BUG: circular locking dependency detected

[Russell King <rmk@arm.linux.org.uk>]

This looks like a bug in the framebuffer/console layers.  Looks like
we have one path where we call the notifier list, and a called
function takes the console lock, and another path where we hold the
console lock while calling the notifier list.

===========================
[ INFO: possible circular locking dependency detected ]
3.8.0-rc4+ #656 Not tainted
-------------------------------------------------------
kworker/0:1/442 is trying to acquire lock:
 ((fb_notifier_list).rwsem){.+.+.+}, at: [<c004ea48>] __blocking_notifier_call_chain+0x34/0x68

but task is already holding lock:
 (console_lock){+.+.+.}, at: [<c01c2b48>] console_callback+0x14/0x138

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (console_lock){+.+.+.}:
       [<c006f36c>] __lock_acquire+0x1d20/0x1e80
       [<c006fa04>] lock_acquire+0x68/0x7c
       [<c002a894>] console_lock+0x5c/0x70
       [<c01c0adc>] register_con_driver+0x44/0x150
       [<c01c1158>] take_over_console+0x24/0x3b4
       [<c019d778>] fbcon_takeover+0x70/0xd4
       [<c01a3108>] fbcon_event_notify+0x7c8/0x818
       [<c004e538>] notifier_call_chain+0x4c/0x8c
       [<c004ea64>] __blocking_notifier_call_chain+0x50/0x68
       [<c004ea9c>] blocking_notifier_call_chain+0x20/0x28
       [<c0196e70>] fb_notifier_call_chain+0x20/0x28
       [<c0198194>] register_framebuffer+0x18c/0x238
       [<c01a6148>] clcdfb_probe+0x2b0/0x3c0
       [<c01a6da4>] amba_probe+0x88/0xa0
       [<c01d1730>] driver_probe_device+0x84/0x218
       [<c01d1960>] __driver_attach+0x9c/0xa0
       [<c01cfe5c>] bus_for_each_dev+0x5c/0x88
       [<c01d1398>] driver_attach+0x20/0x28
       [<c01d0ddc>] bus_add_driver+0xa4/0x244
       [<c01d1ed4>] driver_register+0x80/0x14c
       [<c01a6848>] amba_driver_register+0x48/0x5c
       [<c043aae0>] amba_clcdfb_init+0x28/0x3c
       [<c000868c>] do_one_initcall+0x44/0x1ac
       [<c0425928>] kernel_init_freeable+0x104/0x1c8
       [<c03242ec>] kernel_init+0x10/0xec
       [<c0014510>] ret_from_fork+0x14/0x24

-> #0 ((fb_notifier_list).rwsem){.+.+.+}:
       [<c006bc44>] print_circular_bug+0x84/0x2f0
       [<c006f458>] __lock_acquire+0x1e0c/0x1e80
       [<c006fa04>] lock_acquire+0x68/0x7c
       [<c032b3a0>] down_read+0x34/0x44
       [<c004ea48>] __blocking_notifier_call_chain+0x34/0x68
       [<c004ea9c>] blocking_notifier_call_chain+0x20/0x28
       [<c0196e70>] fb_notifier_call_chain+0x20/0x28
       [<c0197690>] fb_blank+0x40/0xac
       [<c019f874>] fbcon_blank+0x1f4/0x29c
       [<c01c09e0>] do_blank_screen+0x1b8/0x270
       [<c01c2ba8>] console_callback+0x74/0x138
       [<c00408c8>] process_one_work+0x1b4/0x4ec
       [<c0043610>] worker_thread+0x17c/0x4bc
       [<c004891c>] kthread+0xb0/0xbc
       [<c0014510>] ret_from_fork+0x14/0x24

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(console_lock);
                               lock((fb_notifier_list).rwsem);
                               lock(console_lock);
  lock((fb_notifier_list).rwsem);

 *** DEADLOCK ***

3 locks held by kworker/0:1/442:
 #0:  (events){.+.+..}, at: [<c0040854>] process_one_work+0x140/0x4ec
 #1:  (console_work){+.+...}, at: [<c0040854>] process_one_work+0x140/0x4ec
 #2:  (console_lock){+.+.+.}, at: [<c01c2b48>] console_callback+0x14/0x138

stack backtrace:
Backtrace:
[<c00185d8>] (dump_backtrace+0x0/0x10c) from [<c03294c8>] (dump_stack+0x18/0x1c)
 r6:c05323f0 r5:c0524800 r4:c05323f0 r3:cf8f6b80
[<c03294b0>] (dump_stack+0x0/0x1c) from [<c006bda4>] (print_circular_bug+0x1e4/0x2f0)
[<c006bbc0>] (print_circular_bug+0x0/0x2f0) from [<c006f458>] (__lock_acquire+0x1e0c/0x1e80)
[<c006d64c>] (__lock_acquire+0x0/0x1e80) from [<c006fa04>] (lock_acquire+0x68/0x7c)
[<c006f99c>] (lock_acquire+0x0/0x7c) from [<c032b3a0>] (down_read+0x34/0x44)
 r7:00000010 r6:cfb7dd20 r5:00000002 r4:c04715cc
[<c032b36c>] (down_read+0x0/0x44) from [<c004ea48>] (__blocking_notifier_call_chain+0x34/0x68)
 r5:ffffffff r4:c04715cc
[<c004ea14>] (__blocking_notifier_call_chain+0x0/0x68) from [<c004ea9c>] (blocking_notifier_call_chain+0x20/0x28)
 r7:cf0ffc00 r6:00000001 r5:cfb7dd20 r4:cfb5f800
[<c004ea7c>] (blocking_notifier_call_chain+0x0/0x28) from [<c0196e70>] (fb_notifier_call_chain+0x20/0x28)
[<c0196e50>] (fb_notifier_call_chain+0x0/0x28) from [<c0197690>] (fb_blank+0x40/0xac)
[<c0197650>] (fb_blank+0x0/0xac) from [<c019f874>] (fbcon_blank+0x1f4/0x29c)
 r6:00000001 r5:cf80a000 r4:cfb5f800
[<c019f680>] (fbcon_blank+0x0/0x29c) from [<c01c09e0>] (do_blank_screen+0x1b8/0x270)
[<c01c0828>] (do_blank_screen+0x0/0x270) from [<c01c2ba8>] (console_callback+0x74/0x138)
 r7:c0ba8640 r6:c0bac300 r5:c099a51c r4:c099a51c
[<c01c2b34>] (console_callback+0x0/0x138) from [<c00408c8>] (process_one_work+0x1b4/0x4ec)
 r6:c0bac300 r5:cf9489c0 r4:c0472e3c r3:c01c2b34
[<c0040714>] (process_one_work+0x0/0x4ec) from [<c0043610>] (worker_thread+0x17c/0x4bc)
[<c0043494>] (worker_thread+0x0/0x4bc) from [<c004891c>] (kthread+0xb0/0xbc)
[<c004886c>] (kthread+0x0/0xbc) from [<c0014510>] (ret_from_fork+0x14/0x24)
 r8:00000000 r7:00000000 r6:00000000 r5:c004886c r4:cf84dde8

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of: