2 <title>LLVM: bugpoint tool</title>
6 <center><h1>LLVM: <tt>bugpoint</tt> tool</h1></center>
13 <tt>bugpoint [options] [input LLVM ll/bc files] [LLVM passes] --args <program arguments>...</tt>
15 <img src="../Debugging.gif" width=444 height=314 align=right>
18 The <tt>bugpoint</tt> tool narrows down the source of
19 problems in LLVM tools and passes. It can be used to debug three types of
20 failures: optimizer crashes, miscompilations by optimizers, or invalid native
21 code generation. It aims to reduce large test cases to small, useful ones.
23 if <tt><a href="gccas.html">gccas</a></tt> crashes while optimizing a file, it
24 will identify the optimization (or combination of optimizations) that causes the
25 crash, and reduce the file down to a small example which triggers the crash.<p>
27 <a name="designphilosophy">
28 <h4>Design Philosophy</h4>
30 <tt>bugpoint</tt> is designed to be a useful tool without requiring any
31 hooks into the LLVM infrastructure at all. It works with any and all LLVM
32 passes and code generators, and does not need to "know" how they work. Because
33 of this, it may appear to do a lot of stupid things or miss obvious
34 simplifications. <tt>bugpoint</tt> is also designed to trade off programmer
35 time for computer time in the compiler-debugging process; consequently, it may
36 take a long period of (unattended) time to reduce a test case, but we feel it
37 is still worth it. :-) <p>
39 <a name="automaticdebuggerselection">
40 <h4>Automatic Debugger Selection</h4>
42 <tt>bugpoint</tt> reads each <tt>.bc</tt> or <tt>.ll</tt> file
43 specified on the command line and links them together into a single module,
44 called the test program. If any LLVM passes are
45 specified on the command line, it runs these passes on the test program. If
46 any of the passes crash, or if they produce malformed output,
47 <tt>bugpoint</tt> starts the <a href="#crashdebug">crash debugger</a>.<p>
49 Otherwise, if the <a href="#opt_output"><tt>-output</tt></a> option was not
50 specified, <tt>bugpoint</tt> runs the test program with the C backend (which is
51 assumed to generate good code) to generate a reference output. Once
52 <tt>bugpoint</tt> has a reference output for the test program, it tries
53 executing it with the <a href="#opt_run-">selected</a> code generator. If the
54 selected code generator crashes, <tt>bugpoint</tt> starts the <a
55 href="#crashdebug">crash debugger</a> on the code generator. Otherwise, if the
56 resulting output differs from the reference output, it assumes the difference
57 resulted from a code generator failure, and starts the <a
58 href="#codegendebug">code generator debugger</a>.<p>
60 Finally, if the output of the selected code generator matches the reference
61 output, <tt>bugpoint</tt> runs the test program after all of the LLVM passes
62 have been applied to it. If its output differs from the reference output, it
63 assumes the difference resulted from a failure in one of the LLVM passes, and
64 enters the <a href="#miscompilationdebug">miscompilation
65 debugger</a>. Otherwise, there is no problem <tt>bugpoint</tt> can debug.<p>
68 <h4>Crash debugger</h4>
70 If an optimizer or code generator crashes, <tt>bugpoint</tt> will try as hard as
71 it can to reduce the list of passes (for optimizer crashes) and the size of the
72 test program. First, <tt>bugpoint</tt> figures out which combination of
73 optimizer passes triggers the bug. This is useful when debugging a problem
74 exposed by <tt>gccas</tt>, for example, because it runs over 25
77 Next, <tt>bugpoint</tt> tries removing functions from the test program, to
78 reduce its size. Usually it is able to reduce a test program to a single
79 function, when debugging intraprocedural optimizations. Once the number of
80 functions has been reduced, it attempts to delete various edges in the control
81 flow graph, to reduce the size of the function as much as possible. Finally,
82 <tt>bugpoint</tt> deletes any individual LLVM instructions whose absence does
83 not eliminate the failure. At the end, <tt>bugpoint</tt> should tell you what
84 passes crash, give you a bytecode file, and give you instructions on how to
85 reproduce the failure with <tt><a href="opt.html">opt</a></tt>, <tt><a
86 href="analyze.html">analyze</a></tt>, or <tt><a href="llc.html">llc</a></tt>.<p>
88 <a name="codegendebug">
89 <h4>Code generator debugger</h4>
91 The code generator debugger attempts to narrow down the amount of code that is
92 being miscompiled by the <a href="#opt_run-">selected</a> code generator. To do
93 this, it takes the test program and partitions it into two pieces: one piece
94 which it compiles with the C backend (into a shared object), and one piece which
95 it runs with either the JIT or the static LLC compiler. It uses several
96 techniques to reduce the amount of code pushed through the LLVM code generator,
97 to reduce the potential scope of the problem. After it is finished, it emits
98 two bytecode files (called "test" [to be compiled with the code generator] and
99 "safe" [to be compiled with the C backend] respectively), and instructions for
100 reproducing the problem. The code generator debugger assumes that the C
101 backend produces good code.<p>
103 If you are using the code generator debugger and get an error message that
104 says "UNSUPPORTED: external function used as a global initializer!", try using
105 the <tt>-run-llc</tt> option instead of the <tt>-run-jit</tt> option. This is
106 due to an unimplemented feature in the code generator debugger.<p>
108 <a name="miscompilationdebug">
109 <h4>Miscompilation debugger</h4>
111 The miscompilation debugger works similarly to the code generator
112 debugger. It works by splitting the test program into two pieces, running the
113 optimizations specified on one piece, linking the two pieces back together,
114 and then executing the result.
115 It attempts to narrow down the list of passes to the one (or few) which are
116 causing the miscompilation, then reduce the portion of the test program which is
117 being miscompiled. The miscompilation debugger assumes that the selected
118 code generator is working properly.<p>
120 <a name="bugpoint notes">
121 <h4>Advice for using <tt>bugpoint</tt></h4>
123 <tt>bugpoint</tt> can be a remarkably useful tool, but it sometimes works in
124 non-obvious ways. Here are some hints and tips:<p>
127 <li>In the code generator and miscompilation debuggers, <tt>bugpoint</tt> only
128 works with programs that have deterministic output. Thus, if the program
129 outputs the date, time, or any other "random" data, <tt>bugpoint</tt> may
130 misinterpret differences in these data, when output, as the result of a
131 miscompilation. Programs should be temporarily modified to disable
132 outputs that are likely to vary from run to run.
134 <li>In the code generator and miscompilation debuggers, debugging will go
135 faster if you manually modify the program or its inputs to reduce the
136 runtime, but still exhibit the problem.
138 <li><tt>bugpoint</tt> is extremely useful when working on a new optimization:
139 it helps track down regressions quickly. To avoid having to relink
140 <tt>bugpoint</tt> every time you change your optimization however, have
141 <tt>bugpoint</tt> dynamically load your optimization with the <a
142 href="#opt_load"><tt>-load</tt></a> option.
144 <li><tt>bugpoint</tt> can generate a lot of output and run for a long period of
145 time. It is often useful to capture the output of the program to file. For
146 example, in the C shell, you can type:<br>
147 <tt>bugpoint ..... |& tee bugpoint.log</tt>
148 <br>to get a copy of <tt>bugpoint</tt>'s output in the file
149 <tt>bugpoint.log</tt>, as well as on your terminal.
151 <li><tt>bugpoint</tt> cannot debug problems with the linker. If
152 <tt>bugpoint</tt> crashes before you see its "All input ok" message,
153 you might try <tt>llvm-link -v</tt> on the same set of input files. If
154 that also crashes, you may be experiencing a linker bug.
156 <li>If your program is <b>supposed</b> to crash, <tt>bugpoint</tt> will be
157 confused. One way to deal with this is to cause bugpoint to ignore the exit
158 code from your program, by giving it the <tt>-check-exit-code=false</tt>
166 <li><tt>-additional-so <library></tt><br>
167 Load <tt><library></tt> into the test program whenever it is run.
168 This is useful if you are debugging programs which depend on non-LLVM
169 libraries (such as the X or curses libraries) to run.<p>
171 <li><tt>-args <program args></tt><br>
172 Pass all arguments specified after <tt>-args</tt> to the
173 test program whenever it runs. Note that if any of
174 the <tt><program args></tt> start with a '-', you should use:
176 <tt>bugpoint <bugpoint args> -args -- <program args></tt>
178 The "<tt>--</tt>" right after the <tt>-args</tt> option tells
179 <tt>bugpoint</tt> to consider any options starting with <tt>-</tt> to be
180 part of the <tt>-args</tt> option, not as options to <tt>bugpoint</tt>
183 <li><tt>-check-exit-code={true,false}</tt><br>
184 Assume a non-zero exit code or core dump from the test program is
185 a failure. Defaults to true.<p>
187 <li><tt>-disable-{dce,simplifycfg}</tt><br>
188 Do not run the specified passes to clean up and reduce the size of the
189 test program. By default, <tt>bugpoint</tt> uses these passes internally
190 when attempting to reduce test programs. If you're trying to find
191 a bug in one of these passes, <tt>bugpoint</tt> may crash.<p>
193 <li><tt>-enable-correct-eh-support</tt><br>
194 Make the -lowerinvoke pass insert expensive, but correct, exception
197 <li><tt>-internalize-public-api-file <filename></tt><br>
198 Preserve the symbols listed in the file <tt>filename</tt>.<p>
200 <li><tt>-internalize-public-api-list <list></tt><br>
201 Preserve the symbols listed in <tt>list</tt>.<p>
203 <li> <tt>-help</tt><br>
204 Print a summary of command line options.<p>
206 <a name="opt_input"><li><tt>-input <filename></tt><br>
207 Open <tt><filename></tt> and redirect the standard input of the
208 test program, whenever it runs, to come from that file.
211 <a name="opt_load"><li> <tt>-load <plugin></tt><br>
212 Load the dynamic object <tt><plugin></tt> into <tt>bugpoint</tt>
213 itself. This object should register new
214 optimization passes. Once loaded, the object will add new command line
215 options to enable various optimizations. To see the new complete list
216 of optimizations, use the -help and -load options together:
218 <tt>bugpoint -load <plugin> -help</tt>
221 <a name="opt_output"><li><tt>-output <filename></tt><br>
222 Whenever the test program produces output on its standard output
223 stream, it should match the contents of <tt><filename></tt>
224 (the "reference output"). If you do not use this option,
225 <tt>bugpoint</tt> will attempt to generate a reference output by
226 compiling the program with the C backend and running it.<p>
228 <li><tt>-profile-info-file <filename></tt><br>
229 Profile file loaded by -profile-loader.<p>
231 <a name="opt_run-"><li><tt>-run-{int,jit,llc,cbe}</tt><br>
232 Whenever the test program is compiled, <tt>bugpoint</tt> should generate
233 code for it using the specified code generator. These options allow
234 you to choose the interpreter, the JIT compiler, the static native
235 code compiler, or the C backend, respectively.<p>
240 If <tt>bugpoint</tt> succeeds in finding a problem, it will exit with 0.
241 Otherwise, if an error occurs, it will exit with a non-zero value.
244 <a href="opt.html"><tt>opt</tt></a>,
245 <a href="analyze.html"><tt>analyze</tt></a>
248 Maintained by the <a href="http://llvm.cs.uiuc.edu">LLVM Team</a>.