RFR(S): 8016476: PPC64 (part 1): reenable CORE build
Iris Clark
iris.clark at oracle.com
Fri Jun 14 13:23:56 PDT 2013
jcheck is configured to behave exactly the same way that it does for any other JDK release forest. You can install (and configure) jcheck on your client any way you want, but the server-side configuration will determine whether your changeset is accepted for these trees.
(We identical configuration so that when we decide to push from stage to a release, we can push the individual changesets.)
Thanks,
iris
-----Original Message-----
From: Vladimir Kozlov
Sent: Friday, June 14, 2013 1:10 PM
To: Lindenmaier, Goetz
Cc: ppc-aix-port-dev at openjdk.java.net
Subject: Re: RFR(S): 8016476: PPC64 (part 1): reenable CORE build
Hi, Goetz
On 6/14/13 12:37 PM, Lindenmaier, Goetz wrote:
> Hi Vladimir,
>
> Thanks for the honor of the initial change to the staging repo!
> The change should be in there.
>
> Should we next have a look at patch 0004 or 5?
> I assume 3 should wait for 2 as they both affect the closed ports.
Albert tried 3 together with 1+2 and it failed to build very dramatically so I need to look what happens.
It could be because you use TARGET_ARCH_MODEL_ppc_32 but make/linux/platform_ppc was not modified accordingly so the build still use TARGET_ARCH_MODEL_ppc.
> 4 and 5 are feature extensions we use in ppc64. They should not
> affect any existing code, while using them in any port is feasible.
> If you think that's a good idea, give me bug-ids and I'll post
> webrevs.
Did you got JBS notification about Part 2 and Part 3 bugs? I tried to use "share" feature.
And I will file bugs for 4-5 later today.
>
> I intend to lift the staging repository to hs25-b37 once the tag is
> available. Is that ok with you? Or do you want to test the merge
> change?
I think, we (me or Albert when we get committer status) should do the sync because we need to sync closed parts also. It would be nice to do it at the same time and have check point (archived JPRT build) for that.
> As for jcheck, I thought when I push to the staging repository and
> don't pass jcheck the change will not be accepted?
Iris said that it is enabled on server side but I don't know how it works. I would prefer if all people who push into staging repos to have jcheck hook in their ~/.hgrc to avoid problems for all of us:
http://openjdk.java.net/projects/code-tools/jcheck/
Thanks,
Vladimir
>
> Best regards and thanks for the help so far,
> Goetz.
>
>
>
> -----Original Message-----
> From: ppc-aix-port-dev-bounces at openjdk.java.net
> [mailto:ppc-aix-port-dev-bounces at openjdk.java.net] On Behalf Of
> Vladimir Kozlov
> Sent: Friday, June 14, 2013 8:34 PM
> Cc: ppc-aix-port-dev at openjdk.java.net
> Subject: Re: RFR(S): 8016476: PPC64 (part 1): reenable CORE build
>
> Note, only PART 1 is ready to go.
>
> I will leave the honor to push first changeset to you, Goetz :)
>
> Please, make sure you use correct changeset header/descriptor and you pass jcheck (no trailing spaces etc):
>
> 8016476: PPC64 (part 1): reenable CORE build
> Summary: reenable CORE build on PPC64
> Reviewed-by: dholmes, kvn, coleenp
>
>
> About PART 2. Changes are good from reviewers point of view but we
> want to test closed bits on real PPC embedded hardware because you touched #ifdefs around related code. I will let you know the results.
>
> Thanks,
> Vladimir
>
> On 6/14/13 10:58 AM, Albert Noll wrote:
>> Hi,
>>
>> yes, the jprt job for part 1 + part 2 has just finished successfully.
>>
>> Best,
>> Albert
>>
>> On 14.06.2013 17:57, Vladimir Kozlov wrote:
>>> This looks good. Albert tested it and I think it is ready to push.
>>>
>>> Thanks,
>>> Vladimir
>>>
>>> On 6/14/13 12:49 AM, Lindenmaier, Goetz wrote:
>>>> Hi,
>>>>
>>>> I updated the webrev accordingly.
>>>> I'm checking for "ppc" and "64" now.
>>>>
>>>> Best regards,
>>>> Goetz.
>>>>
>>>> -----Original Message-----
>>>> From: Vladimir Kozlov [mailto:vladimir.kozlov at oracle.com]
>>>> Sent: Friday, June 14, 2013 8:37 AM
>>>> To: David Holmes; Lindenmaier, Goetz
>>>> Cc: ppc-aix-port-dev at openjdk.java.net; Chris Plummer;
>>>> hotspot-dev at openjdk.java.net
>>>> Subject: Re: RFR(S): 8016476: PPC64 (part 1): reenable CORE build
>>>>
>>>> Hi, Goetz
>>>>
>>>> I talked with David and we want you to add OS checks as
>>>> generic_buildminimal1 does to avoid a confusion in future if
>>>> someone (as I did) try to build CORE on other platforms. CORE was one of our supported builds and someone can think that this change will reenable it on all platforms which is not true.
>>>>
>>>> Thanks,
>>>> Vladimir
>>>>
>>>> On 6/13/13 2:42 PM, Vladimir Kozlov wrote:
>>>>> On 6/13/13 3:47 AM, David Holmes wrote:
>>>>>> On 13/06/2013 7:53 PM, Lindenmaier, Goetz wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> I fixed the jvmg target and prepared a webrev:
>>>>>>> http://cr.openjdk.java.net/~goetz/webrevs/8016476-CORE/
>>>>>>>
>>>>>>> Thanks for reviewing, Vladimir.
>>>>>>> I need a second reviewer, please.
>>>>>>
>>>>>> So to clarify, this restores the old core targets but will only
>>>>>> actually work on the new ports? If so do we want to validate that
>>>>>> in generic_buildcore similar to the way we validate in generic_buildminimal1?
>>>>>
>>>>> It would be nice to have such validation but on other hand we
>>>>> don't have it for zero and shark, platforms which Oracle does not support. And core will not be in our official builds or supported.
>>>>>
>>>>> thanks,
>>>>> Vladimir
>>>>>
>>>>>>
>>>>>> There is a huge amount of cleanup I'd like to do in this makefile
>>>>>> to remove all the copy'n'paste duplication. :(
>>>>>>
>>>>>> Thanks,
>>>>>> David
>>>>>>
>>>>>>> Best regards,
>>>>>>> Goetz.
>>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: hotspot-dev-bounces at openjdk.java.net
>>>>>>> [mailto:hotspot-dev-bounces at openjdk.java.net] On Behalf Of
>>>>>>> Vladimir Kozlov
>>>>>>> Sent: Donnerstag, 13. Juni 2013 00:58
>>>>>>> To: Volker Simonis
>>>>>>> Cc: ppc-aix-port-dev at openjdk.java.net; Chris Plummer;
>>>>>>> hotspot-dev at openjdk.java.net
>>>>>>> Subject: Re: JEP 175 - Review comments
>>>>>>>
>>>>>>> Hi, Volker
>>>>>>>
>>>>>>> 0001_fix_core_build.patch passed JPRT build/test run so it is
>>>>>>> good, consider reviewed (assuming you fix jvmgcore).
>>>>>>>
>>>>>>> Should we formalize the process to follow our normal openjdk
>>>>>>> review process? It will allow other people to see what is coming
>>>>>>> and to comment on it as you said and I agree.
>>>>>>>
>>>>>>> - I will file rfes and add Volker and Goetz to watch list so you
>>>>>>> get notifications (I hope) about rfes.
>>>>>>>
>>>>>>> - You submit webrev for reviews to ppc-aix-port-dev, hotspot-dev
>>>>>>> mail aliases.
>>>>>>>
>>>>>>> - We do reviews and small testing to make sure changes do not
>>>>>>> break our builds.
>>>>>>>
>>>>>>> - You prepare final patch with correct changeset header after we
>>>>>>> agree on changes.
>>>>>>>
>>>>>>> - I push it to staging repo.
>>>>>>>
>>>>>>> Is this acceptable to you?
>>>>>>>
>>>>>>> About changeset header:
>>>>>>>
>>>>>>> <bugid>: <synopsis-of-symptom>
>>>>>>> Summary: <summary-of-code-change>
>>>>>>> Reviewed-by: <reviewer>+
>>>>>>> Contributed-by: <contributor-email>
>>>>>>>
>>>>>>> I don't think you need "Contributed-by:" since Volker could be author.
>>>>>>> But it is up to you if you want to mention other contributors.
>>>>>>>
>>>>>>> 8016476: PPC64 (part 1): reenable CORE build
>>>>>>> Summary: reenable CORE build for Linux/PPC64 and AIX/PPC64
>>>>>>> Reviewed-by: kvn
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Vladimir
>>>>>>>
>>>>>>> PS: I filed next bug: PPC64 (part 2): Clean up PPC defines.
>>>>>>> Please, check that you get notification. You are on watch list.
>>>>>>>
>>>>>>>
>>>>>>> On 6/12/13 12:18 PM, Vladimir Kozlov wrote:
>>>>>>>> Okay, I will add comment to the rfe that CORE target is only
>>>>>>>> used on
>>>>>>>> Linux/PPC64 and AIX/PPC64 and Oracle will not support it (at
>>>>>>>> least for now).
>>>>>>>>
>>>>>>>> Vladimir
>>>>>>>>
>>>>>>>> On 6/12/13 11:51 AM, Volker Simonis wrote:
>>>>>>>>> On Wed, Jun 12, 2013 at 7:18 PM, Vladimir Kozlov
>>>>>>>>> <vladimir.kozlov at oracle.com
>>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> Thanks, Goetz
>>>>>>>>>>
>>>>>>>>>> I am perfectly fine with such granularity and I can start
>>>>>>>>>> generating RFEs.
>>>>>>>>>> I filed first:
>>>>>>>>>>
>>>>>>>>>> 8016476: PPC64 (part 1): reenable CORE build
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>> Thanks!
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> Are you sure that 0001_fix_core_build.patch is complete? I
>>>>>>>>>> can't build it on my Mac:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>> We haven't fixed the CORE build on all platforms. It should
>>>>>>>>> wok on
>>>>>>>>> Linux/PPC64 and AIX/PPC64 and not break anything else.
>>>>>>>>>
>>>>>>>>> This is the general approach we have taken. So I'm not sure
>>>>>>>>> what will be the best way for you to review the changes. But
>>>>>>>>> all of the patches from the first series of changes (1-9)
>>>>>>>>> won't probably do anything useful on
>>>>>>>>> Linux/x86 and Solaris because we haven't fixed the CORE build
>>>>>>>>> and the C++ interpreter on that platforms. So building a CORE
>>>>>>>>> build or the C++ interpreter on Linux/x86, Solaris or Mac
>>>>>>>>> will probably not succeed (and was not the scope of this
>>>>>>>>> project).
>>>>>>>>>
>>>>>>>>> I think the review for these patches should only make sure
>>>>>>>>> that they don't break anything that worked before an any
>>>>>>>>> supported platforms (and trust us that they are good for our
>>>>>>>>> platforms:).
>>>>>>>>>
>>>>>>>>> So in that sense "reenable CORE build" only means to provide
>>>>>>>>> the appropriate targets in the Makefiles and not the required
>>>>>>>>> source code fixes on all platforms (although they may be
>>>>>>>>> trivial/minimal). But if you'd absolutely also want to have a
>>>>>>>>> core build on Linux/x86 and Solaris, I can have a look at it.
>>>>>>>>>
>>>>>>>>> gnumake[4]: *** No rule to make target `dtrace_stuff'. Stop.
>>>>>>>>>> gnumake[3]: *** [dtrace_stuff] Error 2
>>>>>>>>>> gnumake[2]: *** [debugcore] Error 2
>>>>>>>>>> gnumake[1]: *** [generic_buildcore] Error 2
>>>>>>>>>> gnumake: *** [debugcore] Error 2
>>>>>>>>>>
>>>>>>>>>> And on SPARC:
>>>>>>>>>>
>>>>>>>>>> src/share/vm/runtime/globals.**hpp", line 170: Error:
>>>>>>>>>> Multiple declaration for pd_InlineSmallCode.
>>>>>>>>>>
>>>>>>>>>> And you should used debugcore instead of jvmgcore:
>>>>>>>>>> +all_debugcore: jvmgcore docs export_debug
>>>>>>>>>>
>>>>>>>>>> I want your changes be perfect from the start ;)
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> You're right. The renaming of the 'jvmg' target to 'debug' has
>>>>>>>>> happened recently (
>>>>>>>>> http://hg.openjdk.java.net/hsx/hotspot-main/hotspot/rev/f36e07
>>>>>>>>> 3d56a4) and
>>>>>>>>>
>>>>>>>>> we haven't adapted it. We will fix this.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> And I need to discuss with our embedded group about
>>>>>>>>>> 0002_PPC_defines.patch because it affects them. It may take
>>>>>>>>>> time.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Vladimir
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 6/12/13 8:39 AM, Lindenmaier, Goetz wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi Vladimir,
>>>>>>>>>>>
>>>>>>>>>>> yes, the plan is that each is self contained. When I set up
>>>>>>>>>>> the patches I built and jckecked each. When I updated them I
>>>>>>>>>>> only built them selectively, so there might be minor issues.
>>>>>>>>>>> I had planned to assure this once I make webrevs from the
>>>>>>>>>>> patches.
>>>>>>>>>>>
>>>>>>>>>>> Some of them might apply in any order, but others depend on
>>>>>>>>>>> previous ones, e.g., 0009 containing the ppc files will not
>>>>>>>>>>> work without the changes before.
>>>>>>>>>>>
>>>>>>>>>>> The patches work with hs25-b34.
>>>>>>>>>>>
>>>>>>>>>>> Please tell me if I can do anything to ease your reviewing.
>>>>>>>>>>>
>>>>>>>>>>> Ah, I just saw your mail about closed code ... I tried to
>>>>>>>>>>> keep changes necessary to other platform code to a minimum,
>>>>>>>>>>> but also tried to avoid strange workarounds. Therefore I
>>>>>>>>>>> for example did change
>>>>>>>>>>> 0002
>>>>>>>>>>> renaming the PPC defines, see the comment there.
>>>>>>>>>>>
>>>>>>>>>>> If you agree with the granularity of the changes, it would
>>>>>>>>>>> be great if you could generate bug-ids for them. Maybe at
>>>>>>>>>>> least for the changes up to 0016?
>>>>>>>>>>>
>>>>>>>>>>> Best regards
>>>>>>>>>>> Goetz.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> -----Original Message-----
>>>>>>>>>>> From: Vladimir Kozlov
>>>>>>>>>>> [mailto:vladimir.kozlov@**oracle.com<vladimir.kozlov at oracle.
>>>>>>>>>>> com>
>>>>>>>>>>> ]
>>>>>>>>>>> Sent: Mittwoch, 12. Juni 2013 17:03
>>>>>>>>>>> To: Lindenmaier, Goetz
>>>>>>>>>>> Cc:
>>>>>>>>>>> ppc-aix-port-dev at openjdk.java.**net<ppc-aix-port-dev at openjdk
>>>>>>>>>>> .java.net>;
>>>>>>>>>>>
>>>>>>>>>>> hotspot-dev at openjdk.java.net; Volker Simonis; Azeem Jiva;
>>>>>>>>>>> Chris Plummer
>>>>>>>>>>> Subject: Re: JEP 175 - Review comments
>>>>>>>>>>>
>>>>>>>>>>> Thank you, Goetz.
>>>>>>>>>>>
>>>>>>>>>>> Can I review just 1 patch (for example, 1 from first 1-9),
>>>>>>>>>>> merge it with
>>>>>>>>>>> jdk8 and build? Or I should do review all 1-9 patches and
>>>>>>>>>>> merge them together into jdk8 to be able build? In short, is
>>>>>>>>>>> each patch self-contain?
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>> Vladimir
>>>>>>>>>>>
>>>>>>>>>>> On 6/12/13 7:44 AM, Lindenmaier, Goetz wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Hi,
>>>>>>>>>>>>
>>>>>>>>>>>> With my recent changes I removed some of the problems
>>>>>>>>>>>> Vladimir mentioned.
>>>>>>>>>>>>
>>>>>>>>>>>> I also added the patches queue I maintain into our
>>>>>>>>>>>> jdk8/hotspot repository, at hotspot/ppc_patches.
>>>>>>>>>>>> Applied to the staging hotspot directory, the linuxppc and
>>>>>>>>>>>> aixppc hotspots can be built.
>>>>>>>>>>>>
>>>>>>>>>>>> The queue contains the changes proposed by me before, with
>>>>>>>>>>>> minor changes due to recent development:
>>>>>>>>>>>>
>>>>>>>>>>>> 1-9 linuxppc C-interpreter port (In our plan milestone M2.1)
>>>>>>>>>>>> 11-15 aixppc C-interpreter port (In our plan milestone M2.2)
>>>>>>>>>>>> 101-107 C-interpreter improvements
>>>>>>>>>>>> 111-122 ppc C2 compiler port leading to a vm rudimentarily
>>>>>>>>>>>> working
>>>>>>>>>>>> 200-217 C2 compiler fixes, extensions etc needed for a stable and
>>>>>>>>>>>> performant ppc port.
>>>>>>>>>>>> Altogether currently 49 changes.
>>>>>>>>>>>>
>>>>>>>>>>>> Our plan was to propose the changes in the order of the
>>>>>>>>>>>> queue for review. I'm happy to create webrevs for any of them.
>>>>>>>>>>>>
>>>>>>>>>>>> Vladimir, maybe the queue simplifies reviewing the port, as
>>>>>>>>>>>> the changes are more complete. They include all later
>>>>>>>>>>>> improvements by fixes or adaptions in merge changes.
>>>>>>>>>>>> For why and where I renamed PPC to PPC32 see the second
>>>>>>>>>>>> change in the queue.
>>>>>>>>>>>>
>>>>>>>>>>>> Best regards,
>>>>>>>>>>>> Goetz.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> PS: This can be used as the invokedynamic repository:
>>>>>>>>>>>> hg clone
>>>>>>>>>>>> http://hg.openjdk.java.net/**ppc-aix-port/jdk8/hotspot<http
>>>>>>>>>>>> ://hg.openjdk.java.net/ppc-aix-port/jdk8/hotspot>
>>>>>>>>>>>>
>>>>>>>>>>>> ppc-hotspot
>>>>>>>>>>>> hg clone
>>>>>>>>>>>> http://hg.openjdk.java.net/**ppc-aix-port/stage/hotspot<htt
>>>>>>>>>>>> p://hg.openjdk.java.net/ppc-aix-port/stage/hotspot>stage-ho
>>>>>>>>>>>> tspot
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> cd stage-hotspot
>>>>>>>>>>>> ln -s .../ppc-hotspot/ppc_patches/ .hg/patches
>>>>>>>>>>>> hg qpush -a
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> -----Original Message-----
>>>>>>>>>>>> From: Vladimir Kozlov
>>>>>>>>>>>> [mailto:vladimir.kozlov@**oracle.com<vladimir.kozlov at oracle
>>>>>>>>>>>> .com>
>>>>>>>>>>>> ]
>>>>>>>>>>>> Sent: Dienstag, 11. Juni 2013 18:34
>>>>>>>>>>>> To: Lindenmaier, Goetz
>>>>>>>>>>>> Cc: Volker Simonis; Azeem Jiva; Chris Plummer
>>>>>>>>>>>> Subject: Re: JEP 175 - Review comments
>>>>>>>>>>>>
>>>>>>>>>>>> Here is result of my first attempt to build/test ppc
>>>>>>>>>>>> changes together with our closed sources.
>>>>>>>>>>>>
>>>>>>>>>>>> Small problems:
>>>>>>>>>>>>
>>>>>>>>>>>> src/share/vm/memory/**allocation.hpp:209: Trailing
>>>>>>>>>>>> whitespace
>>>>>>>>>>>> src/share/vm/memory/**allocation.inline.hpp:121: Trailing
>>>>>>>>>>>> whitespace
>>>>>>>>>>>> src/share/vm/opto/escape.cpp:**2207: Trailing whitespace
>>>>>>>>>>>> src/share/vm/memory/**allocation.hpp:212: Trailing
>>>>>>>>>>>> whitespace
>>>>>>>>>>>> src/share/vm/opto/escape.cpp:**2214: Trailing whitespace
>>>>>>>>>>>>
>>>>>>>>>>>> Build on MacOS:
>>>>>>>>>>>>
>>>>>>>>>>>> src/share/vm/opto/**callGenerator.cpp: In member function
>>>>>>>>>>>> 'virtual
>>>>>>>>>>>> JVMState* VirtualCallGenerator::**generate(JVMState*)':
>>>>>>>>>>>> src/share/vm/opto/**callGenerator.cpp:204: error:
>>>>>>>>>>>> 'zero_page_read_protected' is not a member of 'os'
>>>>>>>>>>>>
>>>>>>>>>>>> agent/src/os/bsd/**MacosxDebuggerLocal.m:54:2: error:
>>>>>>>>>>>> #error UNSUPPORTED_ARCH
>>>>>>>>>>>> agent/src/os/bsd/**MacosxDebuggerLocal.m:167:4: error:
>>>>>>>>>>>> #error UNSUPPORTED_ARCH
>>>>>>>>>>>> agent/src/os/bsd/**MacosxDebuggerLocal.m:591:2: error:
>>>>>>>>>>>> #error UNSUPPORTED_ARCH
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> We have several conflict with closed sources builds:
>>>>>>>>>>>>
>>>>>>>>>>>> src/share/vm/utilities/**elfSymbolTable.cpp: In member
>>>>>>>>>>>> function 'bool ElfSymbolTable::lookup(**unsigned char*,
>>>>>>>>>>>> int*, int*, int*)':
>>>>>>>>>>>> src/share/vm/utilities/**elfSymbolTable.cpp:97: error:
>>>>>>>>>>>> comparison between signed and unsigned integer expressions
>>>>>>>>>>>> src/share/vm/utilities/**elfSymbolTable.cpp:124: error:
>>>>>>>>>>>> comparison between signed and unsigned integer expressions
>>>>>>>>>>>>
>>>>>>>>>>>> error: no 'oopDesc** frame::interpreter_frame_**mirror_addr() const'
>>>>>>>>>>>> member function declared in class 'frame'
>>>>>>>>>>>>
>>>>>>>>>>>> error: no matching function for call to
>>>>>>>>>>>> 'SharedRuntime::c_calling_**convention(BasicType*&,
>>>>>>>>>>>> VMRegPair*&, uint&)'
>>>>>>>>>>>>
>>>>>>>>>>>> I think it is due to changes like next:
>>>>>>>>>>>>
>>>>>>>>>>>> -#ifdef PPC
>>>>>>>>>>>> +#ifdef PPC32
>>>>>>>>>>>> oop* interpreter_frame_mirror_addr(**) const;
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Next is easy fix in our closed sources but it requires
>>>>>>>>>>>> efforts from our
>>>>>>>>>>>> side:
>>>>>>>>>>>>
>>>>>>>>>>>> src/share/vm/prims/**methodHandles.cpp: In static member
>>>>>>>>>>>> function 'static void
>>>>>>>>>>>> MethodHandles::generate_**adapters()':
>>>>>>>>>>>> src/share/vm/prims/**methodHandles.cpp:68: error:
>>>>>>>>>>>> 'adapter_code_size'
>>>>>>>>>>>> cannot be used as a function
>>>>>>>>>>>> src/share/vm/prims/**methodHandles.cpp:70: error:
>>>>>>>>>>>> 'adapter_code_size'
>>>>>>>>>>>> cannot be used as a function
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> I would suggest to do such shared changes which affects
>>>>>>>>>>>> different builds later after initial push.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>> Vladimir
>>>>>>>>>>>>
>>>>>>>>>>>> On 6/7/13 3:20 AM, Lindenmaier, Goetz wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> Hi Vladimir,
>>>>>>>>>>>>>
>>>>>>>>>>>>> I updated the repo to jdk8-b92. Our nightly tests built
>>>>>>>>>>>>> and tested it successfully. In case you experience any
>>>>>>>>>>>>> problems please tell me the details so I can fix them.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Best regards,
>>>>>>>>>>>>> Goetz.
>>>>>>>>>>>>>
>>>>>>>>>>>>> http://cr.openjdk.java.net/~**simonis/ppc-aix-port/index.*
>>>>>>>>>>>>> *html<http://cr.openjdk.java.net/~simonis/ppc-aix-port/ind
>>>>>>>>>>>>> ex.html>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> -----Original Message-----
>>>>>>>>>>>>> From: Vladimir Kozlov
>>>>>>>>>>>>> [mailto:vladimir.kozlov@**oracle.com<vladimir.kozlov at oracl
>>>>>>>>>>>>> e.com>
>>>>>>>>>>>>> ]
>>>>>>>>>>>>> Sent: Dienstag, 4. Juni 2013 20:58
>>>>>>>>>>>>> To: Simonis, Volker
>>>>>>>>>>>>> Cc: Iris Clark; Wintergerst, Michael; Lindenmaier, Goetz;
>>>>>>>>>>>>> Bernard Traversat; Jeannette Hung; Azeem Jiva; David
>>>>>>>>>>>>> Therkelsen; Mikael Vidstedt; Neil Richards; Steve Poole;
>>>>>>>>>>>>> luchsh at cn.ibm.com; Tim Ellison; Alan Bateman
>>>>>>>>>>>>> Subject: Re: JEP 175 - Review comments
>>>>>>>>>>>>>
>>>>>>>>>>>>> Volker,
>>>>>>>>>>>>>
>>>>>>>>>>>>> Can you or someone update
>>>>>>>>>>>>> http://hg.openjdk.java.net/**ppc-aix-port/jdk8/hotspot<htt
>>>>>>>>>>>>> p://hg.openjdk.java.net/ppc-aix-port/jdk8/hotspot>to
>>>>>>>>>>>>>
>>>>>>>>>>>>> match latest
>>>>>>>>>>>>> sources in
>>>>>>>>>>>>> http://hg.openjdk.java.net/**jdk8/jdk8/hotspot<http://hg.o
>>>>>>>>>>>>> penjdk.java.net/jdk8/jdk8/hotspot>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> ?
>>>>>>>>>>>>> I just tried to merge them and build Hotspot on x86
>>>>>>>>>>>>> without success.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>> Vladimir
>>>>>>>>>>>>>
>>>>>>>>>>>>> On 6/4/13 7:04 AM, Simonis, Volker wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> We intentionally used 'porters-dev' rather
>>>>>>>>>>>>>> than'ppc-aix-port-dev' in the beginning to address a
>>>>>>>>>>>>>> broader audience for the initial discussions.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> But I'm happy to change that back to 'ppc-aix-port-dev' now.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Regards,
>>>>>>>>>>>>>> Volker
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> ------------------------------**-------------------------
>>>>>>>>>>>>>> -----**
>>>>>>>>>>>>>> ------------
>>>>>>>>>>>>>> *From:* Iris Clark [iris.clark at oracle.com]
>>>>>>>>>>>>>> *Sent:* Tuesday, June 04, 2013 3:50 PM
>>>>>>>>>>>>>> *To:* Simonis, Volker; Wintergerst, Michael; Lindenmaier,
>>>>>>>>>>>>>> Goetz; Bernard Traversat; Jeannette Hung; Azeem Jiva;
>>>>>>>>>>>>>> David Therkelsen; Mikael Vidstedt; Neil Richards; Steve
>>>>>>>>>>>>>> Poole; luchsh at cn.ibm.com; Tim Ellison;
>>>>>>>>>>>>>> iris.clark at oracle.com
>>>>>>>>>>>>>> *Cc:* Alan Bateman; Vladimir Kozlov
>>>>>>>>>>>>>> *Subject:* RE: JEP 175 - Review comments
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi, Volker.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Sorry about this, one more thing about the JEP...
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I think that the "Discussion" list probably needs to be
>>>>>>>>>>>>>> updated to be your Project's mailing list
>>>>>>>>>>>>>> (ppc-aix-port-dev). Right now it's listed as
>>>>>>>>>>>>>> porters-dev.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> iris
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> *From:*Simonis, Volker [mailto:volker.simonis at sap.com**]
>>>>>>>>>>>>>> *Sent:* Tuesday, June 04, 2013 12:12 AM
>>>>>>>>>>>>>> *To:* Iris Clark; Wintergerst, Michael; Lindenmaier,
>>>>>>>>>>>>>> Goetz; Bernard Traversat; Jeannette Hung; Azeem Jiva;
>>>>>>>>>>>>>> David Therkelsen; Mikael Vidstedt; Neil Richards; Steve
>>>>>>>>>>>>>> Poole; luchsh at cn.ibm.com; Tim Ellison
>>>>>>>>>>>>>> *Cc:* Alan Bateman; Vladimir Kozlov
>>>>>>>>>>>>>> *Subject:* RE: JEP 175 - Review comments
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi Iris,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> you're right, the title is too clumsy - I just couldn't
>>>>>>>>>>>>>> come up with something better yesterday in the evening.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> "PowerPC/AIX Port" sounds good to me. If nobody
>>>>>>>>>>>>>> complains, I'll take it.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>> Volker
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> ------------------------------**-------------------------
>>>>>>>>>>>>>> -----**
>>>>>>>>>>>>>> ------------
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> *From:*Iris Clark [iris.clark at oracle.com]
>>>>>>>>>>>>>> *Sent:* Tuesday, June 04, 2013 2:57 AM
>>>>>>>>>>>>>> *To:* Simonis, Volker; Wintergerst, Michael; Lindenmaier,
>>>>>>>>>>>>>> Goetz; Bernard Traversat; Jeannette Hung; Azeem Jiva;
>>>>>>>>>>>>>> David Therkelsen; Mikael Vidstedt; Neil Richards; Steve
>>>>>>>>>>>>>> Poole; luchsh at cn.ibm.com <mailto:luchsh at cn.ibm.com>; Tim
>>>>>>>>>>>>>> Ellison
>>>>>>>>>>>>>> *Cc:* Alan Bateman; Vladimir Kozlov;
>>>>>>>>>>>>>> iris.clark at oracle.com <mailto:iris.clark at oracle.com>
>>>>>>>>>>>>>> *Subject:* RE: JEP 175 - Review comments
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi, Volker.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I've just updated the JEP according to your suggestions.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I'm not a Reviewer however, I think that the phrase
>>>>>>>>>>>>>> "OpenJDK master repositories" in the revised title is not
>>>>>>>>>>>>>> ideal:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> --- a/jep-175.md Mon May 27 23:22:51 2013 +0400
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> +++ b/jep-175.md Mon Jun 03 18:51:18 2013 +0200
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> @@ -1,5 +1,5 @@
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> JEP: 175
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> -Title: Integrate PowerPC/AIX Port into JDK 8
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> +Title: Integrate the PowerPC/AIX Port into the OpenJDK
>>>>>>>>>>>>>> +master
>>>>>>>>>>>>>> repositories
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Author: Volker Simonis
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Organization: SAP AG
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Created: 2013/1/11
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thinking out loud, what about just "PowerPC/AIX Port"?
>>>>>>>>>>>>>> JEPs are all about adding features to JDK Release
>>>>>>>>>>>>>> Projects. There are lots of examples of JEPs [1] which
>>>>>>>>>>>>>> don't begin with verbs, e.g. "133:
>>>>>>>>>>>>>> Unicode
>>>>>>>>>>>>>> 6.2", "148: Small VM", "172: DocLint", etc. The JEP
>>>>>>>>>>>>>> itself contains additional details.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Perhaps others have suggestions?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Am I right with my assumption that the targeted release
>>>>>>>>>>>>>> will still be JDK 8 (or 9/8u respectively) but that the
>>>>>>>>>>>>>> targeted release will be set in the "Release" header
>>>>>>>>>>>>>> field of the JEP by the OpenJDK Lead (as specified in the
>>>>>>>>>>>>>> JEP specification)?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Yes. Once that field is populated, the value will appear
>>>>>>>>>>>>>> in the JEP index [1], see the third column.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Iris
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [1]: http://openjdk.java.net/jeps/0
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> *From:*Simonis, Volker [mailto:volker.simonis at sap.com**]
>>>>>>>>>>>>>> *Sent:* Monday, June 03, 2013 10:10 AM
>>>>>>>>>>>>>> *To:* Iris Clark; Wintergerst, Michael; Lindenmaier,
>>>>>>>>>>>>>> Goetz; Bernard Traversat; Jeannette Hung; Azeem Jiva;
>>>>>>>>>>>>>> David Therkelsen; Mikael Vidstedt; Neil Richards; Steve
>>>>>>>>>>>>>> Poole; luchsh at cn.ibm.com <mailto:luchsh at cn.ibm.com>; Tim
>>>>>>>>>>>>>> Ellison
>>>>>>>>>>>>>> *Cc:* Alan Bateman; Vladimir Kozlov
>>>>>>>>>>>>>> *Subject:* RE: JEP 175 - Review comments
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I've just updated the JEP according to your suggestions.
>>>>>>>>>>>>>> Please find the new version attached to this mail (I
>>>>>>>>>>>>>> haven't checked the new version in until now to give
>>>>>>>>>>>>>> everybody a chance to comment on the changes).
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Am I right with my assumption that the targeted release
>>>>>>>>>>>>>> will still be JDK 8 (or 9/8u respectively) but that the
>>>>>>>>>>>>>> targeted release will be set in the "Release" header
>>>>>>>>>>>>>> field of the JEP by the OpenJDK Lead (as specified in the
>>>>>>>>>>>>>> JEP specification)?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> In addition to the changes proposed by you I've added the
>>>>>>>>>>>>>> contents of the "Approach" section from Azeems "PPCAIX
>>>>>>>>>>>>>> plan" to the "Description"
>>>>>>>>>>>>>> section of the JEP. I've also added links to the new
>>>>>>>>>>>>>> "PowerPC/AIX Port Integration Plan" [2] of our
>>>>>>>>>>>>>> "PowerPC/AIX Port OpenJDK Wiki Space" [3] to the JEP.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> The "PowerPC/AIX Port Integration Plan" in the Wiki is
>>>>>>>>>>>>>> intended to hold Azeems "PPCAIX plan" document.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> @Iris: could you please somehow arrange to give Azeem
>>>>>>>>>>>>>> editing rights to that page?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> @Azeem: could you please be so kind to past the contents
>>>>>>>>>>>>>> of the "PPCAIX plan" into that page (once you have the
>>>>>>>>>>>>>> appropriate rights)? I saw that the document is created
>>>>>>>>>>>>>> from an Atlassian Confluence Wiki anyway and in my
>>>>>>>>>>>>>> unlimited naivety I imagine this could be a simple
>>>>>>>>>>>>>> copy-and-paste
>>>>>>>>>>>>>> operation:) If that doesn't work so easily, please let me
>>>>>>>>>>>>>> know how I could help.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> If there are no objections I plan to checkin the new
>>>>>>>>>>>>>> version of the JEP tomorrow after our telephone call.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thank you and best regards, Volker
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [2]:
>>>>>>>>>>>>>> https://wiki.openjdk.java.net/**pages/viewpage.action?pag
>>>>>>>>>>>>>> eId=**
>>>>>>>>>>>>>> 13729959<https://wiki.openjdk.java.net/pages/viewpage.act
>>>>>>>>>>>>>> ion?pageId=13729959>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [3]:
>>>>>>>>>>>>>> https://wiki.openjdk.java.net/**display/PPCAIXPort<https:
>>>>>>>>>>>>>> //wiki.openjdk.java.net/display/PPCAIXPort>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> ------------------------------**-------------------------
>>>>>>>>>>>>>> -----**
>>>>>>>>>>>>>> ------------
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> *From:*Iris Clark [iris.clark at oracle.com]
>>>>>>>>>>>>>> *Sent:* Friday, May 31, 2013 8:48 PM
>>>>>>>>>>>>>> *To:* Wintergerst, Michael; Simonis, Volker; Lindenmaier,
>>>>>>>>>>>>>> Goetz; Bernard Traversat; Jeannette Hung; Azeem Jiva;
>>>>>>>>>>>>>> David Therkelsen; Mikael Vidstedt; Neil Richards; Steve
>>>>>>>>>>>>>> Poole; luchsh at cn.ibm.com <mailto:luchsh at cn.ibm.com>; Tim
>>>>>>>>>>>>>> Ellison
>>>>>>>>>>>>>> *Cc:* iris.clark at oracle.com
>>>>>>>>>>>>>> <mailto:iris.clark at oracle.com>**; Alan Bateman; Vladimir
>>>>>>>>>>>>>> Kozlov
>>>>>>>>>>>>>> *Subject:* JEP 175 - Review comments
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi, Volker.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> JEP 175: Integrate PowerPC/AIX Port into JDK 8
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> http://openjdk.java.net/jeps/**175
>>>>>>>>>>>>>> <http://openjdk.java.net/jeps/175>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> We're actively working to get your JEP to Funded. We had a few
>>>>>>>>>>>>>> comments:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> -Recommend that "8" be removed from the JEP title, etc.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> -Recommend that the first Motivation bullet clearly
>>>>>>>>>>>>>> indicate that it is only covering PPC/AIX.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> -Recommend that the second Motivation bullet be modified
>>>>>>>>>>>>>> to make it clear that it applies to Hotspot only.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> (Instructions for editing the JEP may be found in the "Mechanics"
>>>>>>>>>>>>>> section at the bottom of JEP 1 [1].)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Vladimir Kozlov (VM) and Alan Bateman (Core Libraries)
>>>>>>>>>>>>>> are lined up to be the JEP's reviewers. Once they're
>>>>>>>>>>>>>> satisfied with your changes/feedback they'll add
>>>>>>>>>>>>>> themselves to the JEP's "Reviewed-by"
>>>>>>>>>>>>>> line.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Iris Clark
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [1]: http://openjdk.java.net/jeps/1
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>
More information about the ppc-aix-port-dev
mailing list