Request for review (m) 8008966: NPG: Inefficient Metaspace counter functions cause large young GC regressions

Jon Masamitsu jon.masamitsu at oracle.com
Fri Mar 29 17:20:07 PDT 2013


Ths webrev has mostly name changes suggested by the code
review comments.  There are additional changes coming that
relate to semantic changes suggested by the code review.

This was created by deleting some of the suggested semantic
changes so will not compile.  The next webrev will be relative
to this one so you won't have to look at the name changes when
looking and the semantic changes.

Ignore the block comment about chunk accounting at the
beginning of metaspace.cpp.  That will be updated or deleted
in the next webrev.

The deletion of some dead code is included in this webrev
(limited number of lines).

http://cr.openjdk.java.net/~jmasa/8008966/webrev.00_1/

Thanks.

Jon

On 3/26/2013 9:13 PM, Jon Masamitsu wrote:
>
> Replace the use of a method that calculated the total capacity of
> the Metaspaces by iterating over all the Metaspaces by maintaining
> the sum of Metaspace capacities as the Metachunks are
> allocated to each Metaspace.  Also maintain a sum for each
> Metaspace as the Metachunks are allocated to that Metaspace.
>
> Change should_expand() and compute_new_space() to
> calculate quantities in bytes.
>
> Remove calls to methods that calculated totals for
> "used" and "free" by iterating over the Metaspaces
> in product mode.  In some cases substitute the use
> of capacity for used.
>
> http://cr.openjdk.java.net/~jmasa/8008966/webrev.00/
>
> Thanks.
>
> Jon



More information about the hotspot-gc-dev mailing list