docs/CommandGuide/bugpoint.pod

=pod

=head1 NAME

bugpoint - automatic test case reduction tool

=head1 SYNOPSIS

bugpoint [options] [input LLVM ll/bc files] [LLVM passes] --args
I<program arguments> ...

=head1 DESCRIPTION

B<bugpoint> 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 B<gccas> 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.

=head2 Design Philosophy

B<bugpoint> 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.  B<bugpoint> 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 B<bugpoint> is generally very quick unless debugging a miscompilation where
each test of the program (which requires executing it) takes a long time.

=head2 Automatic Debugger Selection

B<bugpoint> reads each F<.bc> or F<.ll> 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), B<bugpoint> starts the crash debugger.

Otherwise, if the B<-output> option was not specified, B<bugpoint> runs the test
program with the C backend (which is assumed to generate good code) to generate
a reference output.  Once B<bugpoint> has a reference output for the test
program, it tries executing it with the selected code generator.  If the
selected code generator crashes, B<bugpoint> starts the L</Crash debugger> 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 L</Code generator debugger>.

Finally, if the output of the selected code generator matches the reference
output, B<bugpoint> 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
miscompilation debugger. Otherwise, there is no problem B<bugpoint> can debug.

=head2 Crash debugger

If an optimizer or code generator crashes, B<bugpoint> will try as hard as it
can to reduce the list of passes (for optimizer crashes) and the size of the
test program.  First, B<bugpoint> figures out which combination of optimizer
passes triggers the bug. This is useful when debugging a problem exposed by
B<gccas>, for example, because it runs over 38 passes.

Next, B<bugpoint> 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, B<bugpoint>
deletes any individual LLVM instructions whose absence does not eliminate the
failure.  At the end, B<bugpoint> should tell you what passes crash, give you a
bytecode file, and give you instructions on how to reproduce the failure with
B<opt>, B<analyze>, or B<llc>.

=head2 Code generator debugger

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
C backend (into a shared object), and one piece which it runs with either the
JIT or the static compiler (B<llc>).  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 bytecode files (called
"test" [to be compiled with the code generator] and "safe" [to be compiled with
the C backend], respectively), and instructions for reproducing the problem.
The code generator debugger assumes that the C backend produces good code.

=head2 Miscompilation debugger

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.

=head2 Advice for using bugpoint

B<bugpoint> can be a remarkably useful tool, but it sometimes works in
non-obvious ways.  Here are some hints and tips:

=over

=item *

In the code generator and miscompilation debuggers, B<bugpoint> only
works with programs that have deterministic output.  Thus, if the program
outputs C<argv[0]>, the date, time, or any other "random" data, B<bugpoint> 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.

=item *

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.

=item *

B<bugpoint> is extremely useful when working on a new optimization: it helps
track down regressions quickly.  To avoid having to relink B<bugpoint> every
time you change your optimization, make B<bugpoint> dynamically load
your optimization by using the B<-load> option.

=item *

B<bugpoint> 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 type:

    bugpoint ... |& tee bugpoint.log

to get a copy of B<bugpoint>'s output in the file F<bugpoint.log>, as well as on
your terminal.

=item *

B<bugpoint> cannot debug problems with the LLVM linker. If B<bugpoint> crashes
before you see its C<All input ok> message, you might try running C<llvm-link
-v> on the same set of input files. If that also crashes, you may be
experiencing a linker bug.

=item *

If your program is supposed to crash, B<bugpoint> will be confused. One way to
deal with this is to cause B<bugpoint> to ignore the exit code from your
program, by giving it the B<-check-exit-code=false> option.

=back

=head1 OPTIONS

=over 

=item B<--additional-so> F<library>

Load the dynamic shared object F<library> into the test program whenever it is
run.  This is useful if you are debugging programs which depend on non-LLVM
libraries (such as the X or curses libraries) to run.

=item B<--args> I<program args>

Pass all arguments specified after -args to the test program whenever it runs.
Note that if any of the I<program args> start with a '-', you should use:

    bugpoint [bugpoint args] --args -- [program args]

The "--" right after the B<--args> option tells B<bugpoint> to consider any
options starting with C<-> to be part of the B<--args> option, not as options to
B<bugpoint> itself.

=item B<--tool-args> I<tool args>

Pass all arguments specified after --tool-args to the LLVM tool under test
(B<llc>, B<lli>, etc.) whenever it runs.  You should use this option in the
following way:

    bugpoint [bugpoint args] --tool-args -- [tool args]

The "--" right after the B<--tool-args> option tells B<bugpoint> to consider any
options starting with C<-> to be part of the B<--tool-args> option, not as
options to B<bugpoint> itself. (See B<--args>, above.)

=item B<--check-exit-code>=I<{true,false}>

Assume a non-zero exit code or core dump from the test program is a failure.
Defaults to true.

=item B<--disable-{dce,simplifycfg}>

Do not run the specified passes to clean up and reduce the size of the test
program. By default, B<bugpoint> uses these passes internally when attempting to
reduce test programs.  If you're trying to find a bug in one of these passes,
B<bugpoint> may crash.

=item B<--help>

Print a summary of command line options.

=item B<--input> F<filename>

Open F<filename> and redirect the standard input of the test program, whenever
it runs, to come from that file.

=item B<--load> F<plugin>

Load the dynamic object F<plugin> into B<bugpoint> itself.  This object should
register new optimization passes.  Once loaded, the object will add new command
line options to enable various optimizations.  To see the new complete list of
optimizations, use the B<--help> and B<--load> options together; for example:

    bugpoint --load myNewPass.so --help

=item B<--output> F<filename>

Whenever the test program produces output on its standard output stream, it
should match the contents of F<filename> (the "reference output"). If you
do not use this option, B<bugpoint> will attempt to generate a reference output
by compiling the program with the C backend and running it.

=item B<--profile-info-file> F<filename>

Profile file loaded by B<--profile-loader>.

=item B<--run-{int,jit,llc,cbe}>

Whenever the test program is compiled, B<bugpoint> should generate code for it
using the specified code generator.  These options allow you to choose the
interpreter, the JIT compiler, the static native code compiler, or the C
backend, respectively.

=back

=head1 EXIT STATUS

If B<bugpoint> succeeds in finding a problem, it will exit with 0.  Otherwise,
if an error occurs, it will exit with a non-zero value.

=head1 SEE ALSO

L<opt>, L<analyze>

=head1 AUTHOR

Maintained by the LLVM Team (L<http://llvm.cs.uiuc.edu>).

=cut