RFR: Backport bugfixes to jdk14

[Aleksey Shipilev]

On 1/8/20 4:58 PM, Roman Kennke wrote:
>> This looks fine. The process caveat is that issues should be at least P3 after RDP1, which condition
>> JDK-8234974 fails.
> 
> Would it be justified to raise it to P3?

Not sure, because I cannot gauge the impact of this thing. Skipping class unloading on shortcut
cycles is a regression vs jdk13, right? We used to do this in Final Mark pause, and that would do
unloading even with shortcutting? If so, this should be treated as regression and fixed in 14.

-- 
Thanks,
-Aleksey