Request for approval: 6929067: Stack guard pages should be removed when thread is detached

[Coleen Phillimore]

Great!  Thanks for reporting and diagnosing this.  I was a bit mystified 
because I thought I'd tested it, but after a while the VM just started 
getting "Bus error".  Including java -version.  Do the files stay open 
after the process has exited?  Why did it do this?
thanks,
Coleen

Andreas Kohn wrote:
>
>
> On 18.03.2010, at 18:35, Coleen Phillimore - Sun Microsystems 
> <Coleen.Phillimore at Sun.COM <mailto:Coleen.Phillimore at Sun.COM>> wrote:
>
>>
>> I filed bug  6936168 and have a repository with 3 additional 
>> fcloses() in them for the returns after the file is opened.  See:
>>
>> open webrev at http://cr.openjdk.java.net/~coleenp/6936168/ 
>> <http://cr.openjdk.java.net/%7Ecoleenp/6936168/>
>
> With this change everything seems to work properly again :)
>
> Thanks,
> --
> Andreas
>>
>> Coleen
>>
>> On 03/18/10 06:05, Andrew Haley wrote:
>>> On 03/18/2010 09:10 AM, Andreas Kohn wrote:
>>>   
>>>> On Fri, 2010-03-12 at 09:44 +0000, Andrew Haley wrote:
>>>>     
>>>>> On 03/11/2010 09:06 PM, Coleen Phillimore wrote:
>>>>>       
>>>>>> I've added the test to the changeset and a script to run in our harness.
>>>>>>
>>>>>> Also in os_linux.cpp, I changed the SYS_gettid call to go through our
>>>>>> os::Linux::gettid() because on at least one linux, syscall() returns a
>>>>>> long int which gets a compilation warning with %d.
>>>>>>
>>>>>> open webrev at http://cr.openjdk.java.net/~coleenp/6929067/ <http://cr.openjdk.java.net/%7Ecoleenp/6929067/>
>>>>>> bug link at http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6929067
>>>>>>
>>>>>> Andrew, please have a look since you're the contributor.
>>>>>>         
>>>>> That's OK, but you don't need SYS_gettid.
>>>>> Please look at http://cr.openjdk.java.net/~aph/6929067-jdk7-webrev-4/hotspot.patch <http://cr.openjdk.java.net/%7Eaph/6929067-jdk7-webrev-4/hotspot.patch>
>>>>> I changed to "/proc/self/maps", as you requested.  I think this is better.
>>>>>
>>>>> The copy of my webrev to cr.openjdk.java.net <http://cr.openjdk.java.net> failed for some reason
>>>>> I don't understand.
>>>>>       
>>>> With this change I seem to hit the limit on the number of open files.
>>>> Looking through it, shouldn't get_stack_bounds() close the FILE* it
>>>> opened?
>>>>     
>>>
>>> Oh, how stupid of me!  If this were gcc I'd just push a fix
>>> immediately as obvious/trivial, but I think we need a bug opened to
>>> push the change.
>>>
>>> (BTW, this happened because of a mistake translating the patch I wrote
>>> from using the C++ library into C.  The original patch used an
>>> fstream, whose destructor closes the file.  When I did the translation
>>> I missed the fact that I had to close the file manually.)
>>>
>>> Andrew.
>>>