source: zzuf/trunk/doc/zzuf.1 @ 1666

Last change on this file since 1666 was 1666, checked in by Sam Hocevar, 14 years ago
  • Set --version to -V, not -v. We'll use -v for verbose.
File size: 14.2 KB
Line 
1.TH zzuf 1 "2006-12-22" "zzuf"
2.SH NAME
3zzuf \- multiple purpose fuzzer
4.SH SYNOPSIS
5\fBzzuf\fR [\fB\-AcdiMnqSx\fR] [\fB\-r\fR \fIratio\fR] [\fB\-s\fR \fIseed\fR|\fB\-s\fR \fIstart:stop\fR]
6.br
7                  [\fB\-D\fR \fIdelay\fR] [\fB\-F\fR \fIforks\fR] [\fB\-C\fR \fIcrashes\fR] [\fB\-B\fR \fIbytes\fR]
8.br
9                  [\fB\-T\fR \fIseconds\fR] [\fB\-M\fR \fImegabytes\fR] [\fB\-P\fR \fIprotect\fR] [\fB\-R\fR \fIrefuse\fR]
10.br
11                  [\fB\-I\fR \fIinclude\fR] [\fB\-E\fR \fIexclude\fR] [\fIPROGRAM\fR [\fB\-\-\fR] [\fIARGS\fR]...]
12.br
13\fBzzuf \-h\fR | \fB\-\-help\fR
14.br
15\fBzzuf \-V\fR | \fB\-\-version\fR
16.SH DESCRIPTION
17.PP
18\fBzzuf\fR is a transparent application input fuzzer. It works by intercepting
19file and network operations and changing random bits in the program's input.
20\fBzzuf\fR's behaviour is deterministic, making it easy to reproduce bugs.
21.SH USAGE
22.PP
23\fBzzuf\fR will run an application specified on its command line, one or
24several times, with optional arguments, and will report the application's
25relevant behaviour on the standard output, eg:
26.PP
27\fB    zzuf cat /dev/zero\fR
28.PP
29If you want to specify flags for your application, put a \(oq\fB\-\-\fR\(cq
30marker before them on the command line (otherwise \fBzzuf\fR will try to
31interpret them as arguments for itself), eg:
32.PP
33\fB    zzuf \-B 1000 cat \-\- \-v /dev/zero\fR
34.PP
35When no program is specified, \fBzzuf\fR simply fuzzes the standard input, as
36if the \fBcat\fR utility had been called:
37.PP
38\fB    zzuf < /dev/zero\fR
39.SH OPTIONS
40.TP
41\fB\-A\fR, \fB\-\-autoinc\fR
42Increment random seed each time a new file is opened. This is only required
43if the same application is expected to open the same file several times and
44you want to test a different seed each time.
45.TP
46\fB\-B\fR, \fB\-\-max\-bytes\fR=\fIn\fR
47Automatically terminate child processes that output more than \fIn\fR bytes
48on the standard output and standard error channels. This is useful to detect
49infinite loops. See also the \fB\-T\fR flag.
50.TP
51\fB\-c\fR, \fB\-\-cmdline\fR
52Only fuzz files whose name is specified in the target application's command
53line. This is mostly a shortcut to avoid specifying twice the argument:
54
55\fB    zzuf \-c cat file.txt\fR
56
57has the same effect as
58
59\fB    zzuf \-I \(aq^file\\.txt$\(aq cat file.txt\fR
60
61See the \fB\-I\fR flag for more information on restricting fuzzing to
62specific files.
63.TP
64\fB\-C\fR, \fB\-\-max\-crashes\fR=\fIn\fR
65Stop forking when at least \fIn\fR children have crashed. The default value
66is 1, meaning \fBzzuf\fR will stop as soon as one child has crashed. A process
67is considered to have crashed if any signal (such as, but not limited to,
68\fBSIGSEGV\fR) caused it to exit. If the \fB\-x\fR flag is used, this will
69also include processes that exit with a non-zero status.
70
71This option is only relevant if the \fB\-s\fR flag is used with an interval
72argument.
73.TP
74\fB\-d\fR, \fB\-\-debug\fR
75Activate the display of debug messages.
76.TP
77\fB\-D\fR, \fB\-\-delay\fR=\fIdelay\fR
78Do not launch more than one process every \fIdelay\fR seconds. This option
79should be used together with \fB\-F\fR to avoid fork bombs.
80.TP
81\fB\-E\fR, \fB\-\-exclude\fR=\fIregex\fR
82Do not fuzz files whose name matches the \fIregex\fR regular expression. This
83option supersedes anything that is specified by the \fB\-I\fR flag. Use this
84for instance if you are unsure of what files your application is going to read
85and do not want it to fuzz files in the \fB/etc\fR directory.
86
87Multiple \fB\-E\fR flags can be specified, in which case files matching any one
88of the regular expressions will be ignored.
89.TP
90\fB\-F\fR, \fB\-\-max\-forks\fR=\fIforks\fR
91Specify the number of simultaneous children that can be run.
92
93This option is only relevant if the \fB\-s\fR flag is used with an interval
94argument. See also the \fB\-D\fR flag.
95.TP
96\fB\-i\fR, \fB\-\-stdin\fR
97Fuzz the application's standard input. By default \fBzzuf\fR only fuzzes files.
98.TP
99\fB\-I\fR, \fB\-\-include\fR=\fIregex\fR
100Only fuzz files whose name matches the \fIregex\fR regular expression. Use
101this for instance if your application reads configuration files at startup
102and you only want specific files to be fuzzed.
103
104Multiple \fB\-I\fR flags can be specified, in which case files matching any one
105of the regular expressions will be fuzzed. See also the \fB\-c\fR flag.
106.TP
107\fB\-m\fR, \fB\-\-md5\fR
108Instead of displaying the program's standard output, just print the MD5 digest
109of that output. The standard error channel is left untouched.
110.TP
111\fB\-M\fR, \fB\-\-max-memory\fR=\fImegabytes\fR
112Specify the maximum amount of memory, in megabytes, that children are allowed
113to allocate. This is useful to detect infinite loops that eat up a lot of
114memory. The value should set reasonably high so as not to interfer with normal
115program operation.
116
117\fBzzuf\fR uses the \fBsetrlimit\fR() call to set memory usage limitations and
118relies on the operating system's ability to enforce such limitations.
119.TP
120\fB\-n\fR, \fB\-\-network\fR
121Fuzz the application's network input. By default \fBzzuf\fR only fuzzes files.
122.TP
123\fB\-P\fR, \fB\-\-protect\fR=\fIlist\fR
124Protect a list of characters so that if they appear in input data that would
125normally be fuzzed, they are left unmodified instead.
126
127Characters in \fIlist\fR can be expressed verbatim or through escape sequences.
128The sequences interpreted by \fBzzuf\fR are:
129.RS
130.TP
131\fB\\n\fR
132new line
133.TP
134\fB\\r\fR
135return
136.TP
137\fB\\t\fR
138tabulation
139.TP
140\fB\\\fR\fINNN\fR
141the byte whose octal value is \fINNN\fR
142.TP
143\fB\\x\fR\fINN\fR
144the byte whose hexadecimal value is \fINN\fR
145.TP
146\fB\\\\\fR
147backslash (\(oq\\\(cq)
148.RE
149.IP
150You can use \(oq\fB\-\fR\(cq to specify ranges. For instance, to protect all
151bytes from \(oq\\001\(cq to \(oq/\(cq, use \(oq\fB\-P\ \(dq\\001\-/\(dq\fR\(cq.
152
153The statistical outcome of this option should not be overlooked: if characters
154are protected, the effect of the \(oq\fB\-r\fR\(cq flag will vary depending
155on the data being fuzzed. For instance, asking to fuzz 1% of input bits
156(\fB\-r\ 0.01\fR) and to protect lowercase characters (\fB\-P\ a\-z\fR) will
157result in an actual average fuzzing ratio of 0.9% with truly random data,
1580.3% with random ASCII data and 0.2% with standard English text.
159
160See also the \fB\-R\fR flag.
161.TP
162\fB\-q\fR, \fB\-\-quiet\fR
163Hide the output of the fuzzed application. This is useful if the application
164is very verbose but only its exit code or signaled status is really useful to
165you.
166.TP
167\fB\-r\fR, \fB\-\-ratio\fR=\fIratio\fR
168Specify the proportion of bits that will be randomly fuzzed. A value of 0
169will not fuzz anything. A value of 0.05 will fuzz 5% of the open files'
170bits. A value of 1.0 or more will fuzz all the bytes, theoretically making
171the input files undiscernible from random data. The default fuzzing ratio
172is 0.004 (fuzz 0.4% of the files' bits).
173.TP
174\fB\-R\fR, \fB\-\-refuse\fR=\fIlist\fR
175Refuse a list of characters by not fuzzing bytes that would otherwise be
176changed to a character that is in \fIlist\fR. If the original byte is already
177in \fIlist\fR, it is left unchanged.
178
179See the \fB\-P\fR option for a description of \fIlist\fR.
180.TP
181\fB\-s\fR, \fB\-\-seed\fR=\fIseed\fR
182.PD 0
183.TP
184\fB\-s\fR, \fB\-\-seed\fR=\fIstart:stop\fR
185.PD
186Specify the random seed to use for fuzzing, or an interval of random seeds.
187Running \fBzzuf\fR twice with the same random seed will fuzz the files exactly
188the same way, even with a different target application. The purpose of this is
189to use simple utilities such as \fBcat\fR or \fBcp\fR to generate a file that
190causes the target application to crash.
191
192If an interval is specified, \fBzzuf\fR will run the application several times,
193each time with a different seed, and report the behaviour of each run.
194.TP
195\fB\-S\fR, \fB\-\-signal\fR
196Prevent children from installing signal handlers for signals that usually
197cause coredumps. These signals are \fBSIGABRT\fR, \fBSIGFPE\fR, \fBSIGILL\fR,
198\fBSIGQUIT\fR, \fBSIGSEGV\fR, \fBSIGTRAP\fR and, if available on the running
199platform, \fBSIGSYS\fR, \fBSIGEMT\fR, \fBSIGBUS\fR, \fBSIGXCPU\fR and
200\fBSIGXFSZ\fR. Instead of calling the signal handler, the application will
201simply crash. If you do not want core dumps, you should set appropriate limits
202with the \fBlimit coredumpsize\fR command. See your shell's documentation on
203how to set such limits.
204.TP
205\fB\-T\fR, \fB\-\-max\-time\fR=\fIn\fR
206Automatically terminate child processes that run for more than \fIn\fR
207seconds. This is useful to detect infinite loops or processes stuck in other
208situations. See also the \fB\-B\fR flag.
209.TP
210\fB\-x\fR, \fB\-\-check\-exit\fR
211Report processes that exit with a non-zero status. By default only processes
212that crash due to a signal are reported.
213.TP
214\fB\-h\fR, \fB\-\-help\fR
215Display a short help message and exit.
216.TP
217\fB\-V\fR, \fB\-\-version\fR
218Output version information and exit.
219.SH EXAMPLES
220.PP
221Fuzz the input of the \fBcat\fR program using default settings:
222.PP
223\fB    zzuf cat /etc/motd\fR
224.PP
225Fuzz 1% of the input bits of the \fBcat\fR program using seed 94324:
226.PP
227\fB    zzuf \-s 94324 \-r 0.01 cat /etc/motd\fR
228.PP
229Fuzz the input of the \fBcat\fR program but do not fuzz newline characters
230and prevent non-ASCII characters from appearing in the output:
231.PP
232\fB    zzuf \-P \(aq\\n\(aq \-R \(aq\\x00\-\\x1f\\x7f\-\\xff\(aq cat /etc/motd\fR
233.PP
234Fuzz the input of the \fBconvert\fR program, using file \fBfoo.jpeg\fR as the
235original input and excluding \fB.xml\fR files from fuzzing (because
236\fBconvert\fR will also open its own XML configuration files and we do not
237want \fBzzuf\fR to fuzz them):
238.PP
239\fB    zzuf \-E \(aq\\.xml$\(aq convert \-\- foo.jpeg \-format tga /dev/null\fR
240.PP
241Fuzz the input of \fBVLC\fR, using file \fBmovie.avi\fR as the original input
242and restricting fuzzing to filenames that appear on the command line
243(\fB\-c\fR), then generate \fBfuzzy\-movie.avi\fR which is a file that
244can be read by \fBVLC\fR to reproduce the same behaviour without using
245\fBzzuf\fR:
246.PP
247\fB    zzuf \-c \-s 87423 \-r 0.01 vlc movie.avi\fR
248.br
249\fB    zzuf \-c \-s 87423 \-r 0.01 <movie.avi >fuzzy\-movie.avi\fR
250.br
251\fB    vlc fuzzy\-movie.avi\fR
252.PP
253Fuzz 2% of \fBMPlayer\fR's input bits (\fB\-r\ 0.02\fR) with seeds 0 to 9999
254(\fB\-s\ 0:10000\fR), disabling its standard output messages (\fB\-q\fR),
255launching up to five simultaneous child processes (\fB\-F\ 5\fR) but wait at
256least half a second between launches (\fB\-D\ 0.5\fR), killing \fBMPlayer\fR
257if it takes more than one minute to read the file (\fB\-T\ 60\fR) and
258disabling its \fBSIGSEGV\fR signal handler (\fB\-S\fR):
259.PP
260\fB    zzuf \-c \-r 0.02 \-q \-s 0:10000 \-F 5 \-D 0.5 \-T 60 \-S \\\fR
261.br
262\fB      mplayer \-\- \-benchmark \-vo null \-fps 1000 movie.avi\fR
263.PP
264Create an HTML-like file that loads 1000 times the same \fBhello.gif\fR image
265and open it in \fBFirefox\fR in auto-increment mode (\fB\-A\fR):
266.PP
267\fB    awk \(aqBEGIN { for(i=0; i<1000; i++) { print \\\fR
268.br
269\fB                   "<img src=\\"hello.gif#"i"\\">" }}\(aq > hello.html\fR
270.br
271\fB    zzuf -A -I \(aqhello[.]gif\(aq -r 0.001 firefox hello.html\fR
272.SH RESTRICTIONS
273.PP
274Due to \fBzzuf\fR using shared object preloading (\fBLD_PRELOAD\fR,
275\fB_RLD_LIST\fB, \fBDYLD_INSERT_LIBRARIES\fR, etc.) to run its child
276processes, it will fail in the presence of any mechanism that disables
277preloading. For instance setuid root binaries will not be fuzzed when run
278as an unprivileged user.
279.PP
280For the same reasons, \fBzzuf\fR will also not work with statically linked
281binaries. Bear this in mind when using \fBzzuf\fR on the OpenBSD platform,
282where \fBcat\fR, \fBcp\fR and \fBdd\fR are static binaries.
283.PP
284Though best efforts are made, identical behaviour for different versions of
285\fBzzuf\fR is not guaranteed. The reproducibility for subsequent calls on
286different operating systems and with different target programs is only
287guaranteed when the same version of \fBzzuf\fR is being used.
288.SH BUGS
289.PP
290\fBzzuf\fR probably does not behave correctly with 64-bit offsets.
291.PP
292It is not yet possible to insert or drop bytes from the input, to fuzz
293according to the file format, to swap bytes, etc. More advanced fuzzing
294methods are planned.
295.PP
296As of now, \fBzzuf\fR does not really support multithreaded applications. The
297behaviour with multithreaded applications where more than one thread does file
298descriptor operations is undefined.
299.SH NOTES
300In order to intercept file and network operations, signal handlers and memory
301allocations, \fBzzuf\fR diverts and reimplements the following functions,
302which can be private libc symbols, too:
303.TP
304Unix file descriptor handling:
305\fBopen\fR(), \fBlseek\fR(), \fBread\fR(), \fBaccept\fR(), \fBsocket\fR(),
306\fBclose\fR()
307.TP
308Standard IO streams:
309\fBfopen\fR(), \fBfreopen\fR(), \fBfseek\fR(), \fBfseeko\fR(), \fBrewind\fR(),
310\fBfread\fR(), \fBgetc\fR(), \fBfgetc\fR(), \fBfgets\fR(), \fBungetc\fR(),
311\fBfclose\fR()
312.TP
313Memory management:
314\fBmmap\fR(), \fBmunmap\fR(), \fBmalloc\fR(), \fBcalloc\fR(), \fBvalloc\fR(),
315\fBfree\fR(), \fBmemalign\fR(), \fBposix_memalign\fR()
316.TP
317Linux-specific:
318\fBopen64\fR(), \fBlseek64\fR(), \fBmmap64\fR(), \fB_IO_getc\fR(),
319\fBgetline\fR(), \fBgetdelim\fR(), \fB__getdelim\fR()
320.TP
321BSD-specific:
322\fBfgetln\fR(), \fB__srefill\fR()
323.TP
324Mac OS X-specific:
325\fBmap_fd\fR()
326.TP
327Signal handling:
328\fBsignal\fR(), \fBsigaction\fR()
329.PP
330If an application manipulates file descriptors (reading data, seeking around)
331using functions that are not in that list, \fBzzuf\fR will not fuzz its
332input consistently and the results should not be trusted. You can use a tool
333such as \fBltrace(1)\fR on Linux to know the missing functions.
334.PP
335On BSD systems, such as FreeBSD or Mac OS X, \fB__srefill\fR() is enough to
336monitor all standard IO streams functions. On other systems, such as Linux,
337each function is reimplemented on a case by case basis. One important
338unimplemented function is \fBfscanf\fR(), because of its complexity. Missing
339functions will be added upon user request.
340.SH HISTORY
341.PP
342\fBzzuf\fR started its life in 2002 as the \fBstreamfucker\fR tool, a small
343multimedia stream corrupter used to find bugs in the \fBVLC\fR media player.
344.SH AUTHOR
345.PP
346Copyright \(co 2002, 2007 Sam Hocevar <sam@zoy.org>.
347.PP
348\fBzzuf\fR and this manual page are free software. They come without any
349warranty, to the extent permitted by applicable law. You can redistribute
350them and/or modify them under the terms of the Do What The Fuck You Want
351To Public License, Version 2, as published by Sam Hocevar. See
352\fBhttp://sam.zoy.org/wtfpl/COPYING\fR for more details.
353.PP
354\fBzzuf\fR's webpage can be found at \fBhttp://sam.zoy.org/zzuf/\fR.
Note: See TracBrowser for help on using the repository browser.