review request (1 line) 7162955: Attach api on Solaris, too many open files

Karen Kinnear karen.kinnear at oracle.com
Thu Apr 26 06:39:17 PDT 2012


Dean,

Your suggested comment looks great - thank you.
Adding the fix now and the test later is fine.

thanks,
Karen

On Apr 25, 2012, at 10:18 PM, dean long wrote:

> Thanks Dan and Karen for the reviews.
> 
> Karen, I will add a comment along the lines of:
> 
>      // DOOR_RELEASE flag makes sure fd is closed after passing it to
>      // the client.  See door_return(3DOOR) man page.
> 
> or whatever you suggest.
> 
> There is a test in the bug report, but it has to be run manually, so you can check /proc/<pid>/fd for leaks.
> I can try to come up with an automated test that would do attach/detach repeatedly, but the hard part is knowing when to stop, so it would require some native or script code to get or set the process fd limit.  Adding a test will require another round a reviews, so I'm wondering if I can push the fix now and file a new CR to add a test.
> 
> dl
> 
> On 4/25/2012 12:54 PM, Karen Kinnear wrote:
>> Dean,
>> 
>> Code looks good. Any chance you could add a brief comment on how that flag
>> causes the passed reference to be released?
>> 
>> Is there any way to check in a test for this with the fix?
>> 
>> thanks,
>> Karen
>> 
>> On Apr 25, 2012, at 3:39 PM, dean long wrote:
>> 
>>> Thanks for the reviews, Dmitry and Staffan.  I believe I need one more from a Reviewer.
>>> 
>>> dl
>>> 
>>> On 4/25/2012 1:14 AM, Dmitry Samersoff wrote:
>>>> Dean,
>>>> 
>>>> Looks good for me.
>>>> 
>>>> -Dmitry
>>>> 
>>>> 
>>>> On 2012-04-25 04:53, dean long wrote:
>>>>> http://cr.openjdk.java.net/~dlong/7162955/webrev/
>>>>> Summary of changes: 1 line changed: 0 ins; 0 del; 1 mod; 702 unchg
>>>>> 
>>>>> Release server-side socket after client receives it.
>>>>> 
>>>>> dl



More information about the hotspot-runtime-dev mailing list