[lockdep] cpufreq/ppc: possible circular locking dependency detected
Benjamin Herrenschmidt
benh at kernel.crashing.org
Tue Jun 9 07:10:19 AEST 2015
On Mon, 2015-06-08 at 18:45 +0300, Denis Kirjanov wrote:
> Hi,
>
> I see the following lockdep splat on my RackMac machine on boot:
I don't completely understand what lockdep is trying to tell us here...
I don't "see" the deadlock.
Ben.
> [ 6.002593] Registering G5 CPU frequency driver
> [ 6.011288] Frequency method: i2c/pfunc, Voltage method: i2c/pfunc
> [ 6.023605] Low: 1800 Mhz, High: 2300 Mhz, Cur: 1800 MHz
>
> [ 6.038022] ======================================================
> [ 6.042080] windfarm: CPUs control loops started.
> [ 6.042084] wf_rm31: Backside control loop started.
> [ 6.042086] wf_rm31: Slots control loop started.
> [ 6.078644] [ INFO: possible circular locking dependency detected ]
> [ 6.091149] 4.1.0-rc6-01265-g908e80d-dirty #15 Not tainted
> [ 6.102083] -------------------------------------------------------
> [ 6.114586] swapper/0/1 is trying to acquire lock:
> [ 6.124136] (&bus->mutex){+.+.+.}, at: [<c00000000004d450>]
> .pmac_i2c_open+0x30/0x100
> [ 6.139933]
> but task is already holding lock:
> [ 6.151569] (&policy->rwsem){+.+.+.}, at: [<c000000000620bbc>]
> .__cpufreq_add_dev.isra.29+0x34c/0xad0
> [ 6.170143]
> which lock already depends on the new lock.
>
> [ 6.186464]
> the existing dependency chain (in reverse order) is:
> [ 6.201401]
> -> #4 (&policy->rwsem){+.+.+.}:
> [ 6.210077] [<c0000000000d6380>] .lock_acquire+0x80/0x100
> [ 6.222061] [<c0000000007a411c>] .down_write+0x6c/0x110
> [ 6.233690] [<c000000000620bbc>]
> .__cpufreq_add_dev.isra.29+0x34c/0xad0
> [ 6.248104] [<c000000000489dd8>] .subsys_interface_register+0xc8/0x120
> [ 6.262336] [<c0000000006215f8>] .cpufreq_register_driver+0x138/0x310
> [ 6.276404] [<c000000000adeca8>] .g5_cpufreq_init+0x9cc/0xa28
> [ 6.289079] [<c00000000000a714>] .do_one_initcall+0xd4/0x270
> [ 6.301577] [<c000000000aaa778>] .kernel_init_freeable+0x190/0x270
> [ 6.315120] [<c00000000000b17c>] .kernel_init+0x1c/0x120
> [ 6.326928] [<c0000000000095b4>] .ret_from_kernel_thread+0x58/0xa4
> [ 6.340467]
> -> #3 (subsys mutex#2){+.+.+.}:
> [ 6.349149] [<c0000000000d6380>] .lock_acquire+0x80/0x100
> [ 6.361130] [<c0000000007a0ccc>] .mutex_lock_nested+0xac/0x5a0
> [ 6.373980] [<c00000000048aff4>] .bus_probe_device+0x44/0xf0
> [ 6.386477] [<c000000000488318>] .device_add+0x508/0x730
> [ 6.398284] [<c00000000048fb8c>] .register_cpu+0xfc/0x150
> [ 6.410266] [<c000000000aaf27c>] .topology_init+0x108/0x1c0
> [ 6.422595] [<c00000000000a714>] .do_one_initcall+0xd4/0x270
> [ 6.435092] [<c000000000aaa778>] .kernel_init_freeable+0x190/0x270
> [ 6.448637] [<c00000000000b17c>] .kernel_init+0x1c/0x120
> [ 6.460444] [<c0000000000095b4>] .ret_from_kernel_thread+0x58/0xa4
> [ 6.473983]
> -> #2 (cpu_add_remove_lock){+.+.+.}:
> [ 6.483536] [<c0000000000d6380>] .lock_acquire+0x80/0x100
> [ 6.495517] [<c0000000007a0ccc>] .mutex_lock_nested+0xac/0x5a0
> [ 6.508361] [<c0000000000781ac>] .cpu_maps_update_begin+0x1c/0x30
> [ 6.521732] [<c000000000070220>] .register_power_pmu+0xa0/0x128
> [ 6.534752] [<c000000000abed50>] .init_ppc970_pmu+0x6c/0x8c
> [ 6.547081] [<c00000000000a714>] .do_one_initcall+0xd4/0x270
> [ 6.559578] [<c000000000aaa698>] .kernel_init_freeable+0xb0/0x270
> [ 6.572950] [<c00000000000b17c>] .kernel_init+0x1c/0x120
> [ 6.584758] [<c0000000000095b4>] .ret_from_kernel_thread+0x58/0xa4
> [ 6.598297]
> -> #1 (&host->mutex){+.+.+.}:
> [ 6.606633] [<c0000000000d6380>] .lock_acquire+0x80/0x100
> [ 6.618615] [<c0000000007a0ccc>] .mutex_lock_nested+0xac/0x5a0
> [ 6.631458] [<c00000000004d408>] .kw_i2c_open+0x18/0x30
> [ 6.643092] [<c00000000004d4b4>] .pmac_i2c_open+0x94/0x100
> [ 6.655247] [<c000000000abb174>] .smp_core99_probe+0x260/0x410
> [ 6.668097] [<c000000000ab2f94>] .smp_prepare_cpus+0x1ec/0x220
> [ 6.680941] [<c000000000aaa670>] .kernel_init_freeable+0x88/0x270
> [ 6.694313] [<c00000000000b17c>] .kernel_init+0x1c/0x120
> [ 6.706121] [<c0000000000095b4>] .ret_from_kernel_thread+0x58/0xa4
> [ 6.719660]
> -> #0 (&bus->mutex){+.+.+.}:
> [ 6.727822] [<c0000000000d5304>] .__lock_acquire+0x1194/0x1b70
> [ 6.740672] [<c0000000000d6380>] .lock_acquire+0x80/0x100
> [ 6.752647] [<c0000000007a0ccc>] .mutex_lock_nested+0xac/0x5a0
> [ 6.765497] [<c00000000004d450>] .pmac_i2c_open+0x30/0x100
> [ 6.777652] [<c00000000004da34>] .pmac_i2c_do_begin+0x34/0x120
> [ 6.790501] [<c0000000000517e0>] .pmf_call_one+0x50/0xd0
> [ 6.802302] [<c0000000006265ec>] .g5_pfunc_switch_volt+0x2c/0xc0
> [ 6.815501] [<c00000000062659c>] .g5_pfunc_switch_freq+0x1cc/0x1f0
> [ 6.829040] [<c0000000006262fc>] .g5_cpufreq_target+0x2c/0x40
> [ 6.841716] [<c00000000061de40>] .__cpufreq_driver_target+0x230/0x400
> [ 6.855783] [<c000000000622bec>]
> .cpufreq_governor_performance+0x3c/0x60
> [ 6.870364] [<c00000000061f0d0>] .__cpufreq_governor+0xc0/0x390
> [ 6.883389] [<c00000000061f9fc>] .cpufreq_set_policy+0x25c/0x2c0
> [ 6.896581] [<c0000000006206d4>] .cpufreq_init_policy+0x84/0xe0
> [ 6.909605] [<c0000000006212ec>]
> .__cpufreq_add_dev.isra.29+0xa7c/0xad0
> [ 6.924013] [<c000000000489dd8>] .subsys_interface_register+0xc8/0x120
> [ 6.938253] [<c0000000006215f8>] .cpufreq_register_driver+0x138/0x310
> [ 6.952314] [<c000000000adeca8>] .g5_cpufreq_init+0x9cc/0xa28
> [ 6.964991] [<c00000000000a714>] .do_one_initcall+0xd4/0x270
> [ 6.977488] [<c000000000aaa778>] .kernel_init_freeable+0x190/0x270
> [ 6.991033] [<c00000000000b17c>] .kernel_init+0x1c/0x120
> [ 7.002841] [<c0000000000095b4>] .ret_from_kernel_thread+0x58/0xa4
> [ 7.016380]
> other info that might help us debug this:
>
> [ 7.032354] Chain exists of:
> &bus->mutex --> subsys mutex#2 --> &policy->rwsem
>
> [ 7.047635] Possible unsafe locking scenario:
>
> [ 7.059439] CPU0 CPU1
> [ 7.068467] ---- ----
> [ 7.077495] lock(&policy->rwsem);
> [ 7.084440] lock(subsys mutex#2);
> [ 7.096422] lock(&policy->rwsem);
> [ 7.108400] lock(&bus->mutex);
> [ 7.114829]
> *** DEADLOCK ***
>
> [ 7.126633] 3 locks held by swapper/0/1:
> [ 7.134445] #0: (subsys mutex#2){+.+.+.}, at:
> [<c000000000489d60>] .subsys_interface_register+0x50/0x120
> [ 7.153719] #1: (cpufreq_rwsem){.+.+.+}, at: [<c00000000062093c>]
> .__cpufreq_add_dev.isra.29+0xcc/0xad0
> [ 7.172817] #2: (&policy->rwsem){+.+.+.}, at:
> [<c000000000620bbc>] .__cpufreq_add_dev.isra.29+0x34c/0xad0
> [ 7.192261]
> stack backtrace:
> [ 7.200945] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 4.1.0-rc6-01265-g908e80d-dirty #15
> [ 7.217091] Call Trace:
> [ 7.221953] [c0000000790beab0] [c0000000007ad48c]
> .dump_stack+0x98/0xd4 (unreliable)
> [ 7.237407] [c0000000790beb30] [c0000000007a9094]
> .print_circular_bug+0x364/0x39c
> [ 7.252336] [c0000000790bebe0] [c0000000000d0e40] .check_prev_add+0x8d0/0x8e0
> [ 7.266575] [c0000000790becf0] [c0000000000d5304]
> .__lock_acquire+0x1194/0x1b70
> [ 7.281156] [c0000000790bee20] [c0000000000d6380] .lock_acquire+0x80/0x100
> [ 7.294874] [c0000000790beef0] [c0000000007a0ccc]
> .mutex_lock_nested+0xac/0x5a0
> [ 7.309456] [c0000000790bf010] [c00000000004d450] .pmac_i2c_open+0x30/0x100
> [ 7.323349] [c0000000790bf0a0] [c00000000004da34]
> .pmac_i2c_do_begin+0x34/0x120
> [ 7.337931] [c0000000790bf130] [c0000000000517e0] .pmf_call_one+0x50/0xd0
> [ 7.351477] [c0000000790bf1c0] [c0000000006265ec]
> .g5_pfunc_switch_volt+0x2c/0xc0
> [ 7.366405] [c0000000790bf230] [c00000000062659c]
> .g5_pfunc_switch_freq+0x1cc/0x1f0
> [ 7.381688] [c0000000790bf300] [c0000000006262fc]
> .g5_cpufreq_target+0x2c/0x40
> [ 7.396096] [c0000000790bf370] [c00000000061de40]
> .__cpufreq_driver_target+0x230/0x400
> [ 7.411893] [c0000000790bf430] [c000000000622bec]
> .cpufreq_governor_performance+0x3c/0x60
> [ 7.428219] [c0000000790bf4a0] [c00000000061f0d0]
> .__cpufreq_governor+0xc0/0x390
> [ 7.442975] [c0000000790bf530] [c00000000061f9fc]
> .cpufreq_set_policy+0x25c/0x2c0
> [ 7.457904] [c0000000790bf5c0] [c0000000006206d4]
> .cpufreq_init_policy+0x84/0xe0
> [ 7.472666] [c0000000790bf920] [c0000000006212ec]
> .__cpufreq_add_dev.isra.29+0xa7c/0xad0
> [ 7.488810] [c0000000790bfa10] [c000000000489dd8]
> .subsys_interface_register+0xc8/0x120
> [ 7.504782] [c0000000790bfab0] [c0000000006215f8]
> .cpufreq_register_driver+0x138/0x310
> [ 7.520586] [c0000000790bfb50] [c000000000adeca8]
> .g5_cpufreq_init+0x9cc/0xa28
> [ 7.534993] [c0000000790bfc10] [c00000000000a714] .do_one_initcall+0xd4/0x270
> [ 7.549233] [c0000000790bfd00] [c000000000aaa778]
> .kernel_init_freeable+0x190/0x270
> [ 7.564510] [c0000000790bfdb0] [c00000000000b17c] .kernel_init+0x1c/0x120
> [ 7.578056] [c0000000790bfe30] [c0000000000095b4]
> .ret_from_kernel_thread+0x58/0xa4
>
> Thanks.
More information about the Linuxppc-dev
mailing list