RFR: 8311993: Test serviceability/sa/UniqueVtableTest.java failed: duplicate vtables detected [v2]

[Alex Menkov]

On Tue, 27 Aug 2024 05:18:27 GMT, Chris Plummer <cjplummer at openjdk.org> wrote:

> I found this:
> 
> ```
>     // should we parse DLL symbol table in Java code or use
>     // Windbg's native lookup facility? By default, we use
>     // native lookup so that we can take advantage of '.pdb'
>     // files, if available.
>     useNativeLookup = true;
>     String str = System.getProperty("sun.jvm.hotspot.debugger.windbg.disableNativeLookup");
>     if (str != null) {
>       useNativeLookup = false;
>     }
> ```
> 
> I'm not sure what is meant by "take advantage of '.pbp' files". Is it perhaps a more reliable or complete database of symbols, or perhaps it is faster? In any case, I'd be interested in seeing if all our tests still pass when useNativeLookup is false.

As far as I understand .pdb files contain symbol information.
I run tier1,tier2,hs-tier5-svc with "-Dsun.jvm.hotspot.debugger.windbg.disableNativeLookup=yes" on windows-x64 and windows-x64-debug
Got 110 failures; all failures on debug build, most of them with NPE:


Error: java.lang.NullPointerException: Cannot invoke "sun.jvm.hotspot.debugger.Address.getJBooleanAt(long)" because "address" is null
java.lang.NullPointerException: Cannot invoke "sun.jvm.hotspot.debugger.Address.getJBooleanAt(long)" because "address" is null
	at jdk.hotspot.agent/sun.jvm.hotspot.runtime.VM.isSharingEnabled(VM.java:940)
	at jdk.hotspot.agent/sun.jvm.hotspot.types.basic.BasicTypeDataBase.findDynamicTypeForAddress(BasicTypeDataBase.java:240)
        ...

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

PR Comment: https://git.openjdk.org/jdk/pull/20684#issuecomment-2313681655