RFR 8191102: Incorrect include file use in classLoader.hpp

coleen.phillimore at oracle.com coleen.phillimore at oracle.com
Fri Mar 2 19:10:25 UTC 2018


Hi Harold,
I think this change looks really good.  Thank you for doing this!
Coleen

On 3/2/18 9:03 AM, harold seigel wrote:
> Hi,
>
> Please review this JDK-11 change to fix JDK-8191102.
>
> The fix does the refactoring described in the bug by moving methods in 
> .hpp files that call OrderAccess::load_acquire() or 
> OrderAccess::release_store() to .inline.hpp files.
>
> Note that if a method in an .hpp file called a method that was moved 
> to an .inline.hpp file then it needed to be moved also.  For example, 
> method slot_at() was moved from constantPool.hpp to 
> constantPool.inline.hpp because it calls OrderAccess::load_acquire().  
> So, method is_pseudo_string_at() had to also be moved because it calls 
> slot_at().  And, pseudo_string_at() had to be moved because it calls 
> is_pseudo_string_at(), etc.  A couple of methods were moved from 
> constantPool.hpp to constantPool.cpp to reduce this fan out of changes.
>
> This change also contains a small unrelated cleanup for 
> classLoaderData.inline.hpp.
>
> Open Webrev: 
> http://cr.openjdk.java.net/~hseigel/bug_8191102/webrev/index.html
>
> JBS Bug:  https://bugs.openjdk.java.net/browse/JDK-8191102
>
> The change was tested with Mach5 tiers 1 and 2 tests and builds on all 
> Mach5 platforms and tiers 3-5 tests on Linux-X64.
>
> Thanks, Harold



More information about the hotspot-runtime-dev mailing list