RFR: 8258032: Reconsider LEAF entry restrictions [v2]

[Coleen Phillimore]

On Wed, 6 Jan 2021 16:44:41 GMT, Daniel D. Daugherty <dcubed at openjdk.org> wrote:

>> Coleen Phillimore has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Fix copyrights.
>
> src/hotspot/share/c1/c1_Runtime1.cpp line 1311:
> 
>> 1309:   patch_code(thread, load_klass_patching_id);
>> 1310: 
>> 1311:   // Back in JAVA, use no oops DON'T safepoint
> 
> The baseline here is even more confused.
> debug_only NoHandleMark here on old L1311
> and a ResetNoHandleMark on old L1315.
> I see no reason for old L1311 so glad that it's gone!

Same here.  It was never a JRT_LEAF, which would have necessitated a ResetNoHandleMark.  It was a JRT_ENTRY at one point and then was made a direct call from _thread_in_Java code. That's when the ResetNoHandleMark was added, assuming that even without the NoHandleMark from JRT_LEAF, it was a bad to have Handles when the thread state is in Java.  This is the restriction that this patch removes.

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

PR: https://git.openjdk.java.net/jdk/pull/1846