<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Looks even better!<br>
<br>
- Derek<br>
<br>
On 3/16/16 6:12 PM, Jon Masamitsu wrote:<br>
</div>
<blockquote cite="mid:56E9DA3C.20900@oracle.com" type="cite">
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
<font face="Times New Roman, Times, serif">I updated the webrev to
include Derek's suggestion and seem to have<br>
clobbered the first version of the webrev in the process.
Fortunately,<br>
the changes is so small (hint, hint to potential reviewers) that
it matters<br>
very little.<br>
<br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Ejmasa/8151845/webrev.01/">http://cr.openjdk.java.net/~jmasa/8151845/webrev.01/</a><br>
<br>
Jon<br>
</font><br>
<div class="moz-cite-prefix">On 03/16/2016 12:16 PM, Derek White
wrote:<br>
</div>
<blockquote cite="mid:56E9B10D.1050909@oracle.com" type="cite">Hi
Jon, <br>
<br>
That looks good. <br>
<br>
There's a log message that might need updating also? "initSize"
=> "initThreshold"? <br>
Line 1577: <br>
log_trace(gc, metaspace)(" shrinking: initSize: %.1fK
maximum_desired_capacity: %.1fK", <br>
MetaspaceSize / (double) K,
maximum_desired_capacity / (double) K); <br>
<br>
- Derek <br>
<br>
On 3/15/16 12:56 PM, Jon Masamitsu wrote: <br>
<blockquote type="cite"><a moz-do-not-send="true"
class="moz-txt-link-freetext"
href="https://bugs.openjdk.java.net/browse/JDK-8151845">https://bugs.openjdk.java.net/browse/JDK-8151845</a>
<br>
<br>
Change the description for MetaspaceSize in globals.hpp thus <br>
<br>
- "Initial size of Metaspaces (in
bytes)") \ <br>
+ "Initial threshold (in bytes) at which a garbage
collection " \ <br>
+ "is done to reduce Metaspace
usage") \ <br>
<br>
Garbage collections will be done specifically to recover
metadata <br>
(the GC finds dead classloaders and identifies metadata
associated <br>
with the classloaders that is no longer used). Triggering a
GC for <br>
metadata recovery is done if the amount of metadata in use
exceeds <br>
a threshold that is dynamically adjusted (threshold is raised
or lowered <br>
depending on the amount of Metaspace made available for new
metadata <br>
allocations). MetaspaceSize is the initial threshold. <br>
<br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Ejmasa/8141845/webrev.00/">http://cr.openjdk.java.net/~jmasa/8141845/webrev.00/</a>
<br>
<br>
I am very open to suggestions for a better description. <br>
<br>
My apologies to Yasumasa Suenaga for not settling this issue <br>
when he raised it almost 2 years ago. <br>
<br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2014-April/009916.html">http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2014-April/009916.html</a>
<br>
<br>
Jon <br>
<br>
<br>
</blockquote>
<br>
</blockquote>
<br>
</blockquote>
<br>
</body>
</html>