syzbot


possible deadlock in uprobe_clear_state (2)

Status: fixed on 2020/09/16 22:51
Subsystems: perf
[Documentation on labels]
Reported-by: syzbot+e5344baa319c9a96edec@syzkaller.appspotmail.com
Fix commit: f867c771f988 binder: Don't use mmput() from shrinker function.
First crash: 1610d, last: 1610d
Discussions (6)
Title Replies (including bot) Last reply
[PATCH 5.7 000/179] 5.7.11-rc1 review 187 (187) 2020/07/29 08:22
[PATCH 4.19 00/86] 4.19.135-rc1 review 109 (109) 2020/07/28 21:18
[PATCH 5.4 000/138] 5.4.54-rc1 review 143 (143) 2020/07/28 18:23
[PATCH 4.14 00/64] 4.14.190-rc1 review 68 (68) 2020/07/28 18:22
[PATCH] binder: Don't use mmput() from shrinker function. 9 (9) 2020/07/16 23:53
possible deadlock in uprobe_clear_state (2) 0 (1) 2020/07/15 19:59
Similar bugs (1)
Kernel Title Repro Cause bisect Fix bisect Count Last Reported Patched Status
upstream possible deadlock in uprobe_clear_state perf 2 2150d 2187d 0/28 auto-closed as invalid on 2019/07/18 01:36

Sample crash report:
======================================================
WARNING: possible circular locking dependency detected
5.8.0-rc4-syzkaller #0 Not tainted
------------------------------------------------------
syz-executor.1/4807 is trying to acquire lock:
ffffffff89c3a988 (delayed_uprobe_lock){+.+.}-{3:3}, at: uprobe_clear_state+0x47/0x3b0 kernel/events/uprobes.c:1550

but task is already holding lock:
ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: fs_reclaim_release mm/page_alloc.c:4202 [inline]
ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: fs_reclaim_release mm/page_alloc.c:4198 [inline]
ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: __perform_reclaim mm/page_alloc.c:4227 [inline]
ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: __alloc_pages_direct_reclaim mm/page_alloc.c:4244 [inline]
ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: __alloc_pages_slowpath.constprop.0+0x1554/0x2780 mm/page_alloc.c:4650

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (fs_reclaim){+.+.}-{0:0}:
       __fs_reclaim_acquire mm/page_alloc.c:4183 [inline]
       fs_reclaim_acquire+0x2f/0x40 mm/page_alloc.c:4194
       slab_pre_alloc_hook mm/slab.h:564 [inline]
       slab_alloc mm/slab.c:3306 [inline]
       kmem_cache_alloc_trace+0x29/0x2d0 mm/slab.c:3549
       kmalloc include/linux/slab.h:555 [inline]
       kzalloc include/linux/slab.h:669 [inline]
       delayed_uprobe_add kernel/events/uprobes.c:304 [inline]
       update_ref_ctr+0x4b6/0x700 kernel/events/uprobes.c:438
       uprobe_write_opcode+0xe07/0x1650 kernel/events/uprobes.c:497
       install_breakpoint kernel/events/uprobes.c:915 [inline]
       install_breakpoint.isra.0+0x5a5/0x7c0 kernel/events/uprobes.c:897
       uprobe_mmap+0x5d7/0x1050 kernel/events/uprobes.c:1394
       mmap_region+0x5cf/0x1590 mm/mmap.c:1818
       do_mmap+0xca8/0x1170 mm/mmap.c:1545
       do_mmap_pgoff include/linux/mm.h:2596 [inline]
       vm_mmap_pgoff+0x197/0x200 mm/util.c:506
       ksys_mmap_pgoff+0x455/0x5a0 mm/mmap.c:1595
       do_syscall_64+0x60/0xe0 arch/x86/entry/common.c:384
       entry_SYSCALL_64_after_hwframe+0x44/0xa9

-> #0 (delayed_uprobe_lock){+.+.}-{3:3}:
       check_prev_add kernel/locking/lockdep.c:2496 [inline]
       check_prevs_add kernel/locking/lockdep.c:2601 [inline]
       validate_chain kernel/locking/lockdep.c:3218 [inline]
       __lock_acquire+0x2acb/0x56e0 kernel/locking/lockdep.c:4380
       lock_acquire+0x1f1/0xad0 kernel/locking/lockdep.c:4959
       __mutex_lock_common kernel/locking/mutex.c:956 [inline]
       __mutex_lock+0x134/0x10d0 kernel/locking/mutex.c:1103
       uprobe_clear_state+0x47/0x3b0 kernel/events/uprobes.c:1550
       __mmput+0x73/0x470 kernel/fork.c:1089
       mmput+0x53/0x60 kernel/fork.c:1114
       binder_alloc_free_page+0x441/0xf90 drivers/android/binder_alloc.c:950
       __list_lru_walk_one+0x178/0x5c0 mm/list_lru.c:222
       list_lru_walk_one mm/list_lru.c:266 [inline]
       list_lru_walk_node+0x67/0x2a0 mm/list_lru.c:295
       list_lru_walk include/linux/list_lru.h:215 [inline]
       binder_shrink_scan+0x123/0x190 drivers/android/binder_alloc.c:984
       do_shrink_slab+0x3c6/0xab0 mm/vmscan.c:518
       shrink_slab+0x16f/0x5c0 mm/vmscan.c:679
       shrink_node_memcgs mm/vmscan.c:2658 [inline]
       shrink_node+0x519/0x1b60 mm/vmscan.c:2770
       shrink_zones mm/vmscan.c:2973 [inline]
       do_try_to_free_pages+0x38b/0x1340 mm/vmscan.c:3026
       try_to_free_pages+0x29a/0x8b0 mm/vmscan.c:3265
       __perform_reclaim mm/page_alloc.c:4223 [inline]
       __alloc_pages_direct_reclaim mm/page_alloc.c:4244 [inline]
       __alloc_pages_slowpath.constprop.0+0x949/0x2780 mm/page_alloc.c:4650
       __alloc_pages_nodemask+0x68f/0x930 mm/page_alloc.c:4863
       alloc_pages_current+0x187/0x280 mm/mempolicy.c:2292
       alloc_pages include/linux/gfp.h:545 [inline]
       alloc_mmu_pages+0x7f/0x170 arch/x86/kvm/mmu/mmu.c:5671
       kvm_mmu_create+0x3cb/0x560 arch/x86/kvm/mmu/mmu.c:5704
       kvm_arch_vcpu_create+0x16d/0xb70 arch/x86/kvm/x86.c:9433
       kvm_vm_ioctl_create_vcpu arch/x86/kvm/../../../virt/kvm/kvm_main.c:3060 [inline]
       kvm_vm_ioctl+0x1547/0x23c0 arch/x86/kvm/../../../virt/kvm/kvm_main.c:3621
       vfs_ioctl fs/ioctl.c:48 [inline]
       ksys_ioctl+0x11a/0x180 fs/ioctl.c:753
       __do_sys_ioctl fs/ioctl.c:762 [inline]
       __se_sys_ioctl fs/ioctl.c:760 [inline]
       __x64_sys_ioctl+0x6f/0xb0 fs/ioctl.c:760
       do_syscall_64+0x60/0xe0 arch/x86/entry/common.c:384
       entry_SYSCALL_64_after_hwframe+0x44/0xa9

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(fs_reclaim);
                               lock(delayed_uprobe_lock);
                               lock(fs_reclaim);
  lock(delayed_uprobe_lock);

 *** DEADLOCK ***

3 locks held by syz-executor.1/4807:
 #0: ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: fs_reclaim_release mm/page_alloc.c:4202 [inline]
 #0: ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: fs_reclaim_release mm/page_alloc.c:4198 [inline]
 #0: ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: __perform_reclaim mm/page_alloc.c:4227 [inline]
 #0: ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: __alloc_pages_direct_reclaim mm/page_alloc.c:4244 [inline]
 #0: ffffffff89c6c360 (fs_reclaim){+.+.}-{0:0}, at: __alloc_pages_slowpath.constprop.0+0x1554/0x2780 mm/page_alloc.c:4650
 #1: ffffffff89c46b90 (shrinker_rwsem){++++}-{3:3}, at: shrink_slab+0xc7/0x5c0 mm/vmscan.c:669
 #2: ffff88808ea481e0 (&alloc->mutex){+.+.}-{3:3}, at: binder_alloc_free_page+0x4f/0xf90 drivers/android/binder_alloc.c:923

stack backtrace:
CPU: 0 PID: 4807 Comm: syz-executor.1 Not tainted 5.8.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x18f/0x20d lib/dump_stack.c:118
 check_noncircular+0x324/0x3e0 kernel/locking/lockdep.c:1827
 check_prev_add kernel/locking/lockdep.c:2496 [inline]
 check_prevs_add kernel/locking/lockdep.c:2601 [inline]
 validate_chain kernel/locking/lockdep.c:3218 [inline]
 __lock_acquire+0x2acb/0x56e0 kernel/locking/lockdep.c:4380
 lock_acquire+0x1f1/0xad0 kernel/locking/lockdep.c:4959
 __mutex_lock_common kernel/locking/mutex.c:956 [inline]
 __mutex_lock+0x134/0x10d0 kernel/locking/mutex.c:1103
 uprobe_clear_state+0x47/0x3b0 kernel/events/uprobes.c:1550
 __mmput+0x73/0x470 kernel/fork.c:1089
 mmput+0x53/0x60 kernel/fork.c:1114
 binder_alloc_free_page+0x441/0xf90 drivers/android/binder_alloc.c:950
 __list_lru_walk_one+0x178/0x5c0 mm/list_lru.c:222
 list_lru_walk_one mm/list_lru.c:266 [inline]
 list_lru_walk_node+0x67/0x2a0 mm/list_lru.c:295
 list_lru_walk include/linux/list_lru.h:215 [inline]
 binder_shrink_scan+0x123/0x190 drivers/android/binder_alloc.c:984
 do_shrink_slab+0x3c6/0xab0 mm/vmscan.c:518
 shrink_slab+0x16f/0x5c0 mm/vmscan.c:679
 shrink_node_memcgs mm/vmscan.c:2658 [inline]
 shrink_node+0x519/0x1b60 mm/vmscan.c:2770
 shrink_zones mm/vmscan.c:2973 [inline]
 do_try_to_free_pages+0x38b/0x1340 mm/vmscan.c:3026
 try_to_free_pages+0x29a/0x8b0 mm/vmscan.c:3265
 __perform_reclaim mm/page_alloc.c:4223 [inline]
 __alloc_pages_direct_reclaim mm/page_alloc.c:4244 [inline]
 __alloc_pages_slowpath.constprop.0+0x949/0x2780 mm/page_alloc.c:4650
 __alloc_pages_nodemask+0x68f/0x930 mm/page_alloc.c:4863
 alloc_pages_current+0x187/0x280 mm/mempolicy.c:2292
 alloc_pages include/linux/gfp.h:545 [inline]
 alloc_mmu_pages+0x7f/0x170 arch/x86/kvm/mmu/mmu.c:5671
 kvm_mmu_create+0x3cb/0x560 arch/x86/kvm/mmu/mmu.c:5704
 kvm_arch_vcpu_create+0x16d/0xb70 arch/x86/kvm/x86.c:9433
 kvm_vm_ioctl_create_vcpu arch/x86/kvm/../../../virt/kvm/kvm_main.c:3060 [inline]
 kvm_vm_ioctl+0x1547/0x23c0 arch/x86/kvm/../../../virt/kvm/kvm_main.c:3621
 vfs_ioctl fs/ioctl.c:48 [inline]
 ksys_ioctl+0x11a/0x180 fs/ioctl.c:753
 __do_sys_ioctl fs/ioctl.c:762 [inline]
 __se_sys_ioctl fs/ioctl.c:760 [inline]
 __x64_sys_ioctl+0x6f/0xb0 fs/ioctl.c:760
 do_syscall_64+0x60/0xe0 arch/x86/entry/common.c:384
 entry_SYSCALL_64_after_hwframe+0x44/0xa9
RIP: 0033:0x45cba9
Code: Bad RIP value.
RSP: 002b:00007f45444f8c78 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 00000000004e85c0 RCX: 000000000045cba9
RDX: 0000000000000000 RSI: 000000000000ae41 RDI: 0000000000000004
RBP: 000000000078bf00 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00000000ffffffff
R13: 00000000000003a3 R14: 00000000004c652f R15: 00007f45444f96d4

Crashes (1):
Time Kernel Commit Syzkaller Config Log Report Syz repro C repro VM info Assets (help?) Manager Title
2020/07/11 19:56 upstream a581387e415b 18d18b59 .config console log report ci-upstream-kasan-gce-root
* Struck through repros no longer work on HEAD.