Uploaded image for project: 'JDK'
  1. JDK
  2. JDK-8273635

Attempting to acquire lock StackWatermark_lock/9 out of order with lock tty_lock/3

    XMLWordPrintable

Details

    • b15
    • aarch64
    • os_x

    Description

      compiler/vectorapi/TestVectorShiftImm.java fails with "assert(false) failed: Attempting to acquire lock StackWatermark_lock/9 out of order with lock tty_lock/3 -- possible deadlock detected.

      The following test failed in the JDK18 CI:

      compiler/vectorapi/TestVectorShiftImm.java

      The log file has a garbled portion of the crash mess:

      1034 B140: # out( N1 ) <- in( B23 ) Freq: 4.7682e-07
      1034 + spill R1 -> [sp, #40] # spill size = 32
      1038 + spill R3 -> [sp, #56] # spill size = 32
      103c + spill R11 -> [sp, #44] # spill size = 32
      1040 + spill R17 -> [sp, #32] # spill size = 64
      1044 + spill R19 -> [sp, #24] # spill size = 64
      1044 + mov R1, #-187 # int
      1048 spill R15 -> [sp, #8] # spill size = 64
      104c call,static 0x00000001100d5140 // ==> wrapper for: uncommon_trap(reason='unstable_if' action='reinterpret' debug_id='0')
              # jdk.incubator.vector.VectorOperators$OperatorImpl::opCode @ bci:26 (line 830) L[0]=sp + #24 L[1]=#2048 L[2]=#512#
      # Compiler replay data is saved as:
      # L[3]=sp + #44/System/Volumes/Data/mesos/work_dir/slaves/a2dc162d-743b-4800-9e92-31f85abb45b1-S16/frameworks/1735e8a2-a1db-478c-8104-60c8b0af87dd-0196/executors/95b84b56-03ec-44a6-8c93-1d134ddc92de/runs/c2fdfe43-31e6-4076-ab6c-94c8610f8622/testoutput/test-support/jtreg_open_test_hotspot_jtreg_hotspot_compiler_all_gcs/scratch/2/replay_pid63454.log
       STK[0]=R29 STK[1]=#512
              # jdk.incubator.vector.VectorOperators::opCode @ bci:6 (line 403) L[0]=_ L[1]=_ L[2]=_
              # jdk.incubator.vector.ByteVector::opCode @ bci:7 (line 63) L[0]=_
              # jdk.incubator.vector.ByteVector::lanewiseShiftTemplate @ bci:29 (line 806) L[0]=sp + #48 L[1]=sp + #16 L[2]=#1 L[3]=_
              ##
      # If you would like to submit a bug report, please visit:
      # https://bugreport.java.com/bugreport/crash.jsp
      #
       jdk.incubator.vector.Byte64Vector::
      ----------System.err:(1/55)----------
      WARNING: Using incubator modules: jdk.incubator.vector
      ----------rerun:(41/6424)*----------


      Here's snippets from the hs_err_pid file:

      # Internal Error (/System/Volumes/Data/mesos/work_dir/slaves/a2dc162d-743b-4800-9e92-31f85abb45b1-S22/frameworks/1735e8a2-a1db-478c-8104-60c8b0af87dd-0196/executors/c8ef3af4-816b-4614-a062-bed26f74e970/runs/c297bafa-efc6-40a8-971f-ba31f46b3e2d/workspace/open/src/hotspot/share/runtime/mutex.cpp:415), pid=63454, tid=25603
      # assert(false) failed: Attempting to acquire lock StackWatermark_lock/9 out of order with lock tty_lock/3 -- possible deadlock
      #
      # JRE version: Java(TM) SE Runtime Environment (18.0+15) (fastdebug build 18-ea+15-786)
      # Java VM: Java HotSpot(TM) 64-Bit Server VM (fastdebug 18-ea+15-786, mixed mode, sharing, compressed class ptrs, z gc, bsd-aarch64)
      # Core dump will be written. Default location: core.63454

      <snip>

      --------------- T H R E A D ---------------

      Current thread (0x000000011600a420): JavaThread "C2 CompilerThread0" daemon [_thread_in_vm, id=25603, stack(0x0000000172448000,0x000000017264b000)]


      Current CompileTask:
      C2: 1099 385 % compiler.vectorapi.TestVectorShiftImm::shift_and_accumulate @ 20 (144 bytes)

      Stack: [0x0000000172448000,0x000000017264b000], sp=0x0000000172646660, free space=2041k
      Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
      V [libjvm.dylib+0xfadac4] VMError::report_and_die(int, char const*, char const*, char*, Thread*, unsigned char*, void*, void*, char const*, int, unsigned long)+0x5d8
      V [libjvm.dylib+0xfae200] VMError::report_and_die(Thread*, void*, char const*, int, char const*, char const*, char*)+0x40
      V [libjvm.dylib+0x501198] report_vm_error(char const*, int, char const*, char const*, ...)+0x80
      V [libjvm.dylib+0xc7e84c] Mutex::check_rank(Thread*)+0x30c
      V [libjvm.dylib+0xc7ea14] Mutex::lock_without_safepoint_check(Thread*)+0x60
      V [libjvm.dylib+0xe22514] StackWatermark::on_safepoint()+0x3c
      V [libjvm.dylib+0xdd1544] SafepointMechanism::process(JavaThread*, bool)+0x90
      V [libjvm.dylib+0x41ee90] ThreadStateTransition::transition_from_native(JavaThread*, JavaThreadState, bool)+0x1c0
      V [libjvm.dylib+0x41ec8c] ThreadInVMfromNative::ThreadInVMfromNative(JavaThread*)+0xb4
      V [libjvm.dylib+0x436f48] ciSymbol::as_utf8()+0x40
      V [libjvm.dylib+0xf3cc40] TypeKlassPtr::dump2(Dict&, unsigned int, outputStream*) const+0x58
      V [libjvm.dylib+0xf2f4a4] Type::dump_on(outputStream*) const+0x9c
      V [libjvm.dylib+0x12e750] loadConNKlassNode::format(PhaseRegAlloc*, outputStream*) const+0xec
      V [libjvm.dylib+0xcfcd50] PhaseOutput::dump_asm_on(outputStream*, int*, unsigned int)+0x7dc
      V [libjvm.dylib+0xcf48ec] PhaseOutput::fill_buffer(CodeBuffer*, unsigned int*)+0x2088
      V [libjvm.dylib+0x4ad374] Compile::Code_Gen()+0x480
      V [libjvm.dylib+0x4aad4c] Compile::Compile(ciEnv*, ciMethod*, int, bool, bool, bool, bool, bool, DirectiveSet*)+0x1554
      V [libjvm.dylib+0x3a9c7c] C2Compiler::compile_method(ciEnv*, ciMethod*, int, bool, DirectiveSet*)+0x134
      V [libjvm.dylib+0x4c5900] CompileBroker::invoke_compiler_on_method(CompileTask*)+0x604
      V [libjvm.dylib+0x4c50fc] CompileBroker::compiler_thread_loop()+0x2c0
      V [libjvm.dylib+0xf0f218] JavaThread::thread_main_inner()+0x254
      V [libjvm.dylib+0xf0c4f0] Thread::call_run()+0x21c
      V [libjvm.dylib+0xce1640] thread_native_entry(Thread*)+0x160
      C [libsystem_pthread.dylib+0x7878] _pthread_start+0x140


      I'm starting this bug in hotspot/compiler since this is a compiler
      test and we're crashing in compiler code, but [~coleenp] has
      been recently working on the internal locks so I'm going to
      send this bug her way for an initial evaluation.

      Attachments

        Issue Links

          Activity

            People

              coleenp Coleen Phillimore
              dcubed Daniel Daugherty
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: