RFR: 8243326: Cleanup use of volatile in taskqueue code

Kim Barrett kim.barrett at oracle.com
Wed Apr 29 11:16:40 UTC 2020


> On Apr 29, 2020, at 6:04 AM, Ivan Walulya <ivan.walulya at oracle.com> wrote:
> 
> Looks good.
> 
> Minor:
> Can you please modify some of the comments 
> 
> - “task queue.hpp:227 // Return true if the TaskQueue contains/does not contain any tasks.”
>   Not sure what the comment is supposed to convey
> 
> - “taskqueue.inline.hpp:131  // We lose; a completing pop_global gets the element.  But the queue is empty”
>  'completing' should have been ‘competing’
> 
> - Comments on the pop_global corner case can also be more elaborate, otherwise, it is not clear when “A pop_global operation may read an element that is being concurrently written by a push operation”

Thanks.  I've updated the comments as we discussed.  For the record,
here is the updated version:

full: https://cr.openjdk.java.net/~kbarrett/8243326/open.01/
incr: https://cr.openjdk.java.net/~kbarrett/8243326/open.01.inc/

The incremental contains the changes suggested by Thomas and Ivan.




More information about the hotspot-gc-dev mailing list