Review request: 8026391: The Metachunk header wastes memory - 8026392: Metachunks and Metablocks are using a too large alignment
jon.masamitsu at oracle.com
Mon Oct 14 13:40:00 PDT 2013
Comments only on the first patch.
Looks good. Couple of questions.
Why the change from _is_free to _is_marked_free?
58 bool _is_free;
105 DEBUG_ONLY(bool _is_marked_free;)
Is it to specifically distinguish it from the FreeChunk is_free()?
The "marked" in the name feels a little like GC when it
isn't GC related. Would "_is_tagged_free" make sense?
Did you consider changing the file names from Metachunk.* to
Metamem.*? The file contains the classes Metamem, Metablock,
and Metachunk. It sees more natural to me to have the file named
after the more general Metamem class. Then all the classes in
the file are Metamem (or derived from), yes?
On 10/14/13 4:58 AM, Stefan Karlsson wrote:
> Please, review these two patches to remove some of the wastages
> introduced in the metaspace.
> 8026391: The Metachunk header wastes memory
> There are a couple of fields in Metachunk that store already known
> information, like the beginning and the end of the Metachunk. I
> replaced those fields with code instead.
> To make this easier, I extracted some duplicated code in Metablock and
> Metachunk into a super class named Metamem. The extracted code is
> purely needed to allow these objects to be placed on the FreeList and
> in the BinaryTreeDictionary.
> 8026392: Metachunks and Metablocks are using a too large alignment
> The metaspace code was using ARENA_ALIGN to align the Metablocks and
> Metachunk header. This forced all allocations of metadata to be 16
> bytes aligned. This patch patch changes the restriction to be 8 bytes
> instead, since that's the restriction needed for 64 bit types like
> long, double, pointers and klass pointers,
> Testing: Added unit test for Metachunk. JPRT (before splitting of the
More information about the hotspot-dev