<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix"><br>
Hi Jon,<br>
<br>
In metaspace.hpp maybe capacity_in_bytes should be called
capacity_bytes_slow() or something like that to distinguish it
from capacity_bytes() which is fast.<br>
<pre style="color: rgb(0, 0, 0); font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; widows: auto; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px;"><font color="blue"><b>+ // Total capacity in all Metaspaces</b></font>
<font color="black"> static size_t capacity_in_bytes() {</font>
<s><font color="red"><b> return capacity_in_bytes(Metaspace::ClassType) +</b></font></s>
<s><font color="red"><b> capacity_in_bytes(Metaspace::NonClassType);</b></font></s>
<b><font color="blue">+</font><font color="blue"> #ifdef PRODUCT</font></b>
<b><font color="blue">+</font><font color="blue"> // Use capacity_bytes() in PRODUCT instead of this function.</font></b>
<font color="blue"><b>+ guarantee(false, "Should not call capacity_in_bytes() in the PRODUCT");</b></font>
<font color="blue"><b>+ #endif
</b></font></pre>
Or maybe since capacity_in_bytes() is used for the other counters,
change the capacity_bytes() name to allocated_capacity() or
something like that. Just so the two names are more different
than the presence of "_in".<br>
<br>
metaspace.cpp:<br>
<br>
line 270 and 283 are missing an 'n' in accounting. I like the
promise of a cleanup. Even with the comment, it's hard to keep
these straight.<br>
<br>
1199-1201 is the same code as above it.<br>
<br>
2520 capacity_in_bytes(mdtype) is still called for PrintGCDetails
which iterates over the CLD graph. This seems too expensive for
GC printing. It also calls used_in_bytes() so iterates twice.
Then x2 for class vs. data metaspace. This wasn't part of the GC
slowdown that was observed?<br>
<br>
2935. I don't understand why we are checking UseMallocOnly since
we don't use malloc for metaspaces ever.<br>
<br>
I can't comment on the CMS change. It looks like you just moved
it.<br>
<br>
Coleen<br>
<br>
<br>
On 3/27/2013 12:13 AM, Jon Masamitsu wrote:<br>
</div>
<blockquote cite="mid:515271FB.9090203@oracle.com" type="cite">
<br>
Replace the use of a method that calculated the total capacity of
<br>
the Metaspaces by iterating over all the Metaspaces by maintaining
<br>
the sum of Metaspace capacities as the Metachunks are
<br>
allocated to each Metaspace. Also maintain a sum for each
<br>
Metaspace as the Metachunks are allocated to that Metaspace.
<br>
<br>
Change should_expand() and compute_new_space() to
<br>
calculate quantities in bytes.
<br>
<br>
Remove calls to methods that calculated totals for
<br>
"used" and "free" by iterating over the Metaspaces
<br>
in product mode. In some cases substitute the use
<br>
of capacity for used.
<br>
<br>
<a class="moz-txt-link-freetext" href="http://cr.openjdk.java.net/~jmasa/8008966/webrev.00/">http://cr.openjdk.java.net/~jmasa/8008966/webrev.00/</a>
<br>
<br>
Thanks.
<br>
<br>
Jon
<br>
</blockquote>
<br>
</body>
</html>