openssl/fuzz
Rich Salz 257e9d03b0 Fix issues reported by markdownlint
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11739)
2020-05-08 16:22:02 +02:00
..
corpora Add CMP fuzzing to fuzz/cmp.c, including a couple of helpers in crypto/cmp/ 2020-04-18 19:54:17 +02:00
asn1.c fuzz/asn1.c: Add missing #include 2020-04-28 00:03:17 +02:00
asn1parse.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
bignum.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
bndiv.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
build.info Add CMP fuzzing to fuzz/cmp.c, including a couple of helpers in crypto/cmp/ 2020-04-18 19:54:17 +02:00
client.c In OpenSSL builds, declare STACK for datatypes ... 2020-04-24 16:42:46 +02:00
cmp.c In OpenSSL builds, declare STACK for datatypes ... 2020-04-24 16:42:46 +02:00
cms.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
conf.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
crl.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
ct.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00
driver.c Following the license change, modify the boilerplates in fuzz/ 2018-12-06 15:37:38 +01:00
fuzzer.h Following the license change, modify the boilerplates in fuzz/ 2018-12-06 15:37:38 +01:00
helper.py Following the license change, modify the boilerplates in fuzz/ 2018-12-06 15:37:38 +01:00
mkfuzzoids.pl Update copyright year 2020-04-23 13:55:52 +01:00
oids.txt GOST cipher names adjustment 2020-04-17 16:21:51 +03:00
rand.inc Following the license change, modify the boilerplates in fuzz/ 2018-12-06 15:37:38 +01:00
README.md Fix issues reported by markdownlint 2020-05-08 16:22:02 +02:00
server.c In OpenSSL builds, declare STACK for datatypes ... 2020-04-24 16:42:46 +02:00
test-corpus.c Following the license change, modify the boilerplates in fuzz/ 2018-12-06 15:37:38 +01:00
x509.c Deprecate ERR_get_state() 2019-09-12 18:34:06 +02:00

Fuzzing OpenSSL

OpenSSL can use either LibFuzzer or AFL to do fuzzing.

LibFuzzer

How to fuzz OpenSSL with libfuzzer, starting from a vanilla+OpenSSH server Ubuntu install.

With clang from a package manager

Install clang, which ships with libfuzzer since version 6.0:

sudo apt-get install clang

Configure openssl for fuzzing. For now, you'll still need to pass in the path to the libFuzzer library file while configuring; this is represented as $PATH_TO_LIBFUZZER below. A typical value would be /usr/lib/llvm-7/lib/clang/7.0.1/lib/linux/libclang_rt.fuzzer-x86_64.a.

CC=clang ./config enable-fuzz-libfuzzer \
        --with-fuzzer-lib=$PATH_TO_LIBFUZZER \
        -DPEDANTIC enable-asan enable-ubsan no-shared \
        -DFUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION \
        -fsanitize=fuzzer-no-link \
        enable-ec_nistp_64_gcc_128 -fno-sanitize=alignment \
        enable-weak-ssl-ciphers enable-rc5 enable-md2 \
        enable-ssl3 enable-ssl3-method enable-nextprotoneg \
        --debug

Compile:

sudo apt-get install make
make clean
LDCMD=clang++ make -j4

Finally, perform the actual fuzzing:

fuzz/helper.py $FUZZER

where $FUZZER is one of the executables in fuzz/. It will run until you stop it.

If you get a crash, you should find a corresponding input file in fuzz/corpora/$FUZZER-crash/.

With clang from source/pre-built binaries

You may also wish to use a pre-built binary from the LLVM Download site, or to build clang from source. After adding clang to your path and locating the libfuzzer library file, the procedure for configuring fuzzing is the same, except that you also need to specify a --with-fuzzer-include option, which should be the parent directory of the prebuilt fuzzer library. This is represented as $PATH_TO_LIBFUZZER_DIR below.

CC=clang ./config enable-fuzz-libfuzzer \
        --with-fuzzer-include=$PATH_TO_LIBFUZZER_DIR \
        --with-fuzzer-lib=$PATH_TO_LIBFUZZER \
        -DPEDANTIC enable-asan enable-ubsan no-shared \
        -DFUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION \
        -fsanitize=fuzzer-no-link \
        enable-ec_nistp_64_gcc_128 -fno-sanitize=alignment \
        enable-weak-ssl-ciphers enable-rc5 enable-md2 \
        enable-ssl3 enable-ssl3-method enable-nextprotoneg \
        --debug

AFL

This is an alternative to using LibFuzzer.

Configure for fuzzing:

sudo apt-get install afl-clang
CC=afl-clang-fast ./config enable-fuzz-afl no-shared no-module \
    -DPEDANTIC enable-tls1_3 enable-weak-ssl-ciphers enable-rc5 \
    enable-md2 enable-ssl3 enable-ssl3-method enable-nextprotoneg \
    enable-ec_nistp_64_gcc_128 -fno-sanitize=alignment \
    --debug
make clean
make

The following options can also be enabled: enable-asan, enable-ubsan, enable-msan

Run one of the fuzzers:

afl-fuzz -i fuzz/corpora/$FUZZER -o fuzz/corpora/$FUZZER/out fuzz/$FUZZER

Where $FUZZER is one of the executables in fuzz/.

Reproducing issues

If a fuzzer generates a reproducible error, you can reproduce the problem using the fuzz/-test binaries and the file generated by the fuzzer. They binaries don't need to be build for fuzzing, there is no need to set CC or the call config with enable-fuzz- or -fsanitize-coverage, but some of the other options above might be needed. For instance the enable-asan or enable-ubsan option might be useful to show you when the problem happens. For the client and server fuzzer it might be needed to use -DFUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION to reproduce the generated random numbers.

To reproduce the crash you can run:

fuzz/$FUZZER-test $file

Random numbers

The client and server fuzzer normally generate random numbers as part of the TLS connection setup. This results in the coverage of the fuzzing corpus changing depending on the random numbers. This also has an effect for coverage of the rest of the test suite and you see the coverage change for each commit even when no code has been modified.

Since we want to maximize the coverage of the fuzzing corpus, the client and server fuzzer will use predictable numbers instead of the random numbers. This is controlled by the FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION define.

The coverage depends on the way the numbers are generated. We don't disable any check of hashes, but the corpus has the correct hash in it for the random numbers that were generated. For instance the client fuzzer will always generate the same client hello with the same random number in it, and so the server, as emulated by the file, can be generated for that client hello.

Coverage changes

Since the corpus depends on the default behaviour of the client and the server, changes in what they send by default will have an impact on the coverage. The corpus will need to be updated in that case.

Updating the corpus

The client and server corpus is generated with multiple config options:

  • The options as documented above
  • Without enable-ec_nistp_64_gcc_128 and without --debug
  • With no-asm
  • Using 32 bit
  • A default config, plus options needed to generate the fuzzer.

The libfuzzer merge option is used to add the additional coverage from each config to the minimal set.

Minimizing the corpus

When you have gathered corpus data from more than one fuzzer run or for any other reason want to to minimize the data in some corpus subdirectory fuzz/corpora/DIR this can be done as follows:

mkdir fuzz/corpora/NEWDIR
fuzz/$FUZZER -merge=1 fuzz/corpora/NEWDIR fuzz/corpora/DIR