RFR(S): 8132717: Add tests checking that instances of j.l.Classes of a large size are allocated as Humongous
Jon Masamitsu
jon.masamitsu at oracle.com
Mon Jan 4 23:31:19 UTC 2016
Kirill,
A couple of minor points.
http://cr.openjdk.java.net/~kzhaldyb/webrevs/JDK-8132717/webrev.06/test/gc/g1/humongousObjects/G1SampleClass.java.html
78 * @return a class of specified size loaded in specified class loader
Would this be a slightly more accurate comment?
78 * @return a class with instances of the specified size loaded in specified class loader
http://cr.openjdk.java.net/~kzhaldyb/webrevs/JDK-8132717/webrev.06/test/gc/g1/humongousObjects/TestHumongousNonArrayAllocation.java.html
The @run
64 * @run main/othervm -Xms128M -Xmx128M -XX:+UseG1GC -XX:+UnlockDiagnosticVMOptions -XX:+WhiteBoxAPI -Xbootclasspath/a:.
65 * -XX:G1HeapRegionSize=1M
66 * gc.g1.humongousObjects.TestHumongousNonArrayAllocation TWO_REGION_HUMONGOUS
67 *
68 * @run main/othervm -Xms128M -Xmx128M -XX:+UseG1GC -XX:+UnlockDiagnosticVMOptions -XX:+WhiteBoxAPI -Xbootclasspath/a:.
69 * -XX:G1HeapRegionSize=1M
70 * gc.g1.humongousObjects.TestHumongousNonArrayAllocation MORE_THAN_TWO_REGION_HUMONGOUS
says 2 region and more than 2 region humongous objects are being
created but the comments don't list those.
78 * the following cases:
79 * largest non-humongous object (exactly half a region)
80 * smallest humongous object (half a region + sizeof(long))
81 * humongous object that takes exactly one region
82 * the smallest humongous object that takes more than one region (region + sizeof(long)).
83 *
Since the @run pretty much describes what is being tested, maybe the comments
are not necessary? But they should be updated if kept.
Otherwise, looks good.
Reviewed.
Jon
On 01/04/2016 08:54 AM, Kirill Zhaldbybin wrote:
> Hi!
>
> Could you please review the following test?
>
> The test generates j.l.Classes with instances of different sizes,
> compiles, loads and instantiates them and then checks that instances
> of size larger than half of G1 region are allocated as Humongous.
>
> CR:
> https://bugs.openjdk.java.net/browse/JDK-8132717
>
> Webrev:
> http://cr.openjdk.java.net/~kzhaldyb/webrevs/JDK-8132717/webrev.06/
>
> Thank you.
>
> Regards, Kirill
More information about the hotspot-gc-dev
mailing list