<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">The patch solved some of the messages.
I'm still getting the following:<br>
<br>
==21093== Conditional jump or move depends on uninitialised
value(s)<br>
==21093== at 0x58585E3: arena_run_split (arena.c:442)<br>
==21093== by 0x5858E3F: arena_run_alloc_helper (arena.c:637)<br>
==21093== by 0x585B85D: arena_malloc_large (arena.c:657)<br>
==21093== by 0x5850122: je_calloc (arena.h:883)<br>
<br>
Furthermore, I did some more testing and valgrind is also still
complaining about je_free() calls (with patch without
--enable-munmap OR without the patch with --enable-munmap):<br>
<br>
==21093== Invalid read of size 8<br>
==21093== at 0x5874A31: quarantine (quarantine.c:74)<br>
==21093== by 0x585148F: je_free (jemalloc_internal.h:911)<br>
<br>
==21093== Invalid read of size 8<br>
==21093== at 0x58744E3: quarantine (quarantine.c:100)<br>
==21093== by 0x585148F: je_free (jemalloc_internal.h:911)<br>
<br>
==16789== Invalid read of size 8<br>
==16789== at 0x5874691: quarantine (quarantine.c:103)<br>
==16789== by 0x585148F: je_free (jemalloc_internal.h:911)<br>
<br>
==16789== Invalid read of size 8<br>
==16789== at 0x4C2C40F: memcpy (mc_replace_strmem.c:878)<br>
==16789== by 0x5874C3B: quarantine (string3.h:52)<br>
==16789== by 0x585148F: je_free (jemalloc_internal.h:911)<br>
<br>
The above error messages are present in the full report I sent you
earlier.<br>
<br>
best regards,<br>
<br>
Daniel Mezzatto<br>
<br>
On 12/12/2012 04:17 PM, Jason Evans wrote:<br>
</div>
<blockquote
cite="mid:FF9F96C9-C884-41A8-B81B-A031DE7E03ED@canonware.com"
type="cite">
<meta http-equiv="Content-Type" content="text/html;
charset=ISO-8859-1">
<div>
<div>On Dec 12, 2012, at 10:07 AM, Jason Evans <<a
moz-do-not-send="true" href="mailto:jasone@canonware.com">jasone@canonware.com</a>>
wrote:</div>
<blockquote type="cite">On Dec 12, 2012, at 10:01 AM, Daniel
Mezzatto Rodrigues <<a moz-do-not-send="true"
href="mailto:danielm@buscape-inc.com">danielm@buscape-inc.com</a>>
wrote:<br>
<blockquote type="cite">Configuring jemalloc with
--enable-munmap made valgrind error messages vanish! No more
error messages as far as I could test.<br>
<br>
Are there any disadvantages of using --enable-munmap? I see
that it enables virtual memory deallocation via munmap. It
seens that it was enabled by default in previous version of
jemalloc. Is that correct?<br>
</blockquote>
<br>
The disadvantage of using --enable-munmap on Linux is that the
kernel's heuristics for finding available virtual memory are
quite naive, and they routinely result in virtual memory map
fragmentation. That wouldn't be a big deal, except that the
kernel also uses linear VM map searches for some common
operations, so performance can suffer badly for large-memory
applications.<br>
<br>
<blockquote type="cite">Will you change this
VALGRIND_MAKE_MEM_UNDEFINED() call in chunk_recycle() so
that configuring jemalloc with --enable-munmap won't be in
future versions?<br>
</blockquote>
<br>
Yes, I will fix chunk_recycle() and include the fix in the
next jemalloc release. Thank you for your bug report and help
in diagnosing the cause.<br>
</blockquote>
<div><br>
</div>
If you want to be totally sure that the fix is correct, you can
apply the following and test without the --enable-munmap option:</div>
<div><br>
</div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span><a
moz-do-not-send="true"
href="http://www.canonware.com/cgi-bin/gitweb.cgi?p=jemalloc.git;a=commitdiff;h=1271185b87fcf54afb37dc05e7e0c58e5fb8f06a">http://www.canonware.com/cgi-bin/gitweb.cgi?p=jemalloc.git;a=commitdiff;h=1271185b87fcf54afb37dc05e7e0c58e5fb8f06a</a></div>
<div><br>
</div>
<div>Thanks,</div>
<div>Jason</div>
</blockquote>
<br>
</body>
</html>