Request for reviews (XS): 6866651: Regression: simple int sum crashes jvm (build 1.6.0_14-b08 and 1.7.0-ea-b59)

[John Coomes]

Tom Rodriguez (Thomas.Rodriguez at Sun.COM) wrote:
> No regression test should crash unless there is a bug so I don't see  
> that it makes much difference.  The tests run more quickly without  
> othervm so it's preferred not to use it.

It would make comparing old/new VMs a little easier, since they could
use the same set of tests, which is what I usually do.  I guess it's
not hard to use separate testbases; just need to remember to do it.

-John

> On Aug 15, 2009, at 5:01 PM, John Coomes wrote:
> 
> > Tom Rodriguez (Thomas.Rodriguez at Sun.COM) wrote:
> >> You also don't need the /othervm part.
> >
> > Minor point, but isn't it polite to use /othervm for a test that could
> > cause a crash?  If it's run in the same vm, a crash would bring down
> > the harness.
> >
> > -John
> >
> >> On Aug 14, 2009, at 3:09 PM, Vladimir Kozlov wrote:
> >>
> >>> Looks good.
> >>>
> >>> I would remove -server from the test options to run it with tested
> >>> (client) VM instead of default VM.
> >>>
> >>> Thanks,
> >>> Vladimir
> >>>
> >>> changpeng fang - Sun Microsystems - Santa Clara United States wrote:
> >>>> http://cr.openjdk.java.net/~cfang/6866651/webrev.00/
> >>>> Fixed 6866651: Regression: simple int sum crashes jvm (build
> >>>> 1.6.0_14-b08 and 1.7.0-ea-b59)
> >>>> Problem:
> >>>> set_req_X will do dead code elimination if the original input has
> >>>> no other use. However, it is possible to
> >>>> have the current node (this) removed if a dead loop exists.  So,
> >>>> dead code elimination in set_req_X is not
> >>>> safe, and may  cause undesired consequences (like the segfault in
> >>>> 6866651)
> >>>> Solution:
> >>>> Instead of doing dead code elimination immediately in set_req_X, we
> >>>> put the dead node (original input) into
> >>>> the worklist, and the dead node will eventually be eliminated when
> >>>> it is its turn to be processed.
> >>>> In addition,  before ideal transformations, we replace a node with
> >>>> a constant if we know it computes a constant
> >>>> to skip unneeded optimizations on this node.
> >>>> Tests: JPRT, CompileTheWorld, Test case in bug report (Test.java in
> >>>> webrev)
> >>>> Thanks,
> >>>> Changpeng
> >>
> >
>