RFR(M): 8189113: AARCH64: StringLatin1 inflate intrinsic doesn't use prefetch instruction

[Andrew Haley]

On 05/15/2018 04:07 PM, Dmitrij Pochepko wrote:
> 
> 
> On 15.05.2018 17:33, Andrew Haley wrote:
>> On 05/11/2018 07:14 PM, Dmitrij Pochepko wrote:
>>> please review patch for 8189113: AARCH64: StringLatin1 inflate intrinsic
>>> doesn't use prefetch instruction
>> With this patch applied, I get jmh failures:
>>
>> WARNING: Unknown method to forward: iterationResultorg.openjrg.openjnfra.Bennfra.Benrams,orgrams,org.jmh.inf.jmh.inftionParams,int,org.openjdk.jmh.results.IterationResult,
>> # Warmup Iteration   2: WARNING: Unknown method to forward: iterationResultorg.openjrg.openjnfra.Bennfra.Benrams,orgrams,org.jmh.inf.jmh.inftionParams,int,org.openjdk.jmh.results.IterationResult,
>> # Warmup Iteration   3: WARNING: Unknown method to forward: iterationResultorg.openjrg.openjnfra.Bennfra.Benrams,orgrams,org.jmh.inf.jmh.inftionParams,int,org.openjdk.jmh.results.IterationResult,
>> ...
>>
>> When I run jmh with -XX:DisableIntrinsic=_inflateStringC it's fine.
>> I think your patch is broken.
>>
> I can't reproduce it. Did you use outdated workspace without this patch: 
> http://hg.openjdk.java.net/jdk/jdk/rev/a2322c683d17 ?
> Patched string inflate intrinsic is using zip2 instruction, which was 
> incorrectly encoded before this patch.

Ah, okay, that's a relief.  I must have missed that update.

-- 
Andrew Haley
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671