blob: af6155e8b66f633f2c85f5fbf30b052bd3b59808 [file] [log] [blame]
Reid Spencerd206bc92004-11-01 21:55:46 +00001<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
Brian Gaeke74f470b2004-07-01 20:10:40 +00002<html>
Reid Spencerd206bc92004-11-01 21:55:46 +00003<head>
4 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
5 <title>LLVM: bugpoint tool</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="A tool for automatic test case reduction">
10</head>
11<html>
Brian Gaeke74f470b2004-07-01 20:10:40 +000012
Reid Spencerd206bc92004-11-01 21:55:46 +000013<body>
Brian Gaeke74f470b2004-07-01 20:10:40 +000014<center><h1>LLVM: <tt>bugpoint</tt> tool</h1></center>
15<HR>
16
17<h3>NAME</h3>
18<tt>bugpoint</tt>
19
20<h3>SYNOPSIS</h3>
21<tt>bugpoint [options] [input LLVM ll/bc files] [LLVM passes] --args &lt;program arguments&gt;...</tt>
22
Chris Lattner5706f252004-07-01 20:41:43 +000023<img src="img/Debugging.gif" width=444 height=314 align=right>
Brian Gaeke74f470b2004-07-01 20:10:40 +000024<h3>DESCRIPTION</h3>
25
26The <tt>bugpoint</tt> tool narrows down the source of
27problems in LLVM tools and passes. It can be used to debug three types of
28failures: optimizer crashes, miscompilations by optimizers, or bad native
29code generation (including problems in the static and JIT compilers). It aims
30to reduce large test cases to small, useful ones. For example,
Chris Lattner9500a6e2004-10-15 17:04:28 +000031if <tt><a href="CommandGuide/html/gccas.html">gccas</a></tt> crashes while optimizing a file, it
Brian Gaeke74f470b2004-07-01 20:10:40 +000032will identify the optimization (or combination of optimizations) that causes the
33crash, and reduce the file down to a small example which triggers the crash.<p>
34
35<a name="designphilosophy">
36<h4>Design Philosophy</h4>
37
38<tt>bugpoint</tt> is designed to be a useful tool without requiring any
39hooks into the LLVM infrastructure at all. It works with any and all LLVM
40passes and code generators, and does not need to "know" how they work. Because
41of this, it may appear to do stupid things or miss obvious
42simplifications. <tt>bugpoint</tt> is also designed to trade off programmer
43time for computer time in the compiler-debugging process; consequently, it may
44take a long period of (unattended) time to reduce a test case, but we feel it
45is still worth it. Note that <tt>bugpoint</tt> is generally very quick unless
46debugging a miscompilation where each test of the program (which requires
47executing it) takes a long time.<p>
48
49<a name="automaticdebuggerselection">
50<h4>Automatic Debugger Selection</h4>
51
52<tt>bugpoint</tt> reads each <tt>.bc</tt> or <tt>.ll</tt> file
53specified on the command line and links them together into a single module,
54called the test program. If any LLVM passes are
55specified on the command line, it runs these passes on the test program. If
56any of the passes crash, or if they produce malformed output (which causes the
57verifier to abort),
58<tt>bugpoint</tt> starts the <a href="#crashdebug">crash debugger</a>.<p>
59
60Otherwise, if the <a href="#opt_output"><tt>-output</tt></a> option was not
61specified, <tt>bugpoint</tt> runs the test program with the C backend (which is
62assumed to generate good code) to generate a reference output. Once
63<tt>bugpoint</tt> has a reference output for the test program, it tries
64executing it with the <a href="#opt_run-">selected</a> code generator. If the
65selected code generator crashes, <tt>bugpoint</tt> starts the <a
66href="#crashdebug">crash debugger</a> on the code generator. Otherwise, if the
67resulting output differs from the reference output, it assumes the difference
68resulted from a code generator failure, and starts the <a
69href="#codegendebug">code generator debugger</a>.<p>
70
71Finally, if the output of the selected code generator matches the reference
72output, <tt>bugpoint</tt> runs the test program after all of the LLVM passes
73have been applied to it. If its output differs from the reference output, it
74assumes the difference resulted from a failure in one of the LLVM passes, and
75enters the <a href="#miscompilationdebug">miscompilation
76debugger</a>. Otherwise, there is no problem <tt>bugpoint</tt> can debug.<p>
77
78<a name="crashdebug">
79<h4>Crash debugger</h4>
80
81If an optimizer or code generator crashes, <tt>bugpoint</tt> will try as hard as
82it can to reduce the list of passes (for optimizer crashes) and the size of the
83test program. First, <tt>bugpoint</tt> figures out which combination of
84optimizer passes triggers the bug. This is useful when debugging a problem
85exposed by <tt>gccas</tt>, for example, because it runs over 38 passes.<p>
86
87Next, <tt>bugpoint</tt> tries removing functions from the test program, to
88reduce its size. Usually it is able to reduce a test program to a single
89function, when debugging intraprocedural optimizations. Once the number of
90functions has been reduced, it attempts to delete various edges in the control
91flow graph, to reduce the size of the function as much as possible. Finally,
92<tt>bugpoint</tt> deletes any individual LLVM instructions whose absence does
93not eliminate the failure. At the end, <tt>bugpoint</tt> should tell you what
94passes crash, give you a bytecode file, and give you instructions on how to
Chris Lattner9500a6e2004-10-15 17:04:28 +000095reproduce the failure with <tt><a href="CommandGuide/html/opt.html">opt</a></tt>, <tt><a
96href="CommandGuide/html/analyze.html">analyze</a></tt>, or <tt><a href="CommandGuide/html/llc.html">llc</a></tt>.<p>
Brian Gaeke74f470b2004-07-01 20:10:40 +000097
98<a name="codegendebug">
99<h4>Code generator debugger</h4>
100
101<p>The code generator debugger attempts to narrow down the amount of code that
102is being miscompiled by the <a href="#opt_run-">selected</a> code generator. To
103do this, it takes the test program and partitions it into two pieces: one piece
104which it compiles with the C backend (into a shared object), and one piece which
105it runs with either the JIT or the static LLC compiler. It uses several
106techniques to reduce the amount of code pushed through the LLVM code generator,
107to reduce the potential scope of the problem. After it is finished, it emits
108two bytecode files (called "test" [to be compiled with the code generator] and
109"safe" [to be compiled with the C backend], respectively), and instructions for
110reproducing the problem. The code generator debugger assumes that the C backend
111produces good code.</p>
112
113<a name="miscompilationdebug">
114<h4>Miscompilation debugger</h4>
115
116The miscompilation debugger works similarly to the code generator
117debugger. It works by splitting the test program into two pieces, running the
118optimizations specified on one piece, linking the two pieces back together,
119and then executing the result.
120It attempts to narrow down the list of passes to the one (or few) which are
121causing the miscompilation, then reduce the portion of the test program which is
122being miscompiled. The miscompilation debugger assumes that the selected
123code generator is working properly.<p>
124
125<a name="bugpoint notes">
126<h4>Advice for using <tt>bugpoint</tt></h4>
127
128<tt>bugpoint</tt> can be a remarkably useful tool, but it sometimes works in
129non-obvious ways. Here are some hints and tips:<p>
130
131<ol>
132<li>In the code generator and miscompilation debuggers, <tt>bugpoint</tt> only
133 works with programs that have deterministic output. Thus, if the program
134 outputs <tt>argv[0]</tt>, the date, time, or any other "random" data, <tt>bugpoint</tt> may
135 misinterpret differences in these data, when output, as the result of a
136 miscompilation. Programs should be temporarily modified to disable
137 outputs that are likely to vary from run to run.
138
139<li>In the code generator and miscompilation debuggers, debugging will go
140 faster if you manually modify the program or its inputs to reduce the
141 runtime, but still exhibit the problem.
142
143<li><tt>bugpoint</tt> is extremely useful when working on a new optimization:
144 it helps track down regressions quickly. To avoid having to relink
145 <tt>bugpoint</tt> every time you change your optimization however, have
146 <tt>bugpoint</tt> dynamically load your optimization with the <a
147 href="#opt_load"><tt>-load</tt></a> option.
148
149<li><tt>bugpoint</tt> can generate a lot of output and run for a long period of
150 time. It is often useful to capture the output of the program to file. For
151 example, in the C shell, you can type:<br>
152 <tt>bugpoint ..... |&amp; tee bugpoint.log</tt>
153 <br>to get a copy of <tt>bugpoint</tt>'s output in the file
154 <tt>bugpoint.log</tt>, as well as on your terminal.
155
156<li><tt>bugpoint</tt> cannot debug problems with the LLVM linker. If
157 <tt>bugpoint</tt> crashes before you see its "All input ok" message,
158 you might try <tt>llvm-link -v</tt> on the same set of input files. If
159 that also crashes, you may be experiencing a linker bug.
160
161<li>If your program is <b>supposed</b> to crash, <tt>bugpoint</tt> will be
162 confused. One way to deal with this is to cause bugpoint to ignore the exit
163 code from your program, by giving it the <tt>-check-exit-code=false</tt>
164 option.
165
166</ol>
167
168<h3>OPTIONS</h3>
169
170<ul>
171 <li><tt>-additional-so &lt;library&gt;</tt><br>
172 Load <tt>&lt;library&gt;</tt> into the test program whenever it is run.
173 This is useful if you are debugging programs which depend on non-LLVM
174 libraries (such as the X or curses libraries) to run.<p>
175
176 <li><tt>-args &lt;program args&gt;</tt><br>
177 Pass all arguments specified after <tt>-args</tt> to the
178 test program whenever it runs. Note that if any of
179 the <tt>&lt;program args&gt;</tt> start with a '-', you should use:
180 <p>
181 <tt>bugpoint &lt;bugpoint args&gt; -args -- &lt;program args&gt;</tt>
182 <p>
183 The "<tt>--</tt>" right after the <tt>-args</tt> option tells
184 <tt>bugpoint</tt> to consider any options starting with <tt>-</tt> to be
185 part of the <tt>-args</tt> option, not as options to <tt>bugpoint</tt>
186 itself.<p>
187
188 <li><tt>-tool-args &lt;tool args&gt;</tt><br>
189 Pass all arguments specified after <tt>-tool-args</tt> to the
190 LLVM tool under test (llc, lli, etc.) whenever it runs.
191 You should use this option in the following way:
192 <p>
193 <tt>bugpoint &lt;bugpoint args&gt; -tool-args -- &lt;tool args&gt;</tt>
194 <p>
195 The "<tt>--</tt>" right after the <tt>-tool-args</tt> option tells
196 <tt>bugpoint</tt> to consider any options starting with <tt>-</tt> to be
197 part of the <tt>-tool-args</tt> option, not as options to
198 <tt>bugpoint</tt> itself. (See <tt>-args</tt>, above.)<p>
199
200 <li><tt>-check-exit-code={true,false}</tt><br>
201 Assume a non-zero exit code or core dump from the test program is
202 a failure. Defaults to true.<p>
203
204 <li><tt>-disable-{dce,simplifycfg}</tt><br>
205 Do not run the specified passes to clean up and reduce the size of the
206 test program. By default, <tt>bugpoint</tt> uses these passes internally
207 when attempting to reduce test programs. If you're trying to find
208 a bug in one of these passes, <tt>bugpoint</tt> may crash.<p>
209
210 <li> <tt>-help</tt><br>
211 Print a summary of command line options.<p>
212
213 <a name="opt_input"><li><tt>-input &lt;filename&gt;</tt><br>
214 Open <tt>&lt;filename&gt;</tt> and redirect the standard input of the
215 test program, whenever it runs, to come from that file.
216 <p>
217
218 <a name="opt_load"><li> <tt>-load &lt;plugin&gt;</tt><br>
219 Load the dynamic object <tt>&lt;plugin&gt;</tt> into <tt>bugpoint</tt>
220 itself. This object should register new
221 optimization passes. Once loaded, the object will add new command line
222 options to enable various optimizations. To see the new complete list
223 of optimizations, use the -help and -load options together:
224 <p>
225 <tt>bugpoint -load &lt;plugin&gt; -help</tt>
226 <p>
227
228 <a name="opt_output"><li><tt>-output &lt;filename&gt;</tt><br>
229 Whenever the test program produces output on its standard output
230 stream, it should match the contents of <tt>&lt;filename&gt;</tt>
231 (the "reference output"). If you do not use this option,
232 <tt>bugpoint</tt> will attempt to generate a reference output by
233 compiling the program with the C backend and running it.<p>
234
235 <li><tt>-profile-info-file &lt;filename&gt;</tt><br>
236 Profile file loaded by -profile-loader.<p>
237
238 <a name="opt_run-"><li><tt>-run-{int,jit,llc,cbe}</tt><br>
239 Whenever the test program is compiled, <tt>bugpoint</tt> should generate
240 code for it using the specified code generator. These options allow
241 you to choose the interpreter, the JIT compiler, the static native
242 code compiler, or the C backend, respectively.<p>
243</ul>
244
245<h3>EXIT STATUS</h3>
246
247If <tt>bugpoint</tt> succeeds in finding a problem, it will exit with 0.
248Otherwise, if an error occurs, it will exit with a non-zero value.
249
Brian Gaeke74f470b2004-07-01 20:10:40 +0000250<HR>
251Maintained by the <a href="http://llvm.cs.uiuc.edu">LLVM Team</a>.
252</body>
253</html>