RFR: JDK-8176520: Improve the accuracy of the instance size in hprof heap dumps [v3]

[Serguei Spitsyn]

On Sat, 17 Feb 2024 02:41:20 GMT, Alex Menkov <amenkov at openjdk.org> wrote:

>> The fix updates heap dumpers to report correct instance size value for HPROF_GC_CLASS_DUMP records (currently it's reported as size of all instance fields)
>> 
>> Testing: tier1, tier2, tier5-svc
>
> Alex Menkov has updated the pull request incrementally with one additional commit since the last revision:
> 
>   split test

The fix looks good in general.
I've posted a question about the instance size verification.
It applies to the SA test as well.

test/hotspot/jtreg/serviceability/HeapDump/HeapDumpInstanceSize.java line 39:

> 37:  * @bug 8176520
> 38:  * @summary Test that heap dumper reports correct instance size in HPROF_GC_CLASS_DUMP records
> 39:  * @requires vm.jvmti

Q: Is the `vm.jvmti` really needed?

test/hotspot/jtreg/serviceability/HeapDump/HeapDumpInstanceSize.java line 88:

> 86:             } else {
> 87:                 // non-array should have >0 instance size
> 88:                 if (instSize == 0) {

This check is not that strong. How were the real instance sizes verified?
Is it in the `HprofParser.parseAndVerify(fileDump)`?

test/hotspot/jtreg/serviceability/sa/HeapDumpInstanceSize.java line 41:

> 39:  * @bug 8176520
> 40:  * @summary Test that heap dumpers (VM and SA) report consistent instance size in HPROF_GC_CLASS_DUMP records
> 41:  * @requires vm.jvmti

Q: Why is the `vm.jvmti` needed?

-------------

PR Review: https://git.openjdk.org/jdk/pull/17855#pullrequestreview-1899212106
PR Review Comment: https://git.openjdk.org/jdk/pull/17855#discussion_r1501340467
PR Review Comment: https://git.openjdk.org/jdk/pull/17855#discussion_r1501342401
PR Review Comment: https://git.openjdk.org/jdk/pull/17855#discussion_r1501342661