Request for review 7188233: Deprecate the UseVMInterruptibleIO for JDK8

harold seigel harold.seigel at oracle.com
Mon Oct 15 12:02:56 PDT 2012


Hi Alan,

Bug report 7188233 says the following:

    Deprecate UseVMInterruptibleIO flag.

    JDK7 changed the default value from true to false.
    See 4385444 for long-term plan to gradually deprecate this flag.

    *JDK8: leave code but add flag to deprecated list, default false.*
    JDK9: remove code support

  Hence, I did not totally remove the flag.  However, I can completely 
deprecate the flag entirely if that is a better solution?

Thanks, Harold


On 10/15/2012 2:57 PM, Alan Bateman wrote:
> On 15/10/2012 19:44, harold seigel wrote:
>> Summary:  The -XX:+UseVMInterruptibleIO flag is deprecated for JDK8.  
>> The flag will still enable Interruptible IO on Solaris, but users 
>> will get a warning.
>>
>> open webrev at http://cr.openjdk.java.net/~coleenp/Bug7188233 
>> <http://cr.openjdk.java.net/%7Ecoleenp/Bug7188233>
>>
>> bug link at http://bugs.sun.com/view_bug.do?bug_id=7188233
>>
>> Tested with JCK, JPRT, and tests on Solaris that behave differently 
>> depending on the setting of the UseVMInterruptibleIO flag.
>>
>> Thanks, Harold
> The proposed change looks okay to but if my memory is correct then the 
> original plan (from many years ago) was to finally wield the axe and 
> remove the code in jdk8. This plan started way back in jdk6 when 
> UseVMInterruptibleIO was added, true by default. It was documented at 
> the time to give developers a chance to try it out and see if there 
> was any impact of their applications on Solaris. Then in jdk7 the 
> default was changed to false and the plan was to eventually get rid of 
> it in jdk8. "Getting rid of it" is partly library work where we need 
> to replace the calls to JVM_* with the appropriate syscall function. 
> That will clean things up in a few areas as we've only been using the 
> JVM_* functions because of the legacy interruptible I/O support on 
> Solaris. We haven't got to this clean-up yet, hopefully soon.
>
> -Alan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/hotspot-runtime-dev/attachments/20121015/734acb6b/attachment.html 


More information about the hotspot-runtime-dev mailing list