RFR(S): 8150921: Update Unsafe getters/setters to use double-register variants

[John Rose]

On May 3, 2016, at 1:17 PM, Mikael Vidstedt <mikael.vidstedt at oracle.com> wrote:
> 
>> 
>> I actually wonder why those accessors (especially non-long) are
>> UNSAFE_ENTRY, not UNSAFE_LEAF, if "we can treat those as leaf routines"...
> 
> Ah, yes, now I see what you're saying. In general, since the mutex isn't actually checking for safepoints right now this isn't really a problem, but it would definitely be nice to unify it. How about I file a separate enhancement for moving the other memory operations into MemoryAccess?

+1

>> Another question to pile on: why only LongVolatile is treated specially,
>> but not DoubleVolatile too? Atomicity requirements are the same for both
>> long and double.
> 
> Very good question. I don't have the answer, but maybe somebody else knows.

Probably nobody was using the double version, so we didn't bother to put it in.

— John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20160503/ff48fc38/attachment.html>