New builds from the build-infra team
Kelly O'Hair
kelly.ohair at oracle.com
Tue Nov 6 08:33:23 PST 2012
If you get a chance, could you try this experiment again with the latest in jdk8/build forest?
I fixed a few more libraries that I think had a missing -lc.
I don't have access to a Solaris SPARC 11.1 system, and so far the issues you have run into seem to
be unique to SPARCV9 builds on 11.1, the Solaris X64 11.1 systems I have access to don't seem to reproduce this issue.
I can only assume that the compiler or the system behaves differently between SPARCV9 and X64 on 11.1. :^(
-kto
On Nov 2, 2012, at 1:08 PM, Vincent Ryan wrote:
> Thanks Kelly.
>
> I re-ran the build with your patch and got better results but there
> seems to be another linker problem later in the jdk build.
>
> Here's the tail of the build output:
>
> :
> :
> Undefined first referenced
> symbol in file
> free /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> __iob /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> abort /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> snprintf /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> calloc /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> malloc /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> memcpy /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> memset /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> strcmp /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> strlen /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> realloc /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> strncmp /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> fprintf /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libjava_crw_demo/java_crw_demo.o
> ld: fatal: symbol referencing errors. No output written to /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/lib/sparcv9/libjava_crw_demo.so
> gmake[3]: *** [/export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/lib/sparcv9/libjava_crw_demo.so] Error 1
> gmake[3]: *** Waiting for unfinished jobs....
> Undefined first referenced
> symbol in file
> exit /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> free /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> __iob /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> abort /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf.o
> iconv /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf_md.o
> __ctype /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf.o
> iconv_open /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf_md.o
> calloc /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> memcpy /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf_md.o
> strcmp /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> strdup /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> strlen /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf.o
> nl_langinfo /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf_md.o
> sprintf /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf.o
> setlocale /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf_md.o
> iconv_close /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/utf_md.o
> fprintf /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libnpt/npt.o
> ld: fatal: symbol referencing errors. No output written to /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/lib/sparcv9/libnpt.so
> gmake[3]: *** [/export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/lib/sparcv9/libnpt.so] Error 1
> gmake[3]: Leaving directory `/export/home/vinryan/jdk8-master/jdk/makefiles'
> gmake[2]: *** [libs-only] Error 2
> gmake[2]: Leaving directory `/export/home/vinryan/jdk8-master/jdk/makefiles'
> make[1]: *** [jdk-only] Error 2
> make[1]: Leaving directory `/export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release'
> make: *** [all] Error 2
> t4%
>
>
>
>
>
> On 02/11/2012 18:38, Kelly O'Hair wrote:
>>
>> On Nov 2, 2012, at 11:17 AM, Vincent Ryan wrote:
>>
>>> That attachment got dropped when I re-sent my message (previously I wasn't a member of the build-infra-dev email alias). I've attached it again in case it proves useful.
>>>
>>> <config.status>
>>>
>>>
>>> I realize Solaris 11 is not an official build platform for JDK8 but I thought I'd give it a spin on a reasonably
>>> beefy server.
>>
>> It's great that you tried this, thanks.
>>
>> As far as I can tell, this seems to be a sparcv9 Solaris 11.1 issue, because I tried Solaris 11.1 X64 and it worked fine.
>> However, the link of libhprof.so IS missing the -lc option.
>> So why sparcv9 would complain and amd64 would not is a bit puzzling.
>>
>> The change I think needs to be made is:
>>
>> diff --git a/makefiles/CompileNativeLibraries.gmk b/makefiles/CompileNativeLibraries.gmk
>> --- a/makefiles/CompileNativeLibraries.gmk
>> +++ b/makefiles/CompileNativeLibraries.gmk
>> @@ -1807,7 +1807,7 @@
>> BUILD_LIBHPROF_LDFLAGS:=
>>
>> ifeq ($(OPENJDK_TARGET_OS),solaris)
>> - BUILD_LIBHPROF_LDFLAGS += -lsocket -lnsl
>> + BUILD_LIBHPROF_LDFLAGS += -lsocket -lnsl -lc
>> endif
>>
>> LIBHPROF_OPTIMIZATION:=HIGHEST
>>
>>
>>
>> But I have not tested it yet.
>> Did you want to give this patch a spin?
>>
>> -kto
>>
>>> Thanks.
>>>
>>>
>>> On 2 Nov 2012, at 17:45, Kelly O'Hair wrote:
>>>
>>>> The attachment is missing.
>>>>
>>>> This looks like the -lc is missing from the creation of hprof, we should never let libc be an implicit dependency.
>>>>
>>>> I wonder if Solaris 11.1 has tightened up the rules on implicit dependencies. We have not seen this on Solaris 10.
>>>> We just got svc6 setup as a 11.1 system, I'll see if it reproduces there.
>>>>
>>>> And thank you for reporting this. Although Solaris 11.1 is not one of our critical systems, it's very important that this works.
>>>>
>>>> -kto
>>>>
>>>> On Nov 2, 2012, at 9:38 AM, Vincent Ryan wrote:
>>>>
>>>>>> From: Vincent Ryan <vincent.x.ryan at oracle.com>
>>>>>> Subject: Re: New builds from the build-infra team
>>>>>> Date: 2 November 2012 15:48:03 GMT
>>>>>> To: build-infra-dev at openjdk.java.net
>>>>>>
>>>>>> I tried this out on the latest Solaris 11 Update 1 (sparc) but the build encountered problems locating libc when
>>>>>> building jdk. (BTW the old build works correctly, just slower)
>>>>>>
>>>>>> :
>>>>>> :
>>>>>> strerror /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libhprof_jvmti/hprof_md.o (symbol belongs to implicit dependency /lib/sparcv9/libc.so.1)
>>>>>> vfprintf /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libhprof_jvmti/hprof_error.o (symbol belongs to implicit dependency /lib/sparcv9/libc.so.1)
>>>>>> fprintf /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libhprof_jvmti/hprof_init.o (symbol belongs to implicit dependency /lib/sparcv9/libc.so.1)
>>>>>> gethrvtime /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/objs/libhprof_jvmti/hprof_md.o (symbol belongs to implicit dependency /lib/sparcv9/libc.so.1)
>>>>>> ld: fatal: symbol referencing errors. No output written to /export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/lib/sparcv9/libhprof.so
>>>>>> gmake[3]: *** [/export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release/jdk/lib/sparcv9/libhprof.so] Error 1
>>>>>> gmake[3]: Leaving directory `/export/home/vinryan/jdk8-master/jdk/makefiles'
>>>>>> gmake[2]: *** [libs-only] Error 2
>>>>>> gmake[2]: Leaving directory `/export/home/vinryan/jdk8-master/jdk/makefiles'
>>>>>> make[1]: *** [jdk-only] Error 2
>>>>>> make[1]: Leaving directory `/export/home/vinryan/jdk8-master/build/solaris-sparcv9-normal-server-release'
>>>>>> make: *** [all] Error 2
>>>>>> t4%
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> FYI I've attached the config script that was generated by configure.sh.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 1 Nov 2012, at 18:38, Kelly O'Hair wrote:
>>>>>>
>>>>>>>
>>>>>>> Pardon the wide email, but this impacts everyone building the OpenJDK jdk8/jdk8 derived forests.
>>>>>>>
>>>>>>> Please only reply to the build-infra-dev mailing list, or just me.
>>>>>>>
>>>>>>> With some recent integrations from the build-infra project into jdk8/jdk8 repositories, the build-infra team
>>>>>>> would like to get more exposure of the new builds. These jdk8/jdk8 changes will start showing up in various
>>>>>>> jdk8 and team forests over the next few weeks. The default is still the old builds, but both builds work in most
>>>>>>> cases for OpenJDK as far as we know.
>>>>>>>
>>>>>>> At a very high level, the intent is that once you get a forest:
>>>>>>> hg clone http://hg.openjdk.java.net/jdk8/jdk8 j8
>>>>>>> cd j8
>>>>>>> sh ./get_source.sh
>>>>>>>
>>>>>>> You should be able to simply configure&&make (the ultimate goal is this simple anyway), e.g.
>>>>>>> sh ./configure
>>>>>>> make NEWBUILD=true # The NEWBUILD=true will become the default when we formally switch.
>>>>>>>
>>>>>>> Where "make" is GNU make 3.81, and your system has all the requires packages and PATH contains the
>>>>>>> needed tools. Note that on Windows, MKS unix utilities cannot be used with the new builds, just CYGWIN
>>>>>>> is recommended at this time.
>>>>>>>
>>>>>>> Of course, we know, it's never as easy as a simple configure&&make, and often you will need to pass in
>>>>>>> configure options.
>>>>>>>
>>>>>>> What we would like to know is where a simple configure&&make does not work, and anything people had
>>>>>>> to do to make it work.
>>>>>>>
>>>>>>> I know many of you are quite used to the old builds, so I have a temporary "bridgeBuild" target
>>>>>>> people can try that will attempt to map the ALT_* environment variables to an appropriate configure command
>>>>>>> and then run that configure command and do the build, e.g.
>>>>>>>
>>>>>>> make NEWBUILD=true bridgeBuild
>>>>>>>
>>>>>>> People willing to do comparisons between the old and new builds could:
>>>>>>> rm -f -r build
>>>>>>> time make NEWBUILD=true bridgeBuild
>>>>>>> rm -f -r build
>>>>>>> time make NO_DOCS=true # Old builds do not generate javadocs by default
>>>>>>>
>>>>>>> Any observations about speed of the builds would be appreciated, as will any impressions on what you see.
>>>>>>>
>>>>>>> At this time, we think this is working pretty well with a few caveats:
>>>>>>> * GNU make with the new builds is doing much more parallel processing and this can stress out a system
>>>>>>> - Use "make JOBS=1" if you suspect a problem, then try adjusting it up slowly.
>>>>>>> * Partial builds are limited, right now full builds of the entire OpenJDK is the target
>>>>>>> - Hotspot can still be built on it's own, but everyone else needs to build hotspot at least once
>>>>>>> * Paths with multiple names can cause problems, e.g. being on system svc6, and access an exported share
>>>>>>> area as /net/svc6/export/foobar instead of /export/foobar will cause problems. Use local paths.
>>>>>>>
>>>>>>> We know there are still issues and we will be focusing heavily on the critical ones in the next few weeks, but
>>>>>>> we do need the community to tell us what the critical issues really are.
>>>>>>>
>>>>>>> Our number one priority at this time is that everyone that was able to build the old way, should be able to build
>>>>>>> with the new build-infra makefiles. Please help us verify that.
>>>>>>>
>>>>>>> -kto
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
More information about the build-infra-dev
mailing list