<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 09/13/2013 09:48 AM, Bernd Eckenfels
wrote:<br>
</div>
<blockquote
cite="mid:1F8E7AEF-B1A9-405E-99B7-C20052C5D169@eckenfels.net"
type="cite">
<div><span class="Apple-style-span"
style="-webkit-tap-highlight-color: rgba(26, 26, 26,
0.296875); -webkit-composition-fill-color: rgba(175, 192, 227,
0.230469); -webkit-composition-frame-color: rgba(77, 128, 180,
0.230469); ">
<div>Question: why is the trace block testing for is_humongus
itself?</div>
<div><br>
</div>
<div>I would expect the allocation code to branch somewhere
for this condition anyway - would it make more sense to put
the trace there? Especially if you want to maintain (and
output) additional details/statistics.</div>
</span></div>
</blockquote>
<br>
You are probably right, and I'm not going to defend the current
placement of this trace code. But for this specific change I just
want to fix the crash and the other misplaced print out.<br>
<br>
<blockquote
cite="mid:1F8E7AEF-B1A9-405E-99B7-C20052C5D169@eckenfels.net"
type="cite">
<div><span class="Apple-style-span"
style="-webkit-tap-highlight-color: rgba(26, 26, 26,
0.296875); -webkit-composition-fill-color: rgba(175, 192, 227,
0.230469); -webkit-composition-frame-color: rgba(77, 128, 180,
0.230469); ">
<div><br>
</div>
<div>Is PTR_FORMAT right for word_size()</div>
</span></div>
</blockquote>
<br>
No, it should be SIZE_FORMAT. I'll change it.<br>
<br>
<blockquote
cite="mid:1F8E7AEF-B1A9-405E-99B7-C20052C5D169@eckenfels.net"
type="cite">
<div><span class="Apple-style-span"
style="-webkit-tap-highlight-color: rgba(26, 26, 26,
0.296875); -webkit-composition-fill-color: rgba(175, 192, 227,
0.230469); -webkit-composition-frame-color: rgba(77, 128, 180,
0.230469); ">
<div> and would it be better to output bytes?</div>
</span></div>
</blockquote>
<br>
In my opinion, yes. However, currently, most of print-outs guarded
by TraceMetadata output numbers in words instead of bytes. Changing
that will require another changeset.<br>
<br>
<blockquote
cite="mid:1F8E7AEF-B1A9-405E-99B7-C20052C5D169@eckenfels.net"
type="cite">
<div><span class="Apple-style-span"
style="-webkit-tap-highlight-color: rgba(26, 26, 26,
0.296875); -webkit-composition-fill-color: rgba(175, 192, 227,
0.230469); -webkit-composition-frame-color: rgba(77, 128, 180,
0.230469); ">
<div> Is this also available as a event?</div>
</span></div>
</blockquote>
<br>
Not at the moment. But we plan to send this information out as an
event. Maybe we can clean up some of the issues you have pointed
when that is done.<br>
<br>
thanks,<br>
StefanK<br>
<br>
<blockquote
cite="mid:1F8E7AEF-B1A9-405E-99B7-C20052C5D169@eckenfels.net"
type="cite">
<div><span class="Apple-style-span"
style="-webkit-tap-highlight-color: rgba(26, 26, 26,
0.296875); -webkit-composition-fill-color: rgba(175, 192, 227,
0.230469); -webkit-composition-frame-color: rgba(77, 128, 180,
0.230469); ">
<div><br>
</div>
<div>Bernd</div>
</span><br>
Am 13.09.2013 um 08:57 schrieb Stefan Karlsson <<a
moz-do-not-send="true"
href="mailto:stefan.karlsson@oracle.com">stefan.karlsson@oracle.com</a>>:<br>
<br>
</div>
<blockquote type="cite">
<div>
<meta content="text/html; charset=UTF-8"
http-equiv="Content-Type">
<div class="moz-cite-prefix">On 09/13/2013 08:22 AM, Bengt
Rutisson wrote:<br>
</div>
<blockquote cite="mid:5232AF1E.5030108@oracle.com" type="cite">
<meta content="text/html; charset=UTF-8"
http-equiv="Content-Type">
<div class="moz-cite-prefix"><br>
Hi Stefan,<br>
<br>
On 9/12/13 10:00 PM, Stefan Karlsson wrote:<br>
</div>
<blockquote cite="mid:52321D71.2070008@oracle.com"
type="cite"><a moz-do-not-send="true"
class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Estefank/8024751/webrev.00/">http://cr.openjdk.java.net/~stefank/8024751/webrev.00/</a>
<br>
<br>
Small fixes two fix some issues when TraceMetadata* flags
are turned on. <br>
<br>
- TraceMetadataHumongousAllocation crashes. <br>
- TraceMetadataChunkAllocation prints the same
block_freelist() multiple times. <br>
</blockquote>
<br>
Looks good. I'm fine with pushing this as is, but I think I
would have preferred that this code:<br>
<br>
<meta http-equiv="content-type" content="text/html;
charset=UTF-8">
<br>
2369 if (next != NULL) {<br>
2370 if (TraceMetadataHumongousAllocation &&<br>
2371
SpaceManager::is_humongous(next->word_size())) {<br>
2372 gclog_or_tty->print_cr(" new humongous chunk
word size " PTR_FORMAT,<br>
2373 next->word_size());<br>
2374 }<br>
2375 }<br>
<br>
was more like:<br>
<br>
2370 if (TraceMetadataHumongousAllocation &&<br>
2371 next != NULL &&
SpaceManager::is_humongous(next->word_size())) {<br>
2372 gclog_or_tty->print_cr(" new humongous chunk
word size " PTR_FORMAT,<br>
2373 next->word_size());<br>
2374 }<br>
<br>
To me it makes it clearer that this is only a tracing
section.<br>
</blockquote>
<br>
Fair enough. I'll change it.<br>
<br>
thanks,<br>
StefanK<br>
<blockquote cite="mid:5232AF1E.5030108@oracle.com" type="cite">
<br>
Thanks,<br>
Bengt<br>
<blockquote cite="mid:52321D71.2070008@oracle.com"
type="cite"> <br>
thanks, <br>
StefanK <br>
<br>
<br>
</blockquote>
<br>
</blockquote>
<br>
</div>
</blockquote>
</blockquote>
<br>
</body>
</html>