harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vladimir Ivanov" <ivavladi...@gmail.com>
Subject Re: [drlvm][em64t] build is broken
Date Thu, 07 Dec 2006 04:11:28 GMT
Today the em64t build is able to print version but failed to run something
else:

 thanks, Vladimir


viv> /usr/jdk1.5.0_06/jdk/bin/java  -cp /nfs/site/proj/drl/qe/my/tmp09 test


date = Wed Dec 06 20:07:34 PST 2006, ofset = 480

out = localhost/127.0.0.1

hashCode = 2130706433

viv>
./trunk/cc/projects/drlvm/trunk/build/lnx_em64t_gcc_debug/deploy/jre/bin/java
-version

Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software
Foundation or its licensors, as applicable.

java version "1.5.0"

pre-alpha : not complete or compatible

svn = r483265, (Dec  7 2006), Linux/em64t/gcc 3.3.3, debug build

http://incubator.apache.org/harmony

viv>
./trunk/cc/projects/drlvm/trunk/build/lnx_em64t_gcc_debug/deploy/jre/bin
/java  -cp /nfs/site/proj/drl/qe/my/tmp09 test

Segmentation fault

viv>
./trunk/cc/projects/drlvm/trunk/build/lnx_em64t_gcc_debug/deploy/jre/bin/java
-Xint -cp /nfs/site/proj/drl/qe/my/tmp09 test


Segmentation fault

viv>
./trunk/cc/projects/drlvm/trunk/build/lnx_em64t_gcc_debug/deploy/jre/bin/java
-Xem:opt -cp /nfs/site/proj/drl/qe/my/tmp09 test


Segmentation fault

viv>
./trunk/cc/projects/drlvm/trunk/build/lnx_em64t_gcc_debug/deploy/jre/bin/java
-Xem:jet -cp /nfs/site/proj/drl/qe/my/tmp09 test


Segmentation fault

viv>



On 12/6/06, Geir Magnusson Jr. <geir@pobox.com> wrote:
>
> Paging Dr. Mark... Paging Dr. Mark...
>
>
>
> Alexey Varlamov wrote:
> > Guys,
> >
> > em64t builds of classlib are dead since r482540. That is, DRLVM works
> > with classlib r482539 and fails with segfault on later revisions.
> > The rootcause is hysl_lookup_name returning bad pointers due to
> > incorrect size of UDATA, which become == 4 somehow.
> > I spent several hours trying to understand the reason of the wrong
> > definition, but still miss something. Correct DEFINES seems to be in
> > place...
> > Almighty classlib build gurus, we need to resolve this ASAP, please.
> >
> > NB. Isn't using void* more portable than UDATA for pointers? E.g. for
> > below declaration:
> > hysl_lookup_name (struct HyPortLibrary * portLibrary, UDATA descriptor,
> >                  char *name, UDATA * func, const char *argSignature)
> >
> >
> > 2006/12/5, Vladimir Ivanov <ivavladimir@gmail.com>:
> >> Today the run of classlib tests over the DRLVM leads to the crash
> >> (below).
> >> Could somebody reproduce/fix it?
> >>
> >>  thanks, Vladimir
> >>
> >> vivanov1@nstdrlel1:/export/users/viv/trunk/cc>
> >>
> /export/users/viv/trunk/cc/projects/drlvm/trunk/build/lnx_em64t_gcc_debug/deploy/jre/bin/java
> >>
> >> -version
> >> Apache Harmony Launcher : (c) Copyright 1991, 2006 The Apache Software
> >> Foundation or its licensors, as applicable.
> >>
> >> An unhandled error (4) has occurred.
> >> HyGeneric_Signal_Number=00000004
> >> Signal_Number=0000000b
> >> Error_Value=00000000
> >> Signal_Code=00000001
> >> Handler1=00402D48
> >> Handler2=9568D018
> >> InaccessibleAddress=96B138D0
> >> RDI=BFFFD7E0
> >> RSI=BFFFD7E8
> >> RAX=00000000
> >> RBX=BFFFDB30
> >> RCX=BFFFDB30
> >> RDX=BFFFD800
> >> R8=0053A804
> >> R9=0053A940
> >> R10=732F0000
> >> R11=00000030
> >> R12=BFFFDB30
> >> R13=00010004
> >> R14=00000002
> >> R15=BFFFE378
> >> RIP=96B138D0
> >> GS=00000000
> >> FS=00000000
> >> RSP=BFFFD798
> >> EFlags=00010246
> >> CS=00000020
> >> RBP=00000001
> >> ERR=00000014
> >> TRAPNO=0000000E
> >> OLDMASK=00000000
> >> CR2=96B138D0
> >> Aborted
> >> vivanov1@nstdrlel1:/export/users/viv/trunk/cc>
> >>
> >>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message