Zero and IcedTea7

Andrew John Hughes gnu_andrew at member.fsf.org
Mon Sep 28 14:16:25 PDT 2009


I'm just about to push a changeset which updates IcedTea7 to the
latest build (b72) and Gary's new revision of the Zero patch.  Fans of
Nimbus will also be glad to know that this is the build that drops the
JIBX requirement (though through some autotools hackery we've been
able to get round that in IcedTea7 anyway).

I had to do some hacking on Zero to get it to work with the latest
HotSpot (hs17b01) including a fix to
http://hg.openjdk.java.net/jdk7/jdk7/hotspot/rev/8b46c4d82093 which
has now been committed to hotspot-gc and turning off compressed oops
for Zero (it's now on by default, as is escape analysis).  On finally
getting a successful build, I ran the JTreg tests for HotSpot.  This
gave eight failures and two errors.  A normal build only results in
two failures.

The Zero-specific failures are:

Unexpected exit from test [exit code: 134]
compiler/6711117/Test.java : Assertion in 64bit server vm (flat !=
TypePtr::BOTTOM,"cannot alias-analyze an untyped ptr")
compiler/6826736/Test.java : CMS: core dump with -XX:+UseCompressedOops
compiler/6849574/Test.java : VM crash using NonBlockingHashMap (high_scale_lib)
compiler/6851282/Test.java : JIT miscompilation results in null entry
in array when using CompressedOops
Unexpected exit from test [exit code: 1]
compiler/6795161/Test.java : Escape analysis leads to data corruption
compiler/6865031/Test.java : Application gives bad result (throws bad
exception) with compressed oops

and the two errors are:

Can't load test: java.lang.ClassNotFoundException: bigobj
gc/6845368/bigobj.java : ensure gc updates references > 64K bytes from
the start of the obj
Program `/mnt/builder/zero7/openjdk/build/linux-amd64/j2sdk-image/bin/java'
interrupted! (timed out?)
compiler/6866651/Test.java : delay dead node elimination in set_req_X
to prevent killing the current node when it is in use

6711117, 6826736, 6851282 are all crashes due to compressed oops being
turned on explicitly.
6849574 is an assertion failure:

#  Internal Error (referenceProcessor.cpp:1366), pid=27689, tid=47413998344464
#  Error: guarantee(sentinel_ref() != __null &&
sentinel_ref()->is_oop(),"Lost _sentinelRef")

6795161 fails due to Unrecognized VM option '+DoEscapeAnalysis'
6865031 fails due to Unrecognized VM option '-LoopUnswitching'

The two errors are due to a missing class and a timeout.

I have yet to try Shark with the new build drop.

Cheers,
-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8



More information about the distro-pkg-dev mailing list