-XX:Inline ignored for system classes?
Vitaly Davidovich
vitalyd at gmail.com
Sun Jan 22 17:19:54 UTC 2017
Yes, I believe it's because final instance fields aren't trusted (at the
moment) and interfaces don't receive "single impl" optimization (direct
call + class loading guard) like abstract classes.
On Sun, Jan 22, 2017 at 12:03 PM Peter Veentjer <alarmnummer at gmail.com>
wrote:
Thanks for your answers Andrew. The information you provide is highly
appreciated.
On Sun, Jan 22, 2017 at 6:59 PM, Andrew Haley <aph at redhat.com> wrote:
Actually, no, you could safepoint every thread and forcibly deoptimize
any methods you found which were referring to the class in question.
Sorry. But that's not a lightweight action at all, and in many cases
it would be excessive.
Loading a new class that requires code deoptimization is not any different,
or is it?
Ideally this doesn't happen to often and in case of it not happening at all
(
so no reflection, no override violations) there is no price to pay.
My understanding is correct that the only reason for this typeguard is the
existence of reflection?
Andrew.
--
Sent from my phone
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20170122/bb01c328/attachment.html>
More information about the hotspot-compiler-dev
mailing list