Proposal for adding O_DIRECT support into JDK 9

[Lu, Yingqi]

Hi Brian/Alan,

Thank you very much for all the help! I will apply your patch and send another version tomorrow. At the meantime, please let me know if you see any additional issues.

All, please let us know if you have any comments/feedback in this patch.

Thank you!
Lucy

>-----Original Message-----
>From: Brian Burkhalter [mailto:brian.burkhalter at oracle.com]
>Sent: Monday, September 18, 2017 6:08 PM
>To: Lu, Yingqi <yingqi.lu at intel.com>
>Cc: nio-dev at openjdk.java.net
>Subject: Re: Proposal for adding O_DIRECT support into JDK 9
>
>Hi Lucy,
>
>I ran the tests manually on Linux and macOS and did not see any problems. I ran
>your script for 5000 iterations of DirectIOTest on Solaris and did not see any
>problems. I still need to run the other four tests on Solaris as they do not really
>execute in the test harness which is on ZFS.
>
>I created a small patch http://cr.openjdk.java.net/~bpb/8164900/webrev.15-fix/
>which has a couple of minor cleanups in DirectIOTest.java and trailing whitespace
>fixes in the other files in the this patch. You should be able to apply it directly.
>
>It looks like you fixed the things Alan commented on, viz., the need to do the
>position check in FileChannelImpl with the position lock held, and the use of rem
>instead of dst.remaining() in IOUtil::read. I don't myself see any other problems in
>the code but we shall see what Alan and others think.
>
>I intend to look over the read/write tests one more time.
>
>Thanks,
>
>Brian