RFR: 8334254: Cleanup CGroups initialization code [v2]

[Johan Sjölen]

On Fri, 28 Jun 2024 11:23:59 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:

>> Johan Sjölen has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Review comments
>
> src/hotspot/os/linux/cgroupSubsystem_linux.cpp line 92:
> 
>> 90:    */
>> 91:   assert(is_cgroup_v1(&cg_type_flags), "Cgroup v1 expected");
>> 92:   // The following are required to exist, determine_type will fail and return false otherwise.
> 
> I don't fully follow here. This is changed behavior compared to before?

The behavior is the same. The array `cg_infos` has the memory controller at  the `MEMORY_IDX`, the pids at `PIDS_IDX` and so on, so we don't need a string comparison.

> src/hotspot/os/linux/cgroupSubsystem_linux.cpp line 111:
> 
>> 109:   } else {
>> 110:     log_debug(os, container)("CgroupInfo for %s not complete", cg_controller_name[PIDS_IDX]);
>> 111:   }
> 
> ?? I don't get it. If data_complete is false, we just leak these objects?

That doesn't seem right, we pass them on to `CgroupV1Subsystem` at line 113, I just don't think it's visible in the review as that line is unchanged.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/19709#discussion_r1658612766
PR Review Comment: https://git.openjdk.org/jdk/pull/19709#discussion_r1658603240