Re: [Bug #12574] possible circular locking dependency detected

From: Michael S. Tsirkin
Date: Tue Feb 10 2009 - 17:36:36 EST

Dave, dri guys,
Could you take a look at this circular dependency please (below)? I
observe it when suspending laptop with radeon drm loaded and with
lockdep enabled. It seems that the root of the problem is that
various vm ops such as drm_vm_open, drm_mmap) are called with mm
semaphore taken, and take dev->struct_mutex. On the other hand,
drm_rmmap_locked is called with dev->struct_mutex, and calls mtrr_del
which depends on mm semaphore indirectly.

What do you think?

Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=12574
Subject : possible circular locking dependency detected
Submitter : Michael S. Tsirkin <m.s.tsirkin@xxxxxxxxx>
Date : 2009-01-29 11:35 (11 days old)

/var/log/message dump below.


=======================================================
[ INFO: possible circular locking dependency detected ]
2.6.29-rc4-mst-debug #95
-------------------------------------------------------
sleep.sh/6730 is trying to acquire lock:
(&per_cpu(cpu_policy_rwsem, cpu)){----}, at: [<c02c0da1>] lock_policy_rwsem_write+0x31/0x70

but task is already holding lock:
(&cpu_hotplug.lock){--..}, at: [<c012d89a>] cpu_hotplug_begin+0x1a/0x50

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #6 (&cpu_hotplug.lock){--..}:
[<c0152221>] validate_chain+0xb51/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c0366c5d>] mutex_lock_nested+0x9d/0x2e0
[<c012d8fc>] get_online_cpus+0x2c/0x40
[<c010d96f>] mtrr_del_page+0x2f/0x160
[<c010dada>] mtrr_del+0x3a/0x50
[<f851a342>] drm_rmmap_locked+0xc2/0x180 [drm]
[<f8521d31>] drm_master_destroy+0x151/0x160 [drm]
[<c022a37c>] kref_put+0x2c/0x80
[<f8521af2>] drm_master_put+0x12/0x20 [drm]
[<f851dd1b>] drm_release+0x25b/0x4a0 [drm]
[<c019781d>] __fput+0xbd/0x1d0
[<c0197c09>] fput+0x19/0x20
[<c0194a47>] filp_close+0x47/0x70
[<c0194ada>] sys_close+0x6a/0xc0
[<c0103215>] sysenter_do_call+0x12/0x35
[<ffffffff>] 0xffffffff

-> #5 (&dev->struct_mutex){--..}:
[<c0152221>] validate_chain+0xb51/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c0366c5d>] mutex_lock_nested+0x9d/0x2e0
[<f8522add>] drm_vm_open+0x2d/0x50 [drm]
[<c012a397>] dup_mm+0x227/0x310
[<c012b22f>] copy_process+0xd7f/0x1020
[<c012b5e8>] do_fork+0x78/0x320
[<c01017ef>] sys_clone+0x2f/0x40
[<c0103215>] sysenter_do_call+0x12/0x35
[<ffffffff>] 0xffffffff

-> #4 (&mm->mmap_sem/1){--..}:
[<c0152221>] validate_chain+0xb51/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c01441e8>] down_write_nested+0x48/0x70
[<c012a238>] dup_mm+0xc8/0x310
[<c012b22f>] copy_process+0xd7f/0x1020
[<c012b5e8>] do_fork+0x78/0x320
[<c01017ef>] sys_clone+0x2f/0x40
[<c0103292>] syscall_call+0x7/0xb
[<ffffffff>] 0xffffffff

-> #3 (&mm->mmap_sem){----}:
[<c0152221>] validate_chain+0xb51/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c0183d73>] might_fault+0x73/0x90
[<c022f633>] copy_to_user+0x33/0x60
[<c01a3975>] filldir64+0xb5/0xe0
[<c01e0c2f>] sysfs_readdir+0x11f/0x1f0
[<c01a3b0d>] vfs_readdir+0x8d/0xb0
[<c01a3b99>] sys_getdents64+0x69/0xc0
[<c0103292>] syscall_call+0x7/0xb
[<ffffffff>] 0xffffffff

-> #2 (sysfs_mutex){--..}:
[<c0152221>] validate_chain+0xb51/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c0366c5d>] mutex_lock_nested+0x9d/0x2e0
[<c01e0f0c>] sysfs_addrm_start+0x2c/0xb0
[<c01e14a0>] create_dir+0x40/0x90
[<c01e1556>] sysfs_create_subdir+0x16/0x20
[<c01e2770>] internal_create_group+0x50/0x1a0
[<c01e28ec>] sysfs_create_group+0xc/0x10
[<f81674fc>] cpufreq_stat_notifier_policy+0x9c/0x230 [cpufreq_stats]
[<c036b007>] notifier_call_chain+0x37/0x80
[<c0144d24>] __blocking_notifier_call_chain+0x44/0x60
[<c0144d5a>] blocking_notifier_call_chain+0x1a/0x20
[<c02c0226>] __cpufreq_set_policy+0xd6/0x230
[<c02c14a8>] cpufreq_add_dev+0x4e8/0x6b0
[<c029d5a5>] sysdev_driver_register+0x75/0x130
[<c02bff55>] cpufreq_register_driver+0xb5/0x1c0
[<f808b0bd>] uinput_destroy_device+0x4d/0x60 [uinput]
[<c010111a>] do_one_initcall+0x2a/0x160
[<c015bdf5>] sys_init_module+0x85/0x1b0
[<c0103215>] sysenter_do_call+0x12/0x35
[<ffffffff>] 0xffffffff

-> #1 ((cpufreq_policy_notifier_list).rwsem){----}:
[<c0152221>] validate_chain+0xb51/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c0367441>] down_read+0x41/0x60
[<c0144d0a>] __blocking_notifier_call_chain+0x2a/0x60
[<c0144d5a>] blocking_notifier_call_chain+0x1a/0x20
[<c02c1165>] cpufreq_add_dev+0x1a5/0x6b0
[<c029d5a5>] sysdev_driver_register+0x75/0x130
[<c02bff55>] cpufreq_register_driver+0xb5/0x1c0
[<f808b0bd>] uinput_destroy_device+0x4d/0x60 [uinput]
[<c010111a>] do_one_initcall+0x2a/0x160
[<c015bdf5>] sys_init_module+0x85/0x1b0
[<c0103215>] sysenter_do_call+0x12/0x35
[<ffffffff>] 0xffffffff

-> #0 (&per_cpu(cpu_policy_rwsem, cpu)){----}:
[<c0151cbb>] validate_chain+0x5eb/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c01532d0>] lock_acquire+0x60/0x80
[<c03674a1>] down_write+0x41/0x60
[<c02c0da1>] lock_policy_rwsem_write+0x31/0x70
[<c03655a5>] cpufreq_cpu_callback+0x45/0x80
[<c036b007>] notifier_call_chain+0x37/0x80
[<c0144b49>] __raw_notifier_call_chain+0x19/0x20
[<c03574c9>] _cpu_down+0x79/0x280
[<c012da5c>] disable_nonboot_cpus+0x7c/0x100
[<c015cac5>] suspend_devices_and_enter+0xd5/0x170
[<c015cd40>] enter_state+0x1b0/0x1c0
[<c015cddf>] state_store+0x8f/0xd0
[<c0228cf4>] kobj_attr_store+0x24/0x30
[<c01e02d2>] sysfs_write_file+0xa2/0x100
[<c0196e89>] vfs_write+0x99/0x130
[<c01973cd>] sys_write+0x3d/0x70
[<c0103215>] sysenter_do_call+0x12/0x35
[<ffffffff>] 0xffffffff

other info that might help us debug this:

4 locks held by sleep.sh/6730:
#0: (&buffer->mutex){--..}, at: [<c01e025b>] sysfs_write_file+0x2b/0x100
#1: (pm_mutex){--..}, at: [<c015cbdb>] enter_state+0x4b/0x1c0
#2: (cpu_add_remove_lock){--..}, at: [<c012d83f>] cpu_maps_update_begin+0xf/0x20
#3: (&cpu_hotplug.lock){--..}, at: [<c012d89a>] cpu_hotplug_begin+0x1a/0x50

stack backtrace:
Pid: 6730, comm: sleep.sh Not tainted 2.6.29-rc4-mst-debug #95
Call Trace:
[<c015166c>] print_circular_bug_tail+0x7c/0xe0
[<c0151cbb>] validate_chain+0x5eb/0x1150
[<c0152a66>] __lock_acquire+0x246/0xa50
[<c013cf1e>] ? __cancel_work_timer+0x2e/0x190
[<c01532d0>] lock_acquire+0x60/0x80
[<c02c0da1>] ? lock_policy_rwsem_write+0x31/0x70
[<c03674a1>] down_write+0x41/0x60
[<c02c0da1>] ? lock_policy_rwsem_write+0x31/0x70
[<c02c0da1>] lock_policy_rwsem_write+0x31/0x70
[<c03655a5>] cpufreq_cpu_callback+0x45/0x80
[<c036b007>] notifier_call_chain+0x37/0x80
[<c0144b49>] __raw_notifier_call_chain+0x19/0x20
[<c03574c9>] _cpu_down+0x79/0x280
[<c012d83f>] ? cpu_maps_update_begin+0xf/0x20
[<c012da5c>] disable_nonboot_cpus+0x7c/0x100
[<c02531cb>] ? acpi_disable_all_gpes+0x25/0x2a
[<c015cac5>] suspend_devices_and_enter+0xd5/0x170
[<c015cd40>] enter_state+0x1b0/0x1c0
[<c015cddf>] state_store+0x8f/0xd0
[<c015cd50>] ? state_store+0x0/0xd0
[<c0228cf4>] kobj_attr_store+0x24/0x30
[<c01e02d2>] sysfs_write_file+0xa2/0x100
[<c0196e89>] vfs_write+0x99/0x130
[<c0103247>] ? sysenter_exit+0xf/0x18
[<c01e0230>] ? sysfs_write_file+0x0/0x100
[<c01973cd>] sys_write+0x3d/0x70
[<c0103215>] sysenter_do_call+0x12/0x35

--
MST
--
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/