RFR: 8266369: (se) Add wepoll based Selector [v6]

[Alan Bateman]

On Thu, 6 May 2021 15:51:36 GMT, Daniel Fuchs <dfuchs at openjdk.org> wrote:

>> Alan Bateman has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains 11 additional commits since the last revision:
>> 
>>  - Merge
>>  - Bring back getDescriptor, bzero poll array when initially created
>>  - epoll_create1 returns NULL on error
>>  - Remove unused WEPoll.getDescriptor. Update micro to close listener
>>  - Fix copyright header on SelectOne benchmark
>>  - Merge
>>  - Propagate exception from discardOOB, cleanup
>>  - Merge
>>  - Use 'to' instead of '(int) timeout
>>  - Move SelectorWakeup micro to channels directory
>>  - ... and 1 more: https://git.openjdk.java.net/jdk/compare/b40c5cb7...260b13fe
>
> test/micro/org/openjdk/bench/java/nio/channels/SelectOne.java line 71:
> 
>> 69: 
>> 70:         try (ServerSocketChannel listener = ServerSocketChannel.open()) {
>> 71:             listener.bind(new InetSocketAddress(0));
> 
> Is using the wildcard really necessary here? I am concerned that the test is relying on unspecified behaviour - namely:
>     - listener is bound to the wildcard
>     - listener.getLocalAddress() is *probably* the wildcard
>     - later SocketChannel is connected using that address which is still *probably* the wildcard.
> For the sake of stability - should the test preferably use InetAddress.getLoopbackAddress() for binding and connecting?

This is micro rather than a jtreg test so only suitable for running in very controlled environments. I can of course change it to use the loopback address. The original benchmark did that, but I had to strip it down for openjdk/jdk.

-------------

PR: https://git.openjdk.java.net/jdk/pull/3816