Was using a temporary BTRFS volume to compile mongodb, which is quite intensive and takes quite a bit of time. The volume has been deadlocked for about 12 hours.
Being a temporary volume, I just used mount without options, so it used the defaults: rw,relatime,ssd,space_cache,subvolid=5,subvol=/ Apologies if upgrading to 5.3.5+ will fix this. I didn't see discussions of a deadlock looking like this. See the full sysrq-trigger output here: http://ix.io/1Ysp But, for searchability, here is some of it: systemd-journal D 0 813 1 0x00004320 Call Trace: ? __schedule+0x27f/0x6d0 schedule+0x43/0xd0 schedule_timeout+0x299/0x3d0 io_schedule_timeout+0x19/0x40 wait_for_common_io.constprop.0+0xcf/0x150 ? wake_up_q+0x60/0x60 write_all_supers+0x87f/0x940 [btrfs] btrfs_sync_log+0x71b/0x9f0 [btrfs] ? dput+0xc9/0x2d0 btrfs_sync_file+0x364/0x460 [btrfs] do_fsync+0x38/0x70 __x64_sys_fsync+0x10/0x20 do_syscall_64+0x5f/0x1c0 entry_SYSCALL_64_after_hwframe+0x44/0xa9 Call Trace: ? __schedule+0x27f/0x6d0 schedule+0x43/0xd0 schedule_timeout+0x1cf/0x3d0 ? collect_expired_timers+0xb0/0xb0 flush_space+0x55c/0x710 [btrfs] btrfs_async_reclaim_metadata_space+0xc4/0x4a0 [btrfs] ? __schedule+0x287/0x6d0 process_one_work+0x1d1/0x3a0 worker_thread+0x4a/0x3d0 kthread+0xfb/0x130 ? process_one_work+0x3a0/0x3a0 ? kthread_park+0x80/0x80 ret_from_fork+0x35/0x40 Then 28 of these: Call Trace: ? __schedule+0x27f/0x6d0 schedule+0x43/0xd0 wait_reserve_ticket+0x95/0x150 [btrfs] ? wait_woken+0x70/0x70 btrfs_reserve_metadata_bytes+0x782/0x920 [btrfs] btrfs_block_rsv_add+0x1f/0x50 [btrfs] start_transaction+0x2c2/0x490 [btrfs] btrfs_dirty_inode+0x9d/0xd0 [btrfs] file_update_time+0xfd/0x150 btrfs_page_mkwrite+0xfe/0x4e0 [btrfs] ? lock_page_memcg+0x11/0x90 do_page_mkwrite+0x31/0x90 do_wp_page+0x2c6/0x660 __handle_mm_fault+0xc24/0x15d0 ? call_function_single_interrupt+0xa/0x20 handle_mm_fault+0xce/0x1f0 __do_page_fault+0x216/0x4f0 do_page_fault+0x31/0x130 page_fault+0x3e/0x50