adding rsockets support into JDK
yingqi.lu at intel.com
Wed Nov 21 19:23:44 UTC 2018
We can certainly remove IPv6 support for now if that makes the implementation cleaner. rsocket will be created with AF_INET regardless if IPv6 is supported or not.
I will merge this change together with the suggestions from Chris's email this morning together into version 19.
>From: Alan Bateman [mailto:Alan.Bateman at oracle.com]
>Sent: Wednesday, November 21, 2018 9:50 AM
>To: Lu, Yingqi <yingqi.lu at intel.com>; Chris Hegarty
><chris.hegarty at oracle.com>
>Cc: Brian Burkhalter <brian.burkhalter at oracle.com>; Viswanathan, Sandhya
><sandhya.viswanathan at intel.com>; Aundhe, Shirish
><shirish.aundhe at intel.com>; nio-dev at openjdk.java.net; Kaczmarek, Eric
><eric.kaczmarek at intel.com>
>Subject: Re: adding rsockets support into JDK
>On 20/11/2018 17:08, Lu, Yingqi wrote:
>> The IPv6 addresses work with rsocket when both server and client use IPv6
>addresses. However, the IPv4-mapped addresses does not work. If IPv6 is
>enabled on RDMA adaptors (which I believe it does in majority of the cases),
>current JDK uses IPv6 by default. Then -Djava.net.preferIPv4Stack=true is
>needed if applications use IPv4 addresses.
>I see we originally discussed this issue here in April and at the time you did say
>that rsockets doesn't support IPv4-mapped IPv6 addresses. Do you know if
>there are plans to fix that issue? Requiring -Djava.net.preferIPv4Stack=true
>on the command line will make this feature awkward as it impacts all sockets
>and it may be better to just forgo support for IPv6 addresses in the short term
>(meaning create a AF_INET socket even if IPv6 is enabled). The alternative is
>of course to specify the protocol factory at creation time as Chris suggests but
>that also puts a tax on all usages too.
More information about the nio-dev