On Thu, 20 Oct 2022 09:40:55 GMT, Aleksey Shipilev wrote:
>> Found this when working on JOL support
>> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
>> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
>> would fail with:
>>
>>
>> Caus
On Thu, 20 Oct 2022 09:40:55 GMT, Aleksey Shipilev wrote:
>> Found this when working on JOL support
>> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
>> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
>> would fail with:
>>
>>
>> Caus
On Thu, 20 Oct 2022 09:35:16 GMT, Aleksey Shipilev wrote:
> Ran those too with different object alignments. I am running tests with
> different GCs too.
Ran with Serial, Parallel, G1. Shenandoah and ZGC have problems with SA tests
even with default object alignment, so I skipped those. Serial
On Wed, 19 Oct 2022 22:44:37 GMT, Chris Plummer wrote:
>> Yes. I don't see a need for it.
>
> You could keep in the check to make sure it is aligned to a power of 2, since
> I doubt SA would work with a value that isn't, but then I don't see the
> possibility of hotspot ever supporting it eithe
On Wed, 19 Oct 2022 21:02:13 GMT, Chris Plummer wrote:
> Sorry, I missed that. There are also some SA tests in
> test/jdk/sun/tools/jhsdb.
Ran those too with different object alignments. I am running tests with
different GCs too.
-
PR: https://git.openjdk.org/jdk/pull/10762
> Found this when working on JOL support
> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
> would fail with:
>
>
> Caused by: java.lang.RuntimeException: Object alignment 32 not yet supp
On Wed, 19 Oct 2022 20:59:29 GMT, Chris Plummer wrote:
>> I don't think there is anything specific in SA that prohibits it to follow
>> what Hotspot is capable of. Are you proposing to drop the range check
>> altogether? I can do that.
>
> Yes. I don't see a need for it.
You could keep in the
On Wed, 19 Oct 2022 20:25:32 GMT, Aleksey Shipilev wrote:
>> src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/runtime/VM.java line 505:
>>
>>> 503:
>>> 504: minObjAlignmentInBytes = getObjectAlignmentInBytes();
>>> 505: if (minObjAlignmentInBytes < 8 || minObjAlignmentInBytes > 256 |
On Wed, 19 Oct 2022 20:26:33 GMT, Aleksey Shipilev wrote:
> As I stated in "Additional testing" in the PR description, I did ran
> `serviceability/sa` with different alignments. What other tests we should run?
Sorry, I missed that. There are also some SA tests in test/jdk/sun/tools/jhsdb.
On Wed, 19 Oct 2022 12:04:49 GMT, Aleksey Shipilev wrote:
> Found this when working on JOL support
> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
> would fail with:
>
>
> Caused by:
On Wed, 19 Oct 2022 19:42:31 GMT, Chris Plummer wrote:
>> Found this when working on JOL support
>> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
>> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
>> would fail with:
>>
>>
>> Caused
On Wed, 19 Oct 2022 12:04:49 GMT, Aleksey Shipilev wrote:
> Found this when working on JOL support
> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
> would fail with:
>
>
> Caused by:
On Wed, 19 Oct 2022 12:04:49 GMT, Aleksey Shipilev wrote:
> Found this when working on JOL support
> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
> would fail with:
>
>
> Caused by:
On Wed, 19 Oct 2022 13:09:59 GMT, Aleksey Shipilev wrote:
> Looks fine. Curious, how well are all these alignments tested? E.g. do they
> work with all GCs?
We have gc/TestObjectAlignment.java. I don't see any indication that this test
is not run with all GCs.
-
PR: https://git.o
On Wed, 19 Oct 2022 13:03:53 GMT, Thomas Stuefe wrote:
> Looks fine. Curious, how well are all these alignments tested? E.g. do they
> work with all GCs?
I tested `serviceability/sa` with default (G1) and all acceptable alignments,
seems to work fine. We can run the tests with other GCs, if an
On Wed, 19 Oct 2022 12:04:49 GMT, Aleksey Shipilev wrote:
> Found this when working on JOL support
> ([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
> you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
> would fail with:
>
>
> Caused by:
Found this when working on JOL support
([CODETOOLS-7903364](https://bugs.openjdk.org/browse/CODETOOLS-7903364)). If
you try to attach to VM running with -XX:ObjectAlignmentInBytes=32, then SA
would fail with:
Caused by: java.lang.RuntimeException: Object alignment 32 not yet supported
at jdk.h
17 matches
Mail list logo
