- ppc-aix-port-dev - openjdk.org

hg: ppc-aix-port/jdk7u/hotspot: Opto: adapt assertion in postaloc.cpp for ppc.
by goetz.lindenmaier＠sap.com 30 Nov '12

30 Nov '12

Changeset: 938f9f955e21 Author: Goetz Date: 2012-11-30 07:46 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/938f9f955e21 Opto: adapt assertion in postaloc.cpp for ppc. assert(false, "unexpected yanked node") fired. ! src/share/vm/opto/postaloc.cpp

1 0

hg: ppc-aix-port/jdk7u/hotspot: C-interpreter: Fix for OSR.
by goetz.lindenmaier＠sap.com 29 Nov '12

29 Nov '12

Changeset: 0de4af96141e Author: Goetz Date: 2012-11-29 23:18 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/0de4af96141e C-interpreter: Fix for OSR. ! src/share/vm/interpreter/bytecodeInterpreter.cpp

1 0

hg: ppc-aix-port/jdk7u/hotspot: opto: explicit specification of rematerialization in ad file.
by goetz.lindenmaier＠sap.com 29 Nov '12

29 Nov '12

Changeset: e476a85349be Author: Goetz Date: 2012-11-29 13:43 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/e476a85349be opto: explicit specification of rematerialization in ad file. Introduce attributes "ins_cannot_rematerialize" and "ins_should_rematerialize" to adl which allows/forbids rematerialization. If ins_cannot_rematerialize is set, the instruction will never rematerialize. If ins_should_rematerialize is set, adlc generates a rule for the node that allows rematerialization. See also MachNode::rematerialize(). This fixes a problem with rematerialization on PPC: constants are rematerialized in a way that extends the live range of the MachConstantBaseNode. This causes the assertion 'edge_from_to' in Scheduling::verify_good_schedule() to fire. ! src/share/vm/adlc/formssel.cpp ! src/share/vm/adlc/output_h.cpp

1 0

hg: ppc-aix-port/jdk7u/hotspot: opto: Move static _in_dump_cnt to Compile object.
by goetz.lindenmaier＠sap.com 29 Nov '12

29 Nov '12

Changeset: 6e7831be36b0 Author: Goetz Date: 2012-11-29 10:51 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/6e7831be36b0 opto: Move static _in_dump_cnt to Compile object. If there are two or more compiler threads, a single static variable does not suffice. ! src/share/vm/opto/callnode.cpp ! src/share/vm/opto/compile.cpp ! src/share/vm/opto/compile.hpp ! src/share/vm/opto/idealGraphPrinter.cpp ! src/share/vm/opto/node.cpp ! src/share/vm/opto/node.hpp

1 0

hg: ppc-aix-port/jdk7u/hotspot: C-interpreter: some diffs I missed before
by goetz.lindenmaier＠sap.com 28 Nov '12

28 Nov '12

Changeset: 2248becbd897 Author: Goetz Date: 2012-11-28 16:28 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/2248becbd897 C-interpreter: some diffs I missed before ! src/share/vm/interpreter/bytecodeInterpreter.cpp ! src/share/vm/interpreter/interpreterRuntime.cpp ! src/share/vm/interpreter/interpreterRuntime.hpp

1 0

hg: ppc-aix-port/jdk7u/hotspot: C-interpreter: implement bytecode profiling.
by goetz.lindenmaier＠sap.com 28 Nov '12

28 Nov '12

Changeset: 600ce596eea9 Author: Goetz Date: 2012-11-28 13:26 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/600ce596eea9 C-interpreter: implement bytecode profiling. ! src/share/vm/interpreter/bytecodeInterpreter.cpp + src/share/vm/interpreter/bytecodeInterpreterProfiling.hpp ! src/share/vm/interpreter/interpreterRuntime.cpp ! src/share/vm/interpreter/interpreterRuntime.hpp ! src/share/vm/interpreter/invocationCounter.hpp ! src/share/vm/oops/methodDataOop.cpp ! src/share/vm/oops/methodDataOop.hpp ! src/share/vm/runtime/arguments.cpp ! src/share/vm/runtime/globals.hpp

1 0

Jtreg tests
by Steve Poole 26 Nov '12

26 Nov '12

hi all - I'm in the process of getting jtreg tests running against the port builds. Running the tests seems straightforward. I want to brain dump my thoughts on the testing approach for using jtreg and work out how wide a range of testing we need to do. Ideally I would want to have tests run on: 1 - The powerppc port of Linux PPC 2 - The powerppc port of AIX 3 - The powerppc repo build on Linux Ia64 4 - The powerppc repo build on Windows Ia64 5 - The powerppc repo build on Solaris 6 - The powerppc repo build on Mac OSX 7 - The powerppc port of Linux PPC + IBM J9 JVM 8 - The powerppc port of AIX + IBM J9 JVM And be able to compare the results with the mainstream JDK7 and JDK8 versions That could be a lot of testing. (especially since it takes 24hr+ to run the tests on LinuxPPC right now!) I think its ok at this stage to drop the Windows , Solaris and Mac builds and revisit them later. That gives us 5 sets of builds to test. That seems fairly straightforward. I'm getting some education today on the intricacies of Jtreg curtesy of some knowledgeable Oracle folk - hopefully I can tackle the question of what tests to run and how to compare results over time. The final thought I had was that we'll need a plan get the ppc repos updated to be near current with the mainstream repos and keep them fairly close so that comparing tests across repos is fairly reasonable. We can do that later but we do need to consider it. Cheers Steve

2 1

Linux PPC build failures
by Steve Poole 22 Nov '12

22 Nov '12

hi guys - just done a complete fresh clone and build of linux ppc 7 on REL6. I initially hit a problem with the stack size for javac building corba which I fixed. /home/spoole/boot/jdk1.7.0-ppc-aix-port-b01/bin/javac -J-XX:ThreadStackSize=1536 -J-XX:-PrintVMOptions -J-XX:+UnlockDiagnosticVMOptions -J-XX:-LogVMOutput -J-Xmx512m -J-Xms512m -J-XX:PermSize=32m -J-XX:MaxPermSize=160m -encoding ascii -source 6 -target 6 -d /home/spoole/repos/aix-ppc/jdk7u/build/linux-ppc64/../linux-ppc64-debug/corba/btclasses \ -sourcepath ../../tools/src ../../tools/src/build/tools/stripproperties/StripProperties.java The stack size specified is too small, Specify at least 1856k Error: Could not create the Java Virtual Machine. Error: A fatal exception has occurred. Program will exit. Now I've hit another problem g++ -DLINUX -D_GNU_SOURCE -DPPC64 -DASSERT -DDEBUG -I. -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/prims -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/precompiled -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/cpu/ppc/vm -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/os_cpu/linux_ppc/vm -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/os/linux/vm -I/home/spoole/repos/aix-ppc/jdk7u/hotspot/src/os/posix/vm -I../generated -DHOTSPOT_RELEASE_VERSION="\"23.2-b04\"" -DHOTSPOT_BUILD_TARGET="\"jvmg\"" -DHOTSPOT_BUILD_USER="\"root\"" -DHOTSPOT_LIB_ARCH=\"ppc64\" -DHOTSPOT_VM_DISTRO="\"OpenJDK\"" -DTARGET_OS_FAMILY_linux -DTARGET_ARCH_ppc -DTARGET_ARCH_MODEL_ppc_64 -DTARGET_OS_ARCH_linux_ppc -DTARGET_OS_ARCH_MODEL_linux_ppc_64 -DTARGET_COMPILER_gcc -fPIC -fno-rtti -fno-exceptions -D_REENTRANT -fcheck-new -fvisibility=hidden -m64 -DCC_INTERP -pipe -g -m64 -D_LP64=1 -mminimal-toc -mcpu=powerpc64 -mtune=power5 -minsert-sched-nops=regroup_exact -mno-multiple -mno-string -DSAFEFETCH_STUBS -DINCLUDE_TRACE -Werror -Wpointer-arith -Wsign-compare -c -fpch-deps -MMD -MP -MF ../generated/dependencies/precompiled.hpp.gch.d -x c++-header /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/precompiled/precompiled.hpp -o precompiled.hpp.gch cc1plus: warnings being treated as errors In file included from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/runtime/frame.hpp:500, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/prims/jvmtiExport.hpp:33, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/runtime/thread.hpp:32, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/os/linux/vm/thread_linux.inline.hpp:30, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/memory/resourceArea.hpp:30, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/classfile/classFileParser.hpp:29, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/classfile/classLoader.hpp:28, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/classfile/systemDictionary.hpp:29, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/ci/ciEnv.hpp:30, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/ci/ciUtilities.hpp:28, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/ci/ciNullObject.hpp:30, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/ci/ciConstant.hpp:29, from /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/share/vm/precompiled/precompiled.hpp:36: /home/spoole/repos/aix-ppc/jdk7u/hotspot/src/cpu/ppc/vm/frame_ppc.hpp:96: error: 'typedef' was ignored in this declaration make[7]: *** [precompiled.hpp.gch] Error 1 Are these problems you've seen too? I don't want to go digging if you already know about them :-)

3 12

hg: ppc-aix-port/jdk7u/hotspot: ppc: fix gcc 4.3.4 build
by goetz.lindenmaier＠sap.com 21 Nov '12

21 Nov '12

Changeset: b421abab6db5 Author: Goetz Date: 2012-11-21 15:33 +0100 URL: http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/b421abab6db5 ppc: fix gcc 4.3.4 build ! make/linux/makefiles/gcc.make ! src/cpu/ppc/vm/frame_ppc.hpp

1 0

java8 Repos
by Steve Poole 21 Nov '12

21 Nov '12

hi Volker - thanks for creating the JDK 8 repos. We'll start to add in our AIX changes. We also need to get the new configure process working so we'll look at that too. Cheers Steve

2 2