linux-kernel - Bug when changing cpus

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>] [day] [month] [year] [list]

Message-ID: <193b0f820910260911p218c0c1dn1986c48c7ecf6ee2@mail.gmail.com>
Date:	Mon, 26 Oct 2009 14:11:13 -0200
From:	Lucas De Marchi <lucas.de.marchi@...il.com>
To:	Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...e.hu>
Cc:	linux-kernel@...r.kernel.org
Subject: Bug when changing cpus_allowed of RT tasks?

I think there's a problem when balancing RT tasks: both find_lowest_rq_rt and
select_task_rq_rt check for p->rt.nr_cpus_allowed being greater than 1 to wake
a task in another cpu:

	static int select_task_rq_rt(struct task_struct *p, int sd_flag, int flags)
	{
		[...]
		if (unlikely(rt_task(rq->curr)) &&
		    (p->rt.nr_cpus_allowed > 1)) {
			int cpu = find_lowest_rq(p);

			return (cpu == -1) ? task_cpu(p) : cpu;
		}
		[...]
	}

	static int find_lowest_rq(struct task_struct *task)
	{
		struct sched_domain *sd;
		struct cpumask *lowest_mask = __get_cpu_var(local_cpu_mask);
		int this_cpu = smp_processor_id();
		int cpu      = task_cpu(task);
		cpumask_var_t domain_mask;

		if (task->rt.nr_cpus_allowed == 1)
			return -1; /* No other targets possible */
		[...]
	}


What happens if rt.nr_cpus_allowed continues to be 1, but the allowed cpu is
not the one in which task is being woken up? Since set_cpus_allowed_rt just
changed the mask and updated the value of nr_cpus_allowed_rt, as far as I can
see, it remains on same CPU even if the new mask does not allow.

A possible fix below, and in case it is right:
	Signed-off-by: Lucas De Marchi <lucas.de.marchi@...il.com>



Lucas De Marchi


--

diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c
index a4d790c..531c946 100644
--- a/kernel/sched_rt.c
+++ b/kernel/sched_rt.c
@@ -962,8 +962,7 @@ static int select_task_rq_rt(struct task_struct
*p, int sd_flag, int flags)
 	 * that is just being woken and probably will have
 	 * cold cache anyway.
 	 */
-	if (unlikely(rt_task(rq->curr)) &&
-	    (p->rt.nr_cpus_allowed > 1)) {
+	if (unlikely(rt_task(rq->curr))) {
 		int cpu = find_lowest_rq(p);

 		return (cpu == -1) ? task_cpu(p) : cpu;
@@ -1177,7 +1176,8 @@ static int find_lowest_rq(struct task_struct *task)
 	int cpu      = task_cpu(task);
 	cpumask_var_t domain_mask;

-	if (task->rt.nr_cpus_allowed == 1)
+	if ((task->rt.nr_cpus_allowed == 1) &&
+			likely(cpumask_test_cpu(cpu, &task->cpus_allowed)))
 		return -1; /* No other targets possible */

 	if (!cpupri_find(&task_rq(task)->rd->cpupri, task, lowest_mask))
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/