<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Hi Tao,<br>
<br>
Looks fine except for the following nits....<br>
<br>
concurrentMark.hpp:100<br>
Can you remove the line: <br>
<br>
// XXX Fix these so that offsets are size_t's...<br>
<br>
concurrentMark.hpp:111<br>
It might be safer to return:<br>
<br>
return offsetToHeapWord(heapWordToOffset(addr + obj->size()));<br>
<br>
or assert something like:<br>
<br>
HeapWord* res = <span class="changed">addr + obj->size();<br>
assert(</span><span class="changed">offsetToHeapWord(heapWordToOffset(res))
== res, "sanity");<br>
return res;<br>
</span><span class="changed"><br>
JohnC<br>
</span>
<pre><span class="changed"></span></pre>
<div class="moz-cite-prefix">On 2/21/2013 6:16 PM, Tao Mao wrote:<br>
</div>
<blockquote cite="mid:5126D50B.7050700@oracle.com" type="cite">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
8008079 G1: Add nextObject routine to CMBitMapRO and replace
nextWord<br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="https://jbs.oracle.com/bugs/browse/JDK-8008079">https://jbs.oracle.com/bugs/browse/JDK-8008079</a><br>
<br>
webrev: <br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Etamao/8008079/webrev.00/">http://cr.openjdk.java.net/~tamao/8008079/webrev.00/</a><br>
<br>
changeset:<br>
<meta charset="utf-8">
When concurrent marking scans an object, the task local finger is
updated to point to the start of the object. If the marking task
is asked to abort, the local finger is updated to the next word
where another could possibly start. When the marking task is
restarted, we restart scanning the marking bitmap from the updated
local finger. <br>
<br>
This is a safe implementation but has not been fully exploited for
efficiency because the contents of the marking bitmap should be
all 0's until the offset associated with the actual start of the
next object. When the marking task is restarted, we will scan
these 0's looking for the first set bit. <br>
<br>
We can avoid this redundant scanning by updating the finger to and
re-starting the scan at the actual offset where the next object
starts.<br>
<br>
testing:<br>
passed gc-test-suite with stressing marking cycles<br>
script:<br>
./run.sh /home/tamao/jdk1.8.0 --bits 64 --args -tamao -XX:+UseG1GC
-XX:InitiatingHeapOccupancyPercent=10
-XX:+UnlockDiagnosticVMOptions -XX:+VerifyDuringGC<br>
</blockquote>
<br>
</body>
</html>