Review request: 6765718: Indicate which thread throwing OOME when generating the heap dump at OOME

[Daniel D. Daugherty]

On 9/6/2010 3:24 AM, Tomas Hurka wrote:
> Hi,
> This is a review request for fix of enhancement CR 6765718. I think that the fix is relatively straightforward. I added the new static method dump_heap_from_oome to HeapDumper. The information about OOME is than propagated to VM_HeapDumper and a new fake frame is added to the stacktrace for the thread throwing OOME.
> Webrev is here:
>
> http://cr.openjdk.java.net/~thurka/6765718/webrev.01/
>   

Thanks for taking care of this RFE!

src/share/vm/services/heapDumper.cpp

line 1309 - please line up _oome_constructor with the rest since that's
    the style that in use in this part of the file

line 1902 - typo 'out-of-memorry' -> 'out-of-memory'

line 1974 - please add a comment for the oome parameter, e.g.,
            /* pass along out-of-memory-error flag */

src/share/vm/services/heapDumper.hpp
    Any particular reason for changing the one HeapDumper ctr from
    public to private? Just curious...

src/share/vm/utilities/debug.cpp
    No comments.

Please send this changeset to RT_Baseline through JPRT when you're
ready. Let me know if you need me to take care of the JPRT part.

Again, thanks for taking care of this RFE!

Dan




> Thanks in advance for the review,
> --
> Tomas Hurka   <mailto:tomas.hurka at sun.com>
> NetBeans Profiler http://profiler.netbeans.org
> VisualVM http://visualvm.dev.java.net
> Software Engineer, Developer Platforms Group
> Sun Microsystems, Praha Czech Republic
>
>