Re: possible circular locking dependency

From: Paul E. McKenney
Date: Sun May 06 2012 - 12:43:44 EST

Next message: David Miller: "Re: [PATCH resend] [IPV6] remove sysctl accept_source_route"
Previous message: Sami Liedes: "kmemleak: Cannot insert 0xffff880007fedd28 into the object searchtree (already existing)"
In reply to: Avi Kivity: "Re: possible circular locking dependency"
Next in thread: Sergey Senozhatsky: "Re: possible circular locking dependency"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Sun, May 06, 2012 at 11:55:30AM +0300, Avi Kivity wrote:
> On 05/03/2012 11:02 PM, Sergey Senozhatsky wrote:
> > Hello,
> > 3.4-rc5
>
> Whoa.
>
> Looks like inconsistent locking between cpufreq and
> synchronize_srcu_expedited(). kvm triggered this because it is one of
> the few users of synchronize_srcu_expedited(), but I don't think it is
> doing anything wrong directly.
>
> Dave, Paul?

SRCU hasn't changed much in mainline for quite some time. Holding
the hotplug mutex across a synchronize_srcu() is a bad idea, though.

However, there is a reworked implementation (courtesy of Lai Jiangshan)
in -rcu that does not acquire the hotplug mutex. Could you try that out?

Thanx, Paul

> > [32881.212463] kvm: VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL does not work properly. Using workaround
> > [32882.360505]
> > [32882.360509] ======================================================
> > [32882.360511] [ INFO: possible circular locking dependency detected ]
> > [32882.360515] 3.4.0-rc5-dbg-00932-gfabccd4-dirty #1107 Not tainted
> > [32882.360517] -------------------------------------------------------
> > [32882.360519] qemu-system-x86/15168 is trying to acquire lock:
> > [32882.360521] (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81033a60>] get_online_cpus+0x41/0x55
> > [32882.360532]
> > [32882.360532] but task is already holding lock:
> > [32882.360534] (&sp->mutex){+.+...}, at: [<ffffffff81057f91>] __synchronize_srcu+0x9a/0x104
> > [32882.360542]
> > [32882.360542] which lock already depends on the new lock.
> > [32882.360543]
> > [32882.360545]
> > [32882.360545] the existing dependency chain (in reverse order) is:
> > [32882.360547]
> > [32882.360547] -> #3 (&sp->mutex){+.+...}:
> > [32882.360552] [<ffffffff81085b36>] lock_acquire+0x148/0x1c3
> > [32882.360557] [<ffffffff814aa040>] mutex_lock_nested+0x6e/0x2d1
> > [32882.360562] [<ffffffff81057f91>] __synchronize_srcu+0x9a/0x104
> > [32882.360566] [<ffffffff81058027>] synchronize_srcu+0x15/0x17
> > [32882.360569] [<ffffffff810586e8>] srcu_notifier_chain_unregister+0x5b/0x69
> > [32882.360573] [<ffffffff813e3110>] cpufreq_unregister_notifier+0x22/0x3c
> > [32882.360580] [<ffffffff813e3e42>] cpufreq_governor_dbs+0x322/0x3ac
> > [32882.360584] [<ffffffff813e2075>] __cpufreq_governor+0x6b/0xa8
> > [32882.360587] [<ffffffff813e21a5>] __cpufreq_set_policy+0xf3/0x145
> > [32882.360591] [<ffffffff813e2def>] store_scaling_governor+0x173/0x1a9
> > [32882.360594] [<ffffffff813e1f71>] store+0x5a/0x86
> > [32882.360597] [<ffffffff81181e83>] sysfs_write_file+0xee/0x126
> > [32882.360603] [<ffffffff8111f6b1>] vfs_write+0xa3/0x14c
> > [32882.360607] [<ffffffff8111f959>] sys_write+0x43/0x73
> > [32882.360610] [<ffffffff814b432d>] system_call_fastpath+0x1a/0x1f
> > [32882.360614]
> > [32882.360615] -> #2 (dbs_mutex){+.+.+.}:
> > [32882.360619] [<ffffffff81085b36>] lock_acquire+0x148/0x1c3
> > [32882.360622] [<ffffffff814aa040>] mutex_lock_nested+0x6e/0x2d1
> > [32882.360625] [<ffffffff813e3b9c>] cpufreq_governor_dbs+0x7c/0x3ac
> > [32882.360629] [<ffffffff813e2075>] __cpufreq_governor+0x6b/0xa8
> > [32882.360632] [<ffffffff813e21bb>] __cpufreq_set_policy+0x109/0x145
> > [32882.360636] [<ffffffff813e244e>] cpufreq_add_dev_interface+0x257/0x288
> > [32882.360639] [<ffffffff813e2889>] cpufreq_add_dev+0x40a/0x42a
> > [32882.360643] [<ffffffff81398694>] subsys_interface_register+0x9b/0xdc
> > [32882.360648] [<ffffffff813e1935>] cpufreq_register_driver+0xa0/0x14b
> > [32882.360652] [<ffffffffa00a1086>] store_up_threshold+0x3a/0x50 [cpufreq_ondemand]
> > [32882.360657] [<ffffffff8100020f>] do_one_initcall+0x7f/0x140
> > [32882.360663] [<ffffffff8109158c>] sys_init_module+0x1818/0x1aec
> > [32882.360667] [<ffffffff814b432d>] system_call_fastpath+0x1a/0x1f
> > [32882.360671]
> > [32882.360671] -> #1 (&per_cpu(cpu_policy_rwsem, cpu)){+++++.}:
> > [32882.360675] [<ffffffff81085b36>] lock_acquire+0x148/0x1c3
> > [32882.360679] [<ffffffff814aa661>] down_write+0x49/0x6c
> > [32882.360682] [<ffffffff813e1eaa>] lock_policy_rwsem_write+0x47/0x78
> > [32882.360685] [<ffffffff814a18c9>] cpufreq_cpu_callback+0x57/0x81
> > [32882.360692] [<ffffffff814b0643>] notifier_call_chain+0xac/0xd9
> > [32882.360697] [<ffffffff810582b6>] __raw_notifier_call_chain+0xe/0x10
> > [32882.360701] [<ffffffff81033950>] __cpu_notify+0x20/0x37
> > [32882.360705] [<ffffffff8149091c>] _cpu_down+0x7b/0x25d
> > [32882.360709] [<ffffffff81033b2f>] disable_nonboot_cpus+0x5f/0x10b
> > [32882.360712] [<ffffffff81072e61>] suspend_devices_and_enter+0x197/0x401
> > [32882.360719] [<ffffffff810731cf>] pm_suspend+0x104/0x1bd
> > [32882.360722] [<ffffffff8107213a>] state_store+0xa0/0xc9
> > [32882.360726] [<ffffffff8127115a>] kobj_attr_store+0xf/0x1b
> > [32882.360730] [<ffffffff81181e83>] sysfs_write_file+0xee/0x126
> > [32882.360733] [<ffffffff8111f6b1>] vfs_write+0xa3/0x14c
> > [32882.360736] [<ffffffff8111f959>] sys_write+0x43/0x73
> > [32882.360739] [<ffffffff814b432d>] system_call_fastpath+0x1a/0x1f
> > [32882.360743]
> > [32882.360743] -> #0 (cpu_hotplug.lock){+.+.+.}:
> > [32882.360747] [<ffffffff81084e93>] __lock_acquire+0xf6b/0x1612
> > [32882.360751] [<ffffffff81085b36>] lock_acquire+0x148/0x1c3
> > [32882.360754] [<ffffffff814aa040>] mutex_lock_nested+0x6e/0x2d1
> > [32882.360757] [<ffffffff81033a60>] get_online_cpus+0x41/0x55
> > [32882.360760] [<ffffffff810ab15b>] synchronize_sched_expedited+0x26/0xfa
> > [32882.360766] [<ffffffff81057f9f>] __synchronize_srcu+0xa8/0x104
> > [32882.360769] [<ffffffff81058010>] synchronize_srcu_expedited+0x15/0x17
> > [32882.360773] [<ffffffffa01df109>] __kvm_set_memory_region+0x3d8/0x46a [kvm]
> > [32882.360789] [<ffffffffa01df1d2>] kvm_set_memory_region+0x37/0x50 [kvm]
> > [32882.360798] [<ffffffffa0258a89>] vmx_set_tss_addr+0x4c/0x200 [kvm_intel]
> > [32882.360803] [<ffffffffa01ef732>] kvm_arch_vm_ioctl+0x160/0x9df [kvm]
> > [32882.360816] [<ffffffffa01df571>] kvm_vm_ioctl+0x36a/0x39c [kvm]
> > [32882.360825] [<ffffffff8112fc68>] vfs_ioctl+0x24/0x2f
> > [32882.360829] [<ffffffff81130566>] do_vfs_ioctl+0x412/0x455
> > [32882.360832] [<ffffffff811305ff>] sys_ioctl+0x56/0x7b
> > [32882.360835] [<ffffffff814b432d>] system_call_fastpath+0x1a/0x1f
> > [32882.360839]
> > [32882.360839] other info that might help us debug this:
> > [32882.360840]
> > [32882.360842] Chain exists of:
> > [32882.360842] cpu_hotplug.lock --> dbs_mutex --> &sp->mutex
> > [32882.360847]
> > [32882.360848] Possible unsafe locking scenario:
> > [32882.360849]
> > [32882.360851] CPU0 CPU1
> > [32882.360852] ---- ----
> > [32882.360854] lock(&sp->mutex);
> > [32882.360856] lock(dbs_mutex);
> > [32882.360859] lock(&sp->mutex);
> > [32882.360862] lock(cpu_hotplug.lock);
> > [32882.360865]
> > [32882.360865] *** DEADLOCK ***
> > [32882.360866]
> > [32882.360868] 2 locks held by qemu-system-x86/15168:
> > [32882.360870] #0: (&kvm->slots_lock){+.+.+.}, at: [<ffffffffa01df1c4>] kvm_set_memory_region+0x29/0x50 [kvm]
> > [32882.360882] #1: (&sp->mutex){+.+...}, at: [<ffffffff81057f91>] __synchronize_srcu+0x9a/0x104
> > [32882.360888]
> > [32882.360889] stack backtrace:
> > [32882.360892] Pid: 15168, comm: qemu-system-x86 Not tainted 3.4.0-rc5-dbg-00932-gfabccd4-dirty #1107
> > [32882.360894] Call Trace:
> > [32882.360898] [<ffffffff814a3d1d>] print_circular_bug+0x29f/0x2b0
> > [32882.360901] [<ffffffff81084e93>] __lock_acquire+0xf6b/0x1612
> > [32882.360905] [<ffffffff81085b36>] lock_acquire+0x148/0x1c3
> > [32882.360908] [<ffffffff81033a60>] ? get_online_cpus+0x41/0x55
> > [32882.360911] [<ffffffff814aa040>] mutex_lock_nested+0x6e/0x2d1
> > [32882.360914] [<ffffffff81033a60>] ? get_online_cpus+0x41/0x55
> > [32882.360917] [<ffffffff810ab135>] ? synchronize_sched+0xa8/0xa8
> > [32882.360921] [<ffffffff81033a60>] get_online_cpus+0x41/0x55
> > [32882.360923] [<ffffffff810ab15b>] synchronize_sched_expedited+0x26/0xfa
> > [32882.360927] [<ffffffff810ab135>] ? synchronize_sched+0xa8/0xa8
> > [32882.360930] [<ffffffff81057f9f>] __synchronize_srcu+0xa8/0x104
> > [32882.360933] [<ffffffff81058010>] synchronize_srcu_expedited+0x15/0x17
> > [32882.360942] [<ffffffffa01df109>] __kvm_set_memory_region+0x3d8/0x46a [kvm]
> > [32882.360945] [<ffffffff81086268>] ? mark_held_locks+0xbe/0xea
> > [32882.360954] [<ffffffffa01df1d2>] kvm_set_memory_region+0x37/0x50 [kvm]
> > [32882.360959] [<ffffffffa0258a89>] vmx_set_tss_addr+0x4c/0x200 [kvm_intel]
> > [32882.360971] [<ffffffffa01ef732>] kvm_arch_vm_ioctl+0x160/0x9df [kvm]
> > [32882.360980] [<ffffffffa01df571>] kvm_vm_ioctl+0x36a/0x39c [kvm]
> > [32882.360984] [<ffffffff810855c5>] ? lock_release_non_nested+0x8b/0x241
> > [32882.360987] [<ffffffff8112fc68>] vfs_ioctl+0x24/0x2f
> > [32882.360990] [<ffffffff81130566>] do_vfs_ioctl+0x412/0x455
> > [32882.360993] [<ffffffff81120f8f>] ? fget_light+0x120/0x39b
> > [32882.360996] [<ffffffff811305ff>] sys_ioctl+0x56/0x7b
> > [32882.360999] [<ffffffff814b432d>] system_call_fastpath+0x1a/0x1f
> >
> >
> >
> > -ss
>
>
> --
> error compiling committee.c: too many arguments to function
>

--
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/

Next message: David Miller: "Re: [PATCH resend] [IPV6] remove sysctl accept_source_route"
Previous message: Sami Liedes: "kmemleak: Cannot insert 0xffff880007fedd28 into the object searchtree (already existing)"
In reply to: Avi Kivity: "Re: possible circular locking dependency"
Next in thread: Sergey Senozhatsky: "Re: possible circular locking dependency"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]