• Filipe Manana's avatar
    Btrfs: fix sleeping inside atomic context in qgroup rescan worker · 3b2ba7b3
    Filipe Manana authored
    We are holding a btree path with spinning locks and then we attempt to
    clone an extent buffer, which calls kmem_cache_alloc() and this function
    can sleep, causing the following trace to be reported on a debug kernel:
    
    [107118.218536] BUG: sleeping function called from invalid context at mm/slab.c:2871
    [107118.224110] in_atomic(): 1, irqs_disabled(): 0, pid: 19148, name: kworker/u32:3
    [107118.226120] INFO: lockdep is turned off.
    [107118.226843] Preemption disabled at:[<ffffffffa05ffa22>] btrfs_clear_lock_blocking_rw+0x96/0xea [btrfs]
    
    [107118.229175] CPU: 3 PID: 19148 Comm: kworker/u32:3 Tainted: G        W       4.3.0-rc5-btrfs-next-17+ #1
    [107118.231326] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.8.1-0-g4adadbd-20150316_085822-nilsson.home.kraxel.org 04/01/2014
    [107118.233687] Workqueue: btrfs-qgroup-rescan btrfs_qgroup_rescan_helper [btrfs]
    [107118.236835]  0000000000000000 ffff880424bf3b78 ffffffff812566f4 0000000000000000
    [107118.238369]  ffff880424bf3ba0 ffffffff81070664 ffffffff817f1cd5 0000000000000b37
    [107118.239769]  0000000000000000 ffff880424bf3bc8 ffffffff8107070a 0000000000008850
    [107118.241244] Call Trace:
    [107118.241729]  [<ffffffff812566f4>] dump_stack+0x4e/0x79
    [107118.242602]  [<ffffffff81070664>] ___might_sleep+0x23a/0x241
    [107118.243586]  [<ffffffff8107070a>] __might_sleep+0x9f/0xa6
    [107118.244532]  [<ffffffff8115af70>] cache_alloc_debugcheck_before+0x25/0x36
    [107118.245939]  [<ffffffff8115d52b>] kmem_cache_alloc+0x50/0x215
    [107118.246930]  [<ffffffffa05e627e>] __alloc_extent_buffer+0x2a/0x11f [btrfs]
    [107118.248121]  [<ffffffffa05ecb1a>] btrfs_clone_extent_buffer+0x3d/0xdd [btrfs]
    [107118.249451]  [<ffffffffa06239ea>] btrfs_qgroup_rescan_worker+0x16d/0x434 [btrfs]
    [107118.250755]  [<ffffffff81087481>] ? arch_local_irq_save+0x9/0xc
    [107118.251754]  [<ffffffffa05f7952>] normal_work_helper+0x14c/0x32a [btrfs]
    [107118.252899]  [<ffffffffa05f7952>] ? normal_work_helper+0x14c/0x32a [btrfs]
    [107118.254195]  [<ffffffffa05f7c82>] btrfs_qgroup_rescan_helper+0x12/0x14 [btrfs]
    [107118.255436]  [<ffffffff81063b23>] process_one_work+0x24a/0x4ac
    [107118.263690]  [<ffffffff81064285>] worker_thread+0x206/0x2c2
    [107118.264888]  [<ffffffff8106407f>] ? rescuer_thread+0x2cb/0x2cb
    [107118.267413]  [<ffffffff8106904d>] kthread+0xef/0xf7
    [107118.268417]  [<ffffffff81068f5e>] ? kthread_parkme+0x24/0x24
    [107118.269505]  [<ffffffff8147d10f>] ret_from_fork+0x3f/0x70
    [107118.270491]  [<ffffffff81068f5e>] ? kthread_parkme+0x24/0x24
    
    So just use blocking locks for our path to solve this.
    This fixes the patch titled:
      "btrfs: qgroup: Don't copy extent buffer to do qgroup rescan"
    Signed-off-by: default avatarFilipe Manana <fdmanana@suse.com>
    3b2ba7b3
qgroup.c 65.8 KB