Request for Review: 5049299: (process) Use posix_spawn, not fork, on S10 to avoid swap exhaustion

Alan Bateman Alan.Bateman at oracle.com
Tue Jul 2 15:41:10 UTC 2013


On 23/12/2012 01:19, Rob McKenna wrote:
> Hi Martin, thanks a lot for this.
>
> I've renamed LINKFLAG to the more appropriate (and common) ARCHFLAG. 
> It seems to pop up all around our source but if build-dev know of a 
> better (or canonical) way of doing it I'll take it!
>
> The BUILD_JEXEC differences don't show up in my webrev or hg diff, but 
> I see them in the jdk.patch generated by webrev. I have no idea why 
> this is. (I did use the JEXEC stuff as a template for the JSPAWNHELPER 
> code, but I don't recall altering the JEXEC stuff in any way. It looks 
> like its limited to indentation. Strange.)
>
> W.r.t. useFork, I'll discuss this with Alan once he has a spare 
> minute. I believe he may have had some concerns, but I'll let you know 
> how that goes either way.
>
> The only __APPLE__ should be to get the call to _NSGetEnviron(). 
> Everything else should be bsd.
>
> I've put a webrev at:
>
> http://cr.openjdk.java.net/~robm/5049299/webrev.03/ 
> <http://cr.openjdk.java.net/%7Erobm/5049299/webrev.03/>
>
> I hope this addresses the rest of your comments.
>
>     -Rob
Rob - this was great work but I don't think we brought it to a 
conclusion. Are you planning to re-base the patch and send out a new webrev?

-Alan



More information about the core-libs-dev mailing list