/usr/share/doc/llvm-3.3-doc/html/Bugpoint.html is in llvm-3.3-doc 1:3.3-16ubuntu1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 | <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>LLVM bugpoint tool: design and usage — LLVM 3.3 documentation</title>
<link rel="stylesheet" href="_static/llvm-theme.css" type="text/css" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<script type="text/javascript">
var DOCUMENTATION_OPTIONS = {
URL_ROOT: './',
VERSION: '3.3',
COLLAPSE_INDEX: false,
FILE_SUFFIX: '.html',
HAS_SOURCE: true
};
</script>
<script type="text/javascript" src="_static/jquery.js"></script>
<script type="text/javascript" src="_static/underscore.js"></script>
<script type="text/javascript" src="_static/doctools.js"></script>
<link rel="top" title="LLVM 3.3 documentation" href="index.html" />
<link rel="next" title="The LLVM Target-Independent Code Generator" href="CodeGenerator.html" />
<link rel="prev" title="LLVM Branch Weight Metadata" href="BranchWeightMetadata.html" />
<style type="text/css">
table.right { float: right; margin-left: 20px; }
table.right td { border: 1px solid #ccc; }
</style>
</head>
<body>
<div class="logo">
<a href="index.html">
<img src="_static/logo.png"
alt="LLVM Logo" width="250" height="88"/></a>
</div>
<div class="related">
<h3>Navigation</h3>
<ul>
<li class="right" style="margin-right: 10px">
<a href="genindex.html" title="General Index"
accesskey="I">index</a></li>
<li class="right" >
<a href="CodeGenerator.html" title="The LLVM Target-Independent Code Generator"
accesskey="N">next</a> |</li>
<li class="right" >
<a href="BranchWeightMetadata.html" title="LLVM Branch Weight Metadata"
accesskey="P">previous</a> |</li>
<li><a href="http://llvm.org/">LLVM Home</a> | </li>
<li><a href="index.html">Documentation</a>»</li>
</ul>
</div>
<div class="document">
<div class="documentwrapper">
<div class="body">
<div class="section" id="llvm-bugpoint-tool-design-and-usage">
<h1>LLVM bugpoint tool: design and usage<a class="headerlink" href="#llvm-bugpoint-tool-design-and-usage" title="Permalink to this headline">¶</a></h1>
<div class="contents local topic" id="contents">
<ul class="simple">
<li><a class="reference internal" href="#description" id="id4">Description</a></li>
<li><a class="reference internal" href="#design-philosophy" id="id5">Design Philosophy</a><ul>
<li><a class="reference internal" href="#automatic-debugger-selection" id="id6">Automatic Debugger Selection</a></li>
<li><a class="reference internal" href="#crash-debugger" id="id7">Crash debugger</a></li>
<li><a class="reference internal" href="#code-generator-debugger" id="id8">Code generator debugger</a></li>
<li><a class="reference internal" href="#miscompilation-debugger" id="id9">Miscompilation debugger</a></li>
</ul>
</li>
<li><a class="reference internal" href="#advice-for-using-bugpoint" id="id10">Advice for using bugpoint</a></li>
<li><a class="reference internal" href="#what-to-do-when-bugpoint-isn-t-enough" id="id11">What to do when bugpoint isn’t enough</a></li>
</ul>
</div>
<div class="section" id="description">
<h2><a class="toc-backref" href="#id4">Description</a><a class="headerlink" href="#description" title="Permalink to this headline">¶</a></h2>
<p><tt class="docutils literal"><span class="pre">bugpoint</span></tt> narrows down the source of problems in LLVM tools and passes. It
can be used to debug three types of failures: optimizer crashes, miscompilations
by optimizers, or bad native code generation (including problems in the static
and JIT compilers). It aims to reduce large test cases to small, useful ones.
For example, if <tt class="docutils literal"><span class="pre">opt</span></tt> crashes while optimizing a file, it will identify the
optimization (or combination of optimizations) that causes the crash, and reduce
the file down to a small example which triggers the crash.</p>
<p>For detailed case scenarios, such as debugging <tt class="docutils literal"><span class="pre">opt</span></tt>, or one of the LLVM code
generators, see <a class="reference external" href="HowToSubmitABug.html">How To Submit a Bug Report document</a>.</p>
</div>
<div class="section" id="design-philosophy">
<h2><a class="toc-backref" href="#id5">Design Philosophy</a><a class="headerlink" href="#design-philosophy" title="Permalink to this headline">¶</a></h2>
<p><tt class="docutils literal"><span class="pre">bugpoint</span></tt> is designed to be a useful tool without requiring any hooks into
the LLVM infrastructure at all. It works with any and all LLVM passes and code
generators, and does not need to “know” how they work. Because of this, it may
appear to do stupid things or miss obvious simplifications. <tt class="docutils literal"><span class="pre">bugpoint</span></tt> is
also designed to trade off programmer time for computer time in the
compiler-debugging process; consequently, it may take a long period of
(unattended) time to reduce a test case, but we feel it is still worth it. Note
that <tt class="docutils literal"><span class="pre">bugpoint</span></tt> is generally very quick unless debugging a miscompilation
where each test of the program (which requires executing it) takes a long time.</p>
<div class="section" id="automatic-debugger-selection">
<h3><a class="toc-backref" href="#id6">Automatic Debugger Selection</a><a class="headerlink" href="#automatic-debugger-selection" title="Permalink to this headline">¶</a></h3>
<p><tt class="docutils literal"><span class="pre">bugpoint</span></tt> reads each <tt class="docutils literal"><span class="pre">.bc</span></tt> or <tt class="docutils literal"><span class="pre">.ll</span></tt> file specified on the command line
and links them together into a single module, called the test program. If any
LLVM passes are specified on the command line, it runs these passes on the test
program. If any of the passes crash, or if they produce malformed output (which
causes the verifier to abort), <tt class="docutils literal"><span class="pre">bugpoint</span></tt> starts the <a class="reference internal" href="#crash-debugger">crash debugger</a>.</p>
<p>Otherwise, if the <tt class="docutils literal"><span class="pre">-output</span></tt> option was not specified, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> runs the
test program with the “safe” backend (which is assumed to generate good code) to
generate a reference output. Once <tt class="docutils literal"><span class="pre">bugpoint</span></tt> has a reference output for the
test program, it tries executing it with the selected code generator. If the
selected code generator crashes, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> starts the <a class="reference internal" href="#crash-debugger">crash debugger</a> on
the code generator. Otherwise, if the resulting output differs from the
reference output, it assumes the difference resulted from a code generator
failure, and starts the <a class="reference internal" href="#code-generator-debugger">code generator debugger</a>.</p>
<p>Finally, if the output of the selected code generator matches the reference
output, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> runs the test program after all of the LLVM passes have
been applied to it. If its output differs from the reference output, it assumes
the difference resulted from a failure in one of the LLVM passes, and enters the
<a class="reference internal" href="#miscompilation-debugger">miscompilation debugger</a>. Otherwise, there is no problem <tt class="docutils literal"><span class="pre">bugpoint</span></tt> can
debug.</p>
</div>
<div class="section" id="crash-debugger">
<span id="id1"></span><h3><a class="toc-backref" href="#id7">Crash debugger</a><a class="headerlink" href="#crash-debugger" title="Permalink to this headline">¶</a></h3>
<p>If an optimizer or code generator crashes, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> will try as hard as it
can to reduce the list of passes (for optimizer crashes) and the size of the
test program. First, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> figures out which combination of optimizer
passes triggers the bug. This is useful when debugging a problem exposed by
<tt class="docutils literal"><span class="pre">opt</span></tt>, for example, because it runs over 38 passes.</p>
<p>Next, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> tries removing functions from the test program, to reduce its
size. Usually it is able to reduce a test program to a single function, when
debugging intraprocedural optimizations. Once the number of functions has been
reduced, it attempts to delete various edges in the control flow graph, to
reduce the size of the function as much as possible. Finally, <tt class="docutils literal"><span class="pre">bugpoint</span></tt>
deletes any individual LLVM instructions whose absence does not eliminate the
failure. At the end, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> should tell you what passes crash, give you a
bitcode file, and give you instructions on how to reproduce the failure with
<tt class="docutils literal"><span class="pre">opt</span></tt> or <tt class="docutils literal"><span class="pre">llc</span></tt>.</p>
</div>
<div class="section" id="code-generator-debugger">
<span id="id2"></span><h3><a class="toc-backref" href="#id8">Code generator debugger</a><a class="headerlink" href="#code-generator-debugger" title="Permalink to this headline">¶</a></h3>
<p>The code generator debugger attempts to narrow down the amount of code that is
being miscompiled by the selected code generator. To do this, it takes the test
program and partitions it into two pieces: one piece which it compiles with the
“safe” backend (into a shared object), and one piece which it runs with either
the JIT or the static LLC compiler. It uses several techniques to reduce the
amount of code pushed through the LLVM code generator, to reduce the potential
scope of the problem. After it is finished, it emits two bitcode files (called
“test” [to be compiled with the code generator] and “safe” [to be compiled with
the “safe” backend], respectively), and instructions for reproducing the
problem. The code generator debugger assumes that the “safe” backend produces
good code.</p>
</div>
<div class="section" id="miscompilation-debugger">
<span id="id3"></span><h3><a class="toc-backref" href="#id9">Miscompilation debugger</a><a class="headerlink" href="#miscompilation-debugger" title="Permalink to this headline">¶</a></h3>
<p>The miscompilation debugger works similarly to the code generator debugger. It
works by splitting the test program into two pieces, running the optimizations
specified on one piece, linking the two pieces back together, and then executing
the result. It attempts to narrow down the list of passes to the one (or few)
which are causing the miscompilation, then reduce the portion of the test
program which is being miscompiled. The miscompilation debugger assumes that
the selected code generator is working properly.</p>
</div>
</div>
<div class="section" id="advice-for-using-bugpoint">
<h2><a class="toc-backref" href="#id10">Advice for using bugpoint</a><a class="headerlink" href="#advice-for-using-bugpoint" title="Permalink to this headline">¶</a></h2>
<p><tt class="docutils literal"><span class="pre">bugpoint</span></tt> can be a remarkably useful tool, but it sometimes works in
non-obvious ways. Here are some hints and tips:</p>
<ul>
<li><p class="first">In the code generator and miscompilation debuggers, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> only works
with programs that have deterministic output. Thus, if the program outputs
<tt class="docutils literal"><span class="pre">argv[0]</span></tt>, the date, time, or any other “random” data, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> may
misinterpret differences in these data, when output, as the result of a
miscompilation. Programs should be temporarily modified to disable outputs
that are likely to vary from run to run.</p>
</li>
<li><p class="first">In the code generator and miscompilation debuggers, debugging will go faster
if you manually modify the program or its inputs to reduce the runtime, but
still exhibit the problem.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">bugpoint</span></tt> is extremely useful when working on a new optimization: it helps
track down regressions quickly. To avoid having to relink <tt class="docutils literal"><span class="pre">bugpoint</span></tt> every
time you change your optimization however, have <tt class="docutils literal"><span class="pre">bugpoint</span></tt> dynamically load
your optimization with the <tt class="docutils literal"><span class="pre">-load</span></tt> option.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">bugpoint</span></tt> can generate a lot of output and run for a long period of time.
It is often useful to capture the output of the program to file. For example,
in the C shell, you can run:</p>
<div class="highlight-console"><div class="highlight"><pre><span class="gp">$</span> bugpoint ... |& tee bugpoint.log
</pre></div>
</div>
<p>to get a copy of <tt class="docutils literal"><span class="pre">bugpoint</span></tt>‘s output in the file <tt class="docutils literal"><span class="pre">bugpoint.log</span></tt>, as well
as on your terminal.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">bugpoint</span></tt> cannot debug problems with the LLVM linker. If <tt class="docutils literal"><span class="pre">bugpoint</span></tt>
crashes before you see its “All input ok” message, you might try <tt class="docutils literal"><span class="pre">llvm-link</span>
<span class="pre">-v</span></tt> on the same set of input files. If that also crashes, you may be
experiencing a linker bug.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">bugpoint</span></tt> is useful for proactively finding bugs in LLVM. Invoking
<tt class="docutils literal"><span class="pre">bugpoint</span></tt> with the <tt class="docutils literal"><span class="pre">-find-bugs</span></tt> option will cause the list of specified
optimizations to be randomized and applied to the program. This process will
repeat until a bug is found or the user kills <tt class="docutils literal"><span class="pre">bugpoint</span></tt>.</p>
</li>
</ul>
</div>
<div class="section" id="what-to-do-when-bugpoint-isn-t-enough">
<h2><a class="toc-backref" href="#id11">What to do when bugpoint isn’t enough</a><a class="headerlink" href="#what-to-do-when-bugpoint-isn-t-enough" title="Permalink to this headline">¶</a></h2>
<p>Sometimes, <tt class="docutils literal"><span class="pre">bugpoint</span></tt> is not enough. In particular, InstCombine and
TargetLowering both have visitor structured code with lots of potential
transformations. If the process of using bugpoint has left you with still too
much code to figure out and the problem seems to be in instcombine, the
following steps may help. These same techniques are useful with TargetLowering
as well.</p>
<p>Turn on <tt class="docutils literal"><span class="pre">-debug-only=instcombine</span></tt> and see which transformations within
instcombine are firing by selecting out lines with “<tt class="docutils literal"><span class="pre">IC</span></tt>” in them.</p>
<p>At this point, you have a decision to make. Is the number of transformations
small enough to step through them using a debugger? If so, then try that.</p>
<p>If there are too many transformations, then a source modification approach may
be helpful. In this approach, you can modify the source code of instcombine to
disable just those transformations that are being performed on your test input
and perform a binary search over the set of transformations. One set of places
to modify are the “<tt class="docutils literal"><span class="pre">visit*</span></tt>” methods of <tt class="docutils literal"><span class="pre">InstCombiner</span></tt> (<em>e.g.</em>
<tt class="docutils literal"><span class="pre">visitICmpInst</span></tt>) by adding a “<tt class="docutils literal"><span class="pre">return</span> <span class="pre">false</span></tt>” as the first line of the
method.</p>
<p>If that still doesn’t remove enough, then change the caller of
<tt class="docutils literal"><span class="pre">InstCombiner::DoOneIteration</span></tt>, <tt class="docutils literal"><span class="pre">InstCombiner::runOnFunction</span></tt> to limit the
number of iterations.</p>
<p>You may also find it useful to use “<tt class="docutils literal"><span class="pre">-stats</span></tt>” now to see what parts of
instcombine are firing. This can guide where to put additional reporting code.</p>
<p>At this point, if the amount of transformations is still too large, then
inserting code to limit whether or not to execute the body of the code in the
visit function can be helpful. Add a static counter which is incremented on
every invocation of the function. Then add code which simply returns false on
desired ranges. For example:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="k">static</span> <span class="kt">int</span> <span class="n">calledCount</span> <span class="o">=</span> <span class="mi">0</span><span class="p">;</span>
<span class="n">calledCount</span><span class="o">++</span><span class="p">;</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="k">if</span> <span class="p">(</span><span class="n">calledCount</span> <span class="o"><</span> <span class="mi">212</span><span class="p">)</span> <span class="k">return</span> <span class="nb">false</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="k">if</span> <span class="p">(</span><span class="n">calledCount</span> <span class="o">></span> <span class="mi">217</span><span class="p">)</span> <span class="k">return</span> <span class="nb">false</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="k">if</span> <span class="p">(</span><span class="n">calledCount</span> <span class="o">==</span> <span class="mi">213</span><span class="p">)</span> <span class="k">return</span> <span class="nb">false</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="k">if</span> <span class="p">(</span><span class="n">calledCount</span> <span class="o">==</span> <span class="mi">214</span><span class="p">)</span> <span class="k">return</span> <span class="nb">false</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="k">if</span> <span class="p">(</span><span class="n">calledCount</span> <span class="o">==</span> <span class="mi">215</span><span class="p">)</span> <span class="k">return</span> <span class="nb">false</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="k">if</span> <span class="p">(</span><span class="n">calledCount</span> <span class="o">==</span> <span class="mi">216</span><span class="p">)</span> <span class="k">return</span> <span class="nb">false</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="n">dbgs</span><span class="p">()</span> <span class="o"><<</span> <span class="s">"visitXOR calledCount: "</span> <span class="o"><<</span> <span class="n">calledCount</span> <span class="o"><<</span> <span class="s">"</span><span class="se">\n</span><span class="s">"</span><span class="p">);</span>
<span class="n">DEBUG</span><span class="p">(</span><span class="n">dbgs</span><span class="p">()</span> <span class="o"><<</span> <span class="s">"I: "</span><span class="p">;</span> <span class="n">I</span><span class="o">-></span><span class="n">dump</span><span class="p">());</span>
</pre></div>
</div>
<p>could be added to <tt class="docutils literal"><span class="pre">visitXOR</span></tt> to limit <tt class="docutils literal"><span class="pre">visitXor</span></tt> to being applied only to
calls 212 and 217. This is from an actual test case and raises an important
point—a simple binary search may not be sufficient, as transformations that
interact may require isolating more than one call. In TargetLowering, use
<tt class="docutils literal"><span class="pre">return</span> <span class="pre">SDNode();</span></tt> instead of <tt class="docutils literal"><span class="pre">return</span> <span class="pre">false;</span></tt>.</p>
<p>Now that that the number of transformations is down to a manageable number, try
examining the output to see if you can figure out which transformations are
being done. If that can be figured out, then do the usual debugging. If which
code corresponds to the transformation being performed isn’t obvious, set a
breakpoint after the call count based disabling and step through the code.
Alternatively, you can use “<tt class="docutils literal"><span class="pre">printf</span></tt>” style debugging to report waypoints.</p>
</div>
</div>
</div>
</div>
<div class="clearer"></div>
</div>
<div class="related">
<h3>Navigation</h3>
<ul>
<li class="right" style="margin-right: 10px">
<a href="genindex.html" title="General Index"
>index</a></li>
<li class="right" >
<a href="CodeGenerator.html" title="The LLVM Target-Independent Code Generator"
>next</a> |</li>
<li class="right" >
<a href="BranchWeightMetadata.html" title="LLVM Branch Weight Metadata"
>previous</a> |</li>
<li><a href="http://llvm.org/">LLVM Home</a> | </li>
<li><a href="index.html">Documentation</a>»</li>
</ul>
</div>
<div class="footer">
© Copyright 2003-2013, LLVM Project.
Last updated on 2013-12-23.
Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.
</div>
</body>
</html>
|