RFR: merge of 8203197
Roman Kennke
rkennke at redhat.com
Thu Jun 28 20:42:34 UTC 2018
I cherry-picked:
8204941: Refactor TemplateTable::_new to use MacroAssembler helpers for
tlab and eden
Which means you could go straight to:
hg pull http://hg.openjdk.java.net/jdk/jdk -r fd430e352427
hg merge
hg resolve -l |grep ^U
<fix merge issues>
hg commit -m"Merge"
hg push
This will also pull in a bunch of unrelated changesets, but I don't
think any of the others will cause any issues.
Thanks, Roman
> Hi Roland,
>
> with your changes, I am still getting merge conflicts when pulling from
> upstream. Could you do the following:
>
> hg pull http://hg.openjdk.java.net/jdk/jdk -r 5209d8a6303e
>
> This will pull in the conflicting change. Then:
> hg merge
>
> This will result in a bunch of conflicts:
>
> $ hg resolve -l |grep ^U
> U src/hotspot/share/opto/loopPredicate.cpp
> U src/hotspot/share/opto/loopnode.cpp
>
> Then please resolve those conflicts (I have no idea how), then:
>
> hg resolve -ma # mark conflicts as solved
> hg commit -m"Merge" # commit merge changeset
> hg push # pushes upstream changeset + merge
>
>
> There's another C2 conflict waiting here:
>
> changeset: 50632:fd430e352427
> user: thartmann
> date: Tue Jun 19 12:25:42 2018 +0200
> files: src/hotspot/share/opto/loopPredicate.cpp
> src/hotspot/share/opto/loopTransform.cpp
> src/hotspot/share/opto/loopnode.hpp
> test/hotspot/jtreg/compiler/loopopts/IterationSplitPredicateInconsistency.java
> test/hotspot/jtreg/compiler/loopopts/TestOverunrolling.java
> description:
> 8205033: [REDO] Induction variable of over-unrolled loop conflicts with
> range checks
> Summary: Update skeleton predicates before main loop during unrolling to
> remove dead code.
> Reviewed-by: kvn, roland
>
> Unfortunately, there's another changeset between the two C2 changesets
> that will disturb a straight pull+merge on your side:
>
> changeset: 50628:83b1a52f187c
> user: rkennke
> date: Tue Jun 19 11:27:05 2018 +0200
> files: src/hotspot/cpu/x86/templateTable_x86.cpp
> description:
> 8204941: Refactor TemplateTable::_new to use MacroAssembler helpers for
> tlab and eden
> Reviewed-by: kvn, coleenp
>
> I'll look into cherry-picking this, so that we can make progress faster.
>
> Thanks, Roman
>
>
>
>
>> The upstream change merged with the current shenandoah repo:
>>
>> http://cr.openjdk.java.net/~roland/shenandoah/backport-8203197/webrev.00/
>>
>> Some fixes to the upstream changes that I found are required for
>> shenandoah:
>>
>> http://cr.openjdk.java.net/~roland/shenandoah/fixes-8203197/webrev.00/
>>
>> And a collection of small fixes unrelated to the upstream but I found
>> while testing the change:
>>
>> http://cr.openjdk.java.net/~roland/shenandoah/various-c2-fixes/webrev.00/
>>
>> Roland.
>>
>
>
More information about the shenandoah-dev
mailing list