Name: fh87463 Date: 05/23/2002



Following assertion occur when running Compile The Wrold tests on solsparc platform.

#
# HotSpot Virtual Machine Error, assertion failure
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# Java VM: Java HotSpot(TM) Server VM (optimized+1.4.1+baseline+compiler2+20020521-163236-debug mixed mode)
#
# assert(loop_count < K, "infinite loop in PhaseGVN::transform")
#
# Error ID: /net/jano/export/disk05/hotspot/auto/ws_jano/fpt/src/share/vm/opto/phaseX.cpp, 941 [ Patched ]
#
# Problematic Thread: prio=5 tid=0x11b268 nid=0xb runnable
#
This assertion occured in following log files
========================================================
file:/net/sqesvr/export/vsn/users/ctwresult/solsparc/baseline/compiler2/20020521-163236/Failed/ctw.jlapack.jar.1.txt

Binary used:
  Fastdebug binary of [baseline/compiler2/20020521-163236] from build archive page.
  /net/jano.sfbay/export/disk05/hotspot/auto/ar/baseline/20020521-163236/

Please look into rerun scripts for more detail.

#########################
# To reproduce the bug:
#########################
1. cd /net/jano.sfbay/export/disk20/GammaBase/Bugs/[bug ID]
2. edit script if needed.
3. run one of the shell scripts:
   run.jlapack.jar.1.ksh



From: Mike Paleczny <###@###.###>
To: Francis Hsu <###@###.###>
Subject: Re: BugId 4681949 : (P2/S3) Has been Updated - by jmasa

Good morning <grin>,
    I've confirmed that this is a different bug than 4681949 and have a
prototype source patch which I am testing further. While duplicating
the problem I found that it doesn't reproduce with a 1.4 J2SE even with
the current 1.4.1 JVM. While the bug is in the compiler, it appears to
be made visible by a 1.4.1 change in the libraries.

Curiosity:
    I didn't find this jar file in the list for 4681949. Is this a new
failure, or where there so many jar file failures for 4681949 that the
bug didn't include every jar filename?

    Mike.


Mike Paleczny wrote:
> Thanks Francis,
> This assertion is an indicator for "any" infinite Ideal() call failure.
> After checking the output in ctw.jlapack.jar.1.txt, I believe this should
> be filed as a different bug:
>
> Synopsis: CTW assertion at phaseX.cpp, 941, MergeMem
>
> I will confirm that this is a different bug in a couple hours, but the
> previous bug came from optimizing CMoveINode which does not appear in
> the new printout.
>
> Mike.



======================================================================