EVALUATION

See main CR
                                     

EVALUATION

This scenario should be included in CLI testing.
Verified in HS22 7u2 8 for JDK 7u2 b10 &  HS23 b02 for JDK8 b10.

Here is log:

gtee@sc14160074 7092238 $ /java/re/jdk/7u2/promoted/all/b09/binaries/solaris-sparc/fastdebug/bin/java  -XX:+UseG1GC -XX:+PrintGCDetails -XX:NewRatio=1
VM option '+UseG1GC'
VM option '+PrintGCDetails'
VM option 'NewRatio=1'
# To suppress the following error report, specify this argument
# after -XX: or in .hotspotrc:  SuppressErrorAt=/g1CollectorPolicy.cpp:516
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (/export/HUDSON/workspace/jdk7u2-2-build-solaris-sparc-product/jdk7u2/hotspot/src/share/vm/gc_implementation/g1/g1CollectorPolicy.cpp:516), pid=27715, tid=2
#  assert(initial_region_num <= _max_desired_young_length) failed: Initial young gen size too large
#
# JRE version: 7.0_02-b09
# Java VM: Java HotSpot(TM) Server VM (22.0-b07-fastdebug mixed mode solaris-sparc )
# Core dump written. Default location: /import/gtee/g1/7092238/core or core.27715
#
# An error report file with more information is saved as:
# /import/gtee/g1/7092238/hs_err_pid27715.log
Phoning home... 
Using server: 10.161.186.18, port 4711
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.sun.com/bugreport/crash.jsp
#
Current thread is 2
Dumping core ...
^CAbort
gtee@sc14160074 7092238 $ ^C
gtee@sc14160074 7092238 $ ^C
gtee@sc14160074 7092238 $ /net/sqenfs-1.us.oracle.com/export1/comp/vm/jdk/hsx/pit/8/fastdebug/solaris-sparc/bin/java -XX:+UseG1GC -XX:+PrintGCDetails -XX:NewRatio=1       
VM option '+UseG1GC'
VM option '+PrintGCDetails'
VM option 'NewRatio=1'
Usage: java [-options] class [args...]
.....

net/sqenfs-1.us.oracle.com/export1/comp/vm/jdk/hsx/pit/7u2/fastdebug/solaris-sparc/bin/java -XX:+UseG1GC -XX:+PrintGCDetails -XX:NewRatio=1
VM option '+UseG1GC'
VM option '+PrintGCDetails'
VM option 'NewRatio=1'
...
                                     

EVALUATION

http://hg.openjdk.java.net/hsx/hsx22/hotspot/rev/ac196b091535
                                     

EVALUATION

http://hg.openjdk.java.net/hsx/hotspot-gc/hotspot/rev/d912b598c6c3
                                     

SUGGESTED FIX

See Evaluation.
                                     

EVALUATION

The failure is caused by the following

void G1CollectorPolicy::init() {
  ...
  size_t heap_regions = sizer.size_to_region_num(_g1->n_regions());
  update_young_list_size_using_newratio(heap_regions);

size_to_region_num() takes a byte size and translates it to a region num, so passing a region number to it is incorrect and causes the calculations to be wrong (the min and max desired young length end up being 0).

The fix is to directly pass _g1->n_regions() to update_young_list_size_using_newratio().