linux-kernel - Re: [PATCH V3 04/10] sched/core: Prevent race condition between cpuset and __sched

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]

Date:   Thu, 15 Feb 2018 11:33:53 +0100
From:   Juri Lelli <juri.lelli@...hat.com>
To:     Mathieu Poirier <mathieu.poirier@...aro.org>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Li Zefan <lizefan@...wei.com>, Ingo Molnar <mingo@...hat.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Claudio Scordino <claudio@...dence.eu.com>,
        Daniel Bristot de Oliveira <bristot@...hat.com>,
        Tommaso Cucinotta <tommaso.cucinotta@...tannapisa.it>,
        "luca.abeni" <luca.abeni@...tannapisa.it>, cgroups@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH V3 04/10] sched/core: Prevent race condition between
 cpuset and __sched_setscheduler()

On 14/02/18 17:31, Juri Lelli wrote:

[...]

> Still grabbing it is a no-go, as do_sched_setscheduler calls
> sched_setscheduler from inside an RCU read-side critical section.

I was then actually thinking that trylocking might do.. not sure however
if failing with -EBUSY in the contended case is feasible (and about the
general uglyness of the solution :/).

--->8---
 include/linux/cpuset.h | 4 ++--
 kernel/cgroup/cpuset.c | 6 +++---
 kernel/sched/core.c    | 4 +++-
 3 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
index 4bbb3f5a3020..53c3f4e13cb0 100644
--- a/include/linux/cpuset.h
+++ b/include/linux/cpuset.h
@@ -55,7 +55,7 @@ extern void cpuset_init_smp(void);
 extern void cpuset_force_rebuild(void);
 extern void cpuset_update_active_cpus(void);
 extern void cpuset_wait_for_hotplug(void);
-extern void cpuset_lock(void);
+extern int cpuset_trylock(void);
 extern void cpuset_unlock(void);
 extern void cpuset_cpus_allowed(struct task_struct *p, struct cpumask *mask);
 extern void cpuset_cpus_allowed_fallback(struct task_struct *p);
@@ -178,7 +178,7 @@ static inline void cpuset_update_active_cpus(void)
 
 static inline void cpuset_wait_for_hotplug(void) { }
 
-static inline void cpuset_lock(void) { }
+static inline int cpuset_trylock(void) { return 1; }
 
 static inline void cpuset_unlock(void) { }
 
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 41f8391640e6..995aac5b032d 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -2410,11 +2410,11 @@ void __init cpuset_init_smp(void)
 }
 
 /**
- * cpuset_lock - Grab the cpuset_mutex from another subsysytem
+ * cpuset_trylock - Try to grab the cpuset_mutex atomically from another subsytem
  */
-void cpuset_lock(void)
+int cpuset_trylock(void)
 {
-	mutex_lock(&cpuset_mutex);
+	return mutex_trylock(&cpuset_mutex);
 }
 
 /**
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 0d8badcf1f0f..b491b406a835 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4180,7 +4180,8 @@ static int __sched_setscheduler(struct task_struct *p,
 	 * domains can be rebuilt or modified while operations like DL
 	 * admission checks are carried out.
 	 */
-	cpuset_lock();
+	if (!cpuset_trylock())
+		return -EBUSY;
 
 	/*
 	 * Make sure no PI-waiters arrive (or leave) while we are
@@ -4312,6 +4313,7 @@ static int __sched_setscheduler(struct task_struct *p,
 	/* Avoid rq from going away on us: */
 	preempt_disable();
 	task_rq_unlock(rq, p, &rf);
+	cpuset_unlock();
 
 	if (pi)
 		rt_mutex_adjust_pi(p);