RFR(S): 8170767: Zero fastdebug build triggers assertion
jesper.wilhelmsson at oracle.com
Tue Dec 13 10:50:37 UTC 2016
This is a change in shared VM code. It should go through jprt.
(Please do not interpret this mail as an approval to push. jdk9/hs is still closed.)
Den 13/12/16 kl. 10:38, skrev Severin Gehwolf:
> On Wed, 2016-12-07 at 22:58 -0500, Coleen Phillimore wrote:
>> This looks good and I can sponsor this for you. I haven't built Zero in
>> a while.
> Hi Coleen,
> Is it OK for me to push this change? I'm not entirely clear on which
> changes need to go through JPRT. In theory I should be able to push
> this change with my new committer powers ;-)
>> On 12/7/16 7:57 AM, Severin Gehwolf wrote:
>>> Could somebody please review this Zero only change? The problem was
>>> there before, but only the fix for JDK-8169711 brought it to the
>>> surface. The returned method kind for
>>> java_lang_math_fmaD/java_lang_math_fmaF is no longer guarded by UseFMA
>>> in abstractInterpreter.cpp
>>> The fix is to generate stub entries for
>>> java_lang_math_fmaD/java_lang_math_fmaF which were missing. Thoughts?
>>> bug: https://bugs.openjdk.java.net/browse/JDK-8170767
>>> webrev: http://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-8170767/webrev.01/
More information about the hotspot-dev