FULL PRODUCT VERSION :
A DESCRIPTION OF THE PROBLEM :
Bug http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7130796 describes a BLOCK in nio selector on windows. This bug was incorrectly marked as a duplicate of http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6346658, which is a SPIN!
However 7130796 was correctly analysed and fixed by http://hg.openjdk.java.net/jdk8/tl/jdk/rev/7326971f09af
We have run jdk8 for sometime and it appears to confirm that the problem is fixed. The fixer asked for confirmation before backporting to jdk 7, but the issue had been closed so no further comments could be added.
So I opened http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7173207 to give the report that the fix worked. However the handler of that saw the incorrect link to 6346658 and just said that it had been fixed and back ported to 7.
However, if you look at the fix: http://hg.openjdk.java.net/jdk8/tl/jdk/rev/7326971f09af then at the jdk7 source at http://hg.openjdk.java.net/build-infra/jdk7/jdk/file/0f8da27a3ea3/src/windows/classes/sun/nio/ch/NativeThread.java or http://hg.openjdk.java.net/build-infra/jdk7/jdk/file/0f8da27a3ea3/src/share/classes/sun/nio/ch/NativeThreadSet.java you can see that it has not been backported - at least not to open jdk.
We continue to get jdk7 servers blocking on this issue (not spinning) on windows.
REGRESSION. Last worked in version 7
This bug can be reproduced always.