[aarch64-port-dev ] RFR: 8168909 aarch3264: arm32/64 port contribution cleanups

[Bob Vandette]

> On Nov 1, 2016, at 2:27 PM, Edward Nevill <edward.nevill at gmail.com> wrote:
> 
> Hi Bob,
> 
> 
> On Fri, 2016-10-28 at 11:57 -0400, Bob Vandette wrote:
>> Please review these cleanups of the contributed ARM code.
>> 
>> Let me know what you think of the proposed new configure option for selecting the 64-bit
>> aarch64 port sources.
>> 
>> http://cr.openjdk.java.net/~bobv/8168909/webrev <http://cr.openjdk.java.net/~bobv/8168909/webrev>
> 
> The changes look fine.
> 
> It would be nice to have a list of options for --with-cpu-port
> 
> IE. currently the help reads
> 
>   --with-cpu-port         specify sources to use for Hotspot 64-bit ARM port
>                           (for aarch64)
> 
> It would be better if it read
> 
>   --with-cpu-port         specify sources to use for Hotspot 64-bit ARM port
>                           (arm64, aarch64) [aarch64]
> 
> This shows the options available and the default in the same way, as for example, the --with-jvm-variants. IE
> 
>   --with-jvm-variants     JVM variants (separated by commas) to build
>                           (server,client,minimal,core,zero,zeroshark,custom)
>                           [server]
> 
> This saves having to grub around in hotspot.m4 to find out what the
> available options are.
> 
> The same is true of --with-abi-profile.

Sure, I’ll update the help text.

> 
> BTW: The patch to fix the monitor locking problem encountered in jtreg
> does not seem to have been pushed?
> 
> diff -r e0c5f8cf07f8 src/share/vm/runtime/sharedRuntime.cpp
> --- a/src/share/vm/runtime/sharedRuntime.cpp	Mon Oct 10 09:29:51
> 2016 -0400
> +++ b/src/share/vm/runtime/sharedRuntime.cpp	Sun Oct 30 21:50:31
> 2016 +0000
> @@ -1983,7 +1983,9 @@
>  JRT_BLOCK_ENTRY(void,
> SharedRuntime::complete_monitor_locking_C(oopDesc* _obj, BasicLock*
> lock, JavaThread* thread))
>    // Disable ObjectSynchronizer::quick_enter() in default config
>    // on AARCH64 until JDK-8153107 is resolved.
> -  if (AARCH64_ONLY((SyncFlags & 256) != 0 &&)
> !SafepointSynchronize::is_synchronizing()) {
> +  if (ARM_ONLY((SyncFlags & 256) != 0 &&) 
> +      AARCH64_ONLY((SyncFlags & 256) != 0 &&) 
> +      !SafepointSynchronize::is_synchronizing()) {
>      // Only try quick_enter() if we're not trying to reach a safepoint
>      // so that the calling thread reaches the safepoint more quickly.
>      if (ObjectSynchronizer::quick_enter(_obj, thread, lock)) return;
> 
I was hoping that this change would be pushed directly to JDK 9 and I’d
bring it into jdk9-arm3264 eventually.  There’s been a bit of discussion on this bug since
it doesn’t fix the root problem, it just disables an optimization for ARM.

Bob.


> 
> All the best,
> Ed.
>