rcu warning caused by [sched: Fix race between task_group and sched_task_group]
Hi Kirill,
the problem below has already been reported here:
https://lkml.org/lkml/2014/11/11/505
but seems there is still no fix available upstream?
Would be good to have it soon, since this basically renders lockdep unusable.
[ 0.173906] ===============================
[ 0.173907] [ INFO: suspicious RCU usage. ]
[ 0.173909] 3.18.0-rc5 #15 Not tainted
[ 0.173909] -------------------------------
[ 0.173910] kernel/sched/core.c:7449 suspicious rcu_dereference_check() usage!
[ 0.173912]
other info that might help us debug this:
[ 0.173913]
rcu_scheduler_active = 1, debug_locks = 0
[ 0.173914] 2 locks held by swapper/0/0:
[ 0.173915] #0: (&p->pi_lock){......}, at: [<000000000017baae>] task_rq_lock+0x4e/0xf8
[ 0.173925] #1: (&rq->lock){......}, at: [<000000000017bad6>] task_rq_lock+0x76/0xf8
[ 0.173928]
[ 0.173930] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.18.0-rc5 #15
[ 0.173931] 0000000000b9fb70 0000000000b9fb80 0000000000000002 0000000000000000
0000000000b9fc10 0000000000b9fb88 0000000000b9fb88 0000000000112638
0000000000000000 00000000009d5b8c 00000000009be240 000000000000000b
0000000000b9fbd0 0000000000b9fb70 0000000000000000 0000000000000000
0000000000000000 0000000000112638 0000000000b9fb70 0000000000b9fbd0
[ 0.173944] Call Trace:
[ 0.173947] ([<000000000011251c>] show_trace+0xf4/0x158)
[ 0.173949] [<00000000001125ea>] show_stack+0x6a/0xe8
[ 0.173952] [<0000000000807ac2>] dump_stack+0x82/0xb0
[ 0.173955] [<000000000018700a>] sched_move_task+0x20a/0x210
[ 0.173959] [<0000000000203bf0>] cgroup_post_fork+0x60/0x148
[ 0.173963] [<000000000014c5b8>] copy_process.part.46+0xfe8/0x1bc8
[ 0.173964] [<000000000014d39c>] do_fork+0x104/0x7d8
[ 0.173965] [<000000000014dab6>] kernel_thread+0x46/0x58
[ 0.173968] [<00000000007fe890>] rest_init+0x38/0x170
[ 0.173971] [<0000000000c45b40>] start_kernel+0x460/0x470
[ 0.173972] [<0000000000100020>] _stext+0x20/0x80
[ 0.173973] INFO: lockdep is turned off.caused by (bisected):
commit eeb61e53ea19be0c4015b00b2e8b3b2185436f2b
Author: Kirill Tkhai <[email protected]> Date: Mon Oct 27 14:18:25 2014 +0400
sched: Fix race between task_group and sched_task_group
Quoted 7 lines · Heiko Carstens
Hi Kirill,
the problem below has already been reported here:
https://lkml.org/lkml/2014/11/11/505but seems there is still no fix available upstream?
Would be good to have it soon, since this basically renders lockdep unusable.
This one fixes this bug for me:
https://lkml.org/lkml/2014/11/4/613
Thanx, Paul
Quoted 47 lines
[ 0.173906] =============================== [ 0.173907] [ INFO: suspicious RCU usage. ] [ 0.173909] 3.18.0-rc5 #15 Not tainted [ 0.173909] ------------------------------- [ 0.173910] kernel/sched/core.c:7449 suspicious rcu_dereference_check() usage! [ 0.173912] other info that might help us debug this: [ 0.173913] rcu_scheduler_active = 1, debug_locks = 0 [ 0.173914] 2 locks held by swapper/0/0: [ 0.173915] #0: (&p->pi_lock){......}, at: [<000000000017baae>] task_rq_lock+0x4e/0xf8 [ 0.173925] #1: (&rq->lock){......}, at: [<000000000017bad6>] task_rq_lock+0x76/0xf8 [ 0.173928] [ 0.173930] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.18.0-rc5 #15 [ 0.173931] 0000000000b9fb70 0000000000b9fb80 0000000000000002 0000000000000000 0000000000b9fc10 0000000000b9fb88 0000000000b9fb88 0000000000112638 0000000000000000 00000000009d5b8c 00000000009be240 000000000000000b 0000000000b9fbd0 0000000000b9fb70 0000000000000000 0000000000000000 0000000000000000 0000000000112638 0000000000b9fb70 0000000000b9fbd0 [ 0.173944] Call Trace: [ 0.173947] ([<000000000011251c>] show_trace+0xf4/0x158) [ 0.173949] [<00000000001125ea>] show_stack+0x6a/0xe8 [ 0.173952] [<0000000000807ac2>] dump_stack+0x82/0xb0 [ 0.173955] [<000000000018700a>] sched_move_task+0x20a/0x210 [ 0.173959] [<0000000000203bf0>] cgroup_post_fork+0x60/0x148 [ 0.173963] [<000000000014c5b8>] copy_process.part.46+0xfe8/0x1bc8 [ 0.173964] [<000000000014d39c>] do_fork+0x104/0x7d8 [ 0.173965] [<000000000014dab6>] kernel_thread+0x46/0x58 [ 0.173968] [<00000000007fe890>] rest_init+0x38/0x170 [ 0.173971] [<0000000000c45b40>] start_kernel+0x460/0x470 [ 0.173972] [<0000000000100020>] _stext+0x20/0x80 [ 0.173973] INFO: lockdep is turned off.caused by (bisected):
commit
eeb61e53ea19be0c4015b00b2e8b3b2185436f2bAuthor: Kirill Tkhai <[email protected]> Date: Mon Oct 27 14:18:25 2014 +0400sched: Fix race between
task_groupandsched_task_group--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Quoted 6 lines · Paul E. McKenney
Quoted 5 lines · Heiko Carstens
Hi Kirill,
the problem below has already been reported here:
https://lkml.org/lkml/2014/11/11/505
Quoted 6 lines
Quoted 2 lines
but seems there is still no fix available upstream?
Would be good to have it soon, since this basically renders lockdep unusable.This one fixes this bug for me:
Yes. It's still not in Linus' tree.
josh