RFR (S/M): 8169703: G1 crashes with guarantee(pretouch_gang != NULL) failed: No pretouch gang specified
Dmitry Dmitriev
dmitry.dmitriev at oracle.com
Mon Nov 21 13:51:21 UTC 2016
Hi Thomas,
Test's looks good to me!
Thanks,
Dmitry
On 21.11.2016 16:19, Thomas Schatzl wrote:
> Hi everyone,
>
> thanks for your reviews.
>
> Webrev incorporating your comments available at
> http://cr.openjdk.java.net/~tschatzl/8169703/webrev.0_to_1/
> http://cr.openjdk.java.net/~tschatzl/8169703/webrev.1/
>
> One comment inlined further down:
>
> On Fri, 2016-11-18 at 23:31 +0300, Dmitry Fazunenko wrote:
>> Hi Thomas,
>>
>> A few more comments about the tests:
>>
>> 1) would you fix indention to meet "4-space" rule for java.
>> 2) TestParallelAlwaysPreTouch.java
>>
>> a)
>> @run main ...
>> -->
>> @run main/othervm ...
>>
>> b) You start one class, but declare another:
>> * @run main -Xms10M -Xmx100m ... TestAlwaysPreTouch
>> public class Test2GbHeap {
>> ...
>> }
>>
>> c) Would it make sense to run a group of existing tests with "-
>> XX:+AlwaysPreTouch" regularly (nightly or weekly)
> I think the new tests cover all (CDS allocation and mapping, startup
> heap expansion, concurrent allocation) but one use case, that is pre-
> touch during allocation at the end of GC.
>
> I could try to add a test that particularly tries to test this one, but
> that is a bit tricky (to exactly make this allocation that expands the
> heap occur at end of gc).
>
> So I do not think anything special is necessary.
>
> Thanks,
> Thomas
>
More information about the hotspot-gc-dev
mailing list