Re: INFO: suspicious RCU usage in rcu_torture_writer()

From: Fengguang Wu
Date: Mon Aug 27 2012 - 00:40:57 EST


On Sat, Aug 25, 2012 at 05:01:49PM -0700, Paul E. McKenney wrote:
> On Sat, Aug 25, 2012 at 11:36:23AM +0800, Fengguang Wu wrote:
> > Greetings,
> >
> > I got this warning on 3.6.0-rc2. Full dmesg/config attached.
> >
> > [ 3.051375] Initializing RT-Tester: OK
> > [ 3.052491] rcu-torture:--- Start of test: nreaders=2 nfakewriters=4 stat_interval=0 verbose=0 test_no_idle_hz=0 shuffle_interval=3 stut
> > ter=5 irqreader=1 fqs_duration=0 fqs_holdoff=0 fqs_stutter=3 test_boost=1/1 test_boost_interval=7 test_boost_duration=4 shutdown_secs=0 onoff_interval=0 onoff_holdoff=0
> > [ 3.059084]
> > [ 3.059451] ===============================
> > [ 3.060454] [ INFO: suspicious RCU usage. ]
> > [ 3.061482] 3.6.0-rc2-00010-g4c58c42 #59 Not tainted
> > [ 3.062686] -------------------------------
> > [ 3.063744] /c/kernel-tests/src/stable/kernel/rcutorture.c:990 suspicious rcu_dereference_check() usage!
> >
> > 982 do {
> > 983 schedule_timeout_uninterruptible(1);
> > 984 rp = rcu_torture_alloc();
> > 985 if (rp == NULL)
> > 986 continue;
> > 987 rp->rtort_pipe_count = 0;
> > 988 udelay(rcu_random(&rand) & 0x3ff);
> > 989 old_rp = rcu_dereference_check(rcu_torture_current,
> > >990 current == writer_task);
> > 991 rp->rtort_mbtest = 1;
> > 992 rcu_assign_pointer(rcu_torture_current, rp);
> > 993 smp_wmb(); /* Mods to old_rp must follow rcu_assign_pointer() */
> > 994 if (old_rp) {
>
>
> Does the following clear this up?

Sorry I'm still trying to reproduce this. It must be a rare bug
because it only showed up in several of the tens of thousands of test
boots. To reproduce it, I've done near 1000 boots however still not
caught it yet. Let's run it for more time...

Thanks,
Fengguang

> Thanx, Paul
>
> ------------------------------------------------------------------------
>
> rcu: Prevent initialization race in rcutorture kthreads
>
> When you do something like "t = kthread_run(...)", it is possible that
> the kthread will start running before the assignment to "t" happens.
> If the child kthread expects to find a pointer to its task_struct in "t",
> it will then be fatally disappointed. This commit therefore switches
> such cases to kthread_create() followed by wake_up_process(), guaranteeing
> that the assignment happens before the child kthread starts running.
>
> Reported-by: Fengguang Wu <fengguang.wu@xxxxxxxxx>
> Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
>
> diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c
> index 7a97b5b..8ff4fad 100644
> --- a/kernel/rcutorture.c
> +++ b/kernel/rcutorture.c
> @@ -2028,14 +2028,15 @@ rcu_torture_init(void)
> /* Start up the kthreads. */
>
> VERBOSE_PRINTK_STRING("Creating rcu_torture_writer task");
> - writer_task = kthread_run(rcu_torture_writer, NULL,
> - "rcu_torture_writer");
> + writer_task = kthread_create(rcu_torture_writer, NULL,
> + "rcu_torture_writer");
> if (IS_ERR(writer_task)) {
> firsterr = PTR_ERR(writer_task);
> VERBOSE_PRINTK_ERRSTRING("Failed to create writer");
> writer_task = NULL;
> goto unwind;
> }
> + wake_up_process(writer_task);
> fakewriter_tasks = kzalloc(nfakewriters * sizeof(fakewriter_tasks[0]),
> GFP_KERNEL);
> if (fakewriter_tasks == NULL) {
> @@ -2150,14 +2151,15 @@ rcu_torture_init(void)
> }
> if (shutdown_secs > 0) {
> shutdown_time = jiffies + shutdown_secs * HZ;
> - shutdown_task = kthread_run(rcu_torture_shutdown, NULL,
> - "rcu_torture_shutdown");
> + shutdown_task = kthread_create(rcu_torture_shutdown, NULL,
> + "rcu_torture_shutdown");
> if (IS_ERR(shutdown_task)) {
> firsterr = PTR_ERR(shutdown_task);
> VERBOSE_PRINTK_ERRSTRING("Failed to create shutdown");
> shutdown_task = NULL;
> goto unwind;
> }
> + wake_up_process(shutdown_task);
> }
> i = rcu_torture_onoff_init();
> if (i != 0) {
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/