[crac] RFR: 8350088: [CRaC] Stop & restart JFR recording on checkpoint [v4]

[Timofei Pushkin]

On Tue, 18 Feb 2025 13:09:30 GMT, Radim Vansa <rvansa at openjdk.org> wrote:

>> src/hotspot/share/jfr/jfr.cpp line 179:
>> 
>>> 177:     // trying to register new file descriptors. Instead we just record a request and
>>> 178:     // the recording will be started at the right moment from JDKResource.
>>> 179:     JfrUpcalls::request_start_after_restore(JavaThread::current());
>> 
>> If `request_start_after_restore` remains trapping (see my other comment in its body), `Jfr::after_restore` should also be marked as such so that the caller would do proper exception handling.
>
> I am not sure what you mean by that - the implementation is not propagating any exception, and there's nothing to be done on the caller side. See `check_internal_types()` for reference.

Currently `request_start_after_restore` can propagate an exception if its call to `SystemDictionary::resolve_or_fail` fails. But here this propagation will just silently be ignored (and not even cleared).

Since we've decided that `request_start_after_restore` won't propagate its `call_static` exception I think it should log and clear possible exceptions from `SystemDictionary::resolve_or_fail` and its `TRAPS` mark should be removed.

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

PR Review Comment: https://git.openjdk.org/crac/pull/203#discussion_r1959900646