Re: btrfs: lock inversion between delayed_node->mutex and found->groups_sem

From: David Sterba
Date: Mon Mar 17 2014 - 09:05:59 EST


On Fri, Mar 14, 2014 at 08:12:16PM -0400, Sasha Levin wrote:
> While fuzzing with trinity inside a KVM tools guest running the latest -next
> kernel I've stumbled on the following:
>
> [ 788.458756] CPU0 CPU1
> [ 788.459188] ---- ----
> [ 788.459625] lock(&found->groups_sem);
> [ 788.460041] local_irq_disable();
> [ 788.460041] lock(&delayed_node->mutex);
> [ 788.460041] lock(&found->groups_sem);
> [ 788.460041] <Interrupt>
> [ 788.460041] lock(&delayed_node->mutex);
> [ 788.460041]
> [ 788.460041] *** DEADLOCK ***
> [ 788.460041]
> [ 788.460041] 2 locks held by kswapd3/4199:

I've once (3.14-rc5) seen the same warning also caused by xfstests/generic/224

2 locks held by 224/31203:
#0: (shrinker_rwsem){++++..}, at: [<ffffffff8113be6d>] shrink_slab+0x3d/0x110
#1: (&type->s_umount_key#32){++++++}, at: [<ffffffff8117cd84>] grab_super_passive+0x44/0x90

the shortest dependencies between 2nd lock and 1st lock:
-> (&found->groups_sem){+++++.} ops: 405561 {
HARDIRQ-ON-W at:
[<ffffffff810af476>] __lock_acquire+0x7f6/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a01d3c>] down_write+0x5c/0xc0
[<ffffffffa001f6a6>] __link_block_group+0x46/0x130 [btrfs]
[<ffffffffa0023411>] btrfs_read_block_groups+0x341/0x690 [btrfs]
[<ffffffffa0031c50>] open_ctree+0x1880/0x2310 [btrfs]
[<ffffffffa00065db>] btrfs_mount+0x55b/0x860 [btrfs]
[<ffffffff8117d6f0>] mount_fs+0x20/0xe0
[<ffffffff81199ab6>] vfs_kern_mount+0x76/0x160
[<ffffffff8119c25d>] do_mount+0x31d/0x970
[<ffffffff8119cc30>] SyS_mount+0x90/0xe0
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
HARDIRQ-ON-R at:
[<ffffffff810af265>] __lock_acquire+0x5e5/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a01c8c>] down_read+0x4c/0xa0
[<ffffffffa002da32>] btrfs_calc_num_tolerated_disk_barrier_failures+0x142/0x240 [btrfs]
[<ffffffffa0031c6e>] open_ctree+0x189e/0x2310 [btrfs]
[<ffffffffa00065db>] btrfs_mount+0x55b/0x860 [btrfs]
[<ffffffff8117d6f0>] mount_fs+0x20/0xe0
[<ffffffff81199ab6>] vfs_kern_mount+0x76/0x160
[<ffffffff8119c25d>] do_mount+0x31d/0x970
[<ffffffff8119cc30>] SyS_mount+0x90/0xe0
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
SOFTIRQ-ON-W at:
[<ffffffff810af4aa>] __lock_acquire+0x82a/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a01d3c>] down_write+0x5c/0xc0
[<ffffffffa001f6a6>] __link_block_group+0x46/0x130 [btrfs]
[<ffffffffa0023411>] btrfs_read_block_groups+0x341/0x690 [btrfs]
[<ffffffffa0031c50>] open_ctree+0x1880/0x2310 [btrfs]
[<ffffffffa00065db>] btrfs_mount+0x55b/0x860 [btrfs]
[<ffffffff8117d6f0>] mount_fs+0x20/0xe0
[<ffffffff81199ab6>] vfs_kern_mount+0x76/0x160
[<ffffffff8119c25d>] do_mount+0x31d/0x970
[<ffffffff8119cc30>] SyS_mount+0x90/0xe0
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
SOFTIRQ-ON-R at:
[<ffffffff810af4aa>] __lock_acquire+0x82a/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a01c8c>] down_read+0x4c/0xa0
[<ffffffffa002da32>] btrfs_calc_num_tolerated_disk_barrier_failures+0x142/0x240 [btrfs]
[<ffffffffa0031c6e>] open_ctree+0x189e/0x2310 [btrfs]
[<ffffffffa00065db>] btrfs_mount+0x55b/0x860 [btrfs]
[<ffffffff8117d6f0>] mount_fs+0x20/0xe0
[<ffffffff81199ab6>] vfs_kern_mount+0x76/0x160
[<ffffffff8119c25d>] do_mount+0x31d/0x970
[<ffffffff8119cc30>] SyS_mount+0x90/0xe0
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
RECLAIM_FS-ON-W at:
[<ffffffff810b1c6c>] mark_held_locks+0x8c/0x170
[<ffffffff810b249a>] lockdep_trace_alloc+0x8a/0xd0
[<ffffffff8116f807>] __kmalloc_track_caller+0x47/0x210
[<ffffffff813cdefb>] kvasprintf+0x5b/0x90
[<ffffffff813c166a>] kobject_set_name_vargs+0x2a/0x70
[<ffffffff813c1ffa>] kobject_add+0x5a/0xb0
[<ffffffffa001f75d>] __link_block_group+0xfd/0x130 [btrfs]
[<ffffffffa0023411>] btrfs_read_block_groups+0x341/0x690 [btrfs]
[<ffffffffa0031c50>] open_ctree+0x1880/0x2310 [btrfs]
[<ffffffffa00065db>] btrfs_mount+0x55b/0x860 [btrfs]
[<ffffffff8117d6f0>] mount_fs+0x20/0xe0
[<ffffffff81199ab6>] vfs_kern_mount+0x76/0x160
[<ffffffff8119c25d>] do_mount+0x31d/0x970
[<ffffffff8119cc30>] SyS_mount+0x90/0xe0
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
INITIAL USE at:
[<ffffffff810aefd4>] __lock_acquire+0x354/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a01d3c>] down_write+0x5c/0xc0
[<ffffffffa001f6a6>] __link_block_group+0x46/0x130 [btrfs]
[<ffffffffa0023411>] btrfs_read_block_groups+0x341/0x690 [btrfs]
[<ffffffffa0031c50>] open_ctree+0x1880/0x2310 [btrfs]
[<ffffffffa00065db>] btrfs_mount+0x55b/0x860 [btrfs]
[<ffffffff8117d6f0>] mount_fs+0x20/0xe0
[<ffffffff81199ab6>] vfs_kern_mount+0x76/0x160
[<ffffffff8119c25d>] do_mount+0x31d/0x970
[<ffffffff8119cc30>] SyS_mount+0x90/0xe0
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
}
... key at: [<ffffffffa00d06e0>] __key.40139+0x0/0xfffffffffffe2920 [btrfs]
... acquired at:
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a01c8c>] down_read+0x4c/0xa0
[<ffffffffa001ff78>] find_free_extent+0x7e8/0xbe0 [btrfs]
[<ffffffffa0020434>] btrfs_reserve_extent+0xa4/0x130 [btrfs]
[<ffffffffa0021903>] btrfs_alloc_free_block+0x103/0x4c0 [btrfs]
[<ffffffffa000beb4>] __btrfs_cow_block+0x124/0x550 [btrfs]
[<ffffffffa000c4ab>] btrfs_cow_block+0x12b/0x1e0 [btrfs]
[<ffffffffa0010429>] btrfs_search_slot+0x1d9/0xa00 [btrfs]
[<ffffffffa00123de>] btrfs_insert_empty_items+0x7e/0xe0 [btrfs]
[<ffffffffa008a7a4>] btrfs_insert_delayed_items+0x84/0x460 [btrfs]
[<ffffffffa008b299>] __btrfs_run_delayed_items+0xd9/0x1f0 [btrfs]
[<ffffffffa008b7a3>] btrfs_run_delayed_items+0x13/0x20 [btrfs]
[<ffffffffa0034e6e>] btrfs_commit_transaction+0x3ae/0xa30 [btrfs]
[<ffffffffa006ae72>] btrfs_mksubvol+0x3a2/0x3b0 [btrfs]
[<ffffffffa006b036>] btrfs_ioctl_snap_create_transid+0x1b6/0x1c0 [btrfs]
[<ffffffffa006b1be>] btrfs_ioctl_snap_create_v2+0xfe/0x140 [btrfs]
[<ffffffffa006d1d2>] btrfs_ioctl+0x652/0x1940 [btrfs]
[<ffffffff8118c0a1>] do_vfs_ioctl+0x91/0x560
[<ffffffff8118c5c7>] SyS_ioctl+0x57/0x90
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b

-> (&delayed_node->mutex){+.+.-.} ops: 8191147 {
HARDIRQ-ON-W at:
[<ffffffff810af476>] __lock_acquire+0x7f6/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a00c4e>] mutex_lock_nested+0x6e/0x3a0
[<ffffffffa008c5c9>] btrfs_delayed_update_inode+0x49/0x670 [btrfs]
[<ffffffffa003b76a>] btrfs_update_inode+0x6a/0x100 [btrfs]
[<ffffffffa0043e8e>] btrfs_create+0x16e/0x220 [btrfs]
[<ffffffff81188979>] vfs_create+0x89/0xc0
[<ffffffff81189144>] do_last+0x794/0xd50
[<ffffffff811897c7>] path_openat+0xc7/0x620
[<ffffffff8118a1da>] do_filp_open+0x4a/0xa0
[<ffffffff8117883e>] do_sys_open+0x11e/0x230
[<ffffffff8117896e>] SyS_open+0x1e/0x20
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
SOFTIRQ-ON-W at:
[<ffffffff810af4aa>] __lock_acquire+0x82a/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a00c4e>] mutex_lock_nested+0x6e/0x3a0
[<ffffffffa008c5c9>] btrfs_delayed_update_inode+0x49/0x670 [btrfs]
[<ffffffffa003b76a>] btrfs_update_inode+0x6a/0x100 [btrfs]
[<ffffffffa0043e8e>] btrfs_create+0x16e/0x220 [btrfs]
[<ffffffff81188979>] vfs_create+0x89/0xc0
[<ffffffff81189144>] do_last+0x794/0xd50
[<ffffffff811897c7>] path_openat+0xc7/0x620
[<ffffffff8118a1da>] do_filp_open+0x4a/0xa0
[<ffffffff8117883e>] do_sys_open+0x11e/0x230
[<ffffffff8117896e>] SyS_open+0x1e/0x20
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
IN-RECLAIM_FS-W at:
[<ffffffff810af2ec>] __lock_acquire+0x66c/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a00c4e>] mutex_lock_nested+0x6e/0x3a0
[<ffffffffa0089dff>] __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffffa008bac4>] btrfs_remove_delayed_node+0x24/0x30 [btrfs]
[<ffffffffa0040c97>] btrfs_evict_inode+0x2a7/0x550 [btrfs]
[<ffffffff81195188>] evict+0xb8/0x1c0
[<ffffffff811952df>] dispose_list+0x4f/0x60
[<ffffffff811963dc>] prune_icache_sb+0x4c/0x60
[<ffffffff8117d006>] super_cache_scan+0x126/0x190
[<ffffffff8113a1fe>] shrink_slab_node+0x14e/0x2c0
[<ffffffff8113bea8>] shrink_slab+0x78/0x110
[<ffffffff8113f1bd>] do_try_to_free_pages+0x24d/0x410
[<ffffffff8113f554>] try_to_free_pages+0xe4/0x190
[<ffffffff8113235b>] __alloc_pages_nodemask+0x68b/0xa60
[<ffffffff8116edac>] cache_alloc_refill+0x40c/0x7b0
[<ffffffff8116e98f>] kmem_cache_alloc+0x1ef/0x200
[<ffffffff8105efae>] copy_process+0x13e/0x1910
[<ffffffff810608a5>] do_fork+0x65/0x300
[<ffffffff81060bc6>] SyS_clone+0x16/0x20
[<ffffffff81a0cdb9>] stub_clone+0x69/0x90
INITIAL USE at:
[<ffffffff810aefd4>] __lock_acquire+0x354/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a00c4e>] mutex_lock_nested+0x6e/0x3a0
[<ffffffffa008c5c9>] btrfs_delayed_update_inode+0x49/0x670 [btrfs]
[<ffffffffa003b76a>] btrfs_update_inode+0x6a/0x100 [btrfs]
[<ffffffffa0043e8e>] btrfs_create+0x16e/0x220 [btrfs]
[<ffffffff81188979>] vfs_create+0x89/0xc0
[<ffffffff81189144>] do_last+0x794/0xd50
[<ffffffff811897c7>] path_openat+0xc7/0x620
[<ffffffff8118a1da>] do_filp_open+0x4a/0xa0
[<ffffffff8117883e>] do_sys_open+0x11e/0x230
[<ffffffff8117896e>] SyS_open+0x1e/0x20
[<ffffffff81a0ca92>] system_call_fastpath+0x16/0x1b
}
... key at: [<ffffffffa00d4648>] __key.35416+0x0/0xfffffffffffde9b8 [btrfs]
... acquired at:
[<ffffffff810add8a>] check_usage_forwards+0xaa/0x120
[<ffffffff810ae9f9>] mark_lock+0x1a9/0x430
[<ffffffff810af2ec>] __lock_acquire+0x66c/0x1fb0
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffff81a00c4e>] mutex_lock_nested+0x6e/0x3a0
[<ffffffffa0089dff>] __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffffa008bac4>] btrfs_remove_delayed_node+0x24/0x30 [btrfs]
[<ffffffffa0040c97>] btrfs_evict_inode+0x2a7/0x550 [btrfs]
[<ffffffff81195188>] evict+0xb8/0x1c0
[<ffffffff811952df>] dispose_list+0x4f/0x60
[<ffffffff811963dc>] prune_icache_sb+0x4c/0x60
[<ffffffff8117d006>] super_cache_scan+0x126/0x190
[<ffffffff8113a1fe>] shrink_slab_node+0x14e/0x2c0
[<ffffffff8113bea8>] shrink_slab+0x78/0x110
[<ffffffff8113f1bd>] do_try_to_free_pages+0x24d/0x410
[<ffffffff8113f554>] try_to_free_pages+0xe4/0x190
[<ffffffff8113235b>] __alloc_pages_nodemask+0x68b/0xa60
[<ffffffff8116edac>] cache_alloc_refill+0x40c/0x7b0
[<ffffffff8116e98f>] kmem_cache_alloc+0x1ef/0x200
[<ffffffff8105efae>] copy_process+0x13e/0x1910
[<ffffffff810608a5>] do_fork+0x65/0x300
[<ffffffff81060bc6>] SyS_clone+0x16/0x20
[<ffffffff81a0cdb9>] stub_clone+0x69/0x90


stack backtrace:
CPU: 1 PID: 31203 Comm: 224 Tainted: G W 3.14.0-rc5-default #121
Hardware name: Intel Corporation Santa Rosa platform/Matanzas, BIOS TSRSCRB1.86C.0047.B00.0610170821 10/17/06
ffffffff82c37220 ffff88003faa9460 ffffffff819fd463 0000000000000002
ffffffff82c37220 ffff88003faa94b0 ffffffff810adc97 ffff88003faa9500
ffffffff81f16ba4 ffff88003faa94cc ffff88003faa5730 ffff88003faa94c0
Call Trace:
[<ffffffff819fd463>] dump_stack+0x51/0x6e
[<ffffffff810adc97>] print_irq_inversion_bug+0x1c7/0x210
[<ffffffff810add8a>] check_usage_forwards+0xaa/0x120
[<ffffffff810adce0>] ? print_irq_inversion_bug+0x210/0x210
[<ffffffff810ae9f9>] mark_lock+0x1a9/0x430
[<ffffffff810af2ec>] __lock_acquire+0x66c/0x1fb0
[<ffffffff810adce0>] ? print_irq_inversion_bug+0x210/0x210
[<ffffffff810ae4cf>] ? check_irq_usage+0x9f/0xf0
[<ffffffff810b002b>] ? __lock_acquire+0x13ab/0x1fb0
[<ffffffff810ab11d>] ? trace_hardirqs_off+0xd/0x10
[<ffffffffa0089dff>] ? __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffff810b12e2>] lock_acquire+0x92/0x120
[<ffffffffa0089dff>] ? __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffff81a00c4e>] mutex_lock_nested+0x6e/0x3a0
[<ffffffffa0089dff>] ? __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffff8109a138>] ? sched_clock_cpu+0xa8/0xd0
[<ffffffffa0089dff>] ? __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffff810ab62d>] ? lock_release_holdtime+0x3d/0x1c0
[<ffffffffa0089dff>] __btrfs_release_delayed_node+0x4f/0x220 [btrfs]
[<ffffffffa008bac4>] btrfs_remove_delayed_node+0x24/0x30 [btrfs]
[<ffffffffa0040c97>] btrfs_evict_inode+0x2a7/0x550 [btrfs]
[<ffffffff81a03c6b>] ? _raw_spin_unlock+0x2b/0x40
[<ffffffff81195188>] evict+0xb8/0x1c0
[<ffffffff81195ed0>] ? insert_inode_locked+0x1a0/0x1a0
[<ffffffff811952df>] dispose_list+0x4f/0x60
[<ffffffff811963dc>] prune_icache_sb+0x4c/0x60
[<ffffffff8117d006>] super_cache_scan+0x126/0x190
[<ffffffff8113a1fe>] shrink_slab_node+0x14e/0x2c0
[<ffffffff8109a176>] ? local_clock+0x16/0x30
[<ffffffff8113bea8>] shrink_slab+0x78/0x110
[<ffffffff8113f1bd>] do_try_to_free_pages+0x24d/0x410
[<ffffffff8113f554>] try_to_free_pages+0xe4/0x190
[<ffffffff8113235b>] __alloc_pages_nodemask+0x68b/0xa60
[<ffffffff8116ed6e>] ? cache_alloc_refill+0x3ce/0x7b0
[<ffffffff8116edac>] cache_alloc_refill+0x40c/0x7b0
[<ffffffff8105efae>] ? copy_process+0x13e/0x1910
[<ffffffff8116e98f>] kmem_cache_alloc+0x1ef/0x200
[<ffffffff8105efae>] copy_process+0x13e/0x1910
[<ffffffff8109a138>] ? sched_clock_cpu+0xa8/0xd0
[<ffffffff8109a176>] ? local_clock+0x16/0x30
[<ffffffff810ab62d>] ? lock_release_holdtime+0x3d/0x1c0
[<ffffffff810b0fc5>] ? lock_release_non_nested+0x395/0x3e0
[<ffffffff8114fd86>] ? might_fault+0x66/0xc0
[<ffffffff810608a5>] do_fork+0x65/0x300
[<ffffffff8114fd86>] ? might_fault+0x66/0xc0
[<ffffffff81a0cab7>] ? sysret_check+0x1b/0x56
[<ffffffff81060bc6>] SyS_clone+0x16/0x20
[<ffffffff81a0cdb9>] stub_clone+0x69/0x90
[<ffffffff81a0ca92>] ? system_call_fastpath+0x16/0x1b
--
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/