<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Hi Jon,<br>
<br>
On 3/23/16 2:12 PM, Jon Masamitsu wrote:<br>
</div>
<blockquote cite="mid:56F2DCAB.3070600@oracle.com" type="cite">
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
<font face="Times New Roman, Times, serif">Derek,<br>
<br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Edrwhite/8149405/webrev.01/src/share/vm/oops/methodData.cpp.udiff.html">http://cr.openjdk.java.net/~drwhite/8149405/webrev.01/src/share/vm/oops/methodData.cpp.udiff.html</a><br>
</font><br>
<blockquote type="cite">
<pre> void MethodData::clean_method_data(BoolObjectClosure* is_alive) {
<span class="new">+ ResourceMark rm;</span>
for (ProfileData* data = first_data();
is_valid(data);
data = next_data(data)) {
data->clean_weak_klass_links(is_alive);
}
</pre>
</blockquote>
<font face="Times New Roman, Times, serif"><br>
The clean_weak_klass_links() above has a ResourceMark in it.<br>
<br>
<blockquote type="cite">void
DataLayout::clean_weak_klass_links(BoolObjectClosure* cl) {<br>
ResourceMark m;<br>
data_in()->clean_weak_klass_links(cl);<br>
}</blockquote>
<br>
It's not clear to me where space is being allocated in
clean_method_data() so<br>
that the ResourceMark helps. My understanding of ResourceMarks
is thin<br>
so don't assume there is anything deep about this question.<br>
<br>
Jon<br>
<br>
</font></blockquote>
<font face="Times New Roman, Times, serif">You're right, it isn't
clear where space is being allocated without doing detective work.
I didn't see any conventions w.r.t documentation, naming, type
signatures, etc that would help.<br>
<br>
The problem is that the iterators first_data()/next_data()
allocate also, as well as other functions called further down:
parameters_type_data(), clean_extra_data(), clean_extra_data(). <br>
<br>
- Derek<br>
</font>
<blockquote cite="mid:56F2DCAB.3070600@oracle.com" type="cite"><font
face="Times New Roman, Times, serif"> </font>On 03/23/2016
08:02 AM, Derek White wrote:<br>
<blockquote cite="mid:56F2B016.30304@oracle.com" type="cite">This
is a very small fix that adds ResourceMarks to
MethodData::clean_method_data() (and two similar functions
nearby). <br>
<br>
Basically an iteration over all classes in all methods in all
classes was occurring in one ResourceMark during a full gc, so
we occasionally ran out of malloc space. <br>
<br>
Once again, x86 builds running on Win64 are the "canary in the
coal mine" for these kinds of temp. memory leaks, so it's a
great test case even if not a very realistic one! <br>
<br>
BUG: <a moz-do-not-send="true" class="moz-txt-link-freetext"
href="https://bugs.openjdk.java.net/browse/JDK-8149405">https://bugs.openjdk.java.net/browse/JDK-8149405</a>
<br>
WEBREV: <a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Edrwhite/8149405/webrev.01/">http://cr.openjdk.java.net/~drwhite/8149405/webrev.01/</a>
<br>
TESTS: jprt <br>
<br>
Thanks for looking, <br>
<br>
- Derek <br>
</blockquote>
<br>
</blockquote>
<br>
</body>
</html>