<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta name="Generator" content="Microsoft Word 14 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:Tahoma;
panose-1:2 11 6 4 3 5 4 4 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:12.0pt;
font-family:"Times New Roman","serif";}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-priority:99;
color:purple;
text-decoration:underline;}
span.EmailStyle17
{mso-style-type:personal-reply;
font-family:"Calibri","sans-serif";
color:#1F497D;}
.MsoChpDefault
{mso-style-type:export-only;
font-family:"Calibri","sans-serif";}
@page WordSection1
{size:8.5in 11.0in;
margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang="EN-US" link="blue" vlink="purple">
<div class="WordSection1">
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">It might be a wider problem than libfuzzer. I did want to raise the problem asap and libfuzzer is something we know has the problem.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">If it came across as "libfuzzer is evil" that was not my intent, sorry!<o:p></o:p></span></p>
<p class="MsoNormal"><a name="_MailEndCompose"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">--paulr<o:p></o:p></span></a></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"><o:p> </o:p></span></p>
<div style="border:none;border-left:solid blue 1.5pt;padding:0in 0in 0in 4.0pt">
<div>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in 0in 0in">
<p class="MsoNormal"><b><span style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">From:</span></b><span style="font-size:10.0pt;font-family:"Tahoma","sans-serif""> Kostya Serebryany [mailto:kcc@google.com]
<br>
<b>Sent:</b> Thursday, December 01, 2016 2:53 PM<br>
<b>To:</b> Robinson, Paul<br>
<b>Cc:</b> llvm-dev@lists.llvm.org<br>
<b>Subject:</b> Re: [llvm-dev] Libfuzzer depending on uninitialized debug info<o:p></o:p></span></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal">On Thu, Dec 1, 2016 at 11:08 AM, Robinson, Paul via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:<o:p></o:p></p>
<p class="MsoNormal">TL;DR: LibFuzzer appears to depend on debug-info source locations for<br>
whatever IR instrumentation it uses; however, that instrumentation does<br>
not have proper source locations attached to it, leading to potentially<br>
incorrect reporting. The short-term fix is to make sure the debug info<br>
it needs is actually set up; the long-term fix is not to rely on debug<br>
info, because some optimizations will (correctly) erase it.<o:p></o:p></p>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Why is this libFuzzer-specific? <o:p></o:p></p>
</div>
<div>
<div>
<p class="MsoNormal">We were just [un]lucky to detect the problem early with one of the libFuzzer<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal">tests that required debug info. <o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Any tool that needs debug info will suffer from the same problem. No? <o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<blockquote style="border:none;border-left:solid #CCCCCC 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-right:0in">
<p class="MsoNormal"><br>
The long version:<br>
<br>
When Clang generates IR with debug info, one thing it does is attach a<br>
source location to most IR instructions. This source location (at least<br>
in principle) is carried through optimizations, SelectionDAG, MachineIR,<br>
assembler source, and ultimately ends up in the "line table" in the<br>
object file. The line table describes a mapping from the virtual<br>
addresses of instructions to source locations, which is very useful to<br>
debuggers and other tools.<br>
<br>
Not all IR instructions have a source location attached to them. When<br>
that happens, no specific line-table record is emitted for any machine<br>
instruction produced from that IR instruction. In DWARF, that means you<br>
assume the instruction belongs to the same source location as the<br>
instruction that precedes it in memory.<br>
<br>
This is a problem when the first instruction in a machine-basic-block has<br>
no explicit source location, because it implicitly inherits the source<br>
location of the last instruction of the basic block that precedes it in<br>
memory. That means, the source location is entirely at the mercy of<br>
block layout and other optimizations.<br>
<br>
In effect, the source location for that instruction is UNINITIALIZED.<br>
<br>
In r288283, I committed a patch that explicitly initialized the line<br>
number for some instructions to line 0. The DWARF spec says that line 0<br>
means there is no specific source location for the instruction. Debuggers<br>
and other tools generally respond to this looking *forward* in the<br>
instruction stream to find the *next* instruction with an explicit non-0<br>
location, rather than backward to the *previous* instruction with an<br>
explicit location.<br>
<br>
This caused a libFuzzer test to fail, because it depended on seeing a<br>
real source location for something, and got line 0 instead. This tells<br>
me libFuzzer is depending on an uninitialized source location. Kostya<br>
backed out that patch for me, but we really want to have it for improved<br>
debugger single-stepping behavior.<br>
<br>
I am unclear on what instrumentation the fuzzer is using, although the<br>
instructions for building it suggest it's ASAN instrumentation. Whatever<br>
it is, either the instrumentation should use its own source-location<br>
information scheme, or it should initialize the debug info that it is<br>
depending on.<br>
<br>
Note that debug info is not necessarily reliable in the face of<br>
optimization. If two blocks with different source locations get merged,<br>
most likely the source location will be zeroed (and that's not my patch,<br>
that's optimization-specific behavior). Therefore, I would recommend<br>
that fuzzer/asan/whoever stop relying on debug info for source locations,<br>
if we want all that to work on optimized code.<br>
<br>
In the short term it's probably easier to find places where the<br>
instrumentation is missing debug info, and add it. But that's not going<br>
to be reliable for optimized code.<br>
--paulr<br>
<br>
_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev" target="_blank">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><o:p></o:p></p>
</blockquote>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</div>
</div>
</div>
</body>
</html>