, rcu_note_context_switch() will complain if an RCU read
lock is held when migrate_enable() calls stop_one_cpu().

Signed-off-by: Scott Wood <swood@...hat.com>
---
v2: Added comment.

If my migrate disable changes aren't taken, then pin_current_cpu()
will also need to use sleeping_lock_inc() because calling
__read_rt_lock() bypasses the usual place it's done.

 include/linux/sched.h    | 4 ++--
 kernel/rcu/tree_plugin.h | 2 +-
 kernel/sched/core.c      | 8 ++++++++
 3 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 7e892e727f12..1ebc97f28009 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -673,7 +673,7 @@ struct task_struct {
 	int				migrate_disable_atomic;
 # endif
 #endif
-#ifdef CONFIG_PREEMPT_RT_FULL
+#ifdef CONFIG_PREEMPT_RT_BASE
 	int				sleeping_lock;
 #endif
 
@@ -1881,7 +1881,7 @@ static __always_inline bool need_resched(void)
 	return unlikely(tif_need_resched());
 }
 
-#ifdef CONFIG_PREEMPT_RT_FULL
+#ifdef CONFIG_PREEMPT_RT_BASE
 static inline void sleeping_lock_inc(void)
 {
 	current->sleeping_lock++;
diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
index 23a54e4b649c..7a3aa085ce2c 100644
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -292,7 +292,7 @@ void rcu_note_context_switch(bool preempt)
 	barrier(); /* Avoid RCU read-side critical sections leaking down. */
 	trace_rcu_utilization(TPS("Start context switch"));
 	lockdep_assert_irqs_disabled();
-#if defined(CONFIG_PREEMPT_RT_FULL)
+#if defined(CONFIG_PREEMPT_RT_BASE)
 	sleeping_l = t->sleeping_lock;
 #endif
 	WARN_ON_ONCE(!preempt && t->rcu_read_lock_nesting > 0 && !sleeping_l);
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index e1bdd7f9be05..0758ee85634e 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7405,7 +7405,15 @@ void migrate_enable(void)
 			unpin_current_cpu();
 			preempt_lazy_enable();
 			preempt_enable();
+
+			/*
+			 * sleeping_lock_inc suppresses a debug check for
+			 * sleeping inside an RCU read side critical section
+			 */
+			sleeping_lock_inc();
 			stop_one_cpu(task_cpu(p), migration_cpu_stop, &arg);
+			sleeping_lock_dec();
+
 			return;
 		}
 	}
-- 
1.8.3.1

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ