RFR(S) 8208658: Make CDS archived heap regions usable even if compressed oop encoding has changed

[Jiangli Zhou]

Hi Goetz,

Adding WB_IsZGCIncludedInVmBuild() in whitebox sounds good. It might be 
useful in other cases as well. Have you filed a bug for the 
sharedStrings/IncompatibleOptions.java issue? If not, I can create one.

Thanks,

Jiangli


On 8/29/18 1:18 AM, Lindenmaier, Goetz wrote:
> Hi,
>
> I see appcds/sharedStrings/IncompatibleOptions.java failing since it
> was pushed. Now two more tests for this were added in 8208658.
>
> They fail because we build on linuxx86_64 without ZGC, but the test contains
> if (Platform.isLinux() && Platform.isX64()) {
>              testDump(1, "-XX:+UnlockExperimentalVMOptions", "-XX:+UseZGC", COOPS_DUMP_WARNING, true);
> }
>
> Should we implement WB_IsZGCIncludedInVmBuild() similar to WB_IsCDSIncludedInVmBuild()
> in whitebox?  This could be used instead.
> Or is there already functionality to test this better than checking the platforms?
>
> Best regards,
>    Goetz.
>
>
>
>> -----Original Message-----
>> From: hotspot-gc-dev <hotspot-gc-dev-bounces at openjdk.java.net> On
>> Behalf Of Ioi Lam
>> Sent: Mittwoch, 22. August 2018 17:30
>> To: Thomas Schatzl <thomas.schatzl at oracle.com>; Jiangli Zhou
>> <jiangli.zhou at oracle.com>; hotspot-runtime-dev at openjdk.java.net
>> runtime <hotspot-runtime-dev at openjdk.java.net>; hotspot-gc-dev
>> <hotspot-gc-dev at openjdk.java.net>
>> Subject: Re: RFR(S) 8208658: Make CDS archived heap regions usable even if
>> compressed oop encoding has changed
>>
>> Thanks Thomas. I've fixed the comment, did more hs-tier1-3 testing to
>> validate the patch, and pushed.
>>
>> - Ioi
>>
>> On 8/22/18 1:32 AM, Thomas Schatzl wrote:
>>> Hi Ioi,
>>>
>>> On Tue, 2018-08-21 at 10:38 -0700, Ioi Lam wrote:
>>>> Hi Thomas.
>>>>
>>>> Thanks for the review. I've updated the webrev according to your
>>>> comments. See
>>>>
>>>> http://cr.openjdk.java.net/~iklam/jdk12/8208658-relocate-archived-hea
>>>> p-regions.v03.delta/
>>>>
>>>     looks good sans Jiangli's comment about that log message. I do not
>>> need to see a new webrev for that change either.
>>>
>>> Thanks,
>>>     Thomas
>>>