RFR 8173227: [JVMCI] HotSpotJVMCIMetaAccessContext.fromClass is inefficient

[Tom Rodriguez]

I added a comment explaining the purpose of the loop and updated the 
webrev in place.

tom

Doug Simon wrote:
> +        while (javaType == null) {
> +            WeakReference<ResolvedJavaType>  type = resolvedJavaType.get(javaClass);
> +            javaType = type.get();
> +            if (javaType == null) {
> +                resolvedJavaType.remove(javaClass);
> +            }
>
> Maybe add a small comment as to why the loop is necessary.
>
> Otherwise, looks good.
>
>> On 24 Jan 2017, at 09:50, Tom Rodriguez<tom.rodriguez at oracle.com>  wrote:
>>
>> http://cr.openjdk.java.net/~never/8173227/webrev
>> https://bugs.openjdk.java.net/browse/JDK-8173227
>>
>> HotSpotJVMCIMetaAccessContext.createClass is a synchronized wrapper around a WeakHashMap with WeakReference values. During bytecode parsing this can be a heavy contention point between between multiple threads. Additionally the use of WeakReferences for the keys and values is expensive. A ClassValue<WeakReference<ResolvedJavaType>>  would avoid the contention around the synchronized method and would eliminate one of the two required WeakReferences.  Tested with Graal.
>>
>> tom
>