Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-5219

Btree walk code read the lock WT_REF.addr field without locking

    • Type: Icon: Bug Bug
    • Resolution: Fixed
    • Priority: Icon: Major - P3 Major - P3
    • WT10.0.0, 4.3.4
    • Affects Version/s: None
    • Component/s: None
    • Labels:
    • 8
    • Storage Engines 2020-01-27
    • v4.2, v4.0

      Summary:
      The tree walk code has neither a hazard pointer or a lock to block eviction, change the walk code to lock the WT_REF structure before reading the WT_REF.addr field.


      There's a test failure captured by race condition test/format sanitizer job that the memory used after free during a btree walk with next_random configuration:

      http://build.wiredtiger.com:8080/job/wiredtiger-test-race-condition-stress-sanitizer/33913/

      ==63102==ERROR: AddressSanitizer: heap-use-after-free on address 0x60600066f2e8 at pc 0x000000a2007a bp 0x7fa5370efab0 sp 0x7fa5370efaa8
      READ of size 8 at 0x60600066f2e8 thread T79
          #0 0xa20079 in __wt_ref_info /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/include/btree.i:1073:24
          #1 0xa1fbe6 in __ref_is_leaf /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/btree/bt_walk.c:90:5
          #2 0xa1965a in __tree_walk_skip_count_callback /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/btree/bt_walk.c:607:35
          #3 0xa188e7 in __tree_walk_internal /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/btree/bt_walk.c:482:17
          #4 0xa1941a in __wt_tree_walk_skip /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/btree/bt_walk.c:631:9
          #5 0x970763 in __wt_btcur_next_random /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/btree/bt_random.c:580:9
          #6 0xaca00f in __wt_curfile_next_random /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/cursor/cur_file.c:118:5
          #7 0x52ce91 in random_kv /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/random.c:73:27
          #8 0x4dde52 in __asan::AsanThread::ThreadStart(unsigned long, __sanitizer::atomic_uintptr_t*) (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x4dde52)
          #9 0x7fa54b8e836c in start_thread (/lib64/libpthread.so.0+0x736c)
          #10 0x7fa54aabcb4e in __GI___clone (/lib64/libc.so.6+0x110b4e)
      
      0x60600066f2e8 is located 40 bytes inside of 56-byte region [0x60600066f2c0,0x60600066f2f8)
      freed by thread T71 here:
          #0 0x4d01e8 in __interceptor_free.localalias.0 (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x4d01e8)
          #1 0x6e2299 in __wt_free_int /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/os_common/os_alloc.c:301:5
          #2 0x757922 in __wt_ref_addr_free /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/include/btree.i:640:9
          #3 0x75557c in __wt_ref_block_free /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/include/btree.i:1116:5
          #4 0x7457e7 in __rec_write_wrapup /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/reconcile/rec_write.c:2194:9
          #5 0x735ff6 in __reconcile /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/reconcile/rec_write.c:209:28
          #6 0x734d2c in __wt_reconcile /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/reconcile/rec_write.c:102:11
          #7 0x655f40 in __evict_review /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_page.c:671:11
          #8 0x652249 in __wt_evict /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_page.c:149:5
          #9 0x63552e in __evict_page /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_lru.c:2214:5
          #10 0x63256f in __wt_cache_eviction_worker /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_lru.c:2304:23
          #11 0x84e5ea in __wt_cache_eviction_check /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/include/cache.i:428:13
          #12 0x84f848 in __wt_txn_rollback /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/txn/txn.c:1278:9
          #13 0x7bb28e in __session_rollback_transaction /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/session/session_api.c:1742:5
          #14 0x52b8c2 in rollback_transaction /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:452:5
          #15 0x51f27c in ops /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:978:13
          #16 0x4dde52 in __asan::AsanThread::ThreadStart(unsigned long, __sanitizer::atomic_uintptr_t*) (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x4dde52)
      
      previously allocated by thread T75 here:
          #0 0x4d05a8 in calloc (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x4d05a8)
          #1 0x6e0ae0 in __wt_calloc /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/os_common/os_alloc.c:50:14
          #2 0x657bae in __evict_page_dirty_update /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_page.c:393:13
          #3 0x652c3e in __wt_evict /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_page.c:192:9
          #4 0x63552e in __evict_page /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_lru.c:2214:5
          #5 0x63256f in __wt_cache_eviction_worker /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/evict/evict_lru.c:2304:23
          #6 0x84e5ea in __wt_cache_eviction_check /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/include/cache.i:428:13
          #7 0x84f848 in __wt_txn_rollback /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/txn/txn.c:1278:9
          #8 0x7bb28e in __session_rollback_transaction /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/session/session_api.c:1742:5
          #9 0x52b8c2 in rollback_transaction /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:452:5
          #10 0x51f27c in ops /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:978:13
          #11 0x4dde52 in __asan::AsanThread::ThreadStart(unsigned long, __sanitizer::atomic_uintptr_t*) (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x4dde52)
      
      Thread T79 created by T0 here:
          #0 0x433890 in pthread_create (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x433890)
          #1 0x6ff760 in __wt_thread_create /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/os_posix/os_thread.c:28:5
          #2 0x519ed3 in wts_ops /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:186:9
          #3 0x5355bf in main /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/t.c:212:17
          #4 0x7fa54a9cc889 in __libc_start_main (/lib64/libc.so.6+0x20889)
      
      Thread T71 created by T0 here:
          #0 0x433890 in pthread_create (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x433890)
          #1 0x6ff760 in __wt_thread_create /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/os_posix/os_thread.c:28:5
          #2 0x519896 in wts_ops /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:169:9
          #3 0x5355bf in main /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/t.c:212:17
          #4 0x7fa54a9cc889 in __libc_start_main (/lib64/libc.so.6+0x20889)
      
      Thread T75 created by T0 here:
          #0 0x433890 in pthread_create (/work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/t+0x433890)
          #1 0x6ff760 in __wt_thread_create /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/../src/os_posix/os_thread.c:28:5
          #2 0x519896 in wts_ops /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/ops.c:169:9
          #3 0x5355bf in main /work/jenkins/workspace/wiredtiger-test-race-condition-stress-sanitizer/build_posix/test/format/../../../test/format/t.c:212:17
          #4 0x7fa54a9cc889 in __libc_start_main (/lib64/libc.so.6+0x20889) 

      The configuration:

      ############################################
      #  RUN PARAMETERS
      ############################################
      abort=0
      alter=0
      assert_commit_timestamp=0
      assert_read_timestamp=0
      auto_throttle=1
      backups=0
      bitcnt=4
      bloom=1
      bloom_bit_count=59
      bloom_hash_count=11
      bloom_oldest=0
      cache=440
      cache_minimum=20
      checkpoints=on
      checkpoint_log_size=64
      checkpoint_wait=36
      checksum=uncompressed
      chunk_size=3
      compaction=0
      compression=snappy
      data_extend=0
      data_source=table
      delete_pct=8
      dictionary=1
      direct_io=0
      encryption=none
      evict_max=1
      file_type=row-store
      firstfit=0
      huffman_key=0
      huffman_value=0
      independent_thread_rng=1
      in_memory=0
      insert_pct=46
      internal_key_truncation=0
      internal_page_max=13
      isolation=snapshot
      key_gap=5
      key_max=113
      key_min=26
      leaf_page_max=13
      leak_memory=0
      logging=1
      logging_archive=1
      logging_compression=snappy
      logging_file_max=411791
      logging_prealloc=1
      long_running_txn=0
      lsm_worker_threads=4
      memory_page_max=10
      merge_max=6
      mmap=1
      modify_pct=42
      ops=0
      prefix_compression=0
      prefix_compression_min=4
      prepare=0
      quiet=1
      random_cursor=1
      read_pct=3
      rebalance=1
      repeat_data_pct=56
      reverse=0
      rows=1000000
      runs=1
      salvage=1
      split_pct=67
      statistics=0
      statistics_server=0
      threads=22
      timer=4
      timing_stress_aggressive_sweep=0
      timing_stress_checkpoint=0
      timing_stress_lookaside_sweep=0
      timing_stress_split_1=1
      timing_stress_split_2=0
      timing_stress_split_3=1
      timing_stress_split_4=0
      timing_stress_split_5=0
      timing_stress_split_6=1
      timing_stress_split_7=1
      timing_stress_split_8=0
      transaction_timestamps=1
      transaction-frequency=100
      truncate=1
      value_max=3782
      value_min=17
      verify=1
      wiredtiger_config=
      write_pct=1
      ############################################ 

            Assignee:
            keith.bostic@mongodb.com Keith Bostic (Inactive)
            Reporter:
            luke.chen@mongodb.com Luke Chen
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: