[aarch64-port-dev ] RFR [sh/8u]: Revert accidental OpSpinWait matching

[Aleksey Shipilev]

Now that the merge came in, ping?

-Aleksey

On 03/22/2018 12:06 PM, Aleksey Shipilev wrote:
> We have a weird difference between our sh/jdk8 development repo and aarch64-port-jdk8u-shenandoah
> integration forest: it has a little branch that "matches" OpSpinWait, even though there is no such
> node in 8u at all -- that feature came in JDK 9. This had accidentally leaked during previous merges.
> 
> So, we want to push this to aarch64-port/jdk8u-shenandoah:
> 
> diff -r 966a10251cb4 src/share/vm/adlc/formssel.cpp
> --- a/src/share/vm/adlc/formssel.cpp	Wed Mar 21 20:00:58 2018 +0000
> +++ b/src/share/vm/adlc/formssel.cpp	Thu Mar 22 12:03:44 2018 +0100
> @@ -4066,8 +4066,7 @@
>      !strcmp(_opType,"StoreFence") ||
>      !strcmp(_opType,"MemBarVolatile") ||
>      !strcmp(_opType,"MemBarCPUOrder") ||
> -    !strcmp(_opType,"MemBarStoreStore") ||
> -    !strcmp(_opType, "OnSpinWait");
> +    !strcmp(_opType,"MemBarStoreStore");
>  }
> 
>  bool MatchRule::is_ideal_loadPC() const {
> 
> Testing: x86_64 build, hotspot_gc_shenandoah
> 
> Thanks,
> -Aleksey
>