On Tue, 14 Dec 2021 06:57:36 GMT, Chris Plummer <cjplum...@openjdk.org> wrote:
> SA had a bunch of vtable related code that searched for the vtable pointer of > a hotspot object in a few places. This was all related to weird ways that > Solaris would store the vtable pointer. In the case of > https://bugs.openjdk.java.net/browse/JDK-8269830, this extra vtable searching > code introduced a bug in BasicTypeDataBase.addressTypeIsEqualToType(). Since > we don't support Solaris anymore, this the bug was fixed by removing this > Solaris support. At the time I fixed JDK-8269830, I noticed that > BasicTypeDataBase.findDynamicTypeForAddress() had similar logic for vtable > searching, although didn't seemed to be causing any bugs at the time. I filed > JDK-8269838 to eventually clean it up, and that's what this PR is doing. First word of object is the only sensible place for the vtable pointer... 8-) ------------- Marked as reviewed by kevinw (Committer). PR: https://git.openjdk.java.net/jdk/pull/6827
