[aarch64-port-dev ] RFR: Backports to aarch64/jdk8u

[Yangfei (Felix)]

Hi,

    Webrev: http://cr.openjdk.java.net/~fyang/8u-backports/webrev.00/
    This backport the following patches to the aarch64-jdk8u & aarch64-jdk8u-shenandoah branches:

8191955: AArch64: incorrect prefetch distance causes an internal error
Summary: Fixed incorrect prefetch distance on AArch64.
Reviewed-by: aph

8194256: AARCH64: SIMD shift instructions are incorrectly encoded
Reviewed-by: aph

8195588: AArch64: Codegen fix after JDK-8194256
Summary: Update shift constant usage to align with the assembler change.
Reviewed-by: aph

8202186: AArch64: Debug build VM crashes with PrintC1Statistics option
Reviewed-by: aph

8205421: AARCH64: StubCodeMark should be placed after alignment
Reviewed-by: aph

8209414: AArch64: method handle invocation does not respect JVMTI interp_only mode
Reviewed-by: adinn
Contributed-by: nick.gasson at arm.com

8213134: AArch64: vector shift failed with MaxVectorSize=8
Summary: add vshiftcnt instructions for vector64 and add vsra/vsrl instructions to AArch64 backend. To detect shift failures, MaxVectorSize options are added to jtreg test cases.
Reviewed-by: aph, kvn


    Some of these fix JVM crashes in certain cases, while others just solve some minor issues. 
    Passed JTreg test using a aarch64 fastdebug JDK built from aarch64-jdk8u-shenandoah branch. 
    OK to backport? 

Thanks,
Felix