RFR: 8316967: Correct the scope of vmtimer in UnregisteredClasses::load_class [v2]

[Calvin Cheung]

On Fri, 13 Oct 2023 22:00:19 GMT, Jiangli Zhou <jiangli at openjdk.org> wrote:

>>> Changing the scope seems reasonable to me as well. Taking a close look of the current change after reading @iklam's comment above, I think it might be reasonable to change the `ClassLoader::perf_sys_class_lookup_time` name. Unregistered class loader is for user defined classloader support (continue paging back my memory for this specifically :)). Using `sys` could be confusing.
>> 
>> Thanks for the review and sorry for the delay in replying as I took a few days off.
>> How about changing `ClassLoader::perf_sys_class_lookup_time()` to `ClassLoader::perf_app_class_lookup_time()`?
>> Or simply using the existing `ClassLoader::perf_app_classload_time()`? (Which is being used in https://github.com/openjdk/jdk/blob/master/src/hotspot/share/classfile/systemDictionary.cpp#L1289-#L1299)
>
>> Thanks for the review and sorry for the delay in replying as I took a few days off.
>> How about changing `ClassLoader::perf_sys_class_lookup_time()` to `ClassLoader::perf_app_class_lookup_time()`?
>> Or simply using the existing `ClassLoader::perf_app_classload_time()`? (Which is being used in https://github.com/openjdk/jdk/blob/master/src/hotspot/share/classfile/systemDictionary.cpp#L1289-#L1299)
> 
> `ClassLoader::perf_app_classload_time` seems ok. If there is any interest in differentiating the class loading time for user defined classloader and archived cases, `perf_shared_unregistered_classload_time` might be more specific.

@jianglizhou, @iklam Please re-review the latest version. Thanks!

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

PR Comment: https://git.openjdk.org/jdk/pull/16046#issuecomment-1766975885