possible deadlock in shmem_mfill_atomic

possible deadlock in shmem_mfill_atomic_pte
Status: auto-closed as invalid on 2020/08/21 02:08
Reported-by: syzbot+230bd8477494b8d5d390@syzkaller.appspotmail.com
First crash: 551d, last: 551d
similar bugs (2):
Kernel	Title	Repro	Cause bisect	Fix bisect	Count	Last	Reported	Patched	Status
linux-4.19	possible deadlock in shmem_mfill_atomic_pte				1	586d	586d	0/1	auto-closed as invalid on 2020/07/16 10:10
upstream	possible deadlock in shmem_mfill_atomic_pte	C	done		82	521d	573d	17/22	fixed on 2020/06/18 13:57
Sample crash report:
========================================================
WARNING: possible irq lock inversion dependency detected
4.14.176-syzkaller #0 Not tainted
--------------------------------------------------------
syz-executor.3/7701 just changed the state of lock:
 (&(&info->lock)->rlock){+...}, at: [<ffffffff81745afa>] spin_lock include/linux/spinlock.h:317 [inline]
 (&(&info->lock)->rlock){+...}, at: [<ffffffff81745afa>] shmem_mfill_atomic_pte+0x11ba/0x1790 mm/shmem.c:2333
but this lock was taken by another, HARDIRQ-safe lock in the past:
 (&(&mapping->tree_lock)->rlock){-.-.}


and interrupts could create inverse lock ordering between them.


other info that might help us debug this:
 Possible interrupt unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&(&info->lock)->rlock);
                               local_irq_disable();
                               lock(&(&mapping->tree_lock)->rlock);
                               lock(&(&info->lock)->rlock);
  <Interrupt>
    lock(&(&mapping->tree_lock)->rlock);

 *** DEADLOCK ***

2 locks held by syz-executor.3/7701:
 #0:  (&mm->mmap_sem){++++}, at: [<ffffffff818a0646>] __mcopy_atomic mm/userfaultfd.c:464 [inline]
 #0:  (&mm->mmap_sem){++++}, at: [<ffffffff818a0646>] mfill_zeropage+0x86/0xcd0 mm/userfaultfd.c:604
 #1:  (&(ptlock_ptr(page))->rlock#2){+.+.}, at: [<ffffffff8174551b>] spin_lock include/linux/spinlock.h:317 [inline]
 #1:  (&(ptlock_ptr(page))->rlock#2){+.+.}, at: [<ffffffff8174551b>] shmem_mfill_atomic_pte+0xbdb/0x1790 mm/shmem.c:2320

the shortest dependencies between 2nd lock and 1st lock:
 -> (&(&mapping->tree_lock)->rlock){-.-.} ops: 1348523 {
    IN-HARDIRQ-W at:
                      lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
                      __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
                      _raw_spin_lock_irqsave+0x8c/0xbf kernel/locking/spinlock.c:160
                      test_clear_page_writeback+0x199/0xc20 mm/page-writeback.c:2742
                      end_page_writeback+0xe4/0x210 mm/filemap.c:1153
                      ext4_finish_bio+0x483/0x710 fs/ext4/page-io.c:118
                      ext4_end_bio+0x14b/0x5a0 fs/ext4/page-io.c:342
                      bio_endio+0x315/0x670 block/bio.c:1918
                      req_bio_endio block/blk-core.c:204 [inline]
                      blk_update_request+0x27a/0x9d0 block/blk-core.c:2763
                      scsi_end_request+0x7d/0x9b0 drivers/scsi/scsi_lib.c:659
                      scsi_io_completion+0xa63/0x1590 drivers/scsi/scsi_lib.c:885
                      scsi_softirq_done+0x2c8/0x360 drivers/scsi/scsi_lib.c:1671
                      __blk_mq_complete_request+0x30b/0x6c0 block/blk-mq.c:571
                      blk_mq_complete_request+0x51/0x60 block/blk-mq.c:591
                      virtscsi_vq_done+0xa7/0x130 drivers/scsi/virtio_scsi.c:223
                      vring_interrupt+0x101/0x150 drivers/virtio/virtio_ring.c:951
                      __handle_irq_event_percpu+0x125/0x7e0 kernel/irq/handle.c:147
                      handle_irq_event_percpu+0x66/0x120 kernel/irq/handle.c:187
                      handle_irq_event+0xa2/0x12d kernel/irq/handle.c:204
                      handle_edge_irq+0x215/0x810 kernel/irq/chip.c:770
                      generic_handle_irq_desc include/linux/irqdesc.h:159 [inline]
                      handle_irq+0x35/0x50 arch/x86/kernel/irq_64.c:87
                      do_IRQ+0x93/0x1d0 arch/x86/kernel/irq.c:230
                      ret_from_intr+0x0/0x1e
                      arch_local_irq_restore arch/x86/include/asm/paravirt.h:779 [inline]
                      lock_acquire+0x1ec/0x3f0 kernel/locking/lockdep.c:3997
                      __raw_spin_lock include/linux/spinlock_api_smp.h:142 [inline]
                      _raw_spin_lock+0x2a/0x40 kernel/locking/spinlock.c:152
                      spin_lock include/linux/spinlock.h:317 [inline]
                      list_lru_del+0x80/0x4c0 mm/list_lru.c:131
                      d_lru_del+0xa9/0x100 fs/dcache.c:418
                      __dentry_kill+0x420/0x550 fs/dcache.c:583
                      dentry_kill fs/dcache.c:632 [inline]
                      dput.part.0+0x4be/0x710 fs/dcache.c:847
                      dput+0x1b/0x30 fs/dcache.c:811
                      SYSC_renameat2 fs/namei.c:4648 [inline]
                      SyS_renameat2 fs/namei.c:4533 [inline]
                      SYSC_rename fs/namei.c:4684 [inline]
                      SyS_rename+0x61d/0x6d0 fs/namei.c:4682
                      do_syscall_64+0x1d5/0x640 arch/x86/entry/common.c:292
                      entry_SYSCALL_64_after_hwframe+0x42/0xb7
    IN-SOFTIRQ-W at:
                      lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
                      __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
                      _raw_spin_lock_irqsave+0x8c/0xbf kernel/locking/spinlock.c:160
                      test_clear_page_writeback+0x199/0xc20 mm/page-writeback.c:2742
                      end_page_writeback+0xe4/0x210 mm/filemap.c:1153
                      end_buffer_async_write+0x367/0x6a0 fs/buffer.c:377
                      end_bio_bh_io_sync+0xbf/0x100 fs/buffer.c:3038
                      bio_endio+0x315/0x670 block/bio.c:1918
                      req_bio_endio block/blk-core.c:204 [inline]
                      blk_update_request+0x27a/0x9d0 block/blk-core.c:2763
                      scsi_end_request+0x7d/0x9b0 drivers/scsi/scsi_lib.c:659
                      scsi_io_completion+0xa63/0x1590 drivers/scsi/scsi_lib.c:885
                      scsi_softirq_done+0x2c8/0x360 drivers/scsi/scsi_lib.c:1671
                      __blk_mq_complete_request+0x30b/0x6c0 block/blk-mq.c:571
                      blk_mq_complete_request+0x51/0x60 block/blk-mq.c:591
                      virtscsi_vq_done+0xa7/0x130 drivers/scsi/virtio_scsi.c:223
                      vring_interrupt+0x101/0x150 drivers/virtio/virtio_ring.c:951
                      __handle_irq_event_percpu+0x125/0x7e0 kernel/irq/handle.c:147
                      handle_irq_event_percpu+0x66/0x120 kernel/irq/handle.c:187
                      handle_irq_event+0xa2/0x12d kernel/irq/handle.c:204
                      handle_edge_irq+0x215/0x810 kernel/irq/chip.c:770
                      generic_handle_irq_desc include/linux/irqdesc.h:159 [inline]
                      handle_irq+0x35/0x50 arch/x86/kernel/irq_64.c:87
                      do_IRQ+0x93/0x1d0 arch/x86/kernel/irq.c:230
                      ret_from_intr+0x0/0x1e
                      arch_local_irq_restore arch/x86/include/asm/paravirt.h:779 [inline]
                      kmem_cache_free+0xa9/0x2b0 mm/slab.c:3759
                      __rcu_reclaim kernel/rcu/rcu.h:195 [inline]
                      rcu_do_batch kernel/rcu/tree.c:2699 [inline]
                      invoke_rcu_callbacks kernel/rcu/tree.c:2962 [inline]
                      __rcu_process_callbacks kernel/rcu/tree.c:2929 [inline]
                      rcu_process_callbacks+0x792/0x1190 kernel/rcu/tree.c:2946
                      __do_softirq+0x254/0x9bf kernel/softirq.c:288
                      invoke_softirq kernel/softirq.c:368 [inline]
                      irq_exit+0x15b/0x1a0 kernel/softirq.c:409
                      exiting_irq arch/x86/include/asm/apic.h:648 [inline]
                      smp_apic_timer_interrupt+0x141/0x5e0 arch/x86/kernel/apic/apic.c:1102
                      apic_timer_interrupt+0x8f/0xa0 arch/x86/entry/entry_64.S:792
                      stack_access_ok+0x0/0x130 arch/x86/kernel/unwind_orc.c:149
                      deref_stack_reg+0x56/0xc0 arch/x86/kernel/unwind_orc.c:289
                      unwind_next_frame+0xa06/0x17a0 arch/x86/kernel/unwind_orc.c:425
                      __save_stack_trace+0x6b/0xd0 arch/x86/kernel/stacktrace.c:44
                      save_stack+0x32/0xa0 mm/kasan/kasan.c:447
                      set_track mm/kasan/kasan.c:459 [inline]
                      kasan_kmalloc mm/kasan/kasan.c:551 [inline]
                      kasan_kmalloc+0xbf/0xe0 mm/kasan/kasan.c:529
                      slab_post_alloc_hook mm/slab.h:442 [inline]
                      slab_alloc mm/slab.c:3390 [inline]
                      kmem_cache_alloc+0x114/0x770 mm/slab.c:3550
                      alloc_inode+0xa0/0x170 fs/inode.c:212
                      iget_locked+0x151/0x400 fs/inode.c:1127
                      kernfs_get_inode+0x4c/0x550 fs/kernfs/inode.c:269
                      kernfs_iop_lookup+0x17e/0x200 fs/kernfs/dir.c:1071
                      lookup_open+0x5d1/0x1750 fs/namei.c:3220
                      do_last fs/namei.c:3334 [inline]
                      path_openat+0x1878/0x3c50 fs/namei.c:3569
                      do_filp_open+0x18e/0x250 fs/namei.c:3603
                      do_sys_open+0x29d/0x3f0 fs/open.c:1081
                      do_syscall_64+0x1d5/0x640 arch/x86/entry/common.c:292
                      entry_SYSCALL_64_after_hwframe+0x42/0xb7
    INITIAL USE at:
                     lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
                     __raw_spin_lock_irq include/linux/spinlock_api_smp.h:128 [inline]
                     _raw_spin_lock_irq+0x5b/0x80 kernel/locking/spinlock.c:168
                     spin_lock_irq include/linux/spinlock.h:342 [inline]
                     __add_to_page_cache_locked+0x286/0x870 mm/filemap.c:781
                     add_to_page_cache_lru+0xe9/0x300 mm/filemap.c:830
                     do_read_cache_page+0x6d5/0x12a0 mm/filemap.c:2703
                     read_mapping_page include/linux/pagemap.h:398 [inline]
                     read_dev_sector+0xbf/0x3d0 block/partition-generic.c:659
                     read_part_sector block/partitions/check.h:38 [inline]
                     adfspart_check_ICS+0xf3/0xc10 block/partitions/acorn.c:366
                     check_partition+0x33f/0x62d block/partitions/check.c:167
                     rescan_partitions+0x1e5/0x850 block/partition-generic.c:523
                     __blkdev_get+0x926/0x10d0 fs/block_dev.c:1503
                     blkdev_get+0x84/0x8a0 fs/block_dev.c:1612
                     register_disk block/genhd.c:616 [inline]
                     device_add_disk+0x897/0xd80 block/genhd.c:679
                     add_disk include/linux/genhd.h:396 [inline]
                     brd_init+0x214/0x3c9 drivers/block/brd.c:586
                     do_one_initcall+0x88/0x202 init/main.c:824
                     do_initcall_level init/main.c:890 [inline]
                     do_initcalls init/main.c:898 [inline]
                     do_basic_setup init/main.c:916 [inline]
                     kernel_init_freeable+0x465/0x526 init/main.c:1073
                     kernel_init+0xd/0x15b init/main.c:998
                     ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:404
  }
  ... key      at: [<ffffffff8a5a8020>] __key.49022+0x0/0x40
  ... acquired at:
   __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
   _raw_spin_lock_irqsave+0x8c/0xbf kernel/locking/spinlock.c:160
   shmem_uncharge+0x24/0x250 mm/shmem.c:318
   __split_huge_page mm/huge_memory.c:2387 [inline]
   split_huge_page_to_list+0x1590/0x1f40 mm/huge_memory.c:2649
   split_huge_page include/linux/huge_mm.h:141 [inline]
   shmem_unused_huge_shrink+0x5ca/0x1180 mm/shmem.c:526
   shmem_unused_huge_scan+0x70/0xa0 mm/shmem.c:558
   super_cache_scan+0x333/0x460 fs/super.c:107
   do_shrink_slab mm/vmscan.c:401 [inline]
   shrink_slab.part.0+0x3bb/0xa50 mm/vmscan.c:504
   shrink_slab mm/vmscan.c:468 [inline]
   shrink_node+0x848/0xb60 mm/vmscan.c:2672
   kswapd_shrink_node mm/vmscan.c:3299 [inline]
   balance_pgdat mm/vmscan.c:3405 [inline]
   kswapd+0x88d/0x14b0 mm/vmscan.c:3623
   kthread+0x30d/0x420 kernel/kthread.c:232
   ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:404

-> (&(&info->lock)->rlock){+...} ops: 110062 {
   HARDIRQ-ON-W at:
                    lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
                    __raw_spin_lock include/linux/spinlock_api_smp.h:142 [inline]
                    _raw_spin_lock+0x2a/0x40 kernel/locking/spinlock.c:152
                    spin_lock include/linux/spinlock.h:317 [inline]
                    shmem_mfill_atomic_pte+0x11ba/0x1790 mm/shmem.c:2333
                    shmem_mfill_zeropage_pte+0x81/0xa0 mm/shmem.c:2382
                    mfill_atomic_pte mm/userfaultfd.c:429 [inline]
                    __mcopy_atomic mm/userfaultfd.c:549 [inline]
                    mfill_zeropage+0x773/0xcd0 mm/userfaultfd.c:604
                    userfaultfd_zeropage fs/userfaultfd.c:1760 [inline]
                    userfaultfd_ioctl+0x10da/0x2f80 fs/userfaultfd.c:1861
                    vfs_ioctl fs/ioctl.c:46 [inline]
                    file_ioctl fs/ioctl.c:500 [inline]
                    do_vfs_ioctl+0x75a/0xfe0 fs/ioctl.c:684
                    SYSC_ioctl fs/ioctl.c:701 [inline]
                    SyS_ioctl+0x7f/0xb0 fs/ioctl.c:692
                    do_syscall_64+0x1d5/0x640 arch/x86/entry/common.c:292
                    entry_SYSCALL_64_after_hwframe+0x42/0xb7
   INITIAL USE at:
                   lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
                   __raw_spin_lock_irq include/linux/spinlock_api_smp.h:128 [inline]
                   _raw_spin_lock_irq+0x5b/0x80 kernel/locking/spinlock.c:168
                   spin_lock_irq include/linux/spinlock.h:342 [inline]
                   shmem_getpage_gfp+0x17c6/0x2790 mm/shmem.c:1829
                   shmem_getpage mm/shmem.c:131 [inline]
                   shmem_write_begin+0xf4/0x190 mm/shmem.c:2413
                   generic_perform_write+0x1c9/0x420 mm/filemap.c:3047
                   __generic_file_write_iter+0x227/0x590 mm/filemap.c:3172
                   generic_file_write_iter+0x2fa/0x650 mm/filemap.c:3200
                   call_write_iter include/linux/fs.h:1778 [inline]
                   new_sync_write fs/read_write.c:469 [inline]
                   __vfs_write+0x44e/0x630 fs/read_write.c:482
                   vfs_write+0x192/0x4e0 fs/read_write.c:544
                   SYSC_write fs/read_write.c:590 [inline]
                   SyS_write+0xf2/0x210 fs/read_write.c:582
                   do_syscall_64+0x1d5/0x640 arch/x86/entry/common.c:292
                   entry_SYSCALL_64_after_hwframe+0x42/0xb7
 }
 ... key      at: [<ffffffff8a5968c0>] __key.49220+0x0/0x40
 ... acquired at:
   mark_lock_irq kernel/locking/lockdep.c:2800 [inline]
   mark_lock+0x3c8/0x10b0 kernel/locking/lockdep.c:3190
   mark_irqflags kernel/locking/lockdep.c:3082 [inline]
   __lock_acquire+0xbcf/0x4620 kernel/locking/lockdep.c:3444
   lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
   __raw_spin_lock include/linux/spinlock_api_smp.h:142 [inline]
   _raw_spin_lock+0x2a/0x40 kernel/locking/spinlock.c:152
   spin_lock include/linux/spinlock.h:317 [inline]
   shmem_mfill_atomic_pte+0x11ba/0x1790 mm/shmem.c:2333
   shmem_mfill_zeropage_pte+0x81/0xa0 mm/shmem.c:2382
   mfill_atomic_pte mm/userfaultfd.c:429 [inline]
   __mcopy_atomic mm/userfaultfd.c:549 [inline]
   mfill_zeropage+0x773/0xcd0 mm/userfaultfd.c:604
   userfaultfd_zeropage fs/userfaultfd.c:1760 [inline]
   userfaultfd_ioctl+0x10da/0x2f80 fs/userfaultfd.c:1861
   vfs_ioctl fs/ioctl.c:46 [inline]
   file_ioctl fs/ioctl.c:500 [inline]
   do_vfs_ioctl+0x75a/0xfe0 fs/ioctl.c:684
   SYSC_ioctl fs/ioctl.c:701 [inline]
   SyS_ioctl+0x7f/0xb0 fs/ioctl.c:692
   do_syscall_64+0x1d5/0x640 arch/x86/entry/common.c:292
   entry_SYSCALL_64_after_hwframe+0x42/0xb7


stack backtrace:
CPU: 0 PID: 7701 Comm: syz-executor.3 Not tainted 4.14.176-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:17 [inline]
 dump_stack+0x13e/0x194 lib/dump_stack.c:58
 print_irq_inversion_bug kernel/locking/lockdep.c:152 [inline]
 check_usage_backwards.cold+0x41/0x4a kernel/locking/lockdep.c:2715
 mark_lock_irq kernel/locking/lockdep.c:2800 [inline]
 mark_lock+0x3c8/0x10b0 kernel/locking/lockdep.c:3190
 mark_irqflags kernel/locking/lockdep.c:3082 [inline]
 __lock_acquire+0xbcf/0x4620 kernel/locking/lockdep.c:3444
 lock_acquire+0x170/0x3f0 kernel/locking/lockdep.c:3994
 __raw_spin_lock include/linux/spinlock_api_smp.h:142 [inline]
 _raw_spin_lock+0x2a/0x40 kernel/locking/spinlock.c:152
 spin_lock include/linux/spinlock.h:317 [inline]
 shmem_mfill_atomic_pte+0x11ba/0x1790 mm/shmem.c:2333
 shmem_mfill_zeropage_pte+0x81/0xa0 mm/shmem.c:2382
 mfill_atomic_pte mm/userfaultfd.c:429 [inline]
 __mcopy_atomic mm/userfaultfd.c:549 [inline]
 mfill_zeropage+0x773/0xcd0 mm/userfaultfd.c:604
 userfaultfd_zeropage fs/userfaultfd.c:1760 [inline]
 userfaultfd_ioctl+0x10da/0x2f80 fs/userfaultfd.c:1861
 vfs_ioctl fs/ioctl.c:46 [inline]
 file_ioctl fs/ioctl.c:500 [inline]
 do_vfs_ioctl+0x75a/0xfe0 fs/ioctl.c:684
 SYSC_ioctl fs/ioctl.c:701 [inline]
 SyS_ioctl+0x7f/0xb0 fs/ioctl.c:692
 do_syscall_64+0x1d5/0x640 arch/x86/entry/common.c:292
 entry_SYSCALL_64_after_hwframe+0x42/0xb7
RIP: 0033:0x45c829
RSP: 002b:00007f3f18b8ec78 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 00000000004ef8c0 RCX: 000000000045c829
RDX: 0000000020000080 RSI: 00000000c020aa04 RDI: 0000000000000005
RBP: 000000000078bf00 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00000000ffffffff
R13: 00000000000005b8 R14: 00000000004c8567 R15: 00007f3f18b8f6d4
netlink: 4 bytes leftover after parsing attributes in process `syz-executor.4'.
netlink: 4 bytes leftover after parsing attributes in process `syz-executor.4'.
netlink: 4 bytes leftover after parsing attributes in process `syz-executor.4'.
netlink: 256 bytes leftover after parsing attributes in process `syz-executor.3'.
netlink: 25 bytes leftover after parsing attributes in process `syz-executor.3'.
IPv6: ADDRCONF(NETDEV_UP): veth3: link is not ready
team0: Port device veth3 added
netlink: 256 bytes leftover after parsing attributes in process `syz-executor.3'.
netlink: 25 bytes leftover after parsing attributes in process `syz-executor.3'.
IPv6: ADDRCONF(NETDEV_UP): veth5: link is not ready
team0: Port device veth5 added
netlink: 4 bytes leftover after parsing attributes in process `syz-executor.4'.
netlink: 256 bytes leftover after parsing attributes in process `syz-executor.3'.
netlink: 25 bytes leftover after parsing attributes in process `syz-executor.3'.
IPv6: ADDRCONF(NETDEV_UP): veth7: link is not ready
team0: Port device veth7 added
netlink: 4 bytes leftover after parsing attributes in process `syz-executor.4'.
netlink: 256 bytes leftover after parsing attributes in process `syz-executor.3'.
IPv6: ADDRCONF(NETDEV_UP): veth9: link is not ready
team0: Port device veth9 added
IPv6: ADDRCONF(NETDEV_UP): veth11: link is not ready
team0: Port device veth11 added
IPv6: ADDRCONF(NETDEV_UP): veth13: link is not ready
team0: Port device veth13 added
IPv6: ADDRCONF(NETDEV_UP): veth5: link is not ready
team0: Port device veth5 added
Crashes (1):
Manager	Time	Kernel	Commit	Syzkaller	Config	Log	Report	Syz repro	C repro	VM info	Title
ci2-linux-4-14	2020/04/23 02:07	linux-4.14.y	c10b57a567e4	2e44d63e	.config	log	report