JDK-8024008 : Nashorn V8 (Crypto) benchmark crashes with small CodeCache size
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: hs25
  • Priority: P3
  • Status: Closed
  • Resolution: Duplicate
  • OS: linux
  • CPU: x86
  • Submitted: 2013-08-29
  • Updated: 2013-11-21
  • Resolved: 2013-10-23
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
Other
tbd_minorResolved
Related Reports
Relates :  
JDK-8023465 - JSR292: assert: i2c adapter must return to an interpreter frame
Relates :  
JDK-8015837 - Nashorn crashes with tiered on x86 when running v8 benchmark
Description
Here are the error logs. The crash appears quite reliably.  In 50 runs of the entire V8 suite, there is a high probability that the crash appears.

Here is the run script:

#!/bin/bash
JAVA_HOME=/java/re/jdk/8/promoted/latest/binaries/linux-amd64/fastdebug
NASHORN=${JAVA_HOME}/jre/lib/ext/nashorn.jar
CC=-XX:ReservedCodeCacheSize=24m
for i in {1..50}
do
  echo $i: $JAVA_HOME $CC >&2
  $JAVA_HOME/bin/java $CC -XX:+TieredCompilation -Dengine=nashorn -XX:+TieredCompilation -jar $NASHORN run.js   
done
Comments
If we generate code for all methods that "must be compiled" (comment argument in CompileBroker::compile_method(..)) the bug does not occur.
11-10-2013
With the change above, the following line reproduces the bug. java -XX:-TieredCompilation -XX:-ShowMessageBoxOnError -XX:-BackgroundCompilation -XX:CICompilerCount=1 -jar nashorn.jar run.js
11-10-2013
The bug can be easily triggered by commenting out the following line in CompileBroker::compile_method(...) // (void) AdapterHandlerLibrary::create_native_wrapper(method, compile_id);
11-10-2013
The bug does not seem to be related to sweeping, The bug occurs only if we disable compilation and a certain wrapper is not generated. I could not reproduce the bug when TieredCompilation is disabled.
10-10-2013
ILW=HLL=P4 8-defer-request: This happens only with a small code cache which is not the default mode of running Nashorn or the JVM. Work around is to increase the code cache. Defer to JDK8 update
09-10-2013
Running nashorn with the above configuration and -XX:+UnlockDiagnosticVMOptions -XX:+VerifyAdapterCalls results in the following error:" i2c adapter must return to an interpreter frame
01-10-2013
The sweeping process seems to be broken. If we comment the following line in process_compile(TRAPS) in ciReplay.cpp the crash goes away. // Make sure the existence of a prior compile doesn't stop this one nmethod* nm = (entry_bci != InvocationEntryBci) ? method->lookup_osr_nmethod_for(entry_bci, comp_level, true) : method->code(); if (nm != NULL) { //nm->make_not_entrant(); } Not yet quite sure why.
09-09-2013