RFR (S): 8075288: Remove dictionary NULL check on common path of BlockFreeList methods

Stefan Karlsson stefan.karlsson at oracle.com
Thu Apr 30 14:21:59 UTC 2015


On 2015-04-30 16:05, Jungwoo Ha wrote:
> Well, I just followed what Kim suggested before.
> I did that before for JDK-8075288, and no one responded yet on runtime 
> thread.
> There are too many different voices from Oracle, which confuses me.

Kim suggested that you should involve the Runtime team:

"I'm not sure of this, but I think metaspace belongs to runtime rather
than gc. If so, this should go through the hs-rt repository and have
runtime folks involved in the review."

not that you should create a new, separate thread on the hotspot-rt list.

StefanK

>
> On Thu, Apr 30, 2015 at 6:42 AM, Stefan Karlsson 
> <stefan.karlsson at oracle.com <mailto:stefan.karlsson at oracle.com>> wrote:
>
>     Hi Jungwoo,
>
>
>     On 2015-04-30 15:33, Jungwoo Ha wrote:
>>
>>
>>         I agree. This is a nice cleanup, irrespective of any
>>         potential performance gains.
>>
>>         http://cr.openjdk.java.net/~jwha/8079091/webrev.00/src/share/vm/memory/metaspace.cpp.udiff.html
>>         <http://cr.openjdk.java.net/%7Ejwha/8079091/webrev.00/src/share/vm/memory/metaspace.cpp.udiff.html>
>>
>>         -BlockFreelist::BlockFreelist() : _dictionary(NULL) {}
>>         +BlockFreelist::BlockFreelist()
>>         +    : _dictionary(new BlockTreeDictionary()) {
>>         +  assert(_dictionary != NULL, "Failed to allocate
>>         BlockTreeDictionary");
>>         +}
>>
>>         No need to NULL check CHeapObj allocations, since the JVM
>>         will exit if it failed to get memory. See AllocateHeap:
>>           if (p == NULL && alloc_failmode ==
>>         AllocFailStrategy::EXIT_OOM) {
>>             vm_exit_out_of_memory(size, OOM_MALLOC_ERROR,
>>         "AllocateHeap");
>>           }
>>
>>
>>     http://cr.openjdk.java.net/~jwha/8079091/webrev.02/
>>     <http://cr.openjdk.java.net/%7Ejwha/8079091/webrev.02/>
>>
>>     That part is taken care of on webrev.02.
>>     The conversation is also happening at runtime mailing list as Kim
>>     suggested to pass it to runtime.
>
>     Please don't split up a review request that way in the future. It
>     would have been enough to CC the hotspot-runtime-dev list.
>
>     I'll leave the rest of my comments on that list.
>
>     Thanks,
>     StefanK
>
>>     I think the general agreement is on using webrev.02.
>>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/attachments/20150430/5dc040ac/attachment.html>


More information about the hotspot-gc-dev mailing list