Change "8030763: Validate global memory allocation" breaks the HotSpot build on AIX

[Volker Simonis]

Hi Vladimir, Alejandro,

thanks a lot for the fast help!

Volker

On Friday, July 18, 2014, Vladimir Kozlov <vladimir.kozlov at oracle.com>
wrote:

> Okay, I will push into jdk8u/jdk8u-dev.
>
> Thanks,
> Vladimir
>
> On 7/18/14 11:42 AM, Alejandro E Murillo wrote:
>
>>
>> On 7/18/2014 12:21 PM, Alejandro E Murillo wrote:
>>
>>>
>>> On 7/18/2014 10:03 AM, Seán Coffey wrote:
>>>
>>>> Volker,
>>>>
>>>> Can you work this fix into the jdk8u hotspot team forest :
>>>> http://hg.openjdk.java.net/jdk8u/hs-dev/
>>>>
>>>> Unfortunately it looks like the forest hasn't been synced up with
>>>> 8u11 changes yet. Alejandro - do you have a timeline for that ? If
>>>> it's not going to happen shortly, we should make an exception and
>>>> push the hotspot change to the jdk8u-dev team forest which already
>>>> has the 8u11 changes.
>>>>
>>> I just synched jdk8u/hs-dev with jdk8u/jdk8u.
>>>
>> I just realized the 8u11 changes are only on jdk8u-dev.
>> It's too risky to bring them straight from there to the jdk8u/hs-dev repo,
>> I usually synch from the master (stable snapshot).
>> so given the urgency,  Vladimir, can you push it to jdk8u/jdk8u-dev
>> instead?
>> it will come back down to the hotspot repo along with the 8u11 changes
>> once they are tested and pushed to master.
>>
>> Thanks
>> Alejandro
>>
>>
>>  I wasn't planning on taking a snapshot this week (no new changes)
>>> so if this ready to go in, and need to be in 8u20 next week, let's get it
>>> in now and I can start a snapshot later today. If not, it will have to
>>> go in through 8u-dev.
>>> Volker, do you have the patch? if so send it  out and I (or Vladimir)
>>> can push it
>>> Thanks
>>> Alejandro
>>>
>>>
>>>
>>>> Please mark the bug with '8u20-critical-watch' label when you log it.
>>>> The label can be changed to '8u20-critical-request' once the fix is
>>>> pushed to the 8u40 mainline.
>>>>
>>>> regards,
>>>> Sean.
>>>>
>>>> On 18/07/14 16:53, Vladimir Kozlov wrote:
>>>>
>>>>> File bug P2 on 8u20 and we will try to convince release team to put
>>>>> your fix into 8u20. Only showstopper are allowed now and it looks
>>>>> like you have showstopper.
>>>>>
>>>>> Later for jdk9 it will be forward port.
>>>>>
>>>>> Regards,
>>>>> Vladimir
>>>>>
>>>>> On 7/18/14 8:32 AM, Volker Simonis wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> unfortunately the merging of 8u11 into 8u-dev and 8u20 broke the
>>>>>> HotSpot build on AIX for those code lines.
>>>>>>
>>>>>> While the fix is trivial (see below) I'm not sure how to proceed with
>>>>>> this fix in order to bring it to 8u-dev and 8u20.
>>>>>>
>>>>>> The problem is that 8030763 is not in jdk9 until now so I can't fix it
>>>>>> in 9 and then backport it to 8.
>>>>>>
>>>>>> Should I just open a new bug for 8u and send out a request for review?
>>>>>> Where should this RFR be directed to - to both hotspot-dev and
>>>>>> jdk8u-dev? And who will do the actual push?
>>>>>>
>>>>>> Thank you and best regards,
>>>>>> Volker
>>>>>>
>>>>>>
>>>>>> diff -r f09d1f6a401e src/os/aix/vm/os_aix.cpp
>>>>>> --- a/src/os/aix/vm/os_aix.cpp  Mon Jul 14 10:16:34 2014 -0700
>>>>>> +++ b/src/os/aix/vm/os_aix.cpp  Fri Jul 18 17:22:32 2014 +0200
>>>>>> @@ -1215,10 +1215,6 @@
>>>>>>     ::abort();
>>>>>>   }
>>>>>>
>>>>>> -// Unused on Aix for now.
>>>>>> -void os::set_error_file(const char *logfile) {}
>>>>>> -
>>>>>> -
>>>>>>   // This method is a copy of JDK's sysGetLastErrorString
>>>>>>   // from src/solaris/hpi/src/system_md.c
>>>>>>
>>>>>>
>>>>
>>>
>>