RFR: 8317951: Refactor loading of zip library to help resolve JDK-8315220 [v3]
Markus Grönlund
mgronlun at openjdk.org
Thu Oct 26 09:20:32 UTC 2023
On Wed, 25 Oct 2023 21:56:39 GMT, David Holmes <dholmes at openjdk.org> wrote:
>> It follows that now, with the introduction of the vm_exit construct.
>
> Sorry I am having a hard time seeing the equivalence. The original code doesn't potentially exit after a failed dll_lookup AFAICS - only a failed dl_load does that.
The original code in ClassLoader does not check that the resolved method entry points through dll_lookup are found or resolved correctly. As there are no null checks before using the function pointers, it would instantly crash if the entry points were not resolved correctly. This is semantically equivalent.
The only difference is that ClassLoader, should it be the initializer of the library, also attempts to resolve the init_params() and compress() entry points (used by HeapDumpCompressor).
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/16149#discussion_r1372851655
More information about the hotspot-runtime-dev
mailing list