<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 4/12/16 8:57 AM, Mikael Gerdin
wrote:<br>
</div>
<blockquote cite="mid:570CF09E.6010805@oracle.com" type="cite">Hi
Derek, Tom
<br>
<br>
Are you planning to re-review webrev.1 or can I push this?
<br>
<br>
<a class="moz-txt-link-freetext" href="http://cr.openjdk.java.net/~mgerdin/8153843/webrev.1/">http://cr.openjdk.java.net/~mgerdin/8153843/webrev.1/</a>
<br>
<br>
/Mikael
<br>
</blockquote>
<br>
Hi Mikael,<br>
<br>
My only question is about the issue Thomas pointed out and you
fixed:<br>
g1CollectedHeap.cpp, Line 2946.<br>
<br>
I am surprised that "Dead objects cannot be eager reclaim
candidates" - I thought they were the perfect eager reclaim
candidates. Should it be that we simply declare victory and return
true at line 2951 (bypassing the unsafe stuff)?<br>
<br>
Otherwise everything looks fine to me.<br>
<br>
Thanks,<br>
<br>
- Derek<br>
<br>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
<meta http-equiv="content-type" content="text/html; charset=utf-8">
<blockquote cite="mid:570CF09E.6010805@oracle.com" type="cite">On
2016-04-08 20:10, Derek White wrote:
<br>
<blockquote type="cite">Hi Mikael, Tom,
<br>
<br>
On 4/8/16 12:09 PM, Tom Benson wrote:
<br>
<blockquote type="cite">Hi Mikael,
<br>
It looks like there is an indentation problem at lines
214/215.
<br>
Dumb question: The code now extends the range down to the
object
<br>
start, if current region continues_humongous. Do you want to
extend
<br>
to the true end region of the object, which might not be the
current one?
<br>
Tom
<br>
</blockquote>
<br>
Tom,
<br>
<br>
I don't think this is extending the range to the start of the
HR, it's
<br>
just checking the mark bit at the start of the HR. Not that I'm
<br>
following why we're marking all of the live bytes instead of
just
<br>
relying on the mark for the start of the object, but it looks
like
<br>
that's what we do :-)
<br>
<br>
<br>
Mikael,
<br>
<br>
It looks like the humongous_start_region() of a
"starts_humongous"
<br>
region is itself, so you can remove lines 210 + 212, and merge
209 + 211.
<br>
<br>
Looks good otherwise.
<br>
<br>
- Derek
<br>
<blockquote type="cite">On 4/8/2016 11:47 AM, Mikael Gerdin
wrote:
<br>
<blockquote type="cite">Hi all,
<br>
<br>
Please review this small change to solve an integration
blocker in
<br>
hs-rt which surfaced with the latest integration from main.
<br>
<br>
The problem is that humongous regions are kept around for
longer
<br>
(possibly indefinitely) because the card live data helper
<br>
unconditionally sets the actual size of humongous objects
instead of
<br>
the number of live bytes. Having a non-zero live bytes count
prevents
<br>
the cleanup pause from freeing the humongous objects.
<br>
When the next gc pause then tries to register those objects
with
<br>
eager reclaim it looks at the class pointer of the object
and if the
<br>
class of the humongous object was unloaded then we crash.
<br>
<br>
Bug: <a class="moz-txt-link-freetext" href="https://bugs.openjdk.java.net/browse/JDK-8153843">https://bugs.openjdk.java.net/browse/JDK-8153843</a>
<br>
Webrev:
<a class="moz-txt-link-freetext" href="http://cr.openjdk.java.net/~mgerdin/8153843/webrev.0/">http://cr.openjdk.java.net/~mgerdin/8153843/webrev.0/</a>
<br>
Testing: Crashing testcase, JPRT job running
<br>
<br>
Thanks
<br>
/Mikael
<br>
</blockquote>
<br>
</blockquote>
<br>
</blockquote>
</blockquote>
<br>
</body>
</html>