<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi all,<br>
<br>
please review the following patch:<br>
<a class="moz-txt-link-freetext" href="http://cr.openjdk.java.net/~stefank/8033764/webrev.00/">http://cr.openjdk.java.net/~stefank/8033764/webrev.00/</a><br>
<br>
for the RFE:<br>
<a class="moz-txt-link-freetext" href="https://bugs.openjdk.java.net/browse/JDK-8033764">https://bugs.openjdk.java.net/browse/JDK-8033764</a><br>
<br>
This is a first step to simplify/unify the code root scanning,
something we want for the G1 Class Unloading project:<br>
<a class="moz-txt-link-freetext" href="http://openjdk.java.net/jeps/156">http://openjdk.java.net/jeps/156</a><br>
<br>
Background:<br>
<br>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
G1 uses the BufferingOopClosure to separate the root iteration from
the object copying. The oop*/narrowOop* roots are gathered in a
buffer and then bulk processed. By only timing the processing part
and not the root iteration, the object copying time can be measured.
<br>
<br>
Currently, the BufferingOopClosure uses StarTask, from the taskqueue
code, to differentiate between oop* and narrowOop* roots. The
StarTask uses the least significant bit to mark whether the address
contains an oop or a narrowOop. This works for most of the roots,
since the addresses are aligned and the LSB is always 0. However,
oops can be embedded as immediates in the JITed code and the
addresses for these are not necessarily aligned. This prevents us
from using StarTasks with oops in the CodeCache.
<br>
<br>
See this comment from g1_process_strong_roots:
<br>
...
<br>
BufferingOopClosure buf_scan_non_heap_roots(scan_non_heap_roots);
<br>
<br>
// Walk the code cache/strong code roots w/o buffering, because
StarTask
<br>
// cannot handle unaligned oop locations.
<br>
CodeBlobToOopClosure eager_scan_code_roots(scan_non_heap_roots, true
/* do_marking */);
<br>
...
<br>
<br>
I suggest that we replace the StartTask usage with another
implementation that allows the BufferingOopClosures to be used for
the CodeCache scanning. <br>
<br>
thanks,<br>
StefanK<br>
</body>
</html>